International Soil and Water Conservation 3

7 INTERNATIONAL SOIL AND WATER CONSERVATION 3

RESEARCH (ISWCR) and WaterSoil (ISWCR) International ConservationResearch Vol. 203 (2019) 7 – 316 ISSN 2095-6339 Volume 7, No. 3, September 2019 CN 10-1107/P

CONTENTS

C. Alewell , P. Borrelli , K. Meusburger , P. Panagos Using the USLE: Chances, challenges and limitations of soil erosion modelling ...... 203 C. Kristo , H. Rahardjo , A. Satyanaga Effect of hysteresis on the stability of residual soil slope ...... 226 E. Debie , K.N. Singh , M. Belay Effect of conservation structures on curbing rill erosion in micro-watersheds, northwest Ethiopia ...... 239 K.A. Darkwah , J.D. Kwawu , F. Agyire-Tettey , D.B. Sarpong Assessment of the determinants that inf uence the adoption of sustainable soil and water conservation practices in Techiman Municipality of Ghana ...... 248 M. Ghorbani , H. Asadi , S. Abrishamkesh Effects of rice husk biochar on selected soil properties and nitrate leaching in loamy sand and clay soil ...... 258 M. Norman , H.Z.M. Shafri , S.B. Mansor , B. Yusuf Review of remote sensing and geospatial technologies in estimating rooftop rainwater harvesting (RRWH) quality ...... 266 Volume 7, No. 3, September 2019 A.M. Abdallah The effect of hydrogel particle size on water retention properties and availability under water stress ...... 275 F. Garc í a- Á vila , J. Pati ñ o-Ch á vez , F. Zhin í n-Chimbo , S. Donoso-Moscoso , L. Flores del Pino , A. Avil é s-A ñ azco Performance of Phragmites Australis and Cyperus Papyrus in the treatment of municipal wastewater by vertical f ow subsurface constructed wetlands...... 286 M.V. Slukovskaya , V.I. Vasenev , K.V. Ivashchenko , D.V. Morev , S.V. Drogobuzhskaya , L.A. Ivanova, I.P. Kremenetskaya Technosols on mining wastes in the subarctic: Eff ciency of remediation under Cu-Ni atmospheric pollution...... 297 M. Mohammadi , A. Khaledi Darvishan , N. Bahramifar Spatial distribution and source identif cation of heavy metals (As, Cr, Cu and Ni) at sub-watershed scale using geographically weighted regression ...... 308

Available online at www.sciencedirect.com

ISWCR_v7_i3_COVER.indd 1 30-07-2019 19:09:00 The International Soil and Water Conservation Research (ISWCR) is the ofﬁ cial journal of World Association of Soil and Water Conservation (WASWAC), which is a non-governmental, non-proﬁ t organization. The mission of WASWAC is to promote the wise use of management practices, to improve and safeguard the quality of land and water resources. The Journal is co-owned by the International Research and Training Center on Erosion and Sedimentation, China Water & Power Press, and China Institute of Water Resources and Hydropower Research.

20 Chegongzhuang West Road IRTCES Building 401-514 Beijing 100048, China +86 10 68766413 [email protected] www.irtces.org

The International Research and Training Center on Erosion and Sedimentation (IRTCES) was jointly set up by the Government of China and UNESCO. It aims at the promotion of international exchange of knowledge and cooperation in the studies of erosion and sedimentation.

Building D, No. 1, South Yuyuantan Road, Beijing 100038, China +86 10 68545912 [email protected] http://www.waterpub.com.cn

China Water & Power Press (CWPP) is afﬁ liated to the Ministry of Water Resources of China, which takes leadership of science and technology publishing in China.

A-1 Fuxing Road, Beijing100038, China +86 10 68781650 [email protected] http://www.iwhr.com

China Institute of Water Resources and Hydropower Research is the largest and most comprehensive research institute in the ﬁ eld of water resources and hydropower in China.

Publishing services provided by Elsevier B.V. on behalf of KeAi

ISWCR_v7_i3_COVER.indd 2 30-07-2019 14:47:43 International Soil and Water Conservation Research (ISWCR) Editorial Board Editorial Committee Liu, Ning Kuang, Shangfu Ying, Youfeng Wang, Yujie Ning, Duihu Li, Zhongfeng Liu, Guangquan Advisor Blum, Winfried Li, Rui Sombatpanit, Samran University of Natural Resources and Life Institute of Soil and Water Conservation, Department of Land Development Bangkhen, Sciences, Austria Chinese Academy of Sciences Thailand Critchley, William Liu, Guobin Sun, Honglie CIS-Centre for International Cooperation, Institute of Soil and Water Conservation, Institute of Geographic Science and Natural Netherlands Chinese Academy of Sciences, China Resources Research, Chinese Academy of Dumanski, Julian Shao, Mingan Sciences, China World Bank; Gov’t of Canada, Canada Institute of Geographic Science and Natural Tang, Keli El-Swaify, Samir Resources Research, Chinese Academy of Institute of Soil and Water Conservation, University of Hawai‘I, USA Sciences (CAS), China Chinese Academy of Sciences, China Lal, Rattan Shen, Guofang Walling, Des E. Ohio State University, USA Beijing Forestry University, China University of Exeter, UK Editor-In-Chief Editor in Chief for English Nearing, Mark Laf en, John USDA-ARS Southwest Watershed Research Center, USA Iowa State University, USA [email protected] laf [email protected] Lei, Tingwu China Agricultural University, China [email protected] Associate Editors Al-Hamdan, Osama Gomez Macpherson, Helena Oliveira, Paulo Tarso S. Yu, Xinxiao Texas A&M University, Institute for Sustainable Federal University of Mato Beijing Forestry University, Kingsville, USA Agriculture (IAS-CSIC), Spain Grosso do Sul, Brazil China Baffaut, Claire Guzmán, Gema Osorio, Javier Zhang, Guanghui USDA-ARS Columbia, MO, USA Universidad of Cordoba, Spain Texas A&M University, USA Beijing Normal University, Borrelli, Pasquale Huang, Chi-hua Pla Sentís, Ildefonso China University of Basel, Switzerland USDA-ARS National Soil Erosion Universitat de Lleida, Spain Zhang, Qingwen Bruggeman, Adriana Research Laboratory, USA Polyakov, Viktor CAAS Institute of Agro- Energy, Environment and Water Krecek, Josef USDA-ARS Southwest Watershed Environment and Sustainable Research Center of the Cyprus Czech Technical University, CZ Research Center, USA Development, China Institute, Cyprus Licciardello, Feliciana Reichert, Jose Miguel Zhang, Yongguang Chen, Hongsong University of Catania, Italy Federal University of Santa International Institute for Earth Institute of Subtropical Lorite Torres, Ignacio Maria, Brazil System Science, Nanjing Agriculture, Chinese Academy of Centro Alameda del Obispo, Sadeghi, Seyed Hamidreza University, China Sciences, China IFAPA, Junta de Andalucía, Spain Tarbiat Modares University, Iran Zheng, Fenli Dazzi, Carmelo Merten, Gustavo Wei, Haiyan Institute of Soil and Water University of Palermo, Italy University of Minnesota Duluth, Southwest Watershed Research, Conservation, Chinese Academy Fang, Hongwei USA USDA-ARS, USA of Sciences Tsinghua Univeristy, China Napier, Ted Wagner, Larry Fullen, Mike Ohio State University, USA USDA-ARS, Rangeland Zlatic, Moidrag University of Wolverhampton, UK Nouwakpo, Sayjro Resources & Systems Research University of Belgrade, Golabi, Mohammad University of Nevada, USA Unit, USA Serbia University of Guam, USA Nunes, João Williams, Jason Golosov, Valentin Univeristy of Lisbon, Portugal Southwest Watershed Research, Kazan Federal University, Russia Obando-Moncayo, Franco USDA-ARS, USA Gomez, Jose Alfonso Geographic Institute Agustin Yin, Shuiqing Institute for Sustainable Codazzi, Soil Survey Department, Beijing Normal University, Agriculture, CSIC, Cordob, Spain Colombia China Editorial Board Members Behera, Umakant, India Liang, Yin, China Peiretti, Roberto, Argentina Bhan, Suraj, India Liu, Baoyuan, China Reinert, Dalvan, Brazil Cai, Qiangguo, China Liu, Benli, China Ritsema, Coen, Netherlands Chen, Su-Chin, China (Taiwan) Liu, Xiaoying, China Shahid, Shabbir, UAE Chen, Yongqin David, China (Hong Kong) Lo, Kwong Fai Andrew, China (Taiwan) Shi, Xuezheng, China Cui, Peng, China Mahoney, William B., USA Shi, Zhihua, China Dazzi, Carmelo, Italy Miao, Chiyuan, China Sukvibool, Chinapat, Thailand Delgado, Jorge A., USA Motavalli, Peter, USA Tahir Anwar, Muhammad, Pakistan He, Binghui, China Mrabet, Rachid, France Torri, Dino, Italy Horn, Rainer, Germany Mu, Xingmin, China Wang, Bin, China Kapur, Selim, Turkey Ni, Jinren, China Wang, Tao, China Kukal, Surinder Singh, India Oliveira, Joanito, Brazil Wu, Gaolin, China Kunta, Karika, Thailand Ouyang, Wei, China Zhang, Fan, China Li, Yingkui, USA Owino, James, Kenya Zhang, Kebin, China Li, Zhanbin, China Paz-Alberto, Annie Melinda, Philippines Managing Editors Chyu, Paige Zhang, Tan International Research and Training Center on Erosion China Water and Power Press, China and Sedimentation, China [email protected] [email protected] Assistant Editors Ban, Yunyun Lin, Xingna China Agricultural University, China Beijing Forestry University, China International Soil and Water Conservation Research (ISWCR)

Volume 7, Number 3, September 2019

International Research and Training Center on Erosion and Sedimentation and China Water and Power Press AIMS AND SCOPE

The International Soil and Water Conservation Research (ISWCR), the ofﬁ cial journal of World Association of Soil and Water Conservation (WASWAC) www.waswac.org, is a multidisciplinary journal of soil and water conservation research, practice, policy, and perspectives. It aims to disseminate new knowledge and promote the practice of soil and water conservation. The scope of International Soil and Water Conservation Research includes research, strategies, and technologies for prediction, prevention, and protection of soil and water resources. It deals with identif cation, characterization, and modeling; dynamic monitoring and evaluation; assessment and management of conservation practice and creation and implementation of quality standards. Examples of appropriate topical areas include (but are not limited to): ∑ Soil erosion and its control ∑ Watershed management ∑ Water resources assessment and management ∑ Nonpoint-source pollution ∑ Conservation models, tools, and technologies ∑ Conservation agricultural ∑ Soil health resources, indicators, assessment, and management ∑ Land degradation ∑ Sedimentation ∑ Sustainable development ∑ Literature review on topics related soil and water conservation research © 2019 International Research and Training Center on Erosion and Sedimentation and China Water and Power Press. Production and hosting by Elsevier B.V. Peer review under responsibility of IRTCES and CWPP. Notice No responsibility is assumed by the International Soil and Water Conservation Research (ISWCR) nor Elsevier for any injury and/or damage to persons, property as a matter of product liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. Although all advertising material is expected to conform to ethical standards, inclusion in this publication does not constitute a guarantee or endorsement of the quality or value of such product or of the claims made of it by its manufacturer. Full text available on ScienceDirect

INTERNATIONAL SOIL AND WATER CONSERVATION RESEARCH (ISWCR) Volume 7, No. 3, September 2019

CONTENTS

C. Alewell, P. Borrelli , K. Meusburger , P. Panagos Using the USLE: Chances, challenges and limitations of soil erosion modelling ...... 203 C. Kristo, H. Rahardjo , A. Satyanaga Effect of hysteresis on the stability of residual soil slope ...... 226 E. Debie , K.N. Singh , M. Belay Effect of conservation structures on curbing rill erosion in micro-watersheds, northwest Ethiopia . . . . . 239 K.A. Darkwah , J.D. Kwawu , F. Agyire-Tettey , D.B. Sarpong Assessment of the determinants that inf uence the adoption of sustainable soil and water conservation practices in Techiman Municipality of Ghana...... 248 M. Ghorbani , H. Asadi , S. Abrishamkesh Effects of rice husk biochar on selected soil properties and nitrate leaching in loamy sand and clay soil ...... 258 M. Nor man , H.Z.M. Shafri , S.B. Mansor , B. Yusuf Review of remote sensing and geospatial technologies in estimating rooftop rainwater harvesting (RRWH) quality ...... 266 A.M. Abdallah The effect of hydrogel particle size on water retention properties and availability under water stress ...... 275 F. Garc í a- Á vila, J. Pati ñ o-Ch á vez , F. Zhin í n-Chimbo , S. Donoso-Moscoso, L. Flores del Pino , A. Avil é s-A ñ azco Performance of Phragmites Australis and Cyperus Papyrus in the treatment of municipal wastewater by vertical f ow subsurface constructed wetlands...... 286 M.V. Slukovskaya , V.I. Vasenev , K.V. Ivashchenko , D.V. Morev , S.V. Drogobuzhskaya , L.A. Ivanova , I.P. Kremenetskaya Technosols on mining wastes in the subarctic: Eff ciency of remediation under Cu-Ni atmospheric pollution...... 297 M. Mohammadi , A. Khaledi Darvishan , N. Bahramifar Spatial distribution and source identif cation of heavy metals (As, Cr, Cu and Ni) at sub-watershed scale using geographically weighted regression...... 308

International Soil and Water Conservation Research 7 (2019) 203e225

Contents lists available at ScienceDirect

International Soil and Water Conservation Research

journal homepage: www.elsevier.com/locate/iswcr

Review Article Using the USLE: Chances, challenges and limitations of soil erosion modelling

* Christine Alewell a, , Pasquale Borrelli a, Katrin Meusburger b, Panos Panagos c a Environmental Geosciences, University of Basel, Switzerland b Swiss Federal Institute for Forest, Snow and Landscape Research, Birmensdorf, Switzerland c European Commission, Joint Research Center, Ispra, Italy article info abstract

Article history: To give soils and soil degradation, which are among the most crucial threats to ecosystem stability, social Received 10 January 2019 and political visibility, small and large scale modelling and mapping of soil erosion is inevitable. The most Received in revised form widely used approaches during an 80year history of erosion modelling are Universal Soil Loss Equation 20 May 2019 (USLE)-type based algorithms which have been applied in 109 countries. Addressing soil erosion by Accepted 29 May 2019 water (excluding gully erosion and land sliding), we start this review with a statistical evaluation of Available online 18 June 2019 nearly 2,000 publications). We discuss model developments which use USLE-type equations as basis or side modules, but we also address recent development of the single USLE parameters (R, K, LS, C, P). Keywords: Universal soil loss equation Importance, aim and limitations of model validation as well as a comparison of USLE-type models with RUSLE other erosion assessment tools are discussed. Model comparisons demonstrate that the application of CSLE process-based physical models (e.g., WEPP or PESERA) does not necessarily result in lower uncertainties Soil redistribution compared to more simple structured empirical models such as USLE-type algorithms. We identiﬁed four Soil degradation key areas for future research: (i) overcoming the principally different nature of modelled (gross) versus Water erosion measured (net) erosion rates, in coupling on-site erosion risk to runoff patterns, and depositional regime, Model (ii) using the recent increase in spatial resolution of remote sensing data to develop process based Review models for large scale applications, (iii) strengthen and extend measurement and monitoring programs to build up validation data sets, and (iv) rigorous uncertainty assessment and the application of objective evaluation criteria to soil erosion modelling. © 2019 International Research and Training Center on Erosion and Sedimentation and China Water and Power Press. Production and Hosting by Elsevier B.V. This is an open access article under the CC BY-NC- ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Background the early nineties, it was already estimated that 56% of the global land being degraded and showed light to severe forms of soil In a world of climate and land use change, fertile soils are one of erosion by water (Oldeman,1992, pp. 16e36). Since water erosion is the most essential resources to sustain humankind. In a recent strongly exacerbated by the conversion of natural vegetation to review paper in Science (Amundson et al., 2015) soil is discussed agricultural land, with nearly 40% of Earth's land currently utilized comprehensively as THE essential resource for human security for agricultural production (Foley, 2017), accelerated forms of soil (including climate and food security) in the 21st century with the erosion became a widespread phenomenon representing a major main threat to soils being soil erosion by wind and water ever since challenge to achieving the United Nations Sustainable Develop- humankind had started with agriculture. ment Goals (Keesstra et al., 2016). To date, most of the world's soils are only in fair, poor, or very The fact that soil erosion is a threat to one of the most essential poor condition as was stressed by the latest publications of the resources of humankind is as old as Cain and Abel (Panagos et al., United Nations, i.e. Status of the World's Soil Resources (FAO, 2015) 2016c). In modern ages, Russian soil scientists were probably the where soil erosion was identified as one of the major soil threats. In first to be successful in directing attention towards soil erosion which resulted in governmental mitigation programs at the end of the 19th century (Dokuchaev, 1892). Today, tools to not only map the actual status of soil erosion, but also to test the influence of * Corresponding author. mitigation strategies as well as management or conservations E-mail address: [email protected] (C. Alewell). https://doi.org/10.1016/j.iswcr.2019.05.004 2095-6339/© 2019 International Research and Training Center on Erosion and Sedimentation and China Water and Power Press. Production and Hosting by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). 204 C. Alewell et al. / International Soil and Water Conservation Research 7 (2019) 203e225 practices and last but not least future (climate change) scenarios are Soil Erosion Model (LISEM, (De Roo, Wesseling, & Ritsema, 1996), urgently needed (Tan, Leung, Li, & Tesfa, 2018). Without the pos- the European Soil Erosion Model (EUROSEM; (Morgan et al., 1998) sibility to map and visualize soil erosion on large scales, soil erosion and Pan European Soil Erosion Risk Assessment (PESERA, Kirkby will be out of the focus of all major environmental and agricultural et al. (2008)) resulted in 220 hits (125, 32, 34 and 29, respec- policies such as the European Union's Common Agricultural Policy tively) of the same study period. (CAP), the United Nations Sustainable Development Goals (SDGs), One of the main reasons why USLE type modelling is so widely the United Nations Convention to Combat Desertification (UNCCD) used throughout the world (see also Figs. 1e3 and section 2.1) is and the Intergovernmental Science-Policy Platform on Biodiversity certainly its high degree of flexibility and data accessibility, a and Ecosystem Services (IPBES). Moreover, the lack of large scale parsimonious parametrization, extensive scientific literature and erosion rate data sets also creates knowledge gaps in climate comparability of results allowing to adapt the model to nearly every change and carbon mitigation scenarios, hydrology and flood pre- kind of condition and region of the world. Nevertheless, the USLE diction as well as earth science system modelling. approach is an empirical modelling approach with significant Modelling and prediction of soil erosion by water has a long limitations which already have been addressed in the very first history with first studies published in international journals more publications as there is no simulation of soil deposition (e.g., than seven decades ago using north American data sets (e.g. sedimentation) and that in most cases not enough measured data (Bennett, 1939; Smith, 1941; Zingg, 1940b)). Many mathematical exist to rigorously determine the single factors for all needed sit- models categorized as empirical, conceptual, physically-based or uations and scenarios (Wischmeier and Smith, 1965, 1978, p. 58). process-oriented are available to estimate soil erosion at different The question remains, whether or not research and development of spatial and temporal scales (De Vente & Poesen, 2005; Jetten, the last 5 decades (1965 e today) was able to overcome these main Govers, & Hessel, 2003; Karydas, Panagos, & Gitas, 2014; King & limitations, or, if not, at least improved them to such an extent that Delpont, 1993; Merritt, Letcher, & Jakeman, 2003; Morgan, 2005, is justifiable to apply the model algorithm to large scales and if so, pp. 365e376; Toy T. J, 2002; Vrieling, 2006; Zhang, O'Neill, & Lacey, under which conditions and resolution. 1996a). Most erosion models contain a mix of modules from each of The aim of this review is a thorough and rigorous evaluation of these categories. The research community has been contempora- USLE-type modelling to come to some conclusions on the sense and neously working for both, improving the applicability of complex non-sense of soil erosion modelling with this model concept process-oriented models such as the Water Erosion Prediction depending on specific circumstances and conditions. We do not Project (WEPP) (Boardman, 2006; Morgan & Nearing, 2011) or the aim to provide a user guide or handbook of USLE-type modelling European Soil Erosion Model (EUROSEM; Morgan et al. (1998)) as but ask the interested reader to depend in this respect on the well as updating the existing empirical approaches such as the wealth of literature starting from the original publications Universal Soil Loss Equation (USLE) (Bagarello et al., 2008, 2015; (Wischmeier and Smith, 1965, 1978, p. 58) to reviews on its history Bagarello & Ferro, 2004; Di Stefano, Ferro, & Pampalone, 2017; and past use (e.g. (Laflen & Flanagan, 2013; Laflen & Moldenhauer, Kinnell, 2014) which not only remains attractive from a practical 2003)), on special applications and further development (e.g. point of view (Cao, Zhang, Dai, & Liang, 2015; Gessesse, Bewket, & consideration of runoff and event based modelling (Kinnell, 2010)) Brauning,€ 2015) but also gives estimates on large spatial scales (see to general evaluations of model concepts and usability (Merritt below, (Diodato, Borrelli, Fiener, Bellocchi, & Romano, 2017; et al., 2003; Nearing, 2004; Quinton, 2013). We also do not claim Doetterl, Van Oost, & Six, 2012; Van Oost et al., 2007; Yang, Kanae, or even make an attempt to review all existing studies using USLE- Oki, Koike, & Musiake, 2003a)). type models which is clearly beyond the scope of a manuscript in At present, USLE and the Revised USLE (RUSLE) are by far the ESR but have included a mere numerical geographical meta- most widely applied soil erosion prediction models globally and analysis to embrace the wealth of literature published on USLE- according to (Risse, Nearing, Laflen, & Nicks, 1993) “USLE has been type models quantitatively (see section 2.1). used throughout the world for a variety of purposes and under many different conditions simply because it seems to meet the 2. Historical aspects and evolution of USLE-type modelling need better than any other tool available”. A Web of Science query (http://apps.webofknowledge.com/) for the period 2007e2017 The origin of USLE-type models was in the US to provide a resulted in 1149 hits corresponding to the keywords “Universal Soil management decision support tool and was based upon thousands Loss Equation”, “USLE”, “Revised Universal Soil Loss Equation” or of controlled studies on field plots and small watersheds since 1930 “RUSLE”, with average citations per item of 7.8 (as of August 2017). (Wischmeier & Smith, 1965). As with all empirical methods the A query with the Science Direct tool for the last 40 years resulted in model concept is not based on process description and simulation 1556 studies using USLE or RUSLE with an average citation rate of but rather on understanding a process, capturing the confounding cited publications of 18.8 (studies with no citations were not measureable parameters and delineating a mathematical algorithm considered in this average rate, see meta-analysis in section 2.1). out of the relationship between these parameters and the The most well-known models based on USLE/RUSLE technology are measured output (in this case measured eroded sediments). the Soil and Water Assessment Tool (SWAT) (Arnold, Srinivasan, As such, the USLE was defined as: Muttiah, & Williams, 1998), the AGricultural Non-Point Source Pollution Model (AGNPS) (Cronshey & Theurer, 1998; Young, A ¼ R,K,L,S,C,P (1) Onstad, Bosch, & Anderson, 1989), and the Water and Tillage Erosion and Sediment Model (Watem/Sedem) (Van Rompaey, where: A (Mg ha 1 yr 1) is the annual average soil erosion, R (MJ Verstraeten, Van Oost, Govers, & Poesen, 2001) and the Chinese mm h 1 ha 1 yr 1) is the rainfall-runoff erosivity factor, K (Mg h Soil Loss Equation (CSLE, Liu, Zhang, and Xie (2002)) which total- MJ 1 mm 1) is the soil erodibility factor, L (dimensionless) is the ized 1385, 311, 67 and 35 hits, respectively, in the same period (Web slope length factor, S (dimensionless) is the slope steepness factor, of Science query 2007e2017). Modelling approaches independent C (dimensionless) is the land cover and management factor, P from the USLE such as the Water Erosion Prediction Model (WEPP, (dimensionless) is the soil conservation or prevention practices Laflen, Elliot, Flanagan, Meyer, and Nearing (1997)), the Limburg factor. The impact of the factors R capturing the energy and amount C. Alewell et al. / International Soil and Water Conservation Research 7 (2019) 203e225 205

Fig. 1. Geographic distribution per country of published papers using USLE or RUSLE modelling during the last 40 years (1977 to July 2017).

sediment yield. Even though Wischmeier and Smith (1965) are generally cited as the origin of the equation, the development of the modelling concept actually started much earlier. The soil-loss equation was the result of 20 years of development starting from the so called slope-practice method (Zingg, 1940b) on which Smith (1941) added crop and conservation-practice factors. Eventually soil erodibility, management factors (Browning, Parish, & Glass, 1947) and the rainfall factor (Musgrave, 1947) were added. The resulting “Mus- grave equation” was published as a graphical solution in 1952 (Lloyd & Eley, 1952) but was further improved throughout the 1950's with an improved rainfall-erosion index, a method of evaluating cropping-management effects on the basis of local climatic conditions, a quantitative soil-erodibility factor and a method of accounting for effects of interrelations of such variables as pro- Fig. 2. Number of studies per continent and percentage of total publication number ductivity level, crop sequence and residue management (USDA, (1977 to July 2017). 1961; Wischmeier, 1959; Wischmeier, 1960). Wischmeier and Smith (1965) with an update by Wischmeier and Smith (1978, p. 58) finally put figures, tables and equations for all five single fac- of precipitation, K accounting for the soil parameters determining tors of the USLE together in a guide book explaining application and erosion potential, C describing the vegetation cover and manage- use in detail. This guidebook was published by the US National ment as well as P delineating human management intervention is Runoff and Soil Loss Data Center, in cooperation between the directly related to our process understanding of soil erosion. Un- Agricultural Research Service and Purdue University, and it resulted derstanding erosional processes and driving factors, it might seem from statistical analysis of more than 10,000 plot-years of runoff surprising that neither runoff nor infiltrating was included in the and soil loss data carried out in plots having a length less than or algorithm. However, from a pure statistical perspective the rainfall equal to 122 m and a slope ranging from 3 to 18%. For defining the erosivity was simply the best predictor for the measured erosion mathematical structure of the USLE a reference condition, named output and attempts to include runoff even reduced the quality of as unit plot, was used, which built on the set-up of field experi- the assessment (Wischmeier, 1966; Wischmeier & Smith, 1965). ments which were commonly used. The unit plot was defined as a From a soil scientific perspective infiltration is not a helpful 22.1 m long plot, with a 9% slope, maintained in a continuous parameter for such a modelling endeavour, because it is very prone regularly tilled fallow condition with up-and-down hill tillage. The to measurement errors, extremely variable in soils and we will unit plot was used to compare soil loss data collected on plots that hardly ever be able to capture infiltration at larger scales. The had different slopes, lengths, cropping and management and con- implementation of the L and S are meant to capture the impact of servation practices (Wischmeier and Smith, 1965, 1978, p. 58). runoff energy which is influenced by a mix of processes and pa- The main motivation of the early studies using the USLE were to rameters. However, it has to be noted that in all these endeavours quantify soil erosion rates and their single contributing factors in modelling targeted at on site soil erosion risk, not at catchment 206 C. Alewell et al. / International Soil and Water Conservation Research 7 (2019) 203e225

Fig. 3. Trend of published papers (USLE/RUSLE). The period 2013e2017 covers until July 2017, the projection for the full 5 year period would approximate 600 papers (striped box). comparison to soil-loss tolerance values and assess possible com- Johnson, Savabi, & Loomis, 1984). Johnson et al. (1984) suggested binations of cropping systems and management plans for mitiga- that a more accurate assessment of the cover and management tion (Schwertmann, Vogl, & Kainz, 1987, pp. 1e64; Wischmeier & conditions is needed for applications of the USLE on rangelands. Smith, 1965). Weltz, Kidwell, and Fox (1998) stated that ‘USLE is a lumped Since its first definition, USLE-type modelling was developed empirical model that does not separate factors that influence soil further to meet the multiple needs and conditions of modelled erosion, such as plant growth, decomposition, infiltration, runoff, systems. Renard, Foster, Weesies, McCool, and Yoder (1997) stated soil detachment, or soil transport’. The increasing criticisms with that the USLE for more than four decades has proven that this regard to the limitations of the USLE were met by an increasing technology is valuable as a conservation-planning guide in the US, interest in both the US action agencies and the soil science research providing farmers and conservation planners with a tool to esti- community to update the USLE thus creating a substantial need for mate rates of soil erosion for different cropping systems and land a revision of the USLE which were made possible by the fast im- managements. Even though conceptualized and calibrated for provements in computation capacity. agricultural areas, the USLE has been soon adjusted to extend its Accordingly, the USLE was upgraded to the Revised Universal applicability to undisturbed land. In the early 1970's, as a result of a Soil Loss Equation (RUSLE) during the 1990s (Renard et al., 1991, meeting between the Natural Resources Conservation Service 1997, pp. 1e404). The basic structure of the RUSLE retains the (NRCS) and the Forest Service of the USDA, a first extension of the multiplicative form of the USLE, but it also has process-based USLE was made by developing a sub-factor to compute the cover- auxiliary components such as calculating time-variable soil erod- management factors (C-factor) for woodland, permanent pasture ibility, plant growth, residue management, residue decomposition and rangeland (Spaeth, Pierson, Weltz, & Blackburn, 2003; and soil surface roughness as a function of physical and biological Wischmeier, 1975, pp. 118e124). The proposed method to estimate processes. RUSLE also has updated values for erosivity (R), new the USLE C-factor used relationships to extrapolate three major relationships for the topographical components (L and S factors) sub-factors, canopy, ground cover and below soil effects (Renard & which include ratios of rill and inter-rill erosion, consideration of Foster, 1985), since for these areas an extensive data base of the C seasonality of the K factor and additional P factors for rangelands factor was not available. Later, Dissmeyer and Foster (1980) pro- and subsurface drainage, among other improvements (please see posed further additions and modifications to extend the number of section 3 for description of advancement regarding the single fac- sub-factors operating in woodland, including (i) amount of bare soil tors of RUSLE). (or conversely ground cover), (ii) canopy, (iii) soil reconsolidation, RUSLE combines the advantage of being based on the same (iv) high organic content, (v) fine roots, (vi) residual binding effect, extensive database as was the USLE with some process-based (vii) onsite storage, (viii) steps and (ix) contour tillage (agrofor- computations for time-varying environmental effects on the estry). As reported by Spaeth et al. (2003) early applications of the erosional system (Nearing, 2004). However, it still has the limita- model in rangelands showed little agreement between estimated tions in model structure that allows only for limited interactions and measured soil loss in both catchment (Simanton, Roger, and inter-relationships between the basic multiplicative factors Herbert, & Renard, 1980) and plot experiments (Hart, 1984; (Nearing, 2004). As USLE-type models were designed to predict C. Alewell et al. / International Soil and Water Conservation Research 7 (2019) 203e225 207 long-term average annual soil loss, they have been successful to of the RUSLE published by the Research Branch Agriculture and predict event soil losses reasonably well at some geographic loca- Agri-Food Canada (Wall, Coote, Pringle, & Shelton, 2002, pp. tions (Kinnell, 2010), but often fail to predict event erosion, which is 1e117)); Australia (Ferro & Porto, 1999); Africa (Haileslassie, Priess, highly influenced by the fact that the USLE and its revisions (RUSLE) Veldkamp, Teketay, & Lesschen, 2005; Lufafa, Tenywa, Isabirye, do not consider runoff explicitly. Majaliwa, & Woomer, 2003) and Asia (Lee & Heo, 2011; RUSLE2 was developed to scientifically enhance the USLE/RUSLE Meusburger, Mabit, Park, Sandor, & Alewell, 2013; Xu, Miao, & Liao, equations and offer an improved tool to guide and assist erosion- 2008). In China, the USLE was adapted in substituting the C and the control planning (USDA, 2008). Although the fundamental empir- P factor with three factors considering the biology, engineering and ical equation scheme of the previous equations was retained, tillage practices (Liu et al., 2002). For the application in Europe RUSLE2 uses both empirical and process-based equations that (Schwertmann et al., 1987, pp. 1e64) was the first to rigorously allow it to extend significantly beyond the original USLE structure. evaluate the applicability of the USLE to the different climatic In fact, RUSLE2 can be defined as a hybrid soil erosion prediction conditions and land use management systems of Bavaria (Southern (estimation) approach since it is a combination of the empirical, Germany). Today, USLE-type modelling is being used in Europe index-based USLE and process-based equations for the detach- from the northern Scandinavia to the Southern Mediterranean or ment, transport, and deposition of soil particles (USDA, 2008). The eastern countries like Turkey and Hungary (e.g., (Bagarello & Ferro, RUSLE2 scheme allows to compute net erosion or deposition (mass/ 2004; Ferro & Porto, 1999; Ozsoy, Aksoy, Dirim, & Tumsavas, 2012; area) for each segment in which the overland flow path is divided, Podmanicky et al., 2011; Porto, 2016; Sivertun & Prange, 2003)). sediment load (mass/unit flow width) at the end of each segment Eventually USLE-type modelling was increased to continental and at the end of the overland flow path, and sediment charac- (Europe: Panagos et al. (2015d); China: Yue et al. (2016); Australia teristics at the detachment point and in the sediment load at the Teng et al. (2016)) and finally to global scale (Borrelli et al., 2017b; end of each segment (USDA, 2008). Among other structural im- Diodato et al., 2017; Doetterl et al., 2012; Ito, 2007; Van Oost et al., provements (USDA, 2018), RUSLE2 works on a daily time step and 2007; Yang, Kanae, Oki, Koike, & Musiake, 2003b). introduced the concept of erosivity density which resulted in sig- A database of studies was developed within the frame of this nificant improvements in the calculations and mapping of rainfall review using the Science Direct tool searching for the keywords erosivity (Nearing et al., 2017). Validation of the parameters of the USLE and RUSLE (from 1977 to July 2017) with specific focus on title, original RUSLE2 was achieved with the same 10,000 plot-years data year of publication, study area, journal, paper type and keywords. as the original USLE. Additional 451 studies were not included as they did not propose a As USLE has been frequently criticized of its empirical nature, specific study area or addressed theoretical issues of the model. The Ferro (2010) demonstrated that the original structure of USLE can author keywords of the articles have been processed by a word be theoretically obtained applying the dimensional analysis and the cloud application to produce the most common issues addressed by self-similarity theory (Barenblatt, 1979, 1987) using the same soil authors. In this process, we excluded self-explained keywords erosion representative variables and the reference condition (USLE, RUSLE, soil, erosion) and geographical locations. Besides the adopted by Wischmeier and Smith (1965). In other words, using the most common keywords (GIS, model, factor, map, risk) the highest factor scheme and the reference condition adopted by Wischmeier frequency was reached by keywords such as climate, geo-statistics, and Smith (1965), Ferro (2010) challenges the criticism of the policy, crop, terraces, 137Cs, planning and nutrient (see visualization empirical origin in claiming that “USLE is the subsequent logical in the graphical abstract). structure with respect to the variables used to simulate the physical The USLE/RUSLE model has been extensively used during the soil erosion process”. However, with this statement, we always last 40 years in 109 countries (Fig. 1). need to keep in mind, that USLE edriven modelling targets at the The largest number of publications with the application of USLE/ physical soil erosion process and will thus not be capable of RUSLE model has been found in the United States of America (274 simulating catchment sediment transport or yields. papers), China (218 papers), Brazil (88), Italy (87), India (67), Spain Today, USLE-type modelling has been further advanced to meet (66) Australia (50) and Turkey (43). In an analysis per continent, 519 numerous special requirements and specific needs. E.g., Bagarello, papers (33% of total) have study sites in 32 countries of Asia (Fig. 2). Di Stefano, Ferro, and Pampalone (2017) as well as Larson, In Europe (24% of the total number of studies) USLE/RUSLE is used Lindstrom, and Schumacher (1997) adapted USLE-type models for in 31 countries and in Northern America and the Caribbean in 13 event based soil erosion modelling. USLE-type modelling has also countries (22% of total number of studies). In Africa, 146 publica- been used in all kind of extreme ecosystem types and for various tions (9% of total) used USLE/RUSLE for estimating soil loss by water management scenarios, e.g. from volcanic soils in Chile with in 21 countries. The meta-analysis resulted also in a wide use of the Mediterranean climate (Stolpe, 2005) to the possible mitigation USLE/RUSLE model in East-Southern Asia (South Korea, Malaysia, impact of organic farming on soil erosion rates from mountainous Thailand and Indonesia), Central East Africa (Ethiopia, Kenya, monsoonal watersheds in South Korea (Arnhold et al., 2014) or the Nigeria) and the Mediterranean (Italy, Spain, Greece, Portugal, comparison of conventional with organic farming in northern Morocco, Tunisia, Algeria) (Fig. 1). Bavaria (Auerswald, Kainz, & Fiener, 2003). For a discussion of ad- The 1556 papers that estimated soil loss by water erosion using vancements in model development and extension of the model USLE/RUSLE at local/regional/national and continental scale have concept please see section 3. been published with an average rate of 39 papers per year during the period 1977e2017. The split into time windows of 5-years yielded a strongly increasing trend with the highest number of 2.1. A geographical and temporal meta-analysis of publications published articles (131) in 2016 (extrapolation from July 2017 to the addressing USLE/RUSLE models full year of 2017 is approximating 150 papers). A sharp increase in publication rates started in the late 90ties. The highest relative The USLE concept, originally developed for the US agricultural increase (76%) is noted in the period 2008-12 which is also ex- systems, has been adapted by scientists all over the world and pected to continue in the current period (2013-17). Estimated applied to datasets of different regions, countries and continents. published papers during the last 5 years will be more than 600 Examples would be Canada (e.g., the handbook for the application 208 C. Alewell et al. / International Soil and Water Conservation Research 7 (2019) 203e225 compared to the 414 papers published in 2008-2012 (Fig. 3). implemented in new modelling concepts considering runoff, USLE/RUSLE applications were published in a wide variety of sediment transport and/or sediment delivery. An implementation journals. The 1556 peer reviewed papers have been published in of the USLE where the rainfall erosivity is replaced by runoff vol- more than 500 Science Direct indexed journals with the highest ume (Modified USLE, MUSLE) resulted in a satisfying prediction of frequency in Catena (4.2%), followed by the Journal of Soil & Water measured sediment yield already in the mid-seventies and early Conservation (3.3%) and Land Degradation and Development eighties (Smith, Williams, Menzel, & Coleman, 1984; Williams, (2.6%). More than 83% of the papers are classified as research arti- 1975, pp. 244e252). The latter allows the equation to be applied cles, followed by conference papers (14%), book chapters (2%) and to individual storm events. Several variants of the MUSLE equation review papers (1%). Finally, 1195 articles (77% of the total) were exist and MUSLE was even integrated in GIS (e.g. in ArcMUSLE; cited with an average of 18.8 citations per article (publications with (Zhang, Degroote, Wolter, & Sugumaran, 2009). However, even no citations were not considered in this average rate). though runoff is considered, the restriction that USLE-modelled and observed data represent different parameters (e.g., gross and 3. Developments of USLE-type modelling with extended net erosion rates, respectively, please see box 1 and also section 5) modelling concept are not yet overcome. Another event based derivate of USLE is USLE-M, which also includes event runoff in R, but K-factor values Numerous advanced developments of USLE-type modelling are adjusted accordingly by multiplying them with the ratio of the have been implemented during the last decades (for selections total value of the EI30 index to the total value of the QREI30 index please see Table 1). A substantial change to the USLE concept were (Kinnell & Risse, 1998).

Table 1 Selected models based on the USLE concept or incorporating USLE parameters.

Model Full name Original references USLE components Additional features compared to USLE

RUSLE Revised Universal Soil Loss (Renard et al., 1991, 1997, USLE components to process-based auxiliary components (e.g., time-variable soil Equation pp. 1e404) simulate erosion erodibility, plant growth, residue management, residue decomposition, soil surface roughness, updated values for erosivity (R), new relationships for L and S factors, additional P factors RULSE2 Revised Universal Soil Loss USDA (2008) USLE components to Net erosion and deposition, hybrid soil erosion prediction Equation 2 simulate erosion with combination of the empirical, index-based USLE and process-based equations for the detachment, transport, and deposition of soil particles, calculation on a daily basis, biomass accounting in the C factor, Adding three dimensional components RUSLE3D Revised Universal Soil Loss (Desmet & Govers, 1996; USLE components but Considering the flow convergence of the upslope area thus Equation for Complex Terrain Mitasova et al., 1996; replacing the slope length improving the impact of concentrated flow on increased Mitasova & Mitas, 1999) with contribution of erosion upslope area USPED Unit Stream Power Erosion and Mitasova et al. (1996) USLE components to Considering 3-D topography and prediction erosion and Deposition simulate erosion deposition Adding factors to consider nutrient transport, management measures or replacement of C factors to improve biological dynamic CSLE Chinese Soil Loss Equation Liu et al. (2002) USLE but C and P factor C factor replaced by a biological conservation measures replaced factor (B), and engineering conservation measures factor (E) as well as a tillage conservation measures factor (T) PERFECT Productivity, Erosion and (Littleboy et al., 1992a, Sediment yield is simulated Water balance and runoff predictions, erosion and crop Runoff Functions to Evaluate 1992b) using MUSLE growth and crop yield; including sequences of plantings, Conservation Techniques harvests and stubble management during fallow G2 Geoland 2 erosion model (Karydas & Panagos, 2016; USLE but C and P factor C factor replaced by a vegetation factor retention combining Panagos, Christos, Cristiano, replaced. land use and fractional vegetation coverage; P factor & Ioannis, 2014a) replaced by landscape features factor quantifying the effect of obstacles to interrupt rainfall runoff. Model erosion at monthly step Considering runoff, sediment transport and delivery MUSLE Modified Universal Soil Loss (Smith et al., 1984; USLE components to prediction of sediment yield; simulation of individual storm Equation Williams, 1975, pp. 244 simulate erosion events e252) SWAT Soil and Water Assessment Tool Arnold et al. (1998) Sediment delivery based on Simulation of the hydrological water balance, then using the MUSLE runoff to simulate sediment dynamics WATEM/SEDEM Water and Tillage Erosion and (Van Oost et al., 2000a; Van USLE components to sediment production as well as sediment transport and Sediment Model Rompaey et al., 2001) simulate erosion pathways to derive ultimate source strengths, consideration of tillage erosion SEDD Sediment Delivery Distributed Ferro and Minacapilli USLE-type modelling to Gross erosion rates are coupled to a sediment delivery tool Model (1995) calculate gross erosion to estimate sediment transport and delivery rates Considering runoff and sediment dynamics as well as nutrients and/or pollutants CREAMS Chemical Runoff and Erosion (Knisel, 1980; Silburn & USLE components to Runoff modelling, erosion and sedimentation, gully erosion; from Agricultural Management Freebairn, 1992; Silburn & simulate erosion chemistry and target non-point source pollutants Systems model Loch, 1989) Individual storms/events as well as long term modelling AGNPS Agricultural Non-Point Source Young et al. (1989) USLE or RUSLE used to Runoff and nutrient modelling Model model soil erosion EPIC Erosion Productivity Impact Williams, Renard, and Dyke USLE used to model water in addition to water erosion, hydrology, nutrient dynamics, Calculator (1983) erosion plant growth, soil temperature tillage and economics are simulated with physically based modules C. Alewell et al. / International Soil and Water Conservation Research 7 (2019) 203e225 209

Box 1 erosion modes, namely WaTEM (Water and Tillage Erosion Model; Definition of soil erosion in Universal Soil Loss Equation (USLE)- (Van Oost et al., 2000a) and SEDEM (Sediment Delivery model, Van type algorithms Rompaey et al. (2001). A similar approach is followed with the Sediment Delivery Distributed Modell (SEDD) where USLE-type modelling is used to Soil erosion is, strictly speaking, the detachment of soil calculate spatially distributed gross erosion rates, which are then particles from a source site and its transport to some coupled to a sediment delivery tool to estimate net sediment depositional sink. As such, the processes of erosion and transport and delivery on catchment scale (Ferro & Minacapilli, deposition (often also addressed as sedimentation) are al- 1995; Ferro & Porto, 2000). ways coupled; both terms together might be described as The above discussion clearly demonstrates that USLE e type soil redistribution. models have long ago left the pure empirical modelling realm and Soil redistribution encompasses soil redistribution by water process-based elements have been incorporated into some of the and wind as well as mass transfer of soil particles (e.g. advanced model development approaches. However, the question landslides). Soil redistribution by water can further be arises how far this route can be followed if at the same time large differentiated into transport of single grains such as sheet scale modelling is envisaged, needing spatially discrete and inde- or interrill erosion, rill and gully erosion. In the modelling pendent units. As such, any modelling endeavor must, at the stage concepts of all USLE-type algorithms soil erosion is defined of planning, clearly define envisaged aims and available data, to be as: “Soil loss refers to the amount of sediment that reaches able to choose the best suited modelling concept (see also section 6, the end of a specified area on a hillslope that is experiencing Model Comparison). net loss of soil by water erosion. It is expressed as a mass of soil lost per unit area and time. Soil loss refers to net loss, 4. State of development of the USLE factors and it does not in any way include areas of the slope that experience net deposition over the long term.” (Nearing, As stated above, the origin of the USLE was strictly empirical Yin, Borrelli, & Polyakov, 2017). USLE-Type modelling with the calibration of the single factors from basic erosion plot does not address larger rills or gully erosion (linear struc- treatments (see section 2). However, this basic structure was sub- tures with a depth > 30 cm), but is restricted to sheet/interrill stantially modified over the last decades with many process-based and small rill erosion only (

A combination of RUSLE with sediment transport and delivery Accordingly, the calculation of rainfall erosivity (EI30) of a single processes has been achieved with the Water and Tillage Erosion event was based on the following equation: and Sediment Model (WaTEM/SEDEM (Van Oost, Govers, & Desmet, 2000a; Van Oost, Govers, Van Muysen, & Quine, 2000b; Xk ¼ v Van Rompaey et al., 2001)), which describes not only soil erosion EI30 er r (3) ¼ I30 risk, e.g. risk for sediment production, but also considers sediment r 1 transport and pathways to derive ultimate source strengths. The With vr the rainfall volume (mm) during the rth time period of a spatially distributed soil erosion and sediment delivery model rainfall event divided in k-parts. I30 is the maximum 30-min rainfall WaTEM/SEDEM is a combined version of two empirically-based soil intensity (mm h 1). Finally, the mean annual erosivity sums the 210 C. Alewell et al. / International Soil and Water Conservation Research 7 (2019) 203e225 erosive events during a given period and is divided by the number influence erodibility by water which are those that affect the of years. In RUSLE2 a slight modification of unit rainfall energy infiltration rate, permeability, and total water capacity, as well as function (Yin, Xie, Liu, & Nearing, 2015) was proposed but its use those that might influence the dispersion, splashing, abrasion, and has been mostly limited to the United States. Equation (3) is mostly transporting forces of the rainfall and runoff. The soil erodibility used in R-factor estimations elsewhere in the world. factor K was at first an empirical value determined from 20 years of The varying relationship between rainfall kinetic energy (KE) data on experimental plots from 23 major soil types from the US and intensity (I) have been addressed in a recent study by Angulo- kept fallow for at least 2 years with all other factors kept constant Martinez, Begueria, and Kysely (2016). In the absence of KE mea- (Wischmeier & Smith, 1965) and soil erodibility was regarded as surements (e.g., measuring raindrop size and velocity with dis- the amount of soil loss per unit erosive force with K equal to A/R. drometers) KE is usually estimated with empirical equations from Since the direct measurement of the K-value requires the estab- measured rainfall intensity. Angulo-Martinez et al. (2016) evalu- lishment and maintenance of natural runoff plots over lengthy, ated 14 different KE-I equations to estimate the 1 min KE and event expensive observation periods at various locations, numerous at- total KE and compared these results with 821 observed rainfall tempts have been made to simplify the technique and to establish events recorded by an optical disdrometer. They concluded that (i) estimators for soil erodibility calculation from readily available soil empirical relationships performed well, when complete events property data and standard profile description (Wischmeier, were considered but performed poorly for within-event variation Johnson, & Cross, 1971). (1 min resolution) and (ii) to use local measured data or local ki- The nomograph considering the four most crucial soil parame- netic energy equations. In a recent global R factor assessment 97.7% ters (particle-size, percent organic matter, soil structure and soil of the calculated R-factors stations were based on the original permeability) was suggested to derive the K-factor (Wischmeier equation of Brown and Foster (1987) and many of those were et al., 1971). Later, an approximation equation of the nomograph supported by local observation and validation (Panagos et al., was developed (Wischmeier & Smith, 1978, p. 58): 2017b). The impact of snow and snowmelt was not considered in R- K ¼ 2.77*105*M1,14*(12- a)þ 0.043 (b- 2) þ 0.033 (4- c) (4) factor equations but Schwertmann et al. (1987, pp. 1e64) suggested a very rough approximation in adding 1/10 of the precipitation for where M ¼ the particle-size parameter, which equals % silt (0.1- those month, where significant soil movement due to snow melt is 0.002 mm) times the quantity (with the quantity defined as 100 - % observed. Recent approaches subtracted the precipitation amount clay), a ¼ percent organic matter, b ¼ the soil-structure code used in that falls below 0 C (Meusburger, Steel, Panagos, Montanarella, & soil classification, and c ¼ the profile-permeability class. However, Alewell, 2012). The later approach considers that snowfall does Wischmeier and Smith (1978, p. 58) state that this equation did not not exert erosive energy, but neglects the subsequent process of describe the entire nomograph and differs for soils having high silt snow melt that may trigger soil erosion. content, low erodibility or high organic matter content. An open point of discussion regarding equation (3) is the rain (Auerswald, Fiener, Martin, & Elhaus, 2014) developed a set of drop size. In tropical areas such as Ethiopian highlands the algo- equations mimicking the original K factor nomograph. The de- rithm for unit rainfall energy (equation (2)) may underestimate the viations from the original equation were tested on 19055 soils rainfall erosivity due to large drop sizes in those areas (Nyssen et al., obtained during soil surveys throughout Germany. The set of 2005). equations proposed resulted in deviations compared to the classical In the erosion studies applying USLE/RUSLE in late 1990s and K factor equation in more than 50% of all cases. Thus, the authors early 2000s, many scientists have either applied simplistic multi- recommend using this set of equations, which describe the plications for estimating erosivity (R-factor ¼ 1.3 * precipitation) or nomograph even beyond the limitations of the classical equation, have proposed empirical erosivity equations such as the Fournier but they restrict that predictions may be far from “perfect” since Index (FI) and its modification (MFI) by Arnoldus (1980). The lack of these equations are ignoring seasonality or interaction with stations with sub-hourly data for long-periods and measured climate. Generally, the nomograph should only be used if there is rainfall intensity forced scientists to develop simple empirical not a locally derived relationship or measured data that would also equations correlating R-factor with available daily/monthly/annu- account for seasonal changes in the K-factor. ally rainfall data (Aronica & Ferro, 1997; Bonilla & Vidal, 2011; Seasonal effects on the K-Factor due to freezing and thawing Diodato, Knight, & Bellocchi, 2013; Loureiro & Coutinho, 2001). processes (accompanied by effects on shear strength) or on Recently, the availability of high temporal resolution data compaction (e.g., by rainfall during winter/fall or by life stock (hourly, sub-hourly) for long periods and the development of geo- trampling during summer), and subsequent release processes have statistical algorithms for spatial interpolations contributed to been discussed (Renard et al., 1991; Renard & Ferreira, 1993). regional, national and continental erosivity maps in Spain (Angulo- However, none of these seasonal processes and effects can easily be Martinez, Lopez-Vicente, Vicente-Serrano, & Begueria, 2009), captured for inclusion in the K factor. Kinnell (2010) reviewed Switzerland (Meusburger et al., 2012; Schmidt, Alewell, Panagos, & different approaches to assess the seasonality of the K-factor. Meusburger, 2016), Italy (Borrelli, Diodato, & Panagos, 2016a), Ko- However, none of these approaches include the hardly measurable rea (Risal et al., 2016), Iran (Sadeghi, Zabihi, Vafakhah, & Hazbavi, influencing interactions and effects (e.g., climate influences and 2017), Brazil (Oliveira, Wendland, & Nearing, 2013) and the Euro- seasonality of freeze-thaw, compaction by life stock trampling, pean Union (Panagos et al., 2015a). weathering, human management activities,) simultaneously for a A group of scientists, working with R-factor data all over the proper process-oriented modeling (Leitinger, Tasser, Newesely, world, has collected data on rainfall erosivity from 3,625 meteo- Obojes, & Tappeiner, 2010; Pineiro, Paruelo, Oesterheld, & rological stations in 63 countries and established the Global Rainfall Jobbagy, 2010; Vannoppen, Vanmaercke, De Baets, & Poesen, Erosivity Database (GloReDa) developing a global erosivity map 2015). Furthermore, the divergence of seasonal K-factors to an (Panagos et al., 2017a). annual K-factor is poorly discussed in the literature (e.g.(Wall, Dickinson, Rudra, & Coote, 1988)). In the RUSLE2 User's Reference 4.2. K factor Guide (Foster et al., 2008) it is even stated that no statistical evidence exists for an inconsistency of soil erodibility over time. The K factor is statistically related to the soil properties that Another important parameter affecting the erodibility is stone C. Alewell et al. / International Soil and Water Conservation Research 7 (2019) 203e225 211 content. Originally the effect of surface stones was not included in the factor slope length (L) and slope steepness (S). The original the K factor equation, instead (Wischmeier & Smith, 1978, p. 58) value of S was derived empirically as S ¼ (0.43 þ 0.30s þ 0.043s2) suggested to consider the effect of surface stones in the C factor. The *6.613 with s ¼ slope gradient in percent (Smith & Wischmeier, effect of stone cover on soil loss is scale dependent. At the meso- 1957; Wischmeier & Smith, 1965). Previous parametrizations scale stones and rock fragments on the soil surface may have (Musgrave, 1947; Smith & Whitt, 1948; Zingg, 1940b) and many ambivalent effects while at the micro and macroscale a decrease in posteriori derivatives (Fig. 4) of the original empirical equation sediment yield is predominant (Poesen, Torri, & Bunte, 1994). For with different curve shapes of S against slope angle have been the macroplot scale, Poesen et al. (1994) developed a soil erodibility developed since then (Foster, 1982, pp. 17e35; Liu, Nearing, & Risse, reduction factor expressed as an exponential decay function based 1994; McCool, Brown, Foster, Mutchler, & Meyer, 1987; Nearing, on a world-wide compilation of experimental field data. Imple- 1997; Renard et al., 1997, pp. 1e404). All equations allow the menting the effect of stoniness, reductions of 20-40% of the soil assessment of the S-factor up to a slope angle of 100%, even though erodibility were found to be particularly significant in the Medi- empirical measurements only exist up to a slope angle of 55% (Liu terranean (Panagos, Meusburger, Ballabio, Borrelli, & Alewell, et al., 1994). According to our knowledge only one study explored 2014b). soil loss up to slope angles of 90% using rainfall experiments Today, soil surveys together with large scale data bases such as (Tresch, 2014, p. 132). ISRIC SoilGrids database at a 250 m spatial resolution (Hengl et al., Wischmeier and Smith (1978) defined the slope length (L) as: 2014) and remote sensing give new opportunities for mapping soil “the distance from the point of origin of the surface flow to the erodibility in space and time (Panagos et al., 2014b; Wang, Fang, point where each slope gradient (S) decreases enough for the Teng, & Yu, 2016). Several studies already explored the suitability beginning of deposition or when the flow comes to concentrate in a of hyper-spectral reflectance from soil surfaces to assess the defined channel”. As such, the L factor is defined as the ratio of field chemical and physical properties of soil (Ben-Dor & Banin, 1995; soil loss to that from a 22 m slope and the value of L may be Luleva, van der Werff, Jetten, & van der Meer, 2011; Rossel & expressed as (l/22.1)m, where l is field slope length in meters and Behrens, 2010; Shepherd & Walsh, 2002; Stenberg & Rossel, m is a factor that varies with slope gradient in the USLE and the 2010; Wang et al., 2016). ratio of rill to interrill erosion in the RUSLE. It has to be noted, that the underlying pedotransfer functions to Soil loss is much less sensitive to changes in slope length than to derive soil erodibility estimations are essentially all based on changes in slope steepness (McCool et al., 1987). Most measured American soil-erosion databases, which raises the obvious question slope lengths are less than 120 m and generally do not exceed 300 about their applicability in other geographical locations. Conse- m (McCool, Foster, & Weesies, 1997). Nevertheless, it has to be quently several studies aimed to determine the best suited soil stated, that little research has been conducted to assess the relation erodibility estimation for specific scales, regions or conditions between slope length and soil loss. Moreover, the existing findings (Hussein, Kariem, & Othman, 2007; Romkens,€ Poesen, & Wang, are not ambiguous as with increasing slope length soil erosion was 1988; Torri, Poesen, & Borselli, 1997; Wang, Zheng, & Wang, observed to decrease (Joel, Messing, Seguel, & Casanova, 2002; 2012; Wawer & Nowocien, 2006; Zhang et al., 2004a, 2008). Kara, S¸ensoy, & Bolat, 2010; van de Giesen, Stomph, & de Ridder, However, hardly any study came up with an ultimate conclusion on 2005; Xu et al., 2009; Yair & Raz-Yassif, 2004), increase (Rejman the suitability of different erodibility estimations, because of a lack & Brodowski, 2005; Wischmeier & Smith, 1958; Zingg, 1940a) or of measured plot data (Wang et al., 2016). As such, some new having no remarkable effect (Wischmeier & Smith, 1958). studies, and databases incorporating measured K-values from long- The USLE was firstly developed to predict soil loss on uniform term observations of natural runoff plots are of utmost interest. slopes and fields (Wischmeier, 1976; Wischmeier and Smith, 1965, 1978, p. 58). But already in 1974, Foster and Wischmeier (1974) developed a method to divide an irregular slope into a number of 4.3. LS factor uniform segments and as such accounting for the effect of the shape of the slope on soil loss opening the doors to watershed scale In USLE models the effect of topography is usually considered by

Fig. 4. Dependency of the S factor on inclination (slope steepness) for different parametrizations. 212 C. Alewell et al. / International Soil and Water Conservation Research 7 (2019) 203e225

USLE applications (Griffin, Beasley, Fletcher, & Foster, 1988; management measures (Wischmeier & Smith, 1965). In other Williams & Berndt, 1977). words, it reflects the effects of biomass cover and soil-disturbing However, these manual methods were demanding and by activities. It is expressed as the ratio of soil loss from land under replacing plots with watersheds the slope length loses significance specified conditions to the corresponding loss from clean-tilled, for predicting overland flow and soil loss, since in reality, surface continuous fallow (given by the product of R$K$L$S). The soil loss water flows converge and diverge across the landscape (Desmet & ratio (SLR) is the ratio of the soil loss from the non-bare fallow Govers, 1996). Thus, it was proposed that the unit contributing area surface and the soil loss from the bare fallow surface over a given which is the upslope drainage area per unit of contour length period of time. Accordingly, it is inversely proportional to the soil should substitute slope length (Desmet & Govers, 1996). The latter surface cover which intercepts raindrops and hinders surface approach did not only account for segment length or grid cell size, runoff by slowing it down. but also for the position of the segment or cell within the landscape As nearly all crops can be grown continuously or in rotations and consequently the LS factor was not one-dimensional anymore with varying sequences, the soil loss ratios can vary in cropped land (Kinnell, 2008; Zhang et al., 2013). The first physically based slope- over time as they are a function of canopy, ground cover, roughness, length factor was developed by Moore and Burch (1986). They soil biomass and consolidation change. The C factor depends on the concluded that the theoretical and empirical slope-length factors particular stage of vegetation and vegetation cover development at are equivalent and thus the USLE slope-length factor is indeed a the time the rain event occurs, but also on the prior land use con- measure of the sediment transport capacity of runoff from the ditions. For instance, plant residues can be removed or left on the landscape. However, the USLE slope-length factor fails to fully ac- field, incorporated near the surface or ploughed under, chopped or count for the hydrological processes that affect runoff and erosion omitted by the harvesting operation. All these different measures (Moore & Burch, 1986). can have different effects on soil loss. Originally five crop stages A multiple flow algorithm is implemented by MUSLE, RUSLE, were defined including rough fallow, seedling, establishment, EPIC and G2 models (Karydas et al., 2014). Winchell, Jackson, growing and maturing crop and residue or stubble. From 10,000 Wadley, and Srinivasan (2008) further improved the method and plot years of runoff and soil-loss data assembled from 47 research compared several variations of the GIS approach. A second algo- stations in 24 states of the US empirical tables were created taking rithm was associated with the development of the USPED model by the crop stage, vegetation type, management measures and their Mitasova et al. (1996). The greatest limitation of all these methods timing, and rainfall intensity EI in the specific periods into is (i) the absence of an algorithm for predicting deposition consideration (Wischmeier & Smith, 1965). In RUSLE, the compu- (Mitasova, Hofierka, Zlocha, & Iverson, 1997; Winchell et al., 2008) tation of the C factor follows a scheme proposed by Laflen, Foster, and ii) the neglect of the variability of infiltration along the flow and Onstad (1985) and (Weltz, Renard, & Simanton, 1987, pp. path. The flow path and cumulative cell length-based method (FCL) 104e111) and tables of SLR are no longer provided. Five sub-factors proposed by Dunn and Hickey (1998) and Hickey (2000) and are multiplied together to yield soil loss ratios (prior-land-use advanced by Van Remortel, Hamilton, and Hickey (2001) and Van (PLU), canopy-cover (CC), surface-cover (SC), surface-roughness Remortel, Maichle, and Hickey (2004) who tried to overcome this (SR) and soil-moisture (SM)). Individual values of SLR are disadvantage. In 2013, Zhang et al. (2013) presented an algorithm computed in time intervals within which the five sub-factors that merges both approaches to account for flow convergence remain constant. Subsequently, the SLR values are weighted by based on the contributing area as well as slope cutoff conditions the fraction of rainfall erosivity (EI) of the coinciding time period (i) and thus accomplished an important step towards net erosion and combined to achieve an overall C factor value using estimation. To conclude, increased availability of gridded digital elevation Xn C ¼ ðSLR $ EI Þ = EI (5) data, often referred to as DEMs, improved the estimation of the LS i i t i¼1 factor. The availability of high resolution DEM (Europe 25m but in certain countries much higher (e.g., Switzerland DEM 2m)) has Although it is possible to compute C values for a wide set of outpaced method development of flow algorithms (Seibert & tillage techniques and crop rotations within RUSLE, a large number McGlynn, 2007), which need to define the contributing area. of input data are required (Gabriels, Ghekiere, Schiettecatte, & Currently, four algorithms exist to calculate the upslope contrib- Rottiers, 2003). With the advent of GIS-based modelling and the uting area: 1) single-direction flow algorithm, D8 (e.g., O'Callaghan ambitions to assess the potential impact on soil erosion over larger and Mark (1984), p. 2) the multiple flow direction algorithm (MD8, areas, the C factor, as with most other factors within RUSLE, has Quinn, Beven, Chevallier, and Planchon (1991)), 3) infinite possible experienced a process of computation transformation and simpli- single-direction flow pathways (D∞), and 4) triangular multiple fication. At catchment- and regional-scale, most of the input pa- flow direction algorithm (MD∞, Seibert and McGlynn (2007)). Ac- rameters for the RUSLE C sub-factors became hard, if not impossible cording to our knowledge comparison of the different flow algo- to assess and quantify given that the model simulation is a specific rithm to mapped erosion features is still missing. field plot. In recent years, the methods most commonly employed Another main research gap related to the availability of high to compute the C factor and set large-scale RUSLE-based modelling resolution DEM is the dependence of USLE-type estimates on grid are simple attributions of literature C-factor values to land use size of the DEM. Soil erosion estimates were found to decrease with maps without further land sub-classification (Bakker et al. (2008); decreasing resolution (Mondal et al., 2017). Consequently erosion Marker€ et al. (2008), among others) or satellite remote sensing estimates may be biased depending on the resolution of the DEM. approaches ((Borrelli, Panagos, Marker,€ Modugno, & Schütt, 2017; To investigate resolution dependence seems even more pending in De Jong, 1994; Schonbrodt,€ Saumer, Behrens, Seeber, & Scholten, the light of drone derived DEM with high horizontal resolutions of 2010; Van der Knijff, Jones, & Montanarella, 1999), among others). 5 cm (Peralta et al., 2017). With increasing calls for more knowledge on soil erosion on macro- regional scale, research on C factor based on remote sensing tech- 4.4. C factor niques has become a widely applied method across the globe (Vrieling, 2006; Zhang, Zhang, Liu, Qiao, & Hu, 2011). While these The cover-management factor (C) in USLE-type equations applications often inadequately represent the management aspect measures the combined effect of all interrelated cover and of the C factor, the gained insights nevertheless provide a valuable C. Alewell et al. / International Soil and Water Conservation Research 7 (2019) 203e225 213 contribute to improve the understanding of the soil erosion po- certain slope. Furthermore, the impact of stone walls and grass tential of a given region, especially when considering monthly or margins was modelled using the more than 226,000 observations sub-monthly land use/land cover change. Most recent, regional- from the Land use/cover area frame statistical survey (LUCAS) scale C factor computation approaches have moved towards carried out in 2012 in the European Union (Panagos et al., 2015c). spatially and spatiotemporally descriptions using vegetation based indices for land use classifications (Meusburger, Konz, Schaub, & 5. Model validation Alewell, 2010b; Moller,€ Gerstmann, & Glaßer,€ 2014, pp. 5072e5075; Panagos et al., 2014c; Schmidt, Alewell, & Meusburger, 5.1. General thoughts on model validation 2018; Schonbrodt€ et al., 2010; Teng et al., 2016) or weighted average operations combining crop statistics with remote sensing and GIS ‘Validation’ refers to the testing of the model output to confirm modelling techniques (Borrelli, Panagos, Langhammer, Apostol, & the results that should be produced to reality (Fishman & Kiviat, Schütt, 2016b; Panagos et al., 2015b). A geo-statistical approach 1968). Ecosystem modelers suggested that a ‘valid’ model of a based on weighted average operations was recently introduced to biological system would necessarily be an exact copy of the system thoroughly incorporate the extent, types, spatial distribution of (Oreskes, Shraderfrechette, & Belitz, 1994). Alewell and global croplands, and the effects of the different regional cropping Manderscheid (1998) argued that a direct validation of a model systems into a global RUSLE-based soil erosion model (Diodato would prove that this model is an exact description of the modeled et al., 2017). In these recent approaches, the C factor reflects the system, the model output and the system's output matching under relative effectiveness of the soil and crop management systems in a any circumstance. As such, model validation would only be possible given region in terms of their ability to reduce soil loss. for thermodynamically closed systems (Oreskes et al., 1994), where all parameters and processes are described and measureable (e.g. 4.5. P factor within a mathematical proof). Following this line of thinking, a model describing a thermodynamically open system cannot be As in the case of the C factor, the support practice factor (P) in validated. the RUSLE-type equations is based on the concept of deviation from A very different definition of validation derives from the a standard. It expresses the ratio of soil loss in a field with specific meaning of the Latin word validus (¼ strong, healthy, powerful) and support practice to the loss of soil under conditions of straight-row was used by Martin (1996). In that sense, a model is validated, if it farming up and down the slope. The support practices affecting the has been proven itself as a strong political or economic tool (Martin, erosion process by modifying the amount and dynamics of runoff 1996), e.g. putting a legislation into action or setting up a man- are considered in the P factor. For cultivated land supporting agement plan. Foster, Yoder, Weesies, and Toy (2001), in presenting practices generally considered as P factor are contour tillage/ the RUSLE2 concept, also connects model success to its usefulness, planting, strip-cropping, terrace farming systems and stabilized when stating that a model which provides perfect erosion esti- water ways because improved tillage practices such as sod-based mates but which somehow leads to a poor conservation planning rotations, fertility treatments, and greater quantities of crop resi- decision has failed to meet its objectives. Indubitable, we have left dues left on the field are already part of the C factor (Wischmeier & the realm of natural science with these lines of thinking. However, Smith, 1965). The overall P factor is the product of the individual following the recent critical debate of soil erosion modelling, the support practices. In the original paper of Wischmeier and Smith definition of Martin (1996) or the logic of Foster et al. (2001) has (1965) empirical P values to consider contouring were provided been implicitly applied. E.g., one of the main concluding arguments but the researchers recommended to consult further expert to dismiss the Pan-European modelling of Panagos et al. (2015d) judgement to adapt the P value for specific field conditions espe- was, that “erosion rate figures can be used by politicians/policy cially when small gullies or rills exist or one of the above mentioned makers to advocate measures that may be misguided” or that measures was adapted. Within RUSLE, a list of P values obtained “modelled information can be used to suggest that policies to from experimental data supplemented by analytical experiments is mitigate erosion are successfully tackling soil loss across Europe” provided, offering a wide range of support practice conditions. (Evans & Boardman, 2016a; Fiener & Auerswald, 2016). Nevertheless, according to (McCool et al., 1987) the values of the P If we accept models as tools in natural science, we can use them factor are the least reliable of the USLE/RUSLE factors as the effec- to test hypothesis on process understanding, relative differences tiveness observed in field studies conducted on given slopes between systems or potential development over time. Unlike a showed wide ranges of reductions. mathematical proof, a scientific theory or hypothesis is empirical, Typical P values range from about 0.2 for reverse-slope bench and is always open to falsification if new evidence is presented. A terraces to 1.0 where there are no erosion control practices given model which describes a specific system significantly better (Wischmeier & Smith, 1978, p. 58). The lower the P value, the more might be declared the ‘valid’ model compared to another model successfully the support practice or the combination thereof pro- which might be rejected and the term ‘valid’ in this sense is then motes the deposition of soil particles. From a conservation planning used that any model that could not be proven invalid would be a perspective, the deposition of the soil that is close to the source is valid model for the system (Alewell & Manderscheid, 1998). preferable but the effectiveness of the support practice depends on There is general agreement in modelling that the calibration the characteristics of the agricultural area (e.g., the slope gradient). dataset must be independent from any dataset which is used later With the transition from field-to GIS-based modelling, the lack of to validate the model, and that if the same dataset is used for both it spatial information to compute the P factor implies that in the vast should be no surprise that the model is a perfect predictor (Nearing, majority of the catchment-to regional-scale GIS-based models the 2004). Split sample approaches, in which the available data is crucial effect of support practices on soil erosion has been omitted. separated into a calibration set and a separate validation set, are However, recent large-scale studies proposed alternative ap- usually the solution to this problem. proaches that are able to provide relative estimations of the P factor For the open systems described in ecosystem modelling it was and assess the possible effects of conservation policy (Panagos suggested that a model can only be validated, if it was not cali- et al., 2015c; Yue et al., 2016). Panagos et al. (2015c) considers the brated (Sverdrup, Warfvinge, Blake, & Goulding, 1995), implying latest policy developments in the Common Agricultural Policy, and that measured input parameters implemented in the model will be applies the rules set by Member States for contour farming over a directly compared to measured target values without any inverse 214 C. Alewell et al. / International Soil and Water Conservation Research 7 (2019) 203e225 process step. Interestingly, soil erosion modelling using empirical multi-year erosion totals were used. Values of the coefficient of models, as much as it has been criticized in the past, commonly variability ranged from nearly 150% for a soil loss of 1 Mg ha 1yr 1 follows this most rigorous of all approaches: e.g., USLE-type model to as low as 18% or less for soil-loss values greater than 100 Mg applications derive the input parameters from measured data and, ha 1yr 1 (Nearing et al., 1999). without the process of calibration, compare modelled soil erosion As such, trying to validate USLE-type modelling with measured maps directly with measured data. data, needs to take the high variability of soil erosion measurements into account. In addition, it has been noted all throughout 5.2. Validating USLE-type models with measured data the history of using the USLE that there is a striking lack of data to rigorously determine the single factors for all needed regions, sit- One major obstacle in validating the (R)USLE with measured uations and scenarios (Auerswald et al., 2003; Wischmeier and data is that (R)USLE provides gross erosion rates while most Smith, 1965, 1978, p. 58). Many plot studies measuring soil monitoring programs and measurements (except for experimental erosion rates investigated bare fallow plots as a baseline reference approaches) provide net erosion rates (e.g. on-site measurements which hardly ever exist in reality. While the focus of measurements as the budget of erosion and deposition of sediments or even off- is towards medium sloped (e.g. > 3% < 18%) arable soils, flat areas or site sediment yields in waters). Thus, and as long as USLE is not steeper slopes >20% under grassland or forest used are hardly ever connected to sediment transport models (which is mostly not investigated (Bonta & Sutton,1983; Borst & McCall, 1945; Fan,1987; operational on large scales, see section 3), modelled and observed Kilinc & Richardson, 1973; Liu et al., 1994; Tresch, 2014, p. 132; data represent different fluxes. Or, as (Trimble & Crosson, 2000) Zhang, Wang, & Yang, 2014). Furthermore, measurement plots stated, USLE only presumes to predict the amount of soil moved on were either located in areas with relatively large rainfall erosivity or a field (e.g., gross erosion), not necessarily the amount of soil experimental treatment with high irrigation rates and intensities moved from a field (e.g., net erosion). As such, a RUSLE modelling was used. So, all in all, there is bias in the available measurement was not successfully validated when compared to sediment yields data towards regions or situations with conditions prone to erosion of 454 small catchments with various land covers and uses across (Auerswald et al., 2003; Wischmeier and Smith, 1965, 1978, p. 58). the United States, Puerto Rico, U.S. Virign Island and Guam (Tan Nevertheless, ultimately we need a validation of soil erosion et al., 2018). Instead, models that use the power function of modelling to be able to attribute the relative importance of sedi- runoff, shear stress or stream power such as the Morgan, the tRIBS- ment source strength to specific kinds of land use and management Erosion and the GUEST model were more successful to target as well as other triggering factors (e.g. climatic events, topography, catchment sediment yields (Tan et al., 2018). and vegetation cover (Alewell, Birkholz, Meusburger, Wildhaber, & Nearing (2004) state that observations will always be closer to Mabit, 2016; Meusburger, Banninger, & Alewell, 2010a; the truth than modelling and must remain the most important Meusburger et al., 2010b; Panagos et al., 2015d; Scheurer, Alewell, component of scientific investigation. However, Garcia-Ruiz et al. Banninger, & Burkhardt-Holm, 2009; Schindler Wildhaber, (2015) observed a correlation of measured erosion rates with size Michel, Burkhardt-Holm, Baenninger, & Alewell, 2012). As such, of the study area, measurement method and duration of experi- erosion models have been compared extensively to measured data ment or observation period in a worldwide meta-analysis of 4000 sets. sites. Furthermore, there are many regions of the world being un- Risse et al. (1993) applied the USLE to 1700 plot years of data derrepresented and erosion measurements lack long term obser- from 208 natural runoff plots. Annual values of measured soil loss vations (Garcia-Ruiz et al., 2015). Thus, it has to be taken into averaged 35.1 Mg ha 1 with an average magnitude of prediction account that observations, especially due to the highly heteroge- error of 21.3 Mg ha 1, or approximately 60% of the mean (Risse neous nature of soil erosion and the tendency of erosion mea- et al., 1993). Rapp (1994, pp. 1e95) followed the latter approach surements being very prone to errors (Garcia-Ruiz et al., 2015), and determined the error associated with RUSLE from 21 US sites might also be biased in regional distribution or time periods chosen representing 1704 years of measurements from 206 plots. The by the researchers or due to high systematic errors or random average annual magnitude of error was 11.7 Mg ha 1 and the model uncertainty (Rüttimann, Schaub, Prasuhn, and Rüegg (1995), see efficiency coefficient was 0.73, while the prediction of the 1638 section 5.2). As such, quality and potential biases of modelled and individual annual erosion values had an average magnitude of error measured data has to be equally scrutinized, when model valida- of 20.8 Mg ha 1 and a model efficiency of 0.58 (Rapp, 1994, pp. tion with measured data is aimed at. 1e95). Zhang, Nearing, Risse, and McGregor (1996b) applied the One reason of high modelling uncertainty frequently addressed Water Erosion Prediction Project (WEPP) computer simulation in the literature is certainly that erosion measurements itself are model to 290 annual values and obtained an average of 21.8 Mg connected to considerable uncertainty. Rüttimann et al. (1995) re- ha 1for the measured soil loss, with an average magnitude of ported a statistical analysis of data from four sites, each with five to prediction error of 13.4 Mg ha 1, or approximately 61% of the mean. six reported treatments. Each treatment had three replications. In all cases the relative errors tended to be greater for the lower Reported coefficients of variation of soil loss within treatments soil-loss values with an over prediction on plots with low erosion ranged from 3.4% to 173.2%, with the ploughed sites with higher soil rates and under prediction on plots with high erosion rates. Even losses having a tendency to lower coefficient of variation between though deviations from modelled and measured data might seem 3.4 and 71.4%. Work carried out by Wendt, Alberts, and Hjelmfelt substantial in absolute numbers, the deviation between modelled (1986) using 40 erosion plots, all with the same treatment in and observed erosion rates is not larger than the variability within Minnesota found that coefficients of variation for the 25 storms measured data. ranged from 18% to 91%. The larger erosion events again showed Kinnell (2010) in evaluating data from Rosewell (1993) also less variability. Very little of the variability could be attributed to concludes that even though the RUSLE works well for a number of measured plot properties, and the plots did not perform in the agricultural conditions in New South Wales, Australia, on a wider same manner relative to each other in subsequent events (Wendt geographic area and larger variety of agricultural systems, there is a et al., 1986). Nearing, Govers, and Norton (1999) studied erosion tendency for the USLE and the RUSLE to over-predict low average variability using data from replicated soil-loss plots from the Uni- annual soil losses and to under-predict high average annual soil versal Soil Loss Equation (USLE) database. Data from replicated plot losses. Zhang, Nearing, Risse, and McGregor (1996c) also discussed pairs for 2061 storms, 797 annual erosion measurements and 53 an over prediction of small soil losses as well an under-prediction of C. Alewell et al. / International Soil and Water Conservation Research 7 (2019) 203e225 215

large soil losses for USLE approaches as well as the WEPP model. ha 1yr 1 (Meusburger et al., 2013). Thus, even though the FRN On-site erosion measurements are not available on large scale, method also determines net erosion rates while USLE-type but Cerdan et al. (2010) compiled an extensive database from modelling results in gross erosion estimates, the promising re- literature on short to medium-term erosion rates as measured on sults of these studies might corroborate the assumptions that erosion plots in Europe under natural rainfall conditions. Panagos comparisons may be applicable over long time periods of several et al. (2014c) compared the soil erosion estimates of this decades which are evaluated by the FRN methods. However, compiled plot data base (Cerdan et al., 2010) with modelled values Haciyakupoglu, Hizal, Gokbulak, and Kiziltas (2010) compared 137Cs from the Pan-European Soil Erosion Risk Assessment (PESERA) based erosion rates to USLE assessment in the Turkish Omerli model (Kirkby et al., 2008) and the European Environment Infor- catchment and found substantial deviations between the two mation and Observation Network for soil (EIONET-SOIL) database of methods (higher ULSE rates by a factor of 2-3). The latter over- Europe (which are regional to national USLE-type model applica- estimation of soil losses by USLE compared to 137Cs based erosion tions). The mean values of soil loss reported by the national in- rates was also noted by Warren et al. (2005), but was overcome stitutes (EIONET-SOIL) are larger than the PESERA estimates, with when USPED was applied, which considers net erosion rates (see the main differences being for sloping land (>2) and for the land section 3) and might thus be better comparable to 137Cs based cover type forest and heterogeneous agricultural land cover erosion rates. (Panagos et al., 2014c). A graphical illustration of this comparison Regardless of the latter promising FRN results the question may results in a deviation of both modelling approaches from the plot be asked or, nevertheless, has to be asked: are the model pre- based measurements at the high end of modelled soil losses (Fig. 5). dictions ‘good enough’? And what is actually ’good enough‘ or what Cerdan et al. (2010) warned of extrapolations from local mea- are the reference values considering the high uncertainty and surements to regions and in addition the comparison might illus- biases of measured data? Even though Bagarello et al. (2012) con- trate the above discussed principally different nature of net erosion cludes that soil loss estimates are considered reasonably accurate rates based on plot measurements versus larger scale modelled for most practical purposes, especially for wide spatial scale ap- gross rates. Furthermore, there is no straightforward conclusion plications, when the errors of the predictions do not exceed a factor from these comparisons of modelled and measured data: are the of two or three, erosion modelling remains assailable as long as models invalidated by the measured data or is the measured data there is no general agreement on the ’good enough’. biased and/or not reflecting large spatial scale and long temporal Nearing (2004) concluded that model validation is not just a average situations? matter of comparing measured to modelled data, one must also ask As the direct measurement of soil erosion on plots implies long the question: “How variable is nature?” We would like to add, that term monitoring all year round, the indirect assessment using in bidding farewell to the idea of accurately predicting absolute fallout radionuclides (FRN) as tracers of soil erosion has been used. values with models but rather concentrating on the prediction of Meusburger et al. (2010b) with an adapted C factor implemented relative differences, trends over times and systems reactions to from fractional vegetation cover map derived from QuickBird im- processes and management practices, we can use models as tools to agery, validated USLE derived erosion rates for hot spots of erosion learn about the modelled systems and their reactions. In this con- in Alpine grasslands against FRN derived erosion rates (20.1 Mg ceptual approach, modelling in general and large-scale modelling ha 1yr 1 versus 16 Mg ha 1yr 1 for 137Cs based versus modelled specifically will per se not aim at an accurate prediction of point rates, respectively). RUSLE was also successfully validated at sites of measurements, but will test hypotheses on process understanding, South Korea, with the 137Cs method ranging from 0.9 Mg ha 1yr 1 relative spatial and temporal variations, scenario development and to 7 Mg ha 1yr 1 and RUSLE (considering the steep slopes of up to controlling factors (Oreskes et al., 1994). In an agreement with the 40 and the erosive monsoon events (R factor of 6600 latter and the above discussion, many modelers accept that full MJmmha 1 h 1 yr 1)) yielding 0.02 Mg ha 1yr 1 to 5.1 Mg model validation is a logical impossibility and Morton and Su'arez (2001), Refsgaard and Storm (1996) as well as Senarath, Ogden, Downer, and Sharif (2000) suggest that in most practical contexts the term ‘model’ should be thought of as being synonymous with ‘theory’ or ‘hypothesis’, with the added implication that they are being confronted and evaluated with data. An interesting example of such ’hypothesis testing‘ is a study in northern Bavaria where a comparison of conventional with organic farming practices found measured soil losses to be about an order of magnitude lower than RUSLE based values due to best management practices adopted on both farms (Auerswald et al., 2003). However, modelled and measured values pointed to the same conclusion of organic agriculture resulting in 15% less erosion on arable land despite its localization in areas with substantially higher erosion susceptibility (Auerswald et al., 2003). Garcia-Ruiz et al. (2015) concluded from a global meta-analysis that “Geomorphologists should be more interested in processes, causes, internal relationships, and the role of stream channels than in obtaining absolute erosion rates, the interpretation of which can lead to numerous errors”. However, as USLE-type modelling is often used by land scape managers and stake holders, it is important to communicate that these modelling endeavors should not be used to predict absolute erosion rates (e.g. for comparison to legal limit values) but rather to optimize catchment management and land use in delineating best practice Fig. 5. Erosion estimates from plot based measurements (Cerdan et al., 2010) versus modelled erosion rates from PESERA (Kirkby et al., 2008), USLE-type models (EIONET techniques. Data Base; (Panagos et al., 2014c)) and RUSLE2015 (Panagos et al., 2015d). With the acceptance that at least at large scales we should not 216 C. Alewell et al. / International Soil and Water Conservation Research 7 (2019) 203e225 be solely focused on an accurate prediction of absolute values but parameters (e.g. hydrological as well as vegetation parameters) did rather see models as a tool to test hypotheses of relative differences not invoke a sensitive response from the model. A similar pattern between systems, trends over time, systems reactions to driving can be observed for WEPP in the work of Brazier, Beven, Freer, and factors and management practices, we should, however, not Rowan (2000). Quinton (2013), in discussing the uncertainty of generally exclude the use of models to assess soil erosion magni- complex erosion models like EUROSEM and WEPP, suggests that tudes at least at smaller scales with the aim to assess sustainability this uncertainty is reason enough to describe the erosion system of management. An example would be a study by Alewell, Egli, and with simpler models than those that are currently in vogue. One þ Meusburger (2015) where 137Cs as well as 239 240Pu derived soil reason for the popularity of USLE-type modelling is certainly that it erosion estimates were compared with USLE estimates from is a good compromise between applicability in terms of required Meusburger et al. (2010b) and confronted with soil formation rates input data and relatively good reliability of obtainable soil loss es- in Swiss Alpine grasslands at plot scale. The conclusion from the timates (Risse et al., 1993). As such, it has become the standard latter was, that measured and modelled soil erosion rates compare technique of many soil conservation workers (Morgan, 2005, pp. relatively well but that they are in any case considerably higher 365e376). However, with remote sensing techniques improving than soil formation rates and that thus land management cannot the availability of high resolution data immensely in recent years claim to be sustainable. (dynamic monitoring of vegetation development as well as estimating dynamic of climate variables), the way is paved for process 6. Model comparisons based techniques in the future needing high resolution spatial input data. 6.1. Model choice and suitability The choice of a soil erosion-prediction tool has to be dependent on the spatial and temporal scale of the intended model applica- A complete overview in a recent review on water erosion tion, as the question of scale is crucial in choosing the right models identified 82 models (Karydas et al., 2014). As stated above modelling approach. Process-oriented models require an applica- an Web of Science query (http://apps.isiknowledge.com/) for the tion of the used equations at a given spatial scale, ranging from plot period 2007e2017 resulted in 1149 hits corresponding to the key- to basin, and at event temporal scale. Event and at-point or small words “Universal Soil Loss Equation”, “USLE”, “Revised Universal scale (hillslope) process based models are not suitable to be applied Soil Loss Equation” or “RUSLE”. In comparison, most popular ap- for simulating soil loss on a wide region not only because of the lack proaches independent of the USLE technology, such as WEPP of required input data but also because of the required indepen- (Laflen et al., 1997), LISEM (De Roo et al., 1996), EUROSEM (Morgan dency of spatially discrete units. At a large spatial scale, the area has et al., 1998) and PESERA (Kirkby et al., 2008) totalized 254 hits. to be discretized using, for example, a square grid subdivision Needless to say, that the fact that USLE-type models are the most (raster scheme), choosing a mesh size consistent with the scale of commonly used erosion tools globally, does not rank them auto- the original model deduction. Using a raster scheme applied to the matically as the best available technique and the suitability of USLE model corresponds to hypothesize that each cell is indepen- model concept needs to be proven for each specific case. dent of the others with respect to soil loss. Even though this meets In the choice of the soil erosion model to be applied, and the original requirements postulated by Wischmeier and Smith congruent with the discussion of section 5, it has to be taken into (1965) that for larger scale applications the area needs to be account that it is not always necessary to calculate the exact erosion “broken down into a series of tracts having relatively homogeneous rate for a particular situation (Nearing, 2013, pp. 365e378), but land use and treatment.”, it requires independency of each of these rather to compare among different situations or time periods. units. In other words, at large spatial scales such as a region or even Choosing which model to apply and model development might a global perspective, a simple index-based model able to calculate become a matter of what type of information is available as input an average soil loss at annual or mean annual scale, allows data. As an example, Nearing (2013, pp. 365e378) noted that most computing the involved factors only using spatially distributed applications of WEPP were developed in the United States because input values. However, the cells cannot be assumed to be inde- of the availability of soil, climate and crop information for those pendent of each other when sediment delivery processes have to be areas. Another important criterion for the choice of a soil erosion modeled (e.g., at basin scale). In this case, the USLE scheme has model is the question of spatial and temporal scale of available been applied by coupling it with a mathematical operator input parameters as well as envisaged model output. As most expressing the hillslope transport efficiency (Ferro, 1997; Ferro & erosion parameters are scale-dependent, data collected on plot Porto, 2000). scale are not always appropriate for calibrating or validating models (see also section 5) and upscaling of models through data 6.2. Direct model comparison used implicitly as model validation aggregation may cause biases (Karydas et al., 2014). In today's modelling world, many of the current generation of Model comparison has been used implicitly as a way of model erosion models are getting more complex, with ever more pro- validation. However, for a rigorous model comparison the question cesses being incorporated into them (Quinton, 2013). However, of scale is decisive (see also section 7). In a European application of even the most complex model of erosion and sedimentation pro- the models PESERA and SENSOR both models compared well, cesses can still only be an approximate representation of reality and however when ‘validated’ to the national scale USLE derived not reality itself. The problems with complex models are twofold: erosion map of Hungary, the European applications of SENSOR and (i) unavailability of the complex set of input parameters especially PESERA underestimated (or USLE overestimated) erosion rates at large temporal and spatial scales and (ii) over parameterization considerably (Fig. 6a; (Podmanicky et al., 2011)). The original which results in a non-uniqueness of fit and/or insensitivity of thought of this comparison was, to validate the European applica- model output to some of the input parameters. From a mathe- tions of SENSOR and PESERA with the higher resolution soil erosion matical view point the calibration of ecosystem models in general rates from Hungary derived with USLE (Podmanicky et al., 2011). are hardly ever unique and the non-uniqueness of fit increases with However, if model type and scale are changed simultaneously, it increasing numbers of parameters (Alewell & Manderscheid, 1998). remains unsolved, whether the deviation originates from scaling Quinton (2013) as well as Veihe and Quinton (2000) and Veihe, inconsistencies or from model deviations or even failures. Quinton, and Poesen (2000) showed that many of EUROSEM's input A comparison of national scale USLE-type modelling (EIONET C. Alewell et al. / International Soil and Water Conservation Research 7 (2019) 203e225 217

Fig. 7. Erosion estimates at the Mediterranean climate of Chile from plot based measurements versus modelled rates from USLE, RUSLE, EPIC and WEPP (all data from Stolpe (2005) with two different parameterizations of the WEPP model).

to be not better than either the USLE or the RUSLE (Kinnell, 2010). USLE-type application has been criticized in complex topographic terrain, because the LS factor does not consider terrain shape of the upslope area (Desmet & Govers, 1996; Mitasova et al., 1996). Mitasova et al. (1996) compared the USLE LS factor and the topographic erosion/deposition index incorporated in the so called unit stream power based approach USPED and concluded, that the latter is more appropriate for landscape erosion modelling, especially when erosion risk and deposition potential needs to be evaluated. Using more than 1600 plot years of data from natural Fig. 6. Model comparison of soil loss data (Mg ha 1yr 1) calculated with SENSOR and rainfallerunoff plots in the USA, Tiwari, Risse, and Nearing (2000) PESERA calibrated to European scale with USLE estimates applied to the national data observed that for average annual soil losses, the model efficiency of Hungary (data of Podmanicky et al. (2011), Fig. 6a) and of national USLE-type of WEPP determined using the (Nash & Sutcliffe, 1970) factor was modelling (EIONET data base, Panagos et al. (2014c)) to PESERA (Kirkby et al., 2008) 0.71, whereas it was 0.80 and 0.72 for the USLE and the RUSLE, and RUSLE2015 (Panagos et al., 2015d), Fig. 6b). respectively. However, as the WEPP model derives its strength from being a process based model and has the capability to predict data base, Panagos et al. (2014c)) with the European scale appli- spatial and temporal distribution of soil loss and deposition, it has cation of RUSLE2015 (Panagos et al., 2015d) or PESERA (Kirkby et al., to be considered in the latter study that the data set was biased 2008) resulted in a relatively good agreement for low values but towards the USLE (all uniform natural runoff plots), the parameters high deviations of the models at the higher end of soil losses used in the USLE have undergone more refinement, and the WEPP (Fig. 6b). model was not calibrated at all (Tiwari et al., 2000). WEPP, USLE and Stolpe (2005) compared RUSLE, WEPP and EPIC at the Medi- RUSLE were observed to over-predict low and under-predict high terranean climate of Chile for their consistency with measured annual soil losses by Kinnell (2010). However, the latter was not erosion rates (2*20 m plots) and those calculated using the previ- confirmed for the above presented data from Europe (Fig. 7). ously calibrated USLE. Their conclusion was that WEPP permitted Favis-Mortlock (1997) drew together the results from models the most adequate parameterization for both climate and volcanic simulating annual soil loss for a field site in the USA. The models soils and that WEPP erosion rates were in best agreement with ranged from simple models, such as CSEP (Kirkby and Cox, 1995) measured rates as well as with calibrated USLE values (Fig. 7). and those based on the USLE (Wischmeier & Smith, 1978, p. 58) Nearing (2004) in comparing USLE and WEPP in two case studies, such as EPIC (Williams et al., 1983) and GLEAMS (Leonard, Knisel, & concluded that USLE can provide certain information which WEPP Still, 1987), to the more process based and complex WEPP model. simply cannot provide because of the restrictions of model Favis-Mortlock's results show no significant difference between the complexity, but that the WEPP model can be used in a way where performances of the models. Morgan and Nearing (2011) compared only the complex model interactions will provide the information the performance of the USLE, RUSLE and WEPP using 1700 plot needed regarding system response. years of data from 208 natural runoff plots. Once again it became Kinnell (2010) criticizes that WEPP ignores any variation in the clear that the more complex model, in this case WEPP, performs no kinetic energy per unit quantity of rain, even though it is dependent better than the empirical, five parameter USLE (Morgan & Nearing, on rainfall intensity data. These intensities are generated by the 2011). stochastic daily parameter weather generator CLIGEN (Nicks, Lane, Using the USLE basically comes down to testing the hypothesis, & Gander, 1995), but the meteorological data required to parame- that only a small number of parameters control the response of an terize CLIGEN are not available in many countries outside the USA. erosion model and the remainder become largely inconsequential The ability of WEPP to predict soil losses in the USA has been found (Quinton, 2013). As such, model uncertainty does not seem to be 218 C. Alewell et al. / International Soil and Water Conservation Research 7 (2019) 203e225 smaller in complex, process based models compared to simpler, as a graphical summary in maps. With a spatially discrete model- empirical models. Quinton (2013) when discussing several erosion ling approach as the USLE-type algorithm uncertainty is per se in- model studies increasing the complexity of models and comparing dependent from scale, because the uncertainty within each single it to the original simpler version, concludes that from a scientific discrete pixel is independent from the number of these pixels. The point of view, the inclusion of more processes and parameters uncertainty is merely coupled to the resolution of the modelling make the models a better representation of reality, but it did not approach (e.g. size of pixels) and the connected heterogeneity make the models better in predicting erosion rates. However, within these pixels. However, limitation of the modelling capacity stating this, we have to keep in mind that as noted above, USLE- might be connected to scale, as demands on input data and type modelling will always be less successful when models are computational capacity is of course increasing with increase in compared to catchment sediment yields or off-site measurements resolution and scale. in freshwaters (see discussion in section 5.2 and as good example Again, as discussed above, when aiming at large scales and (Tan et al., 2018)). considering erosion risk of (large) catchments, nations, continents To conclude, there is not one ”superior” model suitable for all or even global scales, often an misunderstanding evolves during kind of situations, temporal and spatial scales as well as data res- this transfer from small to large scale and USLE-type modelling olution. This basically leaves soil erosion modellers with two pos- output is implicitly compared to sediment yield and sediment sibilities: (i) accept that there are no adequate tools to model soil transport capacities. As such, evaluating large-scale model appli- erosion by water and give up modelling until modelling and the cation of USLE-type modelling needs the repeated reminder that required input data sets are improved or (ii) recollect what was on-site erosion risk is simulated and not off-site sediment yield and discussed in section 5: At least at large scales we should bid fare- transport. well to the idea of accurately predicting absolute values with While modelling on catchment scale nowadays operates on a models but should rather aim at the prediction of relative differ- resolution of 2 m (depending on DEM resolution, see section 4), the ences, trends over times and systems reactions to driving factors latest large scale modelling assessments achieved a spatial reso- and management practices and use models as tools to learn about lution of 100 m for European (Panagos et al., 2015d) and 250 m for the modelled systems and their reactions. In this conceptual global scale (Borrelli et al., 2017; Diodato et al., 2017). The latter approach, modelling in general and large-scale modelling specif- approaches use considerably coarser resolution than the original ically will be understood as a tool to test hypotheses on process USLE-type plot of 22.1 m length. However, variability does not understanding, relative spatial and temporal variations, scenario necessarily increase with increase in scale. Even though it might be development and controlling factors (Oreskes et al., 1994). expected that measured soil loss variability should decrease with increasing plot size, Rüttimann et al. (1995) could not confirm an 7. Upscaling: from plot to global and from event to long term effect of plot size on increase in variability but rather noticed a averages striking absence of any effect of an increase in plot size on measurement variability or uncertainty. Application of environmental modelling always involves four So one question to be answered is, how homogenous the model different scales: the geographic scale of a research area, the tem- units (‘tracts’ or better pixels) are and how valid the approach is in poral scale related to the time period of research, the measurement terms of assuming homogenous R, K, LS, C and P factors within each scale of parameters (input data resolution) and the model scale pixel. As spatial data availability is increasing continuously in res- referring to both temporal and spatial scales when a model was olution and remote sensing tools shedding more and more light on established (Zhang, Drage, & Wainwright, 2004b). small scale heterogeneity, the above question might eventually The crucial importance of temporal scale becomes obvious dissipate in the future. In this, there is no principle difference be- when considering that over and under-estimation of modelled tween a global scale modelling from a national scale modelling and erosion rates might be connected merely to the length in obser- not even from large catchments: it always comes down to pixel size vational data: when the observation period is too short to catch the and data quality versus the system's heterogeneity. long-term average, observations either miss out on extreme events, Kinnell (2010) warns at applying the USLE/RUSLE model to or might by chance catch an accumulation of extreme events. The extrapolate data collected at the plot scale to areas much larger area USLE, designed to predict long-term average, will then of course than the field sized areas for which it was originally designed. He overestimate or underestimate these short measurement periods, argues that at the hillslope scale, spatial variability in soil and respectively. vegetation result in spatial variations in runoff, so that the Regarding the spatial scale, the original USLE model and its modelling approach would need to consider those spatial varia- factors were developed for plot scale of 22.1 m length and thus its tions in runoff, which models like WEPP and EUROSEM have an validity for large scale application has been discussed in recent advantage over the USLE, RUSLE and RUSLE2 because they include debates (Auerswald et al., 2015; Evans, 2013; Evans and Boardman, explicit consideration of runoff (Kinnell, 2010). Again, the implicit 2016a, 2016b; Fiener & Auerswald, 2016; Panagos et al., 2016a, transfer from on-site soil erosion risk to off-site sediment yields 2016b; Panagos et al., 2015e; Schwertmann et al., 1987, pp. 1e64). might have slipped into the discussion. However, the problem with Schwertmann et al. (1987, pp. 1e64) when transferring the model increasing variability in vegetation cover at hill slope (or even from the US to sites in Bavaria (Southern Germany) warned that the larger) scales might at least partially be overcome by advanced model approach is limited to plot scale and other equations should approaches of considering high resolution spatial input data from be used for catchments scale applications. However, Wischmeier remote sensing. One example would be a study of Meusburger et al. and Smith (1965) already stated there is principally no logical (2010b) who implemented spatial patterns of fractional vegetation reason why it should not be correct to “break down the drainage cover from Quickbird imagery into the C factor map and achieved area into a series of tracts having relatively homogeneous land use (i) a good agreement of spatial patterns of soil erosion predicted and treatment. The erosion equation is then used to approximate with USLE compared to observed remote sensing patterns and (ii) the average annual rate of soil movement from each tract.” Large erosion rates which are at least in the same order of magnitude scale modelling using GIS tools is basically doing nothing else, compared to 137Cs derived erosion rates (for hot spots of erosion when approximating average annual rates of soil movement for 20.1 Mg ha 1yr 1 versus 16 Mg ha 1yr 1 for 137Cs based and USLE single tracts (¼pixels) presenting the result of each of these pixels derived soil losses, respectively). The latter was neither possible C. Alewell et al. / International Soil and Water Conservation Research 7 (2019) 203e225 219 with an USLE application with uniform C factor rates not consid- climate and rain patterns, vegetation cover and management, as ering the spatial heterogeneous distribution of vegetation cover well as the specific soil characteristics, US type systems and soils with PESERA where the same spatially distributed vegetation cover are not a case sui generis and soil erosion from US soils is regulated was considered but spatial soil loss patterns could not be captured by the same factors and processes as everywhere else in the world. adequately and soil losses on hot spots were considerably under- Long term investigations of Schwertmann et al. (1987, pp. 1e64) estimated (2.9 Mg ha 1yr 1 compared to the above mentioned already supported the latter assumption three decades ago. 137Cs based estimates (Meusburger et al., 2010b)). Schmidt, Alewell, Recent studies confirmed that when appropriately parametrized, and Meusburger (2019) followed the latter approach and combined uncertainty of USLE-type modelling all around the globe is not the spatio-temporal heterogeneity from the C factor map with the bigger than within the US (Kinnell, 2010; Meusburger et al., 2010b, spatio-temporal heterogeneity of the R factor to a soil erosion risk 2013; Stolpe, 2005; Yue et al., 2016). map of Switzerland with monthly temporal and 100 m spatial Apart from the discussion whether or not modelled soil erosion resolution. estimates are actually ‘good enough’ from a scientific perspective and the question what actually ‘good enough’ exactly is, we would 8. Limitations of the USLE concept and the practical need to like to raise another issue: the practical need of soil erosion model soil erosion modelling. Without modelling there will be no large scale erosion assess- Recent studies continue to question the use of models to predict ments. Without large scale erosion assessments, soil erosion as one erosion potential, and argue that instead there is a need for more of the most serious threats to soils will not be taken into consid- field-based assessment and monitoring of water erosion (Evans, eration by any major environmental and agricultural policy pro- 2013). Modelling cannot be an alternative to measurement and grams, as well as integrated land surface models (LSMs). For monitoring but might be a powerful tool in understanding obser- example, the voluntary Guidelines for Sustainable Soil Manage- vations and in developing and testing theories. Understanding ment of Food and Agriculture Organization (FAO, 2017) identify soil processes, regulating parameters, trends and differences between erosion by water and wind as the most significant threat to global spatially or temporally differing data is very often the ultimate goal soils and the ecosystem services they provide. At global scale, the in science and models serve as tools towards that end. As such, Sustainable Development Goals (SDGs) include a sub-indicator on models can be understood as a tool for understanding, as well as estimated soil erosion by water. Europe's Common Agricultural tools for simulation and prediction, a virtual laboratory which Policy sets requirements to protect utilized agricultural areas might as a positive aspect serve as an integrator within and be- against erosion, which established a framework of standards that tween disciplines, as it brings together data, observations and aim, among others, to preventing soil erosion. The Good Agricul- knowledge of different fields (Nearing, 2004). However, as models tural and Environmental conditions (GAEC) implemented in the are also used as means of communicating science and the results of European Common Agricultural Policy reform (Borrelli et al., 2016c) science, there is always the danger, that what they communicate is introduced the prevention of soil erosion and higher residues weak or even wrong and these weaknesses might not be apparent restitution, increased soil cover and lowered tillage disturbance to the model user, or the user of the model output (Nearing, 2004). which are among the most effective mitigation practices in Euro- Trimble and Crosson (2000) criticized the lack of prediction of pean arable lands. All these policy programs can not consider soil sediment delivery ratios by USLE and its failure to predict gully erosion if there are no tools for large scale modelling and mapping erosion, which led them to conclude for the US that “the limitations available. of the USLE (…) are such that we do not seem to have a truly Tolerable soil erosion rates all over the world are given between informed idea of how much soil erosion is occurring in this country, 0.1 and 1 mm yr 1 (for overview of all studies see Li, Du, Wu, and let alone of the processes of sediment movement and deposition.” Liu (2009)). For sake of simplicity, these mm rates might be con- As discussed above, USLE-type models were not designed to predict verted with a bulk density of 1 t m 3 which is typical for the loose gully erosion nor sediment delivery ratios and modelers, model soil structure of surface soils and would result in soil loss rates of users and stakeholders should of course be aware of underlying 1e10 Mg ha 1yr 1. For conditions prevalent in Europe, the upper model concepts and parameters, the above discussed limitations limit of soil formation has been estimated to be approximately and the possible evaluation and interpretation of the model output. 1.4 Mg ha 1 yr 1 while the lower limit is given as 0.3 Mg ha 1 yr 1 Regarding the latter, the simplicity of the USLE-concept might be (Alewell et al., 2015; Verheijen, Jones, Rickson, & Smith, 2009). As regarded as an advantage as each of the five USLE parameters might tolerable soil erosion rates should not exceed soil formation rates, be evaluated separately even by non-expert stakeholders many countries in the world are actually performing non- increasing thus transparency and objectiveness of evaluation. sustainable agriculture. Without large-scale soil erosion mapping The lack of quantifying gully erosion or stream bank erosion has we do not have tools to address this problem and thus no possi- been frequently addressed as a lack or even general failure of USLE- bilities to demand mitigation actions. type modelling (Belyaev, Wallbrink, Golosov, Murray, & Sidorchuk, 2005; Evans and Boardman, 2016a, 2016b; Quinton, 2013; Trimble 9. Conclusions and outlook & Crosson, 2000). Gully erosion and stream bank erosion involve complex and highly heterogeneous hydrological and soil erosion Soil loss occurs not because of any lack of knowledge on how to processes of channel formations difficult to model. We would like protect soils, but a lack in policy governance (Montanarella, 2015; to state that the USLE was never meant to address gully, wind or Panagos et al., 2016b). The lack in policy governance is due to a lack stream bank erosion, as is clearly stated in the very first publica- in research, knowledge and dissemination on spatial distribution of tions of USLE (Wischmeier and Smith,1965, 1978, p. 58). It would be soil erosion and sediment source attribution from water and wind like discrediting a butterfly monitoring tool for missing out on erosion. As such, soil erosion modelling is crucially needed on both important bird species. small and large spatial scales for planning, management and policy USLE-type modelling is frequently discussed to be developed for measures. The USLE-type models have been extensively used and US type soils and systems and thus questioned to be applicable for modified during the last decades reaching the ceiling of improve- other regions (see above). However, as long as model parameters ment in recent years. Recent advancements in remote sensing, are carefully considered and adapted to site or region specific release of climatic datasets, greater availability of earth 220 C. Alewell et al. / International Soil and Water Conservation Research 7 (2019) 203e225 observations data, and increased processing of big datasets are the the exception of the pan-European application of the PESERA promising factors for more dynamic and process-based modelling model (Kirkby et al., 2008). approaches in the near future. Soil erosion modeling moves to- As measured soil loss data itself can be object to high un- wards more dynamic approaches and can use the field parcels as certainties, future research should target reducing uncertainties in objects due to availability of highly accurate spatial and climatic modelling and measurement. Comprehensive and well planned data for monitoring intra-annual variability of climate, vegetation measurement campaigns at small as well as large scale are crucially and management practices (Borrelli, Meusburger, Ballabio, needed for soil erosion modelling validation as well as deepening Panagos, Alewell, 2018). our process understanding. As such, it is recommended to invest on However, improving interpretation of the soil erosion processes monitoring campaigns and experimental designs towards more at different spatial and temporal scales and testing or developing dynamic modelling. In addition, comparison of large scale model- physically oriented modeling approaches is not only of socio- ling with local or regional studies might help to narrow possible economic and political importance but has also an obvious scien- pitfalls, inconsistencies and uncertainty of modelling. A possible tific aim of process understanding and general prediction of future model validation approach might be the comparison of ecosystem dynamics (as soil erosion is an important process of continental scale modeling with local modelling data sets, which ecosystem stability and dynamic, e.g., erosion rates in climate might give important insights into processes, trends and driving change modelling of carbon dynamics, soil nutrient balance or factors of soil erosion. In the upcoming Land Use/Cover Area frame pollution dissemination prediction). statistical Survey Soil (LUCAS Soil) data collection campaign, sur- From our literature evaluation, we conclude that at today's state veyors will make a qualitative assessment of soil erosion in >26,000 of knowledge USLE-type models have no higher uncertainty than locations in Europe (Orgiazzi, Ballabio, Panagos, Jones, & more complex based models provided modelling targets at on-site Fernandez-Ugalde, 2018). As LUCAS is repeated every 3 years, this soil erosion risk (compared to off-site sediment transport and has the potential to evolve into a monitoring program, which can yield). Simultaneously, they can be a choice with regard to data eventually be used to validate modeling on European scale. availability concerns and/or ease of usage and interpretation. This Validation of the USLE-type applications are usually inexplicitly has to be considered with care in case of increased data input carried out in the most rigorous of possible validation procedures as availability in the near future which will allow modelling erosional the model algorithms are per se not calibrated, but the final out- process in a more dynamic and potentially process based way. comes are compared to measured data. Obtaining accurate and reliable soil loss estimates using Despite the large number of soil erosion models available for the spatially distributed models on wide regions depends on both the prediction of both runoff and soil loss at a variety of scales, little resolution (vertical and horizontal) of the input topographic in- quantification is made of uncertainty and error associated with formation and the quality of the land-use input data. USLE-type model output. The latter seems crucially important to not only models are a compromise for wide region application when both advance todays understanding but also the acceptance of soil a mesh-size comparable with the original developed slope-length erosion modelling outputs on larger scales. It might be suggested scale is applied and information on rainfall, soil and land-use sys- that future applications should use objective modelling criteria for tems is available. evaluation (e.g. as has been suggested by Alewell and The meta-analysis of published articles using USLE-type Manderscheid (1998) or Brazier et al. (2000)). However, the latter modelling to estimate soil loss by water erosion from local to would not solve the paradigm of the principally different nature of continental scale showed an increasing trend especially in the last modelled (gross) versus measured (net) erosion rates by USLE-type 20 years. The models are widely used on all continents and within models. As such, parallel to increasing research efforts in soil 109 countries. The long-tradition of articles published mainly in the erosion measurement and investing in rigorous application of US in the early ‘80s and ‘90s, has been continued in Europe, evaluation criteria, the path where soil erosion is connected to Australia and Southern America and more recently increasingly in topography, runoff and flow path dynamic as well as depositional Asia and Africa with scientists addressing also emerging research regime should be further followed and solutions should be found topics connected to soil erosion such as climate change, nutrients for large-scale applications. balance, freshwater pollution, land use and management and protection measures. As discussed above, soil erosion modelling is, like any other References ecosystem model, only a representation of reality and not reality Aiello, A., Adamo, M., & Canora, F. (2015). Remote sensing and GIS to assess soil itself, and is thus prone to deviate from reality and predicted un- erosion with RUSLE3D and USPED at river basin scale in southern Italy. Catena, certainty might be more or less substantial. As such, for each 131,174e185. modelling endeavor a deep understanding of underlying processes, Alewell, C., Birkholz, A., Meusburger, K., Wildhaber, Y. S., & Mabit, L. (2016). Quantitative sediment source attribution with compound-specific isotope assumptions and model structure is necessary to avoid misinter- analysis in a C3 plant-dominated catchment (central Switzerland). Bio- pretation (e.g. in case of USLE-Type modelling the conceptual geosciences, 13(5), 1587e1596. confusion between on-site gross erosion risk and off-site net Alewell, C., Egli, M., & Meusburger, K. (2015). An attempt to estimate tolerable soil erosion rates by matching soil formation with denudation in Alpine grasslands. sediment yields). However, the above discussed validation at- Journal of Soils and Sediments, 15(6), 1383e1399. tempts showed that when appropriately parameterized soil loss Alewell, C., & Manderscheid, B. (1998). Use of objective criteria for the assessment estimates with USLE-type models are within the order of magni- of biogeochemical ecosystem models. Ecological Modelling, 107(2), 213e224. tude compared to measured soil loss rates and relative differences Amundson, R., Berhe, A. A., Hopmans, J. W., Olson, C., Sztein, A. E., & Sparks, D. L. (2015). Soil and human security in the 21st century. Science, 348(6235), 6. between areas, management scenarios and time periods seem to be Angulo-Martinez, M., Begueria, S., & Kysely, J. (2016). Use of disdrometer data to captured correctly in many studies. Based on this, the modelling evaluate the relationship of rainfall kinetic energy and intensity (KE-I). Science e approach can be recommended to be used to target risk areas, of the Total Environment, 568,83 94. Angulo-Martinez, M., Lopez-Vicente, M., Vicente-Serrano, S. M., & Begueria, S. scenario calculation for conservation tools or other management (2009). Mapping rainfall erosivity at a regional scale: A comparison of inter- plans, spatial comparison of regions, countries and continents and polation methods in the Ebro basin (NE Spain). Hydrology and Earth System temporal comparisons of differing time periods. At the current state Sciences, 13(10), 1907e1920. Arnhold, S., Lindner, S., Lee, B., Martin, E., Kettering, J., Nguyen, T. T., et al. (2014). of knowledge, USLE-type models are the only model algorithms Conventional and organic farming: Soil erosion and conservation potential for which have been used at large scales (e.g. continental, global), with row crop cultivation. Geoderma, 219,89e105. C. Alewell et al. / International Soil and Water Conservation Research 7 (2019) 203e225 221

Arnold, J. G., Srinivasan, R., Muttiah, R. S., & Williams, J. R. (1998). Large area hy- Iowa. American Society of Agronomy Journal, 39,65e73. drologic modeling and assesment Part I: Model development. JAWRA Journal of Cao, L., Zhang, K., Dai, H., & Liang, Y. (2015). Modeling interrill erosion on unpaved the American Water Resources Association, 34(1), 73e89. roads in the loess plateau of China. Land Degradation and Development, 26(8), Arnoldus, H. M. J. (1980). An approximation of the rainfall factor in the universal soil 825e832. loss equation. In M. De Boodts, & D. Gabriels (Eds.), Assessments of erosion (pp. Cerdan, O., Govers, G., Le Bissonnais, Y., Van Oost, K., Poesen, J., Saby, N., et al. (2010). 127e132). John Wiley and Sons Ltd. Rates and spatial variations of soil erosion in Europe: A study based on erosion Aronica, G., & Ferro, V. (1997). Rainfall erosivity over the Calabrian region. Hydro- plot data. Studies on Large Volume Landslides, 122(1e2), 167e177. logical Sciences Journal-Journal Des Sciences Hydrologiques, 42(1), 35e48. Cronshey, R. G., & Theurer, F. D. (1998). AnnAGNPS-non-point pollutant loading model Auerswald, K., Fiener, P., Gomez, J. A., Govers, G., Quinton, J. N., & Strauss, P. (2015). (Las Vegas). Comment on "Rainfall erosivity in Europe" by Panagos et al. (Sci. Total Environ., De Jong, S. M. (1994). Derivation of vegetative variables from a landsat tm image for 511, 801-814, 2015). Science of the Total Environment, 532, 849e852. modelling soil erosion. Earth Surface Processes and Landforms, 19,165e178. Auerswald, K., Fiener, P., Martin, W., & Elhaus, D. (2014). Use and misuse of the K De Roo, A. P. J., Wesseling, C. G., & Ritsema, C. J. (1996). Lisem: A single-event factor equation in soil erosion modeling: An alternative equation for deter- physically based hydrological and soil erosion model for drainage basins. I: mining USLE nomograph soil erodibility values. Catena, 118, 220e225. Theory, input and output. Hydrological Processes, 10(8), 1107e1117. Auerswald, K., Kainz, M., & Fiener, P. (2003). Soil erosion potential of organic versus De Vente, J., & Poesen, J. (2005). Predicting soil erosion and sediment yield at the conventional farming evaluated by USLE modelling of cropping statistics for basin scale: Scale issues and semi-quantitative models. Earth Science Reviews, agricultural districts in Bavaria. Soil Use and Management, 19(4), 305e311. 71(1e2), 95. Bagarello, V., Di Piazza, G. V., Ferro, V., & Giordano, G. (2008). Predicting unit plot Desmet, P., & Govers, G. (1996). A GIs procedure for automatically calculating the soil loss in Sicily, south Italy. Hydrological Processes, 22(5), 586e595. USLE LS factor on topographically complex landscape units. Journal of Soil and Bagarello, V., Di Stefano, V., Ferro, V., Giordano, G., Iovino, M., & Pampalone, V. Water Conservation, 51(5), 427e433. (2012). Estimating the USLE soil erodibility factor in Sicily, South Italy. Applied Di Stefano, C., Ferro, V., & Pampalone, V. (2017). Testing the USLE-M family of Engineering in Agriculture, 28(2), 199e206. models at the Sparacia experimental site in South Italy. Journal of Hydrologic Bagarello, V., Di Stefano, C., Ferro, V., & Pampalone, V. (2017). Predicting maximum Engineering, 22(8), 05017012-05017012. annual values of event soil loss by USLE-type models. Catena, 155,10e19. Diodato, N., Borrelli, P., Fiener, P., Bellocchi, G., & Romano, N. (2017). Discovering Bagarello, V., & Ferro, V. (2004). Plot-scale measurement of soil erosion at the historical rainfall erosivity with a parsimonious approach: A case study in experimental area of Sparacia (southern Italy). Hydrological Processes, 18(1), Western Germany. Journal of Hydrology, 544. 141e157. Diodato, N., Knight, J., & Bellocchi, G. (2013). Reduced complexity model for Bagarello, V., Ferro, V., & Pampalone, V. (2015). A new version of the USLE-MM for assessing patterns of rainfall erosivity in Africa. Global and Planetary Change, predicting bare plot soil loss at the Sparacia (South Italy) experimental site. 100,183e193. Hydrological Processes, 29(19), 4210e4219. Dissmeyer, G. E., & Foster, G. R. (1980). A guide for predicting sheet and rill erosion on Bakker, M. M., Govers, G., van Doorn, A., Quetier, F., Chouvardas, D., & Rounsevell, M. forest land. (2008). The response of soil erosion and sediment export to land-use change in Doetterl, S., Van Oost, K., & Six, J. (2012). Towards constraining the magnitude of four areas of Europe: The importance of landscape pattern. Geomorphology, 98, global agricultural sediment and soil organic carbon fluxes. Earth Surface Pro- 213e226. cesses and Landforms, 37(6), 642e655. Barenblatt, G. I. (1979). Similarity, self-similarity and intermediate asymptotics.New Dokuchaev, V. V. (1892). Nashi stepi prezde i teper(Our steppes before and recently). York: Consultants Bureau. Russia(in Russian). Sankt_Peterburg: Published House Evdokimov. Barenblatt, G. I. (1987). Dimensional analysis (Amsterdam). Dunn, M., & Hickey, R. (1998). The effect of slope algorithms on slope estimates Belyaev, V. R., Wallbrink, P. J., Golosov, V. N., Murray, A. S., & Sidorchuk, A. Y. (2005). within a GIS. Cartography, 27(1), 9e15. A comparison of methods for evaluating soil redistribution in the severely Evans, R. (2013). Assessment and monitoring of accelerated water erosion of eroded Stavropol region, southern European Russia. Geomorphology, 65(3e4), cultivated land - when will reality be acknowledged? Soil Use and Management, 173e193. 29(1), 105e118. Ben-Dor, E., & Banin, A. (1995). Near-infrared analysis as a rapid method to Evans, R., & Boardman, J. (2016a). The new assessment of soil loss by water erosion simultaneously evaluate several soil properties. Soil Science Society of America in Europe. Panagos P. et al., 2015 Environmental Science & Policy 54, 438-447-A Journal, 59(2), 364e372. response. Environmental Science & Policy, 58,11e15. Bennett, H. H. (1939). A permanent loss to new england soil erosion resulting from Evans, R., & Boardman, J. (2016b). A reply to panagos et al., 2016 (Environmental the hurricane. Geographical Review, 29(2), 196e204. science & policy 59 (2016) 53-57. Environmental Science & Policy, 60,63e68. Boardman, J. (2006). Soil erosion science: Reflections on the limitations of current Fan, J. C. (1987). Measurements of erosion on highway slopes and of the universal soil approaches. Catena, 68(2e3), 73e86. loss erosion equation. West Lafayette: Purdue Univ. (Ind). Bonilla, C. A., & Vidal, K. L. (2011). Rainfall erosivity in Central Chile. Journal of FAO. (2015). GLADIS - global land degradation information system. Hydrology, 410(1e2), 126e133. FAO. (2017). Voluntary Guidelines for sustainable soil management food and agricul- Bonta, J. V., & Sutton, P. (1983). Erosion and reclamation plots: Research on the ture organization of the united nations. http://www.fao.org/3/a-bl813e.pdfhttp:// hydroogy and water quality of watersheds subjected to surface mining. Bureau of www.fao.org/3/a-bl813e.pdf (Rome, Italy). Mines, U.S: Dept. Interior. Favis-Mortlock, D. T. (1997). Validation of field scale soil erosion models using Borrelli, P., Diodato, N., & Panagos, P. (2016a). Rainfall erosivity in Italy: A national common data sets. In J. Boardman, & D. T. F.-M (Eds.), Modeling soil erosion by scale spatio-temporal assessment. International Journal of Digital Earth, 9(9). WaterNATO ASI series (pp. 89e127). Berlin: Springer-Verlag. Borrelli, P., Meusburger, K., Ballabio, C., Panagos, P., & Alewell, C. (2018). Object- Ferro, V. (1997). Further remarks on a distributed approach to sediment delivery. oriented soil erosion modelling: A possible paradigm shift from potential to Hydrological Sciences Journal, 42(5), 633e647. actual risk assessments in agricultural environments. Land Degradation & Ferro, V. (2010). Deducing the USLE mathematical structure by dimensional analysis Development, 29(4), 1270e1281. and self-similarity theory. Biosystems Engineering, 106(2), 216e220. Borrelli, P., Panagos, P., Langhammer, J., Apostol, B., & Schütt, B. (2016b). Assessment Ferro, V., & Minacapilli, M. (1995). Sediment delivery processes at basin-scale. of the cover changes and the soil loss potential in European forestland: First Hydrological Sciences Journal-Journal Des Sciences Hydrologiques, 40(6), 703e717. approach to derive indicators to capture the ecological impacts on soil-related Ferro, V., & Porto, P. (1999). A comparative study of rainfall erosivity estimation for forest ecosystems. Ecological Indicators, 60. southern Italy and southeastern Australia. Hydrological Sciences Journal-Journal Borrelli, P., Panagos, P., Marker,€ M., Modugno, S., & Schütt, B. (2017). Assessment of Des Sciences Hydrologiques, 44(1), 3e24. the impacts of clear-cutting on soil loss by water erosion in Italian forests: First Ferro, V., & Porto, P. (2000). Sediment delivery distributed (SEDD) model. Journal of comprehensive monitoring and modelling approach. Catena, 149,770e781. Hydrologic Engineering, 5(4), 411e422. Borrelli, P., Paustian, K., Panagos, P., Jones, A., Schütt, B., & Lugato, E. (2016c). Effect Fiener, P., & Auerswald, K. (2016). Comment on "The new assessment of soil loss by of good agricultural and environmental conditions on erosion and soil organic water erosion in Europe" by Panagos et al. (Environmental Science & Policy 54 carbon balance: A national case study. Land Use Policy, 50, 408e421. (2015) 438-447). Environmental Science & Policy, 57,140e142. Borrelli, P., Robinson, D. A., Fleischer, L. R., Lugato, E., Ballabio, C., Alewell, C., Fishman, G. S., & Kiviat, P. J. (1968). The statistics of discrete event simulation. Meusburger, K., Modugno, S., Schütt, B., Ferro, V., Bagarello, V., Oost, K. V., Simulation, 10,185e191. Montanarella, L., & Panagos, P. (2017b). An assessment of the global impact of Foley, J. (2017). Living by the lessons of the planet. Science, 356(6335), 251e252. 21st century land use change on soil erosion. Nature Communications, 8(1), Foster, G. R. (1982). Relation of Usle Factors to Erosion on Rangeland. Science and 2013. Education Administration Publications. ARM(NW-26). Borst, H. L., & McCall, A. G. (1945). Investigations in erosion control and the recla- Foster, G. R., & Wischmeier, W. H. (1974). Evaluating irregular slopes for soil loss mation of eroded land at the Northwest Appalachian conservation experiment prediction. Transactions of the Asae, 17(2), 305e309. station. OH: Zanesville. Foster, G. R., Yoder, D. C., Weesies, G. A., McCool, D. K., McGregor, K. C., & Brazier, R. E., Beven, K. J., Freer, J., & Rowan, J. S. (2000). Equifinality and uncertainty Bingner, R. L. (2008). Draft User's Guide, Revised Universal Soil Loss Equation in physically based soil erosion models: Application of the GLUE methodology Version 2 (RUSLE-2). Washington, DC: USDA-Agricultural Research Service. to WEPPethe water erosion prediction projectefor sites in the UK and USA. Foster, G. R., Yoder, D. C., Weesies, G. A., & Toy, T. J. (2001). The design philosophy Earth Surface Processes and Landforms, 25(8), 825e845. behind RUSLE2: Evolution of an empirical model. In Soil erosion research for the Brown, L. C., & Foster, G. R. (1987). Storm erosivity using idealized intensity dis- 21st century, proceedings (pp. 95e98). tributions. Transactions of the Asae, 30(2), 379e386. Gabriels, D., Ghekiere, G., Schiettecatte, W., & Rottiers, I. (2003). Assessment of USLE Browning, G. M., Parish, C. L., & Glass, J. A. (1947). A method for determining the use cover-management C-factors for 40 crop rotation systems on arable farms in and limitation of rotation and conservation practices in control of soil erosion in the Kemmelbeek watershed, Belgium. Soil and Tillage Research, 74,47e53. 222 C. Alewell et al. / International Soil and Water Conservation Research 7 (2019) 203e225

Garcia-Ruiz, J. M., Begueria, S., Nadal-Romero, E., Gonzalez-Hidalgo, J. C., Lana- Larson, W. E., Lindstrom, M. J., & Schumacher, T. E. (1997). The role of severe storms Renault, N., & Sanjuan, Y. (2015). A meta-analysis of soil erosion rates across the in soil erosion: A problem needing consideration. Journal of Soil and Water world. Geomorphology, 239,160e173. Conservation, 52,90e95. Gessesse, B., Bewket, W., & Brauning,€ A. (2015). Model based characterization and Lee, J. H., & Heo, J. H. (2011). Evaluation of estimation methods for rainfall erosivity monitoring of runoff and soil erosion in response to land use/land cover based on annual precipitation in Korea. Journal of Hydrology, 409(1e2), 30e48. changes in the. Land Degradation & Development, 26,711e724. February 2014. Leitinger, G., Tasser, E., Newesely, C., Obojes, N., & Tappeiner, U. (2010). Seasonal van de Giesen, N., Stomph, T. J., & de Ridder, N. (2005). Surface runoff scale effects in dynamics of surface runoff in mountain grassland ecosystems differing in land west African watersheds: Modeling and management options. Agricultural use. Journal of Hydrology, 385(1e4), 95e104. Water Management, 72(2), 109e130. Leonard, R. A., Knisel, W. G., & Still, D. A. (1987). GLEAMS: Groundwater loading Griffin, M. L., Beasley, D. B., Fletcher, J. J., & Foster, G. R. (1988). Estimating soil loss effects of agricultural management systems. Transactions of the ASAE, 30(5), on topographically non-uniform field and farm units. Journal of Soil and Water 1403e1418. Conservation, 43(4), 326e331. Li, L., Du, S., Wu, L., & Liu, G. (2009). An overview of soil loss tolerance. Catena, 78(2), Haciyakupoglu, S., Hizal, A., Gokbulak, F., & Kiziltas, S. (2010). A comparison of soil 93e99. losses estimated using fallout radionuclides and the USLE for the Omerli Littleboy, M., Freebairn, D. M., Hammer, G. L., & Silburn, D. M. (1992a). Impact of watershed, Turkey. In M. Zlatic (Ed.), Global change: Challenges for soil man- soil-erosion on production in cropping systems .2. simulation of production and agement. Advances in geoecology (pp. 175e183). erosion risks for a wheat cropping system. Australian Journal of Soil Research, Haileslassie, A., Priess, J., Veldkamp, E., Teketay, D., & Lesschen, J. P. (2005). 30(5), 775e788. Assessment of soil nutrient depletion and its spatial variability on smallholders' Littleboy, M., Silburn, D. M., Freebairn, D. M., Woodruff, D. R., Hammer, G. L., & mixed farming systems in Ethiopia using partial versus full nutrient balances. Leslie, J. K. (1992b). Impact of soil-erosion on production in cropping systems .1. Agriculture Ecosystems & Environment, 108(1), 1e16. development and validation of a simulation-model. Australian Journal of Soil Hart, G. E. (1984). Erosion from simulated rainfall on mountain rangeland in Utah. Research, 30(5), 757e774. Journal of Soil and Water Conservation, 39, 330e334. Liu, B. Y., Nearing, M. A., & Risse, L. M. (1994). Slope gradient effects on soil loss for Hengl, T., de Jesus, J. M., MacMillan, R. A., Batjes, N. H., Heuvelink, G. B. M., steep slopes. Transactions of the Asae, 37(6), 1835e1840. Ribeiro, E., et al. (2014). SoilGrids1km-Global soil information based on auto- Liu, B. Y., Zhang, K. L., & Xie, Y. (2002). An empirical soil loss equation. In Proc. of 12th mated mapping. PLoS One, 9(8), 17. ISCO. Beijing: Tsinghua Press. Hickey, R. (2000). Slope angle and slope length solutions for GIS. Cartography, 29(1), Lloyd, C. H., & Eley, G. W. (1952). Graphical solution of probable soil loss formula for 1e8. northeastern region. Journal of Soil and Water Conservation, 7,189e191. Hussein, M. H., Kariem, T. H., & Othman, A. K. (2007). Predicting soil erodibility in Loureiro, N. D., & Coutinho, M. D. (2001). A new procedure to estimate the RUSLE northern Iraq using natural runoff plot data. Soil & Tillage Research, 94(1), EI30 index, based on monthly rainfall data and applied to the Algarve region, 220e228. Portugal. Journal of Hydrology, 250(1e4), 12e18. Ito, A. (2007). Simulated impacts of climate and land-cover change on soil erosion Lufafa, A., Tenywa, M. M., Isabirye, M., Majaliwa, M. J. G., & Woomer, P. L. (2003). and implication for the carbon cycle, 1901 to 2100. Geophysical Research Letters, Prediction of soil erosion in a Lake Victoria basin catchment using a GIS-based 34(9). Universal Soil Loss model. Agricultural Systems, 76(3), 883e894. Jetten, V., Govers, G., & Hessel, R. (2003). Erosion models: Quality of spatial pre- Luleva, M. I., van der Werff, H., Jetten, V., & van der Meer, F. (2011). Can infrared dictions. Hydrological Processes, 17, 887e900. spectroscopy Be used to measure change in potassium nitrate concentration as Joel, A., Messing, I., Seguel, O., & Casanova, M. (2002). Measurement of surface water a proxy for soil particle movement? Sensors (Basel, Switzerland), 11(4), runoff from plots of two different sizes. Hydrological Processes, 16(7), 4188e4206. 1467e1478. Marker,€ M., Angeli, L., Bottai, L., Costantini, R., Ferrari, R., Innocenti, L., et al. (2008). Johnson, C. W., Savabi, M. R., & Loomis, S. A. (1984). Rangeland erosion measure- Assessment of land degradation susceptibility by scenario analysis: A case study ments for the USLE. Transactions of the ASAE, 27,1313e1320. in southern Tuscany, Italy. Geomorphology, 93(1e2), 120e129. Kara, O., S¸ ensoy, H., & Bolat, I. (2010). Slope length effects on microbial biomass and Martin, P. H. (1996). Physics of stamp-collecting? Thoughts on ecosystem model activity of eroded sediments. Journal of Soils and Sediments, 10(3), 434e439. design. Science of the Total Environment, 183(1e2), 7e15. Karydas, C. G., & Panagos, P. (2016). Modelling monthly soil losses and sediment McCool, D. K., Brown, L. C., Foster, G. R., Mutchler, C. K., & Meyer, L. D. (1987). yields in Cyprus. International Journal of Digital Earth, 9(8), 766e787. Revised slope steepness factor for the universal soil loss equation. Transactions Karydas, C. G., Panagos, P., & Gitas, I. Z. (2014). A classification of water erosion of the Asae, 30(5), 1387e1396. models according to their geospatial characteristics. International Journal of McCool, D., Foster, G., & Weesies, G. (1997). Slope length and steepness factors (LS). Digital Earth, 7(3), 229e250. In K. Renard, G. Foster, G. Weesies, D. McCool, & D. Yoder (Eds.), Predicting soil Keesstra, S. D., Bouma, J., Wallinga, J., Tittonell, P., Smith, P., Cerda, A., et al. (2016). erosion by water: A guide to conservation planning with the revised universal soil The significance of soils and soil science towards realization of the United loss equation (RUSLE) (pp. 101e141). USA: Department of Agriculture, Agricul- Nations Sustainable Development Goals. SOIL, 2(2), 111e128. tural Research Service, Washington, District of Columbia. Kilinc, M., & Richardson, E. V. (1973). Mechanics of soil erosion from overland flow Merritt, W. S., Letcher, R. A., & Jakeman, A. J. (2003). A review of erosion and generated by simulated rainfall. Ft. Collins: Colorado State Univ. sediment transport models. Environmental Modelling & Software, 18(8e9), King, C., & Delpont, G. (1993). Spatial assessment of erosion: Contribution of remote 761e799. sensing, a review. Remote Sensing Reviews, 7(3e4), 223e232. Meusburger, K., Banninger, D., & Alewell, C. (2010a). Estimating vegetation Kinnell, P. I. A. (1981). Rainfall intensity kinetic-energy relationships for soil loss parameter for soil erosion assessment in an alpine catchment by means of prediction. Soil Science Society of America Journal, 45(1), 153e155. QuickBird imagery. International Journal of Applied Earth Observation and Geo- Kinnell, P. I. A. (2008). Sediment delivery from hillslopes and the universal soil loss information, 12(3), 201e207. equation: Some perceptions and misconceptions. Hydrological Processes, 22(16), Meusburger, K., Konz, N., Schaub, M., & Alewell, C. (2010b). Soil erosion modelled 3168e3175. with USLE and PESERA using QuickBird derived vegetation parameters in an Kinnell, P. I. A. (2010). Event soil loss, runoff and the universal soil loss equation alpine catchment. International Journal of Applied Earth Observation and Geo- family of models: A review. Journal of Hydrology, 385(1e4), 384e397. information, 12(3), 208e215. Kinnell, P. I. A. (2014). Applying the QREI30 index within the USLE modelling Meusburger, K., Mabit, L., Park, J. H., Sandor, T., & Alewell, C. (2013). Combined use of environment. Hydrological Processes, 28(3), 591e598. stable isotopes and fallout radionuclides as soil erosion indicators in a forested Kinnell, P. I. A., & Risse, L. M. (1998). USLE-M: Empirical modeling rainfall erosion mountain site, South Korea. Biogeosciences, 10(8), 5627e5638. through runoff and sediment concentration. Soil Science Society of America Meusburger, K., Steel, A., Panagos, P., Montanarella, L., & Alewell, C. (2012). Spatial Journal, 62(6), 1667e1672. and temporal variability of rainfall erosivity factor for Switzerland. Hydrology Kirkby, M. J., & Cox, N. J. (1995). A climatic index for soil-erosion potential (csep) and Earth System Sciences, 16. including seasonal and vegetation factors. Catena, 25(1-4), 333e352. Mitasova, H., Hofierka, J., Zlocha, M., & Iverson, L. R. (1996). Modelling topographic Kirkby, M. J., Irvine, B. J., Jones, R. J. A., Govers, G., Boer, M., Cerdan, O., et al. (2008). potential for erosion and deposition using GIS. International Journal of The PESERA coarse scale erosion model for Europe. I. - model rationale and Geographical Information Systems, 10(5), 629e641. implementation. European Journal of Soil Science, 59(6), 1293e1306. Mitasova, H., Hofierka, J., Zlocha, M., & Iverson, L. (1997). Reply to comment by Knisel, W. G. (1980). Creams: A field scale model for chemicals, runoff and erosion from Desmet and Govers. International Journal of Geographical Information Science, agricultural management systems; conservation research report No. 26 (pp. 11(6), 611e618. 1e640). US Departement of Agriculture. Mitasova, H., & Mitas, L. (1999). Modeling soil detachmentwith RUSLE 3D using GIS. Laflen, J. M., Elliot, W. J., Flanagan, D. C., Meyer, C. R., & Nearing, M.a. (1997). WEPP- University of Illinois at Urbana-Champaign. date accessed February 2018 http:// predicting water erosion using a process-based model. Journal of Soil and Water skagit.meas.ncsu.edu/~helena/gmslab/erosion/usle.html. Conservation, 52,96. Moller,€ M., Gerstmann, H., & Glaßer,€ C. (2014). Dyna C : A framework for the dynamic Laflen, J. M., & Flanagan, D. C. (2013). The development of U. S. soil erosion pre- derivation of the usle C -factor using high temporal multi-spectral satellite imagery. diction and modeling. International Soil and Water Conservation Research, 1(2), Mondal, A., Khare, D., Kundu, S., Mukherjee, S., Mukhopadhyay, A., & Mondal, S. 1e11. (2017). Uncertainty of soil erosion modelling using open source high resolution Laflen, J. M., Foster, G. R., & Onstad, C. A. (1985). Simulation of individual- storm soil and aggregated DEMs. Geoscience Frontiers, 8(3), 425e436. loss for modeling the impact of soil erosion on crop productivity. Soil Erosion Montanarella, L. (2015). Govern our soils. Nature, 528(7580), 32e33. and Conservation, 285e295. Moore, I. D., & Burch, G. J. (1986). Physical basis of the length-slope factor in the Laflen, J. M., & Moldenhauer, W. C. (2003). Pioneering soil erosion prediction e the universal soil loss equation. Soil Science Society of America Journal, 50(5), USLE story. Bejing, China: World Association of Soil and Water Conservation. 1294e1298. C. Alewell et al. / International Soil and Water Conservation Research 7 (2019) 203e225 223

Morgan, R. P. C. (2005). Soil loss conservation: Problems and prospects. Chichester: Total Environment, 479e480(1), 189e200. John Wiley & Sons. Panagos, P., Meusburger, K., Ballabio, C., Borrelli, P., Begueria, S., Klik, A., et al. Morgan, R. P. C., & Nearing, M. A. (2011). Handbook of Erosion Modelling (pp. 1e401). (2015e). Reply to the comment on "Rainfall erosivity in Europe" by Auerswald John Wiley and Sons. et al. Science of the Total Environment, 532, 853e857. Morgan, R. P. C., Quinton, J. N., Smith, R. E., Govers, G., Poesen, J. W. A., Auerswald, K., Panagos, P., Meusburger, K., Van Liedekerke, M., Alewell, C., Hiederer, R., & et al. (1998). The European soil erosion model (EUROSEM): A process-based Montanarella, L. (2014c). Assessing soil erosion in Europe based on data approach for predicting soil loss from fields and small catchments. , 527-544. collected through a European network. Soil Science and Plant Nutrition, 60(1), Earth Surface Processes and Landforms, 23(6), 527e544. 15e29. Morton, A., & Súarez, M. (2001). Kinds of models. In M. G. Anderson, & P. D. Bates Peralta, C., Beriain, E., Iglesias, C., Marangunic, C., Marangunic, A., & Casassa, G. (Eds.), Model validation: Perspectives in hydrological science (pp. 11e21). Chi- (2017). Ground topography validation of a high-resolution DEM and ortho- chester: John Wiley & Sons. photos obtained with a commercial UAS. In 2017 first IEEE international sym- Musgrave, G. W. (1947). The quantitative evaluation of factors in water erosion, a posium of geoscience and remote sensing (GRSS-Chile) (Vol. 4, p. 4). first approximation. Journal of Soil and Water Conservation, 2,133e138. Pineiro, G., Paruelo, J. M., Oesterheld, M., & Jobbagy, E. G. (2010). Pathways of Nash, J. E., & Sutcliffe, J. E. (1970). River flow forecasting through conceptual models. grazing effects on soil organic carbon and nitrogen. Rangeland Ecology & Part 1 - a discussion of principles. Journal of Hydrology, 10, 282e290. Management, 63(1), 109e119. Nearing, M. A. (1997). A single, continuous function for slope steepness influence on Podmanicky, L., Balazs, K., Belenyesi, M., Centeri, C., Kristof, D., & Kohlheb, N. (2011). soil loss. Soil Science Society of America Journal, 61(3), 917e919. Modelling soil quality changes in Europe. An impact assessment of land use Nearing, M. (2004). Soil erosion and conservation. In J. Wainwright, & M. Mulligan change on soil quality in Europe. Ecological Indicators, 11(1), 4e15. (Eds.), Environmental modelling: Finding simplicity in complexity (pp. 277e290). Poesen, J. W., Torri, D., & Bunte, K. (1994). Effects of rock fragments on soil-erosion John Wiley & Sons, Ltd. by water at different spatial scales - a review. Catena, 23(1e2), 141e166. Nearing, M. A. (2013). Soil erosion and conservation. Porto, P. (2016). Exploring the effect of different time resolutions to calculate the Nearing, M. A., Govers, G., & Norton, L. D. (1999). Variability in soil erosion data from rainfall erosivity factor R in Calabria, southern Italy. Hydrological Processes, replicated plots. Soil Science Society of America Journal, 63(6), 1829. 30(10), 1551e1562. Nearing, M. A., Yin, S. Q., Borrelli, P., & Polyakov, V. O. (2017). Rainfall erosivity: An Quinn, P., Beven, K., Chevallier, P., & Planchon, O. (1991). The prediction of hillslope historical review. Catena, 157. flow paths for distributed hydrological modelling using digital terrain models. Nicks, A. D., Lane, L. J., & Gander, G. A. (1995). Weather generator (chapter 2). In Hydrological Processes, 5(1), 59e79. D. C. Flanagan, & M. A. Nearing (Eds.), USDA water erosion prediction project: Quinton, J. (2013). Erosion and sediment transport. In J. Wainwright, & M. Mulligan Hillslope profile and watershed model documentation- NSERL report. USDA-ARS (Eds.), Environmental modelling: Finding simplicity in complexity (pp. 187e196). National Soil Erosion Research Laboratory. John Wiley & Sons, Ltd. Nyssen, J., Vandenreyken, H., Poesen, J., Moeyersons, J., Deckers, J., Haile, M., et al. Rapp, J. F. (1994). Error assessment of the revised universal soil loss equation using (2005). Rainfall erosivity and variability in the northern Ethiopian highlands. natural runoff plot data, Master Thesis. Arizona: The University of Arizona. Journal of Hydrology, 311(1e4), 172e187. Refsgaard, J. C., & Storm, B. (1996). Construction, calibration and validation of hy- O'Callaghan, J. F., & Mark, D. M. (1984). The extraction of drainage networks from drological models. In M. B. Abbott, & J. C. Refsgaard (Eds.), Distributed hydro- digital elevation data. Computer Vision, Graphics, and Image Processing, 28(3), logical modelling (pp. 41e45). Dordrecht: Reidel. 323e344. Rejman, J., & Brodowski, R. (2005). Rill characteristics and sediment transport as a Oldeman, L. R. (1992). Global extent of soil degradation. ISRIC Bi-annual report 1991- function of slope length during a storm event on loess soil. Earth Surface Pro- 1992. cesses and Landforms, 30(2), 231e239. Oliveira, P. T. S., Wendland, E., & Nearing, M. A. (2013). Rainfall erosivity in Brazil: A Renard, K. G., & Ferreira, V. A. (1993). Rusle model description and database review. Catena, 100,139e147. sensitivity. Journal of Environmental Quality, 22(3), 458e466. Oreskes, N., Shraderfrechette, K., & Belitz, K. (1994). Verification, validation, and Renard, K. G., & Foster, G. R. (1985). Managing rangeland soil resources: The uni- confirmation of numerical-models in the earth-sciences. Science, 263(5147), versal soil loss equation. Rangelands, 7(3), 118e122. 641e646. Renard, K., Foster, G., Weesies, G., McCool, D., & Yoder, D. (1997). Predicting soil Orgiazzi, A., Ballabio, C., Panagos, P., Jones, A., & Fernandez-Ugalde, O. (2018). LUCAS erosion by water: A guide to conservation planning with the revised universal soil soil, the largest expandable soil dataset for Europe: A review. European Journal loss equation (RUSLE). Agriculture handbook 703. Washington, District of of Soil Science, 69(1), 140e153. Columbia, USA: Department of Agriculture, Agricultural Research Service. Ozsoy, G., Aksoy, E., Dirim, M. S., & Tumsavas, Z. (2012). Determination of soil Renard, K. G., Foster, G. R., Weesies, G. A., & Porter, J. P. (1991). Rusle - revised erosion risk in the mustafakemalpasa river basin, Turkey, using the revised universal soil loss equation. Journal of Soil and Water Conservation, 46(1), 30e33. universal soil loss equation, geographic information system, and remote Risal, A., Bhattarai, R., Kum, D., Park, Y. S., Yang, J. E., & Lim, K. J. (2016). Application sensing. Environmental Management, 50(4), 679e694. of Web ERosivity module (WERM) for estimation of annual and monthly R Panagos, P., Ballabio, C., Borrelli, P., Meusburger, K., Klik, A., Rousseva, S., et al. factor in Korea. Catena, 147, 225e237. (2015a). Rainfall erosivity in Europe. Science of the Total Environment, 511, Risse, L. M., Nearing, M. A., Laflen, J. M., & Nicks, A. D. (1993). Error assessment in 801e814. the universal soil loss equation. Soil Science Society of America Journal, 57(3), Panagos, P., Borrelli, P., Meusburger, K., Alewell, C., Lugato, E., & Montanarella, L. 825. (2015b). Estimating the soil erosion cover-management factor at the European Rodriguez, J. L. G., & Suarez, M. C. G. (2012). Methodology for estimating the scale. Land Use Policy, 48,38e50. topographic factor LS of RUSLE3D and USPED using GIS. Geomorphology, 175, Panagos, P., Borrelli, P., Meusburger, K., van der Zanden, E. H., Poesen, J., & Alewell, C. 98e106. (2015c). Modelling the effect of support practices (P-factor) on the reduction of Romkens,€ M., Poesen, J., & Wang, J. (1988). Relationship between the USLE soil soil erosion by water at European scale. Environmental Science and Policy, 51. erodibility factor and soil properties. Proceedings,371e385. Panagos, P., Borrelli, P., Meusburger, K., Yu, B., Klik, A., Jae Lim, K., et al. (2017a). Rosewell, C. J. (1993). SOILOSS: A program to assist in the selection of management Global rainfall erosivity assessment based on high-temporal resolution rainfall practices to reduce erosion. Technical handbook (Vol. 11). Soil Conservation Ser- records. Scientific Reports, 7(1), 4175. vice of New South Wales. Panagos, P., Borrelli, P., Meusburger, K., Yu, B., Klik, A., Lim, k.j., et al. (2017b). Global Rossel, R. A. V., & Behrens, T. (2010). Using data mining to model and interpret soil rainfall erosivity assessment based on high-temporal resolution rainfall records. diffuse reflectance spectra. Geoderma, 158(1), 46e54. Scientific Reports. In review(February): 1-12. Rüttimann, M., Schaub, D., Prasuhn, V., & Rüegg, W. (1995). Measurement of runoff Panagos, P., Borrelli, P., Poesen, J., Ballabio, C., Lugato, E., Meusburger, K., et al. and soil erosion on regularly cultivated fields in Switzerland d some critical (2015d). The new assessment of soil loss by water erosion in Europe. Environ- considerations. Catena, 25(1), 127e139. mental Science & Policy, 54, 438e447. Sadeghi, S. H., Zabihi, M., Vafakhah, M., & Hazbavi, Z. (2017). Spatiotemporal Panagos, P., Borrelli, P., Poesen, J., Meusburger, K., Ballabio, C., Lugato, E., et al. mapping of rainfall erosivity index for different return periods in Iran. Natural (2016b). Reply to the comment on "The new assessment of soil loss by water Hazards, 87(1), 35e56. erosion in Europe" by Fiener & Auerswald. Environmental Science & Policy, 57, Scheurer, K., Alewell, C., Banninger, D., & Burkhardt-Holm, P. (2009). Climate and 143e150. land-use changes affecting river sediment and brown trout in alpine countries- Panagos, P., Borrelli, P., Poesen, J., Meusburger, K., Ballabio, C., Lugato, E., a review. Environmental Science and Pollution Research, 16(2), 232e242. Montanarella, L., & Alewell, C. (2016a). Reply to "The new assessment of soil loss Schindler Wildhaber, Y., Michel, C., Burkhardt-Holm, P., Baenninger, D., & Alewell, C. by water erosion in Europe. In P. Panagos, et al. (Eds.), 2015 Environ. Sci. Policy 54 (2012). Measurement of spatial and temporal fine sediment dynamics in a small (Vol. 59, pp. 438e447). Environmental Science & Policy. A response" by Evans river. Hydrology and Earth System Sciences, 16(5), 1501e1515. and Boardman Environ. Sci. Policy 58, 11-15. 53-57. Schmidt, S., Alewell, C., & Meusburger, K. (2018). Mapping spatio-temporal dy- Panagos, P., Christos, K., Cristiano, B., & Ioannis, G. (2014a). Seasonal monitoring of namics of the cover and management factor (C-factor) for grasslands in soil erosion at regional scale: An application of the G2 model in Crete focusing Switzerland. Remote Sensing of Environment, 211,89e104. on agricultural land uses. International Journal of Applied Earth Observation and Schmidt, S., Alewell, C., & Meusburger, K. (2019). Monthly RUSLE soil erosion risk of Geoinformation, 27,147e155. Swiss grasslands. Journal of Maps,1e10. Panagos, P., Imeson, A., Meusburger, K., Borrelli, P., Poesen, J., & Alewell, C. (2016c). Schmidt, S., Alewell, C., Panagos, P., & Meusburger, K. (2016). Regionalization of Soil conservation in Europe: Wish or reality? Land Degradation & Development, monthly rainfall erosivity patterns in Switzerland. Hydrology and Earth System 27(6), 1547e1551. Sciences, 20(10), 4359e4373. Panagos, P., Meusburger, K., Ballabio, C., Borrelli, P., & Alewell, C. (2014b). Soil Schonbrodt,€ S., Saumer, P., Behrens, T., Seeber, C., & Scholten, T. (2010). Assessing the erodibility in Europe: A high-resolution dataset based on LUCAS. Science of the USLE crop and management factor C for soil erosion modeling in a large 224 C. Alewell et al. / International Soil and Water Conservation Research 7 (2019) 203e225

mountainous watershed in Central China. Journal of Earth Science, 21, 835e845. processing of digital elevation data using a Cþþ executable. Computers & Schwertmann, U., Vogl, W., & Kainz, M. (1987). Bodenerosion durch Wasser - Geosciences, 30(9e10), 1043e1053. vorhersage des Abtrags und Bewertung von Gegenmaßnahmen. Stuttgart: Verlag Van Rompaey, A. J. J., Verstraeten, G., Van Oost, K., Govers, G., & Poesen, J. (2001). Eugen Ulmer. Modelling mean annual sediment yield using a distributed approach. Earth Seibert, J., & McGlynn, B. L. (2007). A new triangular multiple flow direction algo- Surface Processes and Landforms, 26(11), 1221e1236. rithm for computing upslope areas from gridded digital elevation models. Van der Knijff, J., Jones, R. R. J., & Montanarella, L. (1999). Soil erosion risk assessment Water Resources Research, 43(4). in Italy. Luxemburg: Off. Publ. Eur. Communities EUR 19022. Senarath, S. U. S., Ogden, F., Downer, C. W., & Sharif, H. O. (2000). On the calibration Vannoppen, W., Vanmaercke, M., De Baets, S., & Poesen, J. (2015). A review of the and verification of two-dimensional, distributed, Hortonian, continuous mechanical effects of plant roots on concentrated flow erosion rates. Earth- watershed models. Water Resources Research, 36,1495e1510. Science Reviews, 150,666e678. Shepherd, K. D., & Walsh, M. G. (2002). Development of reflectance spectral libraries Veihe, A., & Quinton, J. (2000). Sensitivity analysis of EUROSEM using Monte Carlo for characterization of soil properties. Soil Science Society of America Journal, simulation I: Hydrological, soil and vegetation parameters. Hydrological Pro- 66(3), 988e998. cesses, 14(5), 915e926. Silburn, D. M., & Freebairn, D. M. (1992). Evaluations of the creams model .3. Veihe, A., Quinton, J., & Poesen, J. (2000). Sensitivity analysis of EUROSEM using simulation of the hydrology of vertisols. Australian Journal of Soil Research, Monte Carlo simulation II: The effect of rills and rock fragments. Hydrological 30(5), 547e564. Processes, 14(5), 927e939. Silburn, D. M., & Loch, R. J. (1989). Evaluation of the creams model .1. sensitivity Verheijen, F. G. A., Jones, R. J. A., Rickson, R. J., & Smith, C. J. (2009). Tolerable versus analysis of the soil-erosion sedimentation component for aggregated clay soils. actual soil erosion rates in Europe. Earth-Science Reviews, 94(1e4), 23e38. Australian Journal of Soil Research, 27(3), 545e561. Vrieling, A. (2006). Satellite remote sensing for water erosion assessment: A review. Simanton, J., Roger, H. B., Herbert, B., & Renard, K. G. (1980). Application of the USLE Catena, 65(1), 2e18. to southwestern rangelands. Hydrology and water resources in Arizona and the Wall, G. J., Coote, D. R., Pringle, E. A., & Shelton, I. J. (2002). RUSLEFAC d revised Southwest. universal soil loss equation for application in Canada: A handbook for estimating Sivertun, A., & Prange, L. (2003). Non-point source critical area analysis in the soil loss from water erosion in Canada. ECORC contribution number 02-92, AAFC Gisselo watershed using GIS. Environmental Modelling & Software, 18(10), contribution number AAFC/AAC2244E. Ottawa, Ontario: Research Branch Agri- 887e898. culture and Agri-Food Canada. Smith, D. D. (1941). Interpretation of soil conservation data for field use. Agricultural Wall, G. J., Dickinson, W. T., Rudra, R. P., & Coote, D. R. (1988). Seasonal soil erod- Engineering, 22,173e175. ibility variation in southwestern ontario. Canadian Journal of Soil Science, 68(2), Smith, D. D., & Whitt, D. M. (1948). Evaluating soil losses from field areas. Agri- 417e424. cultural Engineering, 29, 394e396. Wang, G. Q., Fang, Q. Q., Teng, Y. G., & Yu, J. S. (2016). Determination of the factors Smith, S. J., Williams, J. R., Menzel, R. G., & Coleman, G. A. (1984). Prediction of governing soil erodibility using hyperspectral visible and near-infrared reflec- sediment yield from southern plains grasslands with the modified universal soil tance spectroscopy. International Journal of Applied Earth Observation and Geo- loss equation. Journal of Range Management, 37(4), 295e297. information, 53,48e63. Smith, D. D., & Wischmeier, W. H. (1957). Factors affecting sheet and rill erosion. Wang, B., Zheng, F., & Wang, Y. (2012). Adaptability analysis on soil erodibility American Geophysical Union Transaction, 38, 889e896. models in typical thin layer black soil area of Northeast China. Nongye Gong- Spaeth, K. E., Pierson, F. B., Weltz, M.a., & Blackburn, W. H. (2003). Evaluation of cheng Xuebao/Transactions of the Chinese Society of Agricultural Engineering, USLE and RUSLE estimated soil loss on rangeland. Journal of Range Managment, 28(6), 126e131. 56, 234e246. Warren, S. D., Mitasova, H., Hohmann, M. G., Landsberger, S., Iskander, F. Y., Stenberg, B., & Rossel, R. A. V. (2010). Diffuse reflectance spectroscopy for high- Ruzycki, T. S., et al. (2005). Validation of a 3-D enhancement of the Universal resolution soil sensing. In R. A. Viscarra Rossel, A. B. McBratney, & B. Minasny Soil Loss Equation for prediction of soil erosion and sediment deposition. Ca- (Eds.), Proximal soil sensing (pp. 29e47). Dordrecht: Springer Netherlands. tena, 64(2e3), 281e296. Stolpe, N. B. (2005). A comparison of the RUSLE, EPIC and WEPP erosion models as Wawer, R., & Nowocien, E. (2006). Cyfrowa mapa wystepowania erozji wodnej calibrated to climate and soil of south-central Chile. Acta Agriculturae Scandi- powierzchniowej obszaru Polski z wykorzystaniem CORINE Land Cover 2000. navica Section B Soil and Plant Science, 55(1), 2e8. Rozdzielczosc przestrzenna 500 m. Roczniki Akademii Rolniczej w Poznaniu. Sverdrup, H., Warfvinge, P., Blake, L., & Goulding, K. (1995). Modeling recent and Rolnictwo, 65, 207e213. historic soil data from the rothamsted-experimental-station, UK using safe. Weltz, M. A., Kidwell, M., & Fox, H. D. (1998). Influence of abiotic and biotic factors Agriculture Ecosystems & Environment, 53(2), 161e177. in measuring and modeling soil erosion on rangelands: State of knowledge. Tan, Z., Leung, L. R., Li, H.-Y., & Tesfa, T. (2018). Modeling sediment yield in land Journal of Range Management, 51, 482e495. surface and earth system models: Model comparison, development, and eval- Weltz, M. A., Renard, K. G., & Simanton, J. R. (1987). Revised universal soil loss uation. Journal of Advances in Modeling Earth Systems, 0(0). equation for western rangelands, General technical report RM-Rocky Mountain Teng, H. F., Rossel, R. A. V., Shi, Z., Behrens, T., Chappell, A., & Bui, E. (2016). Forest and Range Experiment Station. USA): US Department of Agriculture, Forest Assimilating satellite imagery and visible-near infrared spectroscopy to model Service. and map soil loss by water erosion in Australia. Environmental Modelling & Wendt, R. C., Alberts, E. E., & Hjelmfelt, A. T. (1986). Variability of runoff and soil loss Software, 77,156e167. from fallow experimental plots. Soil Science Society of America Journal, 50(3), Tiwari, A. K., Risse, L. M., & Nearing, M. A. (2000). Evaluation of WEPP and its 730e736. comparison with USLE and RUSLE. Transactions of the Asae, 43(5), 1129e1135. Williams, J. R. (1975). Sediment-yield prediction with universal equation using runoff Torri, D., Poesen, J., & Borselli, L. (1997). Predictability and uncertainty of the soil energy factor, present and prospective technology for predicting sediment yield and erodibility factor using a global dataset. Catena, 31(1), 1e22. sources. US Department of Agriculture. Toy, T. J., F, G. R., & R, K. G. (2002). Soil erosion: Processes, prediction, measurement, Williams, J. R., & Berndt, H. D. (1977). Sediment yield prediction based on watershed and control. New York: John Wiley & Sons, Inc. hydrology. Transactions of the Asae, 20(6), 1100e1104. Tresch, S. (2014). Determination of the S-factor in steep slopes. Master Thesis,. Basel: Williams, J. R., Renard, K. G., & Dyke, P. T. (1983). Epic: A new method for assessing University Basel. erosion's effect on soil productivity. Journal of Soil and Water Conservation, Trimble, S. W., & Crosson, P. (2000). U.S. Soil erosion rates–myth and reality. Science, 38(5), 381e383. 289(5477), 248e250. Winchell, M., Jackson, S., Wadley, A., & Srinivasan, R. (2008). Extension and vali- USDA. (1961). US agricultural research service. a universal equation for predicting dation of a geographic information system-based method for calculating the rainfall-erosion losses. U.S. Dept. Agr. ARS, 22e66,1e11. Revised Universal Soil Loss Equation length-slope factor for erosion risk as- USDA. (2008). Revised universal soil loss equation version 2 (RUSLE2). Science docu- sessments in large watersheds. Journal of Soil and Water Conservation, 63(3), mentation. Washington, D.C: USDA-Agricultural Research Service. 105e111. USDA. (2018). Revised universal soil loss equation 2 - download RUSLE2. https://www. Wischmeier, W. H. (1959). A rainfall erosion index for a universal SOIL-loss equa- ars.usda.gov/southeast-area/oxford-ms/national-sedimentation-laboratory/ tion. Soil Science Society of America Proceedings, 23, 246e249. watershed-physical-processes-research/research/rusle2/revised-universal-soil- Wischmeier, W. H. (1960). Cropping-management factor evaluations for a universal loss-equation-2-download-rusle2/. soil-loss equation. Soil Science Society of America Proceedings, 24, 322e326. Van Oost, K., Govers, G., & Desmet, P. (2000a). Evaluating the effects of changes in Wischmeier, W. H. (1966). Relation of field-plot runoff to management and physical landscape structure on soil erosion by water and tillage. Landscape Ecology, Factors1. Soil Science Society of America Journal, 30(2), 272e277. 15(6), 577e589. Wischmeier, W. H. (1975). Estimating the soil loss equations cover and management Van Oost, K., Govers, G., Van Muysen, W., & Quine, T. A. (2000b). Modeling trans- factor for undisturbed areas, Present and prospective technology for predicting location and dispersion of soil constituents by tillage on sloping land. Soil Sci- sediment yields and sources. Oxford. ence Society of America Journal, 64(5), 1733e1739. Wischmeier, W. H. (1976). Use and misuse of the universal soil loss equation. Journal Van Oost, K., Quine, T. A., Govers, G., De Gryze, S., Six, J., Harden, J. W., et al. (2007). of Soil and Water Conservation, 3,5e9. The impact of agricultural soil erosion on the global carbon cycle. Science, 318, Wischmeier, W. H., Johnson, C. B., & Cross, B. V. (1971). A soil erodibility nomograph 626e629. for farmland and construction sites. Journal of Soil and Water Conservation, 26, Van Remortel, R. D., Hamilton, M. E., & Hickey, R. J. (2001). Estimating the LS factor 189e193. for RUSLE through iterative slope length processing of digital elevation data Wischmeier, W. H., & Smith, D. D. (1958). Evaluation of factors in the soil-loss within arclnfo grid. Cartography, 30(1), 27e35. equation. Agricultural Engineering, 39(8), 458e462. Van Remortel, R. D., Maichle, R. W., & Hickey, R. J. (2004). Computing the LS factor Wischmeier, W. H., & Smith, D. D. (1965). Predicting rainfall-erosion losses from for the Revised Universal Soil Loss Equation through array-based slope cropland east of the Rocky Mountains. Agriculture Handbook (Vol 282). C. Alewell et al. / International Soil and Water Conservation Research 7 (2019) 203e225 225

Washington DC: US Department of Agriculture. risk. Land Degradation & Development, 20(1), 84e91. Wischmeier, W. H., & Smith, D. D. (1978). Predicting rainfall erosion losses - a guide to Zhang, X., Drage, N. A., & Wainwright, J. (2004b). Scaling issues in environmental conservation planning, agriculture handbook No. 537. USDA/Science and Educa- modelling. In J. Wainwright, & M. Mulligan (Eds.), Environmental modelling: tion Administration, US. Govt. Printing Office, Washington D.C. Finding simplicity in complexity (pp. 319e334). John Wiley & Sons, Ltd. Xu, X. L., Liu, W., Kong, Y. P., Zhang, K. L., Yu, B., & Chen, J. D. (2009). Runoff and Zhang, K., Li, S., Peng, W., & Yu, B. (2004a). Erodibility of agricultural soils on the water erosion on road side-slopes: Effects of rainfall characteristics and slope loess plateau of China. Soil and Tillage Research, 76(2), 157e165. length. Transportation Research Part D: Transport and Environment, 14(7), Zhang, X. C., Nearing, M. A., Risse, L. M., & McGregor, K. C. (1996b). Evaluation of 497e501. runoff and soil loss predictions using natural runoff plot data. Transactions of the Xu, X., Miao, C., & Liao, Y. (2008). Quantitative remote sensing study on regional soil American Society of Agricultural Engineers, 39(3), 855e863. erosion. Wuhan University Journal of Natural Sciences, 13(3), 377e384. Zhang, X. C., Nearing, M. A., Risse, L. M., & McGregor, K. C. (1996c). Evaluation of Yair, A., & Raz-Yassif, N. (2004). Hydrological processes in a small arid catchment: WEPP runoff and soil loss predictions using natural runoff plot data. Trans- Scale effects of rainfall and slope length. Geomorphology, 61(1e2), 155e169. actions of the Asae, 39(3), 855e863. Yang, D., Kanae, S., Oki, T., Koike, T., & Musiake, K. (2003a). Global potential soil Zhang, L., O'Neill, A. L., & Lacey, S. (1996a). Modelling approaches to the prediction erosion with reference to land use and climate changes. Hydrological Processes, of soil erosion in catchments. Environmental Software, 11(1), 123e133. 17(14), 2913e2928. Zhang, K. L., Shu, A. P., Xu, X. L., Yang, Q. K., & Yu, B. (2008). Soil erodibility and its Yang, D. W., Kanae, S., Oki, T., Koike, T., & Musiake, K. (2003b). Global potential soil estimation for agricultural soils in China. Journal of Arid Environments, 72(6), erosion with reference to land use and climate changes. Hydrological Processes, 1002e1011. 17(14), 2913e2928. Zhang, F. B., Wang, Z. L., & Yang, M. Y. (2014). Validating and improving interrill Yin, S., Xie, Y., Liu, B., & Nearing, M. A. (2015). Rainfall erosivity estimation based on erosion equations. PLoS One, 9(2), 11. rainfall data collected over a range of temporal resolutions. Hydrology and Earth Zhang, H., Yang, Q., Li, R., Liu, Q., Moore, D., He, P., et al. (2013). Extension of a GIS System Sciences, 19(10), 4113e4126. procedure for calculating the RUSLE equation LS factor. Computers & Geo- Young, R. A., Onstad, C. A., Bosch, D. D., & Anderson, W. P. (1989). AGNPS - A sciences, 52,177e188. nonpoint-source pollution model for evaluating agricultural watersheds. Jour- Zhang, W., Zhang, Z., Liu, F., Qiao, Z., & Hu, S. (2011). Estimation of the USLE cover nal of Soil and Water Conservation, 44(2), 168e173. and management factor C using satellite remote sensing: A review. In Pro- Yue, Y., Ni, J. R., Ciais, P., Piao, S. L., Wang, T., Huang, M. T., et al. (2016). Lateral ceedings - 2011 19th international conference on geoinformatics, geoinformatics transport of soil carbon and land-atmosphere CO2 flux induced by water 2011. erosion in China. Proceedings of the National Academy of Sciences of the United Zingg, A. W. (1940a). Degree and length of land slope as it affects soil loss in runoff. States of America, 113(24), 6617e6622. Agricultural Engineering, 21(2), 59e64. Zhang, Y., Degroote, J., Wolter, C., & Sugumaran, R. (2009). Integration of modified Zingg, A. W. (1940b). Degree and length of land slope as it affects soil loss in runoff. universal soil loss equation (musle) into a gis framework to assess soil erosion Agricultural Engineering, 21,59e64. International Soil and Water Conservation Research 7 (2019) 226e238

Contents lists available at ScienceDirect

International Soil and Water Conservation Research

journal homepage: www.elsevier.com/locate/iswcr

Original Research Article Effect of hysteresis on the stability of residual soil slope

* Christofer Kristo, Harianto Rahardjo , Alfrendo Satyanaga

School of Civil and Environmental Engineering, Nanyang Technological University, Hall of Residence 9, Block 46, #04-867, Singapore, 639811, Singapore article info abstract

Article history: Soil-water characteristic curve (SWCC) is an essential parameter in unsaturated soil mechanics. Matric Received 25 April 2018 suction in unsaturated residual soils changes with varying climatic conditions associated with cyclic Received in revised form drying-wetting conditions that result in hysteresis in the SWCC. The soil mechanical behaviour under 6 May 2019 wetting is crucial since numerous rainfall-induced slope failures occur during the wetting process. Accepted 15 May 2019 However, many slope stability analyses were carried out using drying SWCC. Consequently, the factor of Available online 24 May 2019 safety (FoS) calculation may not represent the actual ﬁeld condition. This paper presents the effect of hysteresis in SWCC on the stability of unsaturated residual soil slopes from Bukit Timah Granite in Keywords: SWCC Singapore. The study focused on the analyses considering the differences in pore-water pressure and Hysteresis water content variations under drying and wetting conditions by performing numerical seepage and Rainfall stability analyses. Each analysis was carried out on a slope subjected to dry and rainy periods under three Seepage different conditions: i) using only drying SWCC; ii) using only wetting SWCC; iii) using combined drying Slope stability and wetting SWCCs. The results indicated that the FoS variations obtained from the numerical analyses incorporating combined SWCC matched more closely those obtained by incorporating only wetting SWCC than those obtained by incorporating only drying SWCC, regardless of the wetting or drying processes that the soil experienced. Moreover, the numerical analyses under high rainfall intensity by incorporating only wetting SWCC gave a more conservative FoS as compared to those obtained by incorporating combined SWCC. Numerical analyses incorporating only drying SWCC gave the most conservative FoS regardless of rainfall intensity. © 2019 International Research and Training Center on Erosion and Sedimentation and China Water and Power Press. Production and Hosting by Elsevier B.V. This is an open access article under the CC BY-NC- ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction and matric suction (ua uw), are generally used in practice to distinguish the effects of normal stress and pore-water pressure 1.1. Unsaturated residual soils (uw) changes, respectively (Fredlund & Morgenstern, 1977; Fredlund & Rahardjo, 1993; Matyas & Radhakrishna, 1968). The net Residual soils are commonly found to have in-situ negative normal stress is equivalent to the total normal stress ðsÞ less the pore-water pressures within the unsaturated zone above the pore-air pressure. groundwater table (Fredlund & Rahardjo, 1993). In addition to air, Matric suction can be defined as a negative pore-water pressure water and solid phases, unsaturated residual soils have an air-water referenced to the pore-air pressure. An increase in matric suction interface or contractile skin as a fourth phase (Fredlund & causes a decrease in the soil volumetric water content as described Morgenstern, 1977). This contractile skin acts like an elastic in the SWCC. SWCC is the most important soil property in unsat- membrane and influences the mechanical behaviour of the soil by urated residual soil mechanics. It can be used to predict other un- pulling the soil particles together through surface tension. Since saturated residual soil properties, such as permeability function. atmospheric pore-air pressure (ua) is adopted in most geotechnical The movement of water through the soil will be affected due to less engineering practice, pore-air pressure is taken as a reference and water-filled spaces available for water flow. Thus, the permeability two independent stress state variables, net normal stress (s ua) of the soil also decreases with increasing matric suction, represented by the permeability function. Moreover, as matric suction increases, the air/water interface exerts a tension force on the soil * Corresponding author. particles giving the soil an added value of shear strength. In E-mail addresses: [email protected] (C. Kristo), [email protected] contrast, when infiltration occurs, the matric suction decreases, and (H. Rahardjo), [email protected] (A. Satyanaga). https://doi.org/10.1016/j.iswcr.2019.05.003 2095-6339/© 2019 International Research and Training Center on Erosion and Sedimentation and China Water and Power Press. Production and Hosting by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). C. Kristo et al. / International Soil and Water Conservation Research 7 (2019) 226e238 227 the shear strength of the soil is reduced (Kim, Jeong, Park, & drying or wetting SWCC) should be used in accordance with the Sharma, 2004; Krahn, Fredlund, & Klassen, 1989; Rahardjo et al. process that the soils actually experience (i.e., drying process or 1995, 2004). Therefore, infiltration has a significant effect on the wetting process) even under a steady-state rainfall condition. unsaturated shear strength of soil. (Gasmo, Rahardjo, & Leong, When both drying and wetting SWCCs are available, numerical 2000). Historical events have shown that many slope failures analysis for simulating the infiltration process should incorporate occurred during or shortly after rainfall (Gavin & Xue, 2008). Thus, the wetting SWCC since soil gains moisture, while the drying SWCC SWCC is indispensable in conducting seepage and slope stability should be used when simulating evaporation and drainage since analyses for unsaturated residual soils. soil loses moisture in the process. In addition, Zhai, Rahardjo, and Satyanaga (2016) suggested that the upper limit of SWCC pro- 1.2. Hysteresis in SWCC duces the most conservative result for rainfall-induced slope failures analyses, while the lower limit of SWCC is recommended to be The SWCC describes the relationship between the amount of used for conservative shear strength estimation. water in a soil and matric suction. It is commonly presented in a Common laboratory experiments using Tempe Cell and Pressure graph of either gravimetric water content (w), volumetric water Plate to determine SWCC only yield few data points along the content (qw) or degree of saturation (S) in the vertical axis against drying and wetting SWCCs. For the purpose of numerical analyses, matric suction in the horizontal axis in logarithmic scale (Fredlund, a continuous SWCC along the matric suction range is needed. This Rahardjo, & Fredlund, 2012). Among the three possible represen- could be obtained by best-fitting the discrete data points from tations of SWCC, SWCC in the form of degree of saturation laboratory experiments using the Fredlund and Xing (1994) equa- (S-SWCC) incorporates soil volume change calculations and is tion, developed using the least-square method: considered to be analogous to the pore-size distribution function. Thus, S-SWCC is adopted as the probability function of random pore þ j q ln 1 j ¼ ð Þ s ; ð Þ¼ r connections (Zhai & Rahardjo, 2015), since the area under the pore- qw C j n h iom C j 1 (1) n 106 fi ln e þ j ln 1 þ suction distribution function de nes the degree of saturation. a jr An important parameter of the SWCC for unsaturated residual fi soil mechanics is the air-entry value (AEV), de ned as the matric where: qw ¼ Volumetric water content at any matric suction value, suction where air starts to enter the largest pores in the soil (Brooks qs ¼ Saturated volumetric water content, j ¼ Suction of the soil, a & Corey, 1966). Alternatively, AEV could be equivalently defined as ¼ Fitting parameter related to the soil suction at the inflection the matric suction that breaks the meniscus formed by water sur- point, related to AEV of the soil, n ¼ Fitting parameter related to the face tension in the largest pores (Fredlund & Xing, 1994). The AEV slope of SWCC, m ¼ Fitting parameter related to residual water has been found to depend on the grain size distribution of the soil. content of the soil, CðjÞ¼Correction function that forces SWCC A larger proportion of fine particles implies smaller intra-particle 6 through a suction of 10 kPa and zero water content, and jr pore spaces between soil particles resulting in a higher AEV ¼ Suction at which residual water content occurs. (Aung, Rahardjo, Leong, & Toll, 2001). On the other hand, the re- Leong and Rahardjo (1997a) suggested that CðjÞ¼1 in Equation sidual suction refers to the matric suction beyond which there is no (1) to obtain a better SWCC fitting in the low-suction range (less significant removal of water from the soil. than 500 kPa), albeit compromising the fitting accuracy at the It is observed that different degrees of saturation can exist in an higher suction range. unsaturated residual soil under the same matric suction when In addition, due to the close relationship between SWCC and subjected to cyclic drying and wetting in nature due to climatic coefficient of permeability, ks (Childs & Collis-George, 1950), it is conditions (Rahardjo et al., 2004). This indicates a hysteretic possible to estimate the permeability function of unsaturated re- behaviour in SWCC of soil. The volumetric water content is found to sidual soils from the available SWCC and ks. Mualem (1976) clas- be lower along the wetting curve than that along the drying curve sified various prediction models for the permeability functions by for the same matric suction and the difference between the drying different researchers into three groups: empirical, macroscopic, and wetting SWCC is termed hysteresis. The four mechanisms and statistical model. Leong and Rahardjo (1997b) later found that proposed to be the causes of hysteresis include: statistical model is the most rigorous and yields the most accurate permeability function. The permeability function can be estimated 1. The capillary theory or the “ink bottle effect” due to irregularity statistically using the equation developed by Kunze, Uehara, and in the cross section of the voids (Fredlund & Rahardjo, 1993; Graham (1968) and best-fitted using the equation developed by Taylor, 1948); Leong and Rahardjo (1997b) to form a continuous permeability 2. The swelling and shrinking of “aged” soil (Hillel & Mottas, 1966); function curve. 3. The contact-angle (between air-water and soil solids) effect It was noted, however, that the permeability estimated from a (Bear, 1979; Hillel, 1971); and SWCC can lead to an underestimation of permeability at low water 4. The presence of entrapped air in a soil (Fredlund & Rahardjo, contents (Choo & Yanful, 2000). Since SWCC exhibits hysteresis, the 1993; Hillel, 1971). estimated permeability function also exhibits hysteresis, i.e. the permeability of soil at a given suction on a wetting curve is less than Numerous laboratory experiments using Tempe cell for the on a drying curve. drying path and capillary rise measurements for the wetting path (Fredlund & Rahardjo, 1993; Yang, 2002) have confirmed the hysteretic behaviour in SWCC. The differences between the drying and 1.3. Seepage and slope stability analyses wetting SWCC could affect the accuracy of the numerical seepage analyses. Dye et al. (2011) concluded that variability in SWCC had a Analysing the passage of water through unsaturated residual significant effect on the computed pore-water pressure. Currently, soil slopes could be a complex and difficult task for hand calcula- numerical analyses are usually performed using the adsorption or tions using traditional flow net method due to the variability of the drying SWCC since the SWCC is more commonly obtained in a input parameters with space and time. Engineers commonly resort drying process. However, Tami, Rahardjo, and Leong (2004) sug- to finite element software, such as the commercially available SEEP/ gested that the appropriate hydraulic properties of the soils (i.e., W (Geo-Slope International Ltd., 2012a), to perform more accurate 228 C. Kristo et al. / International Soil and Water Conservation Research 7 (2019) 226e238 and efficient seepage numerical analyses. The essential input pa- and too thin a sample will pose a difficulty in handling and rameters for SEEP/W include the SWCC, permeability function, weighing the specimen. The axis-translation technique (Hilf, 1956) boundary conditions, and the slope geometry. To model the infil- was used to apply matric suction to the soil specimens, where air tration of rainfall into a slope, flux is controlled on the slope surface pressure was supplied to the pressure chamber and the lower part as a boundary condition. When the rainfall intensity exceeds the of the high-air entry disk was connected to a burette of water under infiltration capacity and ponding occurs, zero constant water atmospheric pressure. The drying SWCC was first obtained by pressures are prescribed to the slope surface to simulate runoff, increasing the matric suction to let water to drain out of the soil which is common in a sloping field situations. The output of the specimen. Subsequently, the wetting SWCC was obtained by seepage analyses is the distribution of pore-water pressure at reducing the matric suction gradually, allowing water to seep into different points in the slope and at different time steps, which are the soil specimen by capillary action. used as an input to SLOPE/W for slope stability analyses. Limit A shrinkage test was also carried out to monitor the void ratio equilibrium analyses using SLOPE/W (Geo-Slope International Ltd., changes in the specimen during the application of matric suction. 2012b) make use of unsaturated shear strength parameters to As a result, the SWCC in the form of degree of saturation and determine the FoS of a slope. The complex nature of unsaturated volumetric water content could be plotted after incorporating the residual soil slopes warrants the need to simplify the analyses and volume change data. more efficiently find the critical slip surface and the associated minimum FoS with time. 2.3. Drying and wetting permeability functions Presently, the effect of hysteresis of SWCC on seepage analyses of slope under drying and wetting conditions is not well under- The drying and wetting SWCCs from the laboratory experiments stood. Drying SWCC is commonly used in practice to conduct carried out in this study and the upper limits of the drying and seepage analyses on slope during dry and rainy periods. However, wetting SWCCs from the experimental studies (Rahardjo et al., drying SWCC does not quite represent the hydraulic processes that 2012) were chosen to derive the permeability function by the sta- the soil slope actually experiences during dry and rainy periods. tistical method. The upper limits were selected to give a more Therefore, the objective of this study is to evaluate the accuracy of conservative factor of safety, as suggested by Zhai et al. (2016), since the seepage analyses using only drying SWCC as compared with the they correspond to a higher permeability function, allowing more analyses using drying and wetting SWCCs. water to infiltrate through the soil slope. This study focused on the effect of hysteresis in SWCC on the stability of residual soil slope during rainfall. The scope of the study 2.4. Saturated and unsaturated shear strength involved numerical seepage and slope stability analyses using data from laboratory experiments conducted in this study, as well as the Furthermore, Saturated Consolidated Undrained (CU) with data from the experimental studies by Rahardjo, Satyanaga, Leong, pore-water pressure measurements and Multistage Unsaturated and Ng (2012). The study also considered incorporating drying Consolidated Drained (CD) triaxial tests were carried out to deter- SWCC during drying period and wetting SWCC during rainy period mine the saturated (ASTM D4767-11, 2011) and unsaturated to observe the variations of factor of safety due to changes in the (Fredlund & Rahardjo, 1993) shear strength parameters, rainfall intensity and SWCC. Results from this study could be useful respectively. in designing safe yet economical residual soil slopes by considering the unsaturated residual soil properties without being overly con- 2.5. Seepage and slope stability analyses servative as compared to the conventional saturated soil slope stability analyses. Transient seepage analyses of a slope model in SEEP/W were carried out with input parameters (SWCC and permeability func- 2. Methodology tion) from the experimental studies and the laboratory experiments of this study, as illustrated in Fig. 1. The slope geometry 2.1. Soil sample preparation followed a typical slope in Singapore with a height of 15 m inclined at an angle of 27 (Rahardjo et al., 2007). A typical groundwater Laboratory experiments were carried out to verify data obtained table level was also adopted at a depth of 10 m and 5 m below the from the experimental studies by Rahardjo et al. (2012). The un- crest and toe of the slope, respectively and extending at an angle of disturbed soil samples were obtained from Orchard Boulevard 7 beyond the crest and toe (Rahardjo et al., 2007). The side (Bukit Timah Granite) at a depth of 7e7.5 m using a 70 mm diam- boundaries were at a distance of 45 m from the crest and the toe (3 eter Mazier sampler. The specimens were trimmed to a diameter of times the height of the slope) to avoid any influence of the 50 mm and cut to the required heights for different tests (30 mm boundary conditions on the seepage process within the slope. The height for SWCC, saturated permeability, and shrinkage tests; following mesh distribution was adopted to balance accuracy and 100 mm height for saturated and unsaturated shear strength tests). time required for the analyses as suggested by Tsaparas, Rahardjo, Specific gravity, grain-size distribution and Atterberg limits testing Toll, and Leong (2002): fine 8-noded quadrilateral elements as well as soil classification were carried out according to ASTM (0.5 0.5 m) near the slope surface up to a depth of 3 m, and D854-02 (2002), ASTM D422-63 (2002), ASTM D4318-00 (2000), coarser 4-noded quadrilateral elements (1 1 m) from 3 m depth and ASTM D2487-00 (2000), respectively. and below, as well as at a distance of 15 m from the crest and toe. In total, the mesh consisted of 9212 nodes and 8976 elements. 2.2. Drying and wetting SWCCs The boundary conditions adopted were as follows: zero total flux without potential seepage face review was applied to the Tempe Cell and Pressure Plate were used to obtain the drying bottom boundary and the side boundaries above the groundwater and wetting SWCCs experimentally up to 60 kPa and up to 490 kPa table; constant total head was applied to the side boundaries below matric suctions, respectively, according to ASTM D6838-02 (2008). the groundwater table; and unit flux boundary of 14 mm/h or The soil thickness of 30 mm was chosen because too thick a soil 31 mm/h or 57 mm/h for 10 h with potential seepage face review specimen will lead to longer time required for the water content of was applied to slope surface during wetting periods to simulate the specimen to reach equilibrium at the applied matric suction, rainfall infiltration with no ponding. These rainfall intensities were C. Kristo et al. / International Soil and Water Conservation Research 7 (2019) 226e238 229

Fig. 1. Slope model for seepage and slope stability numerical analyses. obtained from the linear correlation developed by Kristo, Rahardjo, Casagrande's plasticity chart and could be described as inorganic and Satyanaga (2017) by considering historical rainfall trends from silts of high plasticity (Casagrande, 1948). This was confirmed with Seletar weather station, projecting it into the years of 2017, 2050, the grain-size distribution (GSD) of the soil which indicated a well- and 2100. Drying periods of 24 h were modelled after the wetting graded soil with high fines content of 60.3% and no gravel content. periods for drainage analysis, during which zero total flux boundary condition with potential seepage face review was applied to the 3.2. Drying and wetting SWCCs slope surface. The time step adopted was 0.5 h to obtain accurate results. Moreover, the hydrostatic pore-water pressure was limited The relationship between the gravimetric water content and to a lower threshold of 75 kPa as suggested by Rahardjo, Leong, void ratio in the shrinkage curve is shown in Fig. 2. Deutcher, Gasmo, and Tang (2000) to avoid unrealistically high The relationship was fitted using the Fredlund et al. equation negative pore-water pressure near the ground surface. (2002). The SWCC data in terms of volumetric water content Three sets of three numerical transient seepage analyses each (qw-SWCC) and degree of saturation (S-SWCC) were subsequently were carried out using different SWCCs from the experimental plotted and fitted with the Fredlund and Xing (1994) equation to studies, to study the effects of hysteresis on seepage and slope obtain continuous drying and wetting SWCCs for the full range of stability: using drying SWCC in rainy and dry conditions (Case 1); matric suction. The correction factor of the fitting equation, CðjÞ, using wetting SWCC in rainy and dry conditions (Case 2); and using was set to unity as recommended by Leong and Rahardjo (1997a). wetting and drying SWCC during rainy and dry conditions, Fig. 3 shows the S-SWCC with the AEV indicated, whereas Fig. 4 respectively (Case 3). Each set of three analyses was distinguished shows the qw-SWCC together with the upper and lower drying by the applied rainfall intensity, ranging from 14 mm/h, 31 mm/h to and wetting SWCC envelopes from the experimental studies by 57 mm/h (Kristo et al., 2017). The pore-water pressure profiles with Rahardjo et al. (2012). depth were also extracted at a cross section on the sloping face of The drying and wetting qw-SWCCs from the laboratory experi- the slope. ments of this study corroborate well within the upper and lower A similar set of three analyses was also conducted using SWCCs envelopes from the experimental studies by Rahardjo et al. (2012), and permeability functions from the laboratory experiments of this as illustrated in Fig. 4. However, the saturated water contents were study for an applied rainfall intensity of 14 mm/h only. An example slightly lower than the lower envelopes for both the drying and of the results from SEEP/W is presented in Fig. 9, where the pore- wetting SWCCs, resulting in lower volumetric water contents at water pressure variation with depth was extracted from the mid- low suctions (less than 1 kPa) than the lower envelopes. However, dle of the sloping face of the slope with SWCCs and permeability the AEV for the drying SWCC and the water entry value for the functions from the laboratory experiments of this study under an wetting SWCC lie within the envelopes from the experimental applied rainfall intensity of 14 mm/h. These pore-water pressure studies. Furthermore, the data points are closer to the upper en- variations were then imported to SLOPE/W for numerical slope velopes of the drying and wetting SWCCs for Bukit Timah Granite stability analyses. from the experimental studies.

3. Results 3.3. Drying and wetting permeability functions

3.1. Soil index properties The average coefficient of saturated permeability of 1 105 m/ s was obtained from the constant-head permeability test. The Basic soil properties were determined by the Index Property resulting permeability functions and the fitting parameters are tests. The specific gravity of the soil was 2.61. It had a Liquid Limit shown in Fig. 5 and Table 1, respectively. The fitting parameters p (LL) and Plastic Limit (PL) of 56.5% and 41.8%, respectively, with a was found to be between 2.4 and 5.6, confirming the findings by Plasticity Index (PI) of 14.7%. The soil was classified as MH based on Fredlund, Fredlund, and Zakerzadeh (2001). 230 C. Kristo et al. / International Soil and Water Conservation Research 7 (2019) 226e238

Fig. 2. Shrinkage curve obtained the laboratory experiments of this study at zero conﬁning pressure.

Fig. 3. Drying and wetting SWCCs with respect to degree of saturation for soil from Orchard Boulevard.

Fig. 5 indicates that the drying coefficients of permeability from drawn tangent to all Mohr circles inclined at 26 and intercepted the laboratory experiments in this study are lower than the co- the vertical axis at 11 kPa, which correspond to the effective angle efficients obtained from the experimental studies at suctions less of internal friction (f0) and the effective cohesion (c0), respectively. than 150 kPa, but are higher at suctions above 150 kPa. Moreover, Similarly, the unsaturated shear strength parameter, fb, could the wetting coefficients of permeability from the laboratory ex- be determined from the results of the Multistage Unsaturated CD periments in this study are much lower than the coefficients from triaxial tests. Mohr circles for each matric suction could be drawn the experimental studies at any suctions within the range. This may by considering the corresponding maximum deviatoric stresses as be caused by the large difference between the water-entry values of shown in Fig. 7(a)-(d) and the failure envelopes for each matric the wetting SWCCs from the experimental studies and from the suction were determined by drawing tangents inclined at 26, laboratory experiments in this study. equivalent to f0. A graph of shear strength against matric suction at zero net confining pressure could be plotted to investigate the variation of shear strength with matric suction (Fig. 8). The data 3.4. Saturated and unsaturated shear strength parameters points were then best-fitted using Goh, Rahardjo, and Leong (2010) equation. By considering the maximum deviatoric stress for each consol- It was observed that the shear strength increased linearly with idation pressure during the Saturated CU triaxial test, three Mohr matric suction before the AEV at a rate equivalent to f0. However, circles could be drawn as illustrated in Fig. 6. The failure envelope C. Kristo et al. / International Soil and Water Conservation Research 7 (2019) 226e238 231

Fig. 4. Drying and wetting volumetric water content SWCCs obtained from the laboratory experiments of this study, superimposed with SWCCs from the experimental studies by Rahardjo et al. (2012).

Fig. 5. Permeability functions derived from SWCCs using the statistical method.

Table 1 Fitting parameters for permeability functions after Leong and Rahardjo (1997b).

a (kPa) b c p ks (m/s)

Laboratory Experiment Drying 410.38 0.36 5.48 4.21 1 105 Wetting 0.41 1.03 1.64 5.60 Experimental Studies (Rahardjo et al., 2012) Upper Drying 94.33 0.37 6.39 3.32 5 105 Upper Wetting 33.31 0.41 4.50 4.32

beyond the AEV, the shear strength increased non-linearly with Fredlund, & Vanapalli, 1992, pp. 531e537) was used as a conser- matric suction at a slower rate. The unsaturated shear strength vative numerical slope stability analysis. Furthermore, it was parameter fb for numerical analyses could be obtained by ﬁnding a observed that fb approaches zero at matric suctions beyond line of best ﬁt of the three data points beyond AEV (70 kPa), i.e. at 280 kPa. Table 2 compares the unit weight and shear strength pa- suctions equal to 75 kPa, 100 kPa, and 200 kPa. Due to the limitation rameters obtained from the experimental studies and the labora- in SLOPE/W that only allows a single value of fb as an input inde- tory experiments of this study. pendent of matric suction, the Total Cohesion method (Rahardjo, 232 C. Kristo et al. / International Soil and Water Conservation Research 7 (2019) 226e238

Fig. 8. Variation of shear strength with matric suction before and after the Air-Entry Value.

Fig. 6. Mohr Circles from Multistage Saturated CU triaxial tests at different consoli- matric suction increases) during drying periods. This is especially dation pressures with the failure envelope drawn tangent to all the Mohr circles. evident in Fig. 9(a) when only the drying SWCC is used. Moreover, the water table is observed to rise during rainfall and fall during 3.5. Seepage and slope stability analyses drying in Case 1, reaching the surface of the slope at t ¼ 39 h to t ¼ 44 h (Fig. 9(a)). However, the water table continued to rise In general, for the same depth, pore-water pressure increases regardless rainfall or drying period in Cases 2 and 3, creating a layer (or matric suction decreases) during rainfall, and decreases (or

Fig. 7. Mohr Circles and failure envelopes from Multistage Unsaturated CD triaxial tests at different matric suctions. C. Kristo et al. / International Soil and Water Conservation Research 7 (2019) 226e238 233

Fig. 9. Pore-water pressure variation with depth at the middle of the slope face when 14 mm/h of rainfall intensity was applied for different SWCC combinations: (a) Drying SWCC e Case 1, (b) Wetting SWCC e Case 2, and (c) Combined SWCC e Case 3, from the laboratory experiments of this study.

Table 2 Comparison of shear strength parameters from the laboratory experiments in this study with laboratory data from Rahardjo et al. (2012).

Shear strength parameters Laboratory Data (Rahardjo et al., 2012) Laboratory Experiment

c0 (kPa) 12 11 f0 (degrees) 32 26 fb (degrees) 20 14 g (kN/m3) 18.6 19.0 234 C. Kristo et al. / International Soil and Water Conservation Research 7 (2019) 226e238 of wetting and drying bands across the depths as evident in Fig. 11(a)-(c) show that FoS variations obtained from stability Fig. 9(b). A wetting band was observed between depths of 2 me4m analyses using the combined SWCC match more closely with those when the wetting SWCC was used. using the wetting SWCC as compared to those using the drying For slope stability analyses, the shear strength parameters in SWCC, independent of the hydraulic processes the soil actually Table 2 were inputted into SLOPE/W with the corresponding pore- experienced (i.e. drying or wetting). However, at 57 mm/h rainfall water pressure variations imported from SEEP/W. The SEEP/W intensity, the FoS variation using the combined SWCC during the outputs were obtained using SWCCs and permeability functions drying periods matches the FoS variation using the drying SWCC either from the experimental studies or from the laboratory ex- better, as seen in Fig. 11(c). Similar observation can be seen towards periments of this study. Bishop's simplified method of slices for the end of the second drying cycle at 31 mm/h rainfall intensity circular slip surfaces was used in calculating the factor of safety by (Fig. 11(b)). dividing the slope into 30 slices. No tension crack was allowed to Furthermore, the FoS obtained from using the wetting SWCC at develop in the model. The grid of potential centers of rotation was 14 mm/h rainfall intensity is higher than the FoS obtained from drawn such that the critical centers of rotation were located in the using the combined SWCC, as seen in Fig. 11(a). Thus, using the middle of the grid. wetting SWCC at 14 mm/h rainfall intensity might be under- Fig. 10 shows the comparison of FoS variation between those conservative in design leading to unsafe designs. Similar observa- obtained from the experimental studies and the laboratory exper- tion can be seen for 31 mm/h rainfall intensity during the first iments of this study. Moreover, Fig. 11 shows the FoS variation wetting-drying cycle (Fig. 11(b)). In contrast, during the second obtained from the experimental studies for different rainfall in- wetting-drying cycle at 31 mm/h rainfall intensity (Fig. 11(b)) and tensities and different SWCC combinations. during both wetting-drying cycles at rainfall intensity of 57 mm/h The FoS variations between those obtained from the experi- (Fig. 11(c)), the FoS obtained from using the wetting SWCC is lower mental studies and the laboratory experiments of this study indi- than FoS obtained from using the combined SWCC. Using the cate similar trends. Using the drying SWCC in Case 1, the FoS falls wetting SWCC in this case is conservative and leads to safer designs. during rainfall and recovers during drying period for both analyses It seems that with higher rainfall intensity, using the wetting SWCC based on the experimental studies and the laboratory experiments leads to more conservative design than using the combined SWCC. of this study as seen in Fig. 10. In contrast, the FoS falls during both In addition, higher rainfall intensity leads to a faster reduction in rainfall and drying periods in Cases 2 and 3 for both analyses based FoS and a faster recovery of FoS, resulting in a lower minimum FoS on the experimental studies and the laboratory experiments of this at higher rainfall intensity for the same soil permeability. This can study. One exception is the slight FoS recovery in the second drying be observed from Fig. 11(d), where the FoS drops from an initial period in the experimental studies. Moreover, due to the difference value of 2.20 to 1.94, 1.66, and 1.17 under rainfall intensities of in the initial FoS, the FoS falls below 1.00 during the second rainfall 14 mm/h, 31 mm/h, and 57 mm/h, respectively. Additionally, during period in the analyses based on the laboratory experiments of this the second drying cycle when using the drying SWCC (Fig. 11(d)), study, indicating a slope failure. The lower initial FoS for the results the FoS rises from 1.68 to 1.77, from 1.43 to 1.73, and from 1.17 to from the analyses based on the laboratory experiments of this 1.72 after the 14 mm/h, 31 mm/h, and 57 mm/h rainfall has stopped study is attributable to the lower shear strength parameters used, for 24 h, respectively. as illustrated in Table 2. Furthermore, the minimum FoS during the Similarly, using the drying SWCC also leads to faster reduction first drying period in Case 1 occurred some time after the rainfall and recovery of FoS as compared to using the wetting and com- stopped, whereas the FoS recovered immediately in the second bined SWCCs. For instance, the FoS dropped to 1.17 when using the drying period after the rainfall stopped. drying SWCC under 57 mm/h rainfall intensity (Fig. 11(d)), while

Fig. 10. Factor of safety variation comparing different combinations of SWCC from the experimental studies and the laboratory experiments of this study under applied rainfall intensity of 14 mm/h. Vertical dotted lines mark the start and end of rainfall as indicated. C. Kristo et al. / International Soil and Water Conservation Research 7 (2019) 226e238 235

Fig. 11. Factor of safety variations from the experimental studies comparing different rainfall intensities (a)e(c) and different SWCC combinations (d)e(f). Vertical dotted lines mark the start and end of rainfall as in Fig. 10. the FoS only dropped to about 1.39 and 1.44 when using the wet- 4. Discussions ting and the combined SWCC, respectively, under the same rainfall intensity (Fig. 11(e)&(f)). Moreover, it seems that there is a SLOPE/W restricts user to only input one fb value regardless of threshold minimum FoS at 1.17 observed at the end of the first and suction for slope stability analyses. This limitation does not reflect second wetting cycle when the rainfall intensity is 57 mm/h the actual variation of fb with matric suction in reality as shown in (Fig. 11(d)). Fig. 8. Therefore, a fb value of 14 (after AEV) was chosen following 236 C. Kristo et al. / International Soil and Water Conservation Research 7 (2019) 226e238 the Total Cohesion method suggested by Rahardjo et al. (1992, pp. the infiltration capacity has not been reached. This results in a 531e537). This is a conservative estimate since at every suction, the greater increase in pore-water pressure, a greater reduction in shear strengths are lower than or equal to the actual shear strength matric suction, and a greater reduction in shear strength, thus indicated by the fitting by Goh et al. (2010) in Fig. 8. As a result, the reducing the FoS to a greater extent. Another consequence of a actual FoS variations will also be higher than what was presented in higher volume of water infiltrating into the soil is the higher degree Fig. 10. This limitation could improve the FoS and prevent the slope of saturation in the soil at the end of the rainfall event, where more failure when using the drying SWCC. voids in the soil being filled with more saturated and continuous It is also evident from Fig. 9 that the variation of pore-water conduits of water. This could explain the faster recovery of FoS after pressures in the slope with depth did not strictly follow the the rainfall event, since water flows more easily through conduits of drying-wetting cycle periods. This is more pronouncedly seen in water as it encounters less friction from the soil particles. Hence, a the analyses using the wetting and the combined SWCCs as more saturated soil after a more intense rainfall will lead to a more compared to using the drying SWCC. This is because the wetting rapid discharge of water from the soil. This leads to a faster SWCC corresponds to a lower permeability function at a given reduction in pore-water pressure, increasing the matric suction and suction (refer to Fig. 5), which causes water to take a longer time to shear strength. As a result, the FoS recovers at a faster rate. More- flow into or out of the soil when the wetting SWCC is used. For over, this argument could also explain the reduction of the rate of example, the wetting front reached a depth of about 5 m at t ¼ 5h FoS recovery further into the drying period as the soil becomes when the drying SWCC is used in Fig. 9(a), while it only reached a increasingly less saturated. depth of about 0.6 m at the same time when the wetting SWCC is The results from the experimental studies in Fig. 11(a)-(c) sug- used in Fig. 9(b)-(c), indicating that most of the rainfall resulted in gest that the FoS variation using the combined SWCC matched the surface runoff. Similarly during the drying period between t ¼ 10 h FoS variation using the wetting SWCC more closely than the FoS and t ¼ 22 h, the water table dropped from a depth of 1.5 me3m variation using the drying SWCC, regardless the wetting or drying when using the drying SWCC (Fig. 9(a)), whereas the water table processes that the soil experienced. The FoS variation using the rose from a depth of 7 me6.8 m and from a depth of 7 me6.3 m combined SWCC shall be taken as the standard reference of the when using the wetting and the combined SWCC, respectively actual FoS variation of the residual soil slope in field conditions for (Fig. 9(b)-(c)). Moreover, during the same drying period, the pore- this study. It is not surprising that using the wetting SWCC during water pressure decreased near the surface when using the drying the first and second wetting periods will result in a more repre- and the combined SWCC, while the pore-water pressure continued sentative FoS variation than using the drying SWCC, since the soil is to increase when using the wetting SWCC. Hence, some parts of the experiencing wetting. However, using the wetting SWCC during soil still undergo wetting during the drying period. This explains drying still results in a fairly representative FoS variation than using why the minimum FoS might not coincide with the end of the the drying SWCC. This is evident from both drying periods in rainfall, since more water might still percolate into the soil during Fig. 11(a) and from the first drying period in Fig. 11(b). This might be the drying period. This causes an increase in pore-water pressure, attributable to the fact that certain parts of the soil still undergo a reducing matric suction, and in turn, the shear strength of the soil. wetting process despite no additional rainfall falling to the ground In this study, the whole slope was assigned wetting SWCC during the drying period. This is confirmed by the results of the during rainfall and drying SWCC during drying for the combined numerical seepage analyses. However, after some time into the SWCC case, without considering wetting and drying sections that drying period, most parts of the soil are now undergoing a drying were contrary to the applied unit flux boundary. Future works process and no longer experiencing wetting process. This explains could investigate a method for proper assignment of SWCCs why using the drying SWCC gives a better representative FoS depending on the hydraulic processes that the slope sections variation towards the end of the second drying cycle in Fig. 11(b) actually experience. This can be done by considering the progres- and in both drying cycles in Fig. 11(c). sion of the wetting front from the surface as well as the rise in Moreover, the FoS variation using the drying SWCC gives the groundwater table. This is especially important for seepage ana- most conservative result as the FoS is the lowest compared to using lyses through repeated cycles of drying and wetting using soils of the wetting and the combined SWCC. This can be explained by the low permeability, which has a high likelihood of showing bands of abovementioned discussion on the fact that using the drying SWCC drying and wetting layers, as illustrated in Fig. 9(b). leads to a faster reduction in FoS, and thus the lowest possible FoS. It is reasonable to expect that using the drying SWCC will lead to However, this might be overly conservative in designing the slope faster reduction and recovery of FoS during and after a rainfall and might not be economical. event as compared to using the wetting and the combined SWCC. In addition, comparing the FoS variation using the wetting and This is because the drying SWCC corresponds to higher coefficients the combined SWCC suggests that using the combined SWCC re- of permeability than the wetting SWCC at the same suction, as seen sults in a lower and more conservative FoS at lower rainfall in- in the permeability function in Fig. 5. Hence, water will infiltrate tensities, while using the wetting SWCC is more conservative at faster and deeper into the soil when using the drying SWCC, higher rainfall intensities. This can be explained by considering the causing an increase in pore-water pressure, a reduction in matric water flow during the drying periods. At the end of a rainfall with suction, and a decrease in shear strength to a greater extent. As a low intensities, the soil is only partially saturated and further result, the residual soil slope is weaker and has a lower FoS. On the infiltration of water could reduce the FoS further. Therefore, using other hand, water will also flow out of soil at a faster rate when the drying SWCC during drying period (as in the case of the com- using the drying SWCC, causing a more rapid decrease in pore- bined SWCC), which corresponds to higher coefficients of perme- water pressure, an increase in matric suction, and an increase in ability, will allow water to seep through the soil more easily. This shear strength, resulting in a faster recovery of FoS of the residual increases the pore-water pressure to a greater extent, reduces soil slope. matric suction and shear strength, resulting in a lower and more Similarly, the observation that a higher rainfall intensity leads to conservative FoS. On the contrary, at the end of a rainfall with high faster reduction and recovery of FoS could be explained by intensity, the soil is approaching a fully saturated state and further considering the amount of water available for infiltration and the infiltration of water will only reduce the FoS marginally. A more permeability function of the soil. A higher rainfall intensity trans- significant effect, rather, is the rapid recovery of FoS when water lates into a higher amount of water infiltrating into the soil when flows out of the soil after the rainfall event. Therefore, using the C. Kristo et al. / International Soil and Water Conservation Research 7 (2019) 226e238 237 drying SWCC during drying period (as in the case of the combined Brooks, R. H., & Corey, A. T. (1966). Properties of porous media affecting fluid flow. e SWCC), will allow water to flow out more easily. This reduces the Journal of Irrigation and Drainage Division. ASCE, 92(2), 61 68. Casagrande, A. (1948). Classification and identification of soils. Transactions, ASCE, pore-water pressure, increases the matric suction and shear 113,901e930. strength, resulting in a higher FoS. This is less conservative since the Childs, E. C., & Collis-George, N. (1950). The permeability of porous materials. & slope may be thought to be safe when it actually is not as stable as Proceedings of the Royal Society A: Mathematical, Physical Engineering Sciences, 201(1066), 392e405. expected. This corroborates with the observations that using the Choo, L. P., & Yanful, E. K. (2000). Water flow through cover soils using modeling combined SWCC at lower rainfall intensities and using the wetting and experimental methods. Journal of Geotechnical and Geoenvironmental En- SWCC at higher rainfall intensities will give a more conservative gineering, 126(4), 324e334. Dye, H. B., Houston, S. L., & W, B. D. (2011). Influence of unsaturated soil properties estimate of the FoS. uncertainty on moisture flow modeling. Geotechnical Geology Engineering, 29, Looking forward to the likelihood of more intense rainfall events 161e169. due to climate change in the future, it is therefore advisable to use Fredlund, D. G., Fredlund, M. D., & Zakerzadeh, N. (2001). Predicting the permeability function for unsaturated soils. In Proceedings of the international sym- the wetting SWCC instead of the combined SWCC to perform nu- posium on suction, swelling, permeability and structural clays, Japan (pp. merical seepage and slope stability analyses to obtain a conserva- 215e222). tive estimate of the FoS, while not being overly conservative as in Fredlund, D. G., & Morgenstern, N. R. (1977). Stress state variables for unsaturated e the case of using the drying SWCC. The results of this study align soils. Journal of Geotechnical Engineering, 103,447 466. Fredlund, D. G., & Rahardjo, H. (1993). Soil mechanics for unsaturated soils. New York: with the results from previous works which indicated the crucial John Wiley and Sons Inc. role of SWCC on seepage and stability analyses. Rahardjo et al. Fredlund, D. G., Rahardjo, H., & Fredlund, M. D. (2012). Unsaturated soil mechanics in & (2012) indicated that SWCC and permeability vary gradually with engineering practice (Vol. 926). New York: John Wiley Sons, Inc. ISBN 978-1- 118-13359-0. depth, controlling both local seepage response to rainfall infiltra- Fredlund, M. D., Wilson, G. W., & Fredlund, D. G. (2002). Representation and esti- tion and location of the shear surface. Studies by Gitirana, Fredlund, mation of the shrinkage curve. In Paper presented at the proceedings, 3rd international conference on unsaturated soils, UNSAT 2002. and Fredlund (2005), Krahn et al. (1989) and Ng and Shi (1998) e fi fl Fredlund, D. G., & Xing, A. (1994). Equations for the soil water characteristic curve. indicated that the rainfall in ltration produces a downward ux Canadian Geotechnical Journal, 31(3), 521e532. that changes the water content and pore-water pressure gradients Gasmo, J. M., Rahardjo, H., & Leong, E. C. (2000). Infiltration effects on stability of a with depth. This rainwater infiltration increases the moisture residual soil slope. Computers and Geotechnics, 26,145e165. Gavin, K., & Xue, J. (2008). A simple method to analyze infiltration into unsaturated content and reduces the matric suction, hence reducing the soil soil slopes. Computers and Geotechnics, 35(2), 223e230. shear strength and subsequently triggering the slope failure. Many Geo-Slope International Ltd. (2012a). Seepage modeling with SEEP/W. Calgary, research works have confirmed the adverse effect of infiltration of Alberta: July 2012 Edition. & Geo-Slope International Ltd. (2012b). Slope modeling with SLOPE/W. Calgary, rainwater into slope to its stability (Alonso, Gens, Delahaye, Alberta: July 2012 Edition. 2003; Gasmo et al., 2000; Melinda, Rahardjo, Han, & Leong, 2004). Gitirana, G. J., Fredlund, M. D., & Fredlund, D. G. (2005). Inflitrationerunoff boundary conditions in seepage analysis. In 58th Canadian geotechnical con- 5. Conclusions ference and 6th joint IAH-CGS conference, september 19-21, saskatoon, SK, Canada. Goh, S. G., Rahardjo, H., & Leong, E. C. (2010). Shear strength equations for unsaturated soil under drying and wetting. Journal of Geotechnical and Geo- The following conclusions are drawn from the numerical ana- environmental Engineering, 136(4), 594e606. lyses performed in this study. Firstly, using only the drying SWCC in Hilf, J. W. (1956). An investigation of pore-water pressure in compacted cohesive soils. Ph.D. Dissertation, Technical Memo. No. 654. Denver, CO: U.S. Dept. of the performing numerical seepage and slope stability analyses through Interior, Bureau of Land Reclamation, Design and Construction Division. cycles of wetting and drying lead to a faster reduction and recovery Hillel, D. (1971). Soil and water. San Diego: Academic. of FoS, resulting in the lowest and most conservative FoS in Hillel, D., & Mottas, J. (1966). Effect of plate impedance, wetting method and aging on soil moisture retention. Soil Science, 102,135e139. assessing residual soil slopes. Similarly, applying a higher rainfall Kim, J., Jeong, S., Park, S., & Sharma, J. (2004). Influence of rainfall-induced wetting intensity will also result in rapid reduction and recovery of FoS. on the stability of slopes in weathered soils. Engineering Geology, 75,251e262. Secondly, the FoS variation using the combined SWCC matches the Krahn, J., Fredlund, D. G., & Klassen, G. M. (1989). Effect of soil suction on slope stability at notch hill. Canadian Geotechnical Journal, 26, 269e278. FoS variation using the wetting SWCC more closely than the FoS Kristo, C., Rahardjo, H., & Satyanaga, A. (2017). Effect of variations in rainfall in- variation using the drying SWCC, regardless the wetting or drying tensity on slope stability in Singapore. International Soil and Water Conservation processes that the soil experienced. Finally, with increasing rainfall Research, 5(4), 258e264. Kunze, R. J., Uehara, G., & Graham, K. (1968). Factors important in the calculation of intensity, using only the wetting SWCC gives a more conservative hydraulic conductivity. Proceedings - Soil Science Society of America, 32, FoS as compared to using the combined SWCC, without being 760e765. overly conservative as using only the drying SWCC. Leong, E. C., & Rahardjo, H. (1997a). Review of soil-water characteristic curve equations. Journal of Geotechnical and Geoenvironmental Engineering, 123(12), 1106e1117. References Leong, E. C., & Rahardjo, H. (1997b). Permeability function for unsaturated soils. Journal of Geotechnical and Geoenvironmental Engineering, 123(12), 1118e1126. Alonso, E. E., Gens, A., & Delahaye, C. H. (2003). Influence of rainfall on the defor- Matyas, E. L., & Radhakrishna, H. S. (1968). Volume change characteristics of mation and stability of A slope in overconsolidated clays: A case study. Hy- partially saturated soils. Geotechnique, 18(4), 432e448. drogeology Journal, 11,174e192. Melinda, F., Rahardjo, H., Han, K. K., & Leong, E. C. (2004). Shear strength of com- American Society for Testing and Materials (ASTM) D2487e00. (2000a). Standard pacted soil under infiltration condition. Journal of Geotechnical and Geo- practice for classification of soils for engineering purposes (unified soil classification environmental Engineering, 130(8), 807e817. system). Mualem, Y. (1976). A new model for predicting the hydraulic conductivity of un- American Society for Testing and Materials (ASTM) D422e63. (2002). Standard test saturated porous media. Water Resources Research, 12,513e522. method for particleesize analysis of soils. Ng, C. W. W., & Shi, Q. (1998). A numerical investigation of the stability of unsat- American Society for Testing and Materials (ASTM) D4318e00. (2000b). Standard urated soil slopes subjected to transient seepage. Computers and Geotechnics, test methods for Liquid limit, plastic limit, and plasticity Index of soils. 22(1), 1e28. American Society for Testing and Materials (ASTM) D4767-11. (2011). Standard test Rahardjo, H., Fredlund, D. G., & Vanapalli, S. K. (1992). Use of linear and nonlinear method for consolidated undrained triaxial compression test for cohesive soils. shear strength versus matric suction relations in slope stability analyses. In American Society for Testing and Materials (ASTM) D6838-02. (2008). Standard test Sixth international symposium on landslides. methods for the soilewater characteristic curve for desorption using hanging col- Rahardjo, H., Krisdani, H., Leong, E. C., Ng, Y. S., Foo, M. D., & Wang, C. L. (2007). umn, pressure extractor, chilled mirror hygrometer, or centrifuge. Unsaturated soil mechanics for solving seepage and slope stability problems. American Society for Testing and Materials ASTM D854e02. (2002). Standard test Keynote lecture. In Proceedings of twelfth international colloquium on structural methods for specific gravity of soil solids by water pycnometer. and geotechnical engineering. Ain shams University, cairo, Egypt, 10 e 12 december Aung, K. K., Rahardjo, H., Leong, E. C., & Toll, D. G. (2001). Relationship between (pp. 1e38). porosimetry measurement and soil-water characteristic curve for an unsatu- Rahardjo, H., Leong, E. C., Deutcher, M. S., Gasmo, J. M., & Tang, S. K. (2000). Rainfall- rated residual soil. Geotechnical & Geological Engineering, 19,401e416. induced slope failures, geotechnical engineering monograph 3. NTU-PWD Bear, J. (1979). Hydraulics of groundwater (p. 567). New York: McGraw-Hill. Geotechnical Research Centre, Nanyang Technological University Singapore. 238 C. Kristo et al. / International Soil and Water Conservation Research 7 (2019) 226e238

ISBN 981-04-2928-2. Engineering, 130(9), 956e967. Rahardjo, H., Lim, T. T., Chang, M. F., & Fredlund, D. G. (1995). Shear strength Taylor, D. W. (1948). Fundamentals of soil mechanics. New York: Wiley. characteristic of residual soils. Canadian Geotechnical Journal, 32,60e77. Tsaparas, I., Rahardjo, H., Toll, D. G., & Leong, E. C. (2002). Controlling parameters for Rahardjo, H., Ong, B. H., & Leong, E. C. (2004). Shear strength of a compacted re- rainfall-induced landslides. Computers and Geotechnics, 29,1e27. sidual soil from consolidated drained and constant water content triaxial tests. Yang, H. (2002). Inﬁltration characteristic of capillary barrier. Master of Engineering Canadian Geotechnical Journal, 41,421e436. thesis. Singapore: Nanyang Technological University. Rahardjo, H., Satyanaga, A., Leong, E. C., & Ng, Y. S. (2012). Variability of residual soil Zhai, Q., & Rahardjo, H. (2015). Estimation of permeability function from the properties. Journal of Engineering Geology, 141e142,124e140. soilewater characteristic curve. Engineering Geology, 199,148e156. Tami, D., Rahardjo, H., & Leong, E. C. (2004). Effects of hysteresis on steady-state Zhai, Q., Rahardjo, H., & Satyanaga, A. (2016). Variability in unsaturated hydraulic inﬁltrationin unsaturated slopes. Journal of Geotechnical and Geoenvironmental properties of residual soil in Singapore. Engineering Geology, 209,21e29. International Soil and Water Conservation Research 7 (2019) 239e247

Contents lists available at ScienceDirect

International Soil and Water Conservation Research

journal homepage: www.elsevier.com/locate/iswcr

Original Research Article Effect of conservation structures on curbing rill erosion in micro- watersheds, northwest Ethiopia

* Ermias Debie a, Kailash N. Singh b, Mehretie Belay c, a Department of Geography and Environmental Studies, Debre Markos University, P. O. Box 169, Debre Markos, Ethiopia b Department of Geography and Environmental Studies Addis Ababa University, P. O. Box, 1176, Addis Ababa, Ethiopia c Department of Geography and Environmental Studies, Bahir-Dar University, P. O. Box 79, Bahir-Dar, Ethiopia article info abstract

Article history: Accelerated soil erosion by water is a critical problem in Ethiopia, where population is rapidly growing Received 8 May 2018 and extensive farming systems are very common with no or less preventive measures. The study aimed Received in revised form to evaluate effects of conservation structures (terraces) in reducing the magnitude of rill erosion by using 15 May 2019 rill survey technique at cultivated field scales in relation to crop type and location in hill slope direction Accepted 21 June 2019 of the fields in the Beribere micro-watershed (catchment), northwest Ethiopia. The study assessed factors Available online 27 June 2019 accelerating rill erosion in the micro-watershed. Rill erosion in the terraced fields was reduced by 53.5% as compared to the non-terraced control fields. In the terraced areas, terracing damage is the main Keywords: fi Rill erosion contributing factor to soil loss due to rills. In the non-terraced elds, entering of high erosive runoff from Conservation structures uphill areas and damage of drainage ditches were observed as the principal accelerating factors for Rill surveying initiation and development of rill erosion, whereas these processes had insignificant effect on rill erosion Micro-watersheds in the terraced fields. Moreover, rill erosion shows spatial differences in terms of crop covers and location Ethiopian highlands in hill slope direction of the study site. Hence, there should be adequate criteria for effective communication between farmers and extension staff to design site-specific erosion control practices for addressing the problem. The study verifies that rill surveying is a significant method to pragmatic assessment of the effectiveness of terracing practices in controlling soil erosion by water. © 2019 International Research and Training Center on Erosion and Sedimentation and China Water and Power Press. Production and Hosting by Elsevier B.V. This is an open access article under the CC BY-NC- ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction generates an average soil loss rate of 27.5 t ha 1yr 1 (range from 0to200)tha 1yr 1 of which, at least 10% comes from gully erosion Accelerated soil erosion is a critical problem and will remain too and 26.7% leaves Ethiopia. They further noted that soil loss varied during the 21st century in developing countries (Lal, 2001; across the basin in response to different agroecologies and slopes. Pimentel et al., 1995). Pimentel (2006) estimated that on an Highest soil loss rates were measured in Gojjam, in the north- average, the ratio of soil erosion is the highest (ranging between 30 west highlands of Ethiopia due to combined factors of high rainfall and 40 t ha 1yr 1) in Africa. Accelerated soil erosion by water is too and erosivity, population density and intensive cultivation along a critical problem for Ethiopia, where population is rapidly growing with non-conservation based land management and cropping and extensive farming system is the backbone of the country's systems (Amdihun, Gebremariam, Rebelo, & Zeleke, 2014; Haile, economy (Hurni, 1998). Hurni (1993) estimated that soil loss due to Herweg, & Stillhardt, 2006; Hargeweyen et al., 2017). Tadesse, erosion by water from slopes amounts about 1493 million tons per Suryabhagvan, Sridhar, and Legesse (2017) indicates that soil annum, out of which 45% is from cropland only. He further esti- erosion is more sensitive to changes in vegetation cover, implying mated mean soil loss from cultivated fields to be about 42 t that deterioration of vegetation covers with encroachment of ha 1yr 1. Hargeweyen et al. (2017) reported that Blue Nile basin cultivated fields in hilly and steep slope results increasing soil loss. For example, based on the field assessment of rill erosion, Bewket and Sterk (2003) indicate that the rate of soil loss from cultivated fields ranges between 18 and 79 t ha 1yr 1 due to rill and * Corresponding author. sheet erosion in the Chemoga watershed. By using the same E-mail addresses: [email protected] (E. Debie), [email protected] 1 1 (M. Belay). method, Bewket (2003) also estimated around 37 t ha yr of soil https://doi.org/10.1016/j.iswcr.2019.06.001 2095-6339/© 2019 International Research and Training Center on Erosion and Sedimentation and China Water and Power Press. Production and Hosting by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). 240 E. Debie et al. / International Soil and Water Conservation Research 7 (2019) 239e247 loss due to rills and sheet from croplands in the Digil watershed. simulation models covering larger areas. Fisseha, Gebrekidan, Kibret, Bedadi, and Yitaferu (2011) measured Participatory identification of soil erosion indicators and two years amount of soil loss of 141.9 t ha 1 from experimental computing soil loss approaches, like what was done by Belay and cultivated plots of Debre-Mewi watershed. Belay and Bewket Bewket (2012, 2015) in northwest Ethiopia, are quite important (2012) also estimated 1.26 mm year 1 or a 16 t ha 1 yr 1 soil loss to enhance awareness on types of soil erosion by water, magnitude from cultivated fields in northwest Ethiopia through measurement of soil loss and conservation attention to key actors such as farmers of exposed tree roots. Subhatu et al. (2017) noted that the annual and development agents (Evans, 2002; Okoba & De Graaff, 2005). soil loss ranged from 32 to 37 t ha-1 in the terraced croplands in Bewket and Sterk (2003) suggest that a rill survey approach is Minchet catchment, northwest Ethiopia. commendable for obtaining good semi-quantitative information on Other studies estimated the rates of soil loss from cultivated magnitudes of soil loss in the real situation of different land uses plots in semi-arid environments of the Ethiopian highlands. For and conservation practices. Therefore, this study conducted with instance, Gebremichael et al. (2005) estimated the mean annual respect to the actual soil loss from cultivated fields, the principal soil loss rate of 57 t ha 1yr 1 by sheet and rill erosion from 202 units for evaluating the effectiveness of terracing practices by the plots in 12 representative sites of Dogu’ Tembien District, northern farmers, using rill survey technique in high rainfall highlands of Ethiopia. Based on eight years (1982e1989) Soil Conservation Ethiopia in which studies have been few in number as far as to the Research Project (SCRP) data, Tegene (2000) estimated the average knowledge of the authors. rate of soil erosion by water as 35 t ha 1yr 1 on cultivated slopes of This study presents the results of the effects of conservation South Wollo Zone, North-eastern Ethiopia. structures (terracing) on reducing the magnitude of rill erosion by In general, all those studies of soil erosion conducted in varied using rill survey methodology at cultivated field scales with a focus agro-ecologies of the Ethiopian highlands endorse that the to crop type and location at hill slope direction of valley sides of the magnitude of soil loss is high and revealed urgent need of inter- Beribere micro-watershed (catchment) in northwest Ethiopia. The vention in Soil and Water Conservation (SWC) practices. The basic paper assessed the factors accelerating rill erosion. paradigm and approach to SWC has developed through time. The Ministry of Agriculture of Ethiopia and international development agencies like World Food Program (WFP) have been investing considerable resources in encouraging and scaling up of SWC 2. Materials and methods practices like terracing for curbing severe soil erosion by water (Zeleke, Kassie, Pender, & Yesuf, 2006). The undertaking of effective 2.1. Description of the study area SWC practices should consider using interdependence between biophysical factors (erosivity, erodibility and slope) and land use The study was conducted in the Beribere micro-watershed - patterns (Tegene, 2000). located between 380042.171" & 384054.788"E longitudes, and Numerous studies at micro-watershed level estimated and 1055028.874" & 11 0034.617"N latitudes in northwest highlands of agreed upon the substantial role of conservation practices (in Ethiopia (Figure-1). The major landforms of the catchment are particular stone terraces) in reducing the magnitude of soil loss in volcanic in origin characterized by gently sloping undulating plains the semi-arid environments of the Ethiopian highlands to slightly dissected and rugged plateaus. The mean elevation in the (Gebremichael et al., 2005; Haregeweyn, Berhe, Tsunekawa, Tsubo, micro-watershed is 2175.5 masl, with ranges from 1677 to 2674 & Tsegaye, 2012; Nyssen et al., 2009, 2007; Taye et al., 2013). masl. Moreover, few plot level studies indicated contribution of conser- The local climatic condition is moist sub-tropical (Woina-Dega), vation practices in reducing soil loss in the high rainfall areas of the where high annual rainfalls and moderate temperatures are Ethiopian highlands (Amare et al., 2014; Fisseha et al., 2011; recorded. The mean annual rainfall distribution in the catchment Herweg & Ludi, 1999) with regardless to emphasis on the actual soil estimated by interpolating a raster surface from points’ discrete loss from cultivated fields, the principal units for evaluating the data (average value of monthly rainfall records from National effectiveness of SWC practices by the farmers. Metrological Agency) using Arc GIS-version-10.2 software is Despite substantial progresses in assessing soil erosion using approximately 1312 mm. More than three-fourth of the total rain- different models since the 1930s, field validation of these models fall occurs in the summer season (JuneeSeptember). Large areas of remained to be accomplished (Lal, 2001). There are limitations and the study site are covered by steep slopes and reddish soils (locally advantages of the different approaches of soil erosion measure- named as Boda implying that they are infertile). Under the mixed ments. Because of spatially aggregated nature, measurements using farming systems of the smallholder community, tef (Eragrostis tef) macro-order modelling often do not provide critical information to and wheat (Triticum vulgare) are predominantly grown in the significantly understand the spatial patterns of erosion and effec- catchment. Tef is more likely to have sparse canopies with little tiveness of conservation practices (Stocking, 2001). Contrary to this, cover against soil erosion by water, compared to wheat. others argue that micro-level measurements like at experimental Conventional tilling through hand press with ox-pulling tradi- plot levels may not be relevant to draw conclusions at field or large tional technique is common to all farmers for tef, wheat and other spatial scales as it represents a particular confined area (Boix-Fayos crop production in the study areas. Multiple ploughing runs et al., 2006). However, erosion shows large spatial variation due to (ranging from 3 to 9 times) are commonly undertaken starting from influence of erosivity of rain, effect of relief on such erosion, soil mid-September to end of July for seedbed preparation and then profile, vegetative or plant residue cover and conservation prac- sowing. The cropland is ploughed along the contour. As tef seed is tices. In accordance with this, field assessment techniques like rill very small and requires very loose soil for seedling, fine seed bed surveys provide substantial advantages as they are more realistic, preparation for sowing is required with more intensive labour costs simple and practically done at field conditions. Such methods of than for wheat. The tilled soil is compacted by animal trampling for measurement of soil erosion by water and the effectiveness of the tef seedbed preparation. Tef seed is sown on compacted seedbed conservation practices at actual field level by integrating views of and left uncovered. However, wheat seed is sown on seedbed that is the ultimate clients for the work, the farmers (Evans, 2002; tilled one week before. Unlike tef, seedbed trampling by animals is Morgan, 2005; Stocking, 2001), provide more accurate information not practiced before sowing of wheat seed. The spread wheat seed about soil losses than that information gathered through is covered by tilling of the seedbed soil surface. E. Debie et al. / International Soil and Water Conservation Research 7 (2019) 239e247 241

Fig. 1. Location of Beribere micro-watershed.

2.2. Methods and procedures of data collection and analysis July/2015. Second, during the early growing season, between end of July and beginning of September/2015 frequent visits performed to Monthly rainfall records covering the period between 1994 and identify the prevalence of rills following surface area covered with 2013 were collected from stations in close proximity to the study crops. Last, rapid and timely measurements of rills made between area and then used to select the conserved catchment under the mid-November to end of December/2015, before disappearance of highest condition of erosive factor. To estimate soil loss on terraced rills due to farming practices like free grazing and tillage following and adjacent non-terraced cultivated fields, primary field data crops harvesting from the fields. Section of traditional ditches that generated using informal discussions, observations and field rill washed run-off accompanied with uncovered crop seedlings also survey methods. During the transect walks informal discussions identified, numbered and measured as rills. with farmers working on their farm plots undertaken concerning In field assessment of rill erosion, detailed measurements of soil erosion indicators and magnitude on both terraced and non- length, and cross-section of rills using tape are considered terraced fields. Frequent observations also carried out regarding (Stroosnijder, 2005; Casalı, Loizu, Campo, De Santisteban, & lvarez- land use types, soil properties (including soil colour and texture), Mozos, 2006). Hence, for a study rills identified and numbered, and slope length & gradient, erosion indicators (broken terraces and their depths, widths and lengths measured by using tape meter. To traditional ditches, rills, sheet wash, deposition occurrence, pres- assure the precision and reducing potential error from total volume ence of surface runoff entering into the fields from uphill areas) and estimation, measurements of depth and width of each rill channel the existing status of the SWC practices. along its length (Herweg & Ostrowski, 1997) carried out. Mea- Sample form of 24 representative cultivated fields (commonly surements repeated and averaged with emphasis on keeping track identified as infertile and steeply sloping) under terraced and of the actual shape of the rill (often a point of changing width and adjacent non-terraced zones selected in reference to relative loca- depth of rill) as widths and depths of rills are seldom constant tion in hill slope direction of the valley sides and fields covered with throughout the length. Moreover, the widths and depths of rills principal crops (tef and wheat) in the catchment for the rill erosion measured up to 3 times and averaged at a point due to the occur- survey. The total area of the selected fields was 76,500 m2 (7.65 ha) rence of different values. The length of a rill taken between the (Table-1), about 2% of 389.9 ha, the total area of the Beriberi origin area of rill formation on uphill and on downhill where main catchment. rill disappeared due to occurrence of sedimentation and confluence From the selected fields, measuring and recording of the fields’ with other rills. area and rills in the fields undertaken in the presence of cropland Based on the measured depths, widths and lengths of rills, the owners and sometimes DAs as well. Rill survey conducted in quantitative data analysis on magnitude of rill erosion like the different periods of the growing season to increase the reliability of actual damage of surface area covered by rills, the rill density and the data. First, frequent visits undertaken to identify and measure volume of soil lost due to rills’ estimated (Bewket & Sterk, 2003; the rills before planting by crops i.e. between mid-June and early Herweg, 1996). The distributions of the approximate values of 242 E. Debie et al. / International Soil and Water Conservation Research 7 (2019) 239e247

Table 1 Distributions of surveyed ﬁelds by topographic position, conservation status and crop types cultivated in the Beribere micro-watershed.

Location in hill slope direction in the valley sides No. of cropped ﬁelds of No. of cropped ﬁelds of non- Total terraced zone terraced zone

Tef Wheat Tef Wheat

Uphill-slope 3(.92) 2(.88) 2(.66) 1(.26) 8(2.72) Middle-hill 4(1.32) 3(1) 2(.8) 2(.61) 11(3.73) Downhill 2(.5) 1(.25) 1(.2) 1 (.25) 5(1.2) Total 9(2.74) 6(2.13) 5(1.66) 4(1.12) 24(7.65)

Note: Values in parentheses represent total surface area of the surveyed fields in ha. actual damage, rills density and volume of soils loss of all identified transitional phase between sheet and gully erosion. Hence, it can be rills across the surveyed fields used to analyse spatial variation assessed by using intermediary field study in contrast to sheet and between homogeneous segments in terms of status of conservation gully erosion that need micro and macro modelling, respectively. practices, location in hill slope direction of the valley sides, crop Accordingly, results in Tablee2 reveal that total numbers and patterns and accelerating factors of rills initiation and lengths of rills, and damaged areas and volumes of eroded soils by development. rills under different conditions of conservation statuses and crop The actual damage of surface area covered by rills calculated patterns of 24 surveyed fields in the Beribere catchment in the using equation: study year of 2015. For instance, the total numbers of rills were 172 in 4.87 ha terraced fields and 126 in 2.78 ha non-terraced fields. The total length of rills was 406.5 m ha 1 in the terraced fields and 710.4 m ha 1 in the non-terraced fields. This indicates that rill density was reduced by 42.8% in terraced fields, compared with non-terraced fields. The total actual surface area damaged was where ADSA-is Actual Damage of Surface Area covered with rills (in 2 1 2 1 133.41 m ha in terraced fields and 302.7 m ha in the non- m2 ha 1) in the study year; L is rill length in meter; W is the mean i i terraced fields, implying that terracing contributed to reducing width of rill in meter; ni is the number of rills of each homogeneous damaged area by 55.93%, relative to non-terraced fields. Further- segment; and A is total area of homogeneous fields in ha. Rill more, damaged area as percent of total in the non-terraced fields density (in m ha 1) in the study year was obtained by: dividing the was 2.31 times that of the terraced fields. This shows the direct total length of all rills (in meter) by the total area of all surveyed fi effect of terracing practices on reducing the damage of productive homogeneous elds (in ha). fi 3 1 1 cropping area in the eld. Moreover, to calculate the approximate volume (m ha yr )of The total volume of all rills was 22.8 m3 ha 1 in the non-terraced soil removed by the rill channel: fields and 10.6 m3 ha 1 in the terraced fields (Table-2). This entails that rill erosion was reduced by 53.5% in terraced fields when compared with non-terraced (control) fields. In agreement with this, different experimental studies have reported significant reduction in soil loss due to implementation of conservation where: V is total volume of soil loss due to formation of rills from practices in the high rainfall areas of northwest highlands Ethiopia. 3 1 homogeneous fields (in m ha ) in the study year; Wi is the mean For instance, Herweg and Ludi (1999) reported that the majority of fi width of rill in meter; Di is the mean depth of rill in meter; Li is the SWC treatments showed a signi cant effect on reduction of soil length of rill in meter; ni is the number of rills of each homoge- loss. Results of the study conducted by Amare et al. (2014) in Debre- neous segment; and A is total area of homogeneous fields in Mewi watershed indicate that 63% of soil loss reduced due to hectare. combined effect of soil bunds with elephant grass, compared to the The magnitude of soil loss is normally expressed either in unit of non-conserved plots in similar setting. In the same study site, mass or in volume per unit area per unit of time (Morgan, 2005). Fisseha et al. (2011) also reported that rill erosion due to the Fanya- For this study, the rate of soil loss is described in unit of volume per juu with elephant grass, Fanya-juu with Vetiver grass and sole ha (m3 ha 1) in the study year. Quantitative analysis is substanti- Fanya-juu reduced by 75.1, 80.3 and 63.6% respectively, when ated with qualitative data generated using observations and compared with the non-conserved plots in a similar setting. informal discussions. Nevertheless, as sheet erosion is often so Nevertheless, for this study, the measured magnitudes of rill inconspicuous and difficult to measure using experimental erosion underestimated the actual rates of soil loss since inter-rill/ methods, its measurement is excluded from this survey study. sheet/erosion excluded. Sheet wash is the most effective sediment Moreover, analysis of temporal variation of soil erosion due to rills transport agent, in particular when the tilled cultivated field is is not incorporated in this study since the survey was planned to be uncovered before and early after crop sowing time (Morgan, 2005). completed within one year. Data computed using the above models Govers and Poesen (1988) found that the relative importance of was further supported with qualitative information obtained from inter-rill erosion ranges between 22 and 46% whereas soil materials field observation and informal discussion with farmers. transported in rill erosion accounted for 54%e78% of the total erosion. In contrast to this estimates, Vandaele and Poesen (1995) 3. Results and discussion found that rill erosion in the hill slopes accounted for 33% of the erosion coverage for the three years period in the hammerveld-1 3.1. Effects of conservation structures on magnitude of rill erosion catchment, central Belgium. In the rill survey methodology, it is probably to underestimate rill erosion by 10e30% since it ignores Terraced SWC structures stabilized with vegetative supports can the contribution of inter-rill erosion to the sediment carried in the reduce the volume, speed and direction of concentrated soil rills and depends upon being able to identify conspicuously the movement along narrow water courses like rills. Run-off in rills is a edge of the rills (Morgan, 2005). Accordingly, for this study by E. Debie et al. / International Soil and Water Conservation Research 7 (2019) 239e247 243

Table-2 Numbers and lengths of rills, and damaged area and volume of eroded soils by rills in 24 surveyed ﬁelds of Beribere micro-watershed in 2015.

Parameter of rill erosion Surveyed cultivated ﬁelds (7.65 ha)

Terraced (4.87 ha) Non-terraced (2.78 ha) Tef cropped (4.4 ha) Wheat cropped (3.25 ha)

Total number of rills 172 126 149 149 Total length (m ha 1) 406.5 710.4 441 619.6 Total damaged area (m2 ha 1) 133.41 302.73 193.61 215.3 Damaged area as % of total 1.3 3 1.94 2.2 Eroded soil volume (m3 ha 1) 10.64 22.84 15.34 16.9

Note: Minimum mean depth of rill was 2cm(influenced by improper constructed ditches) in the terraced and wheat-cropped fields, situated in down-slope of valley sides of the catchment. assuming the contribution of inter-rill/sheet/erosion as 28% in stimulated surface run-off. In association to this, the field obser- terraced fields and 30% in non-terraced fields, the annual actual soil vations discovered the trampling of tilled surfaces by animals for tef loss due to rills and sheet erosion estimated at 32.6 m3 ha 1 in the seedbed preparation in the study district. This probably results non-terraced fields and 14.8 m3 ha 1 in the terraced fields (Table 2). potential prevalence of splash and sheet erosion during the heavy This, perhaps, indicates that the magnitude of soil loss in the rain storms when the fields are uncovered before and early crop- terraced fields was not likely at an acceptable level, compared with ping in the sowing time. the mean soil loss of 11 t ha 1yr 1, which is generally considered as tolerance level (Hudson, 1981; cited in Morgan, 2005). In agree- 3.3. Location at the hill slope of the fields and rill erosion ment to this, Subhatu et al. (2017) reported that net soil loss was fi above tolerable level on wider spaced terraced elds in Minchet Figure-2 demonstrates that surface area damaged by the rill and catchment, northwest Ethiopia. However, for this study, the esti- volume of soils eroded due to rill from location in hill slope di- mated values of soil loss due to rills and sheet erosion (Table-2) are rection of the valley sides of the study site. In both terraced and not more likely presented as total actual soil loss including sedi- non-terraced areas, fields in middle-hill were more susceptible to ments leaving the hill slopes. In relation to this, Subhatu et al. rill erosion than other location in hill slope direction of the valley (2017) indicated that from the total soil loss from terraced crop- sides. For instance, under the terraced areas, the surface area of e lands, about 54 74% sediment was deposited on the deposition actual damaged from the middle-hill fields was 1.47 times that of fi zone of terraced crop elds. the uphill and 1.79 times that of downhill fields. Likewise, the volume of soil eroded from middle-hill fields was 1.47 times that of fi 3.2. Crop type and rill erosion the uphill and 1.53 times that of downhill elds. In the non-terraced area, the surface area actual damage from fi When the root and shoot densities of crops increase on the the middle-hill elds was 1.72 times that of the uphill and 1.16 fi topsoil, the erosive force of the run-off exponentially decrease, times that of the downhill elds. As well, the volume of soil fi particularly for sheet and rill erosion in the early plant growth removed from middle-hill elds was also 1.47 times that of the fi stages (Gyssels & Poesen, 2003). Table-2 shows the density of rills, uphill and 1.4 times that of the downhill elds. This is, perhaps, due affected area and volume of soil removed due to rills in the tef and to increasing of the potential of concentrated run-off that entering fi wheat cropped fields in the Beribere catchment in 2015. Total rill from the uphill elds. In agreement with this, Bewket and Sterk density, proportion of damaged area out of the total area and total (2003) reported that middle-hill in non-terraced area was more volume of soil removed slightly increased in the wheat-cropped vulnerable to rill erosion. fi fi fields when compared to tef -cropped fields. For instance, the to- Next to elds from middle hill, elds from the uphill in the fi tal length of rills was 619.6 m ha 1 in the wheat fields and terraced area and elds from the downhill in the non-terraced area 441 m ha 1 in the tef cropped fields, implying that rill densities in found to be vulnerable for rill erosion. This may be due to the wheat fields were 1.4 times that of the tef fields. Out of the total contribution of terraces to reduce hydrological connectivity (the area, the proportion of the actual damaged cropping area due to processes of surface run-off concentration) by reducing surface rills in the wheat fields was also 1.34 times that of the tef fields. gradient and diverting course on uphill and middle-hill of the Moreover, the annual total soil loss due to rills was 16.9 m3 ha 1 valley sides of the catchment. in the wheat fields and 15.34 m3 ha 1in the tef fields; indicating that soil loss due to rills in the wheat cropped fields was 1.1 times 3.4. Factors accelerating rill erosion that of the tef cropped fields. This is probably, due to more culmi- nation of sheet flow into rills on cropping horizons because of Based on field observations and measurements, concentrated constructing wide spaced ditches and inappropriate gradient of run-off entering from areas of uphill direction, damaging of terraces row crop sowing in the wheat fields than in the tef fields. There was and ditches, improper constructed ditches (which easily expose for initiation of more rills due to improper constructed ditches and cut- bed erosion of tilled soils of channel of ditches), growing of sheet off drains in the wheat fields than tef fields. Even the dimensions of flow in-situ were found to be the major factors accelerating rill rills of wheat fields were larger than tef that might be due to tearing erosion in the study site. of tilled surfaces by highly higher erosive power run-off generated Table-3 reveals that out of the total volume of soils removed by from the widely spaced drainage ditches. In contrast to this study rills in the terraced fields; almost the three-fourth was from results, some studies indicated that soil loss rate in tef-cropped damaged terraces. Therefore, in the terraced areas, terrace damage fields was larger than wheat fields. For example, Tefera and Sterk is the most important contribut ing factor for soil loss due to rills in (2010) reported that the highest rill densities observed in the tef both tef and wheat cropped fields (Figure-3). fields compared to other fields was because the fields were com- As checked from field observation and informal discussion with pacted by animal trampling that reduced infiltration rates and farmers, lack of proper maintenance of bunds and diversion ditches 244 E. Debie et al. / International Soil and Water Conservation Research 7 (2019) 239e247

Figure-2. Rill erosion distribution in terraced and non-terraced ﬁelds by location at the hill slope of the valley sides of Beribere Micro-watershed.

Table-3 Factors accelerating rill erosion in terraced, non-terraced, tef and wheat cropped ﬁelds.

Factors accelerating rill erosion (in m3) in the ﬁelds Field types

Terraced Non-terraced Tef cropped Wheat cropped

Uphill-ﬁeld entering 1(1.9) 27.4(43.2) 15.1(22.4) 13.3(24.2) Terrace damage 38.7(74.7) e 23.3(34.5) 15.4(28.1) Ditch damage 2.8(5.4) 27.3(43) 22.3(33) 7.8(14.2) Improper-constructed ditches 8.1(15.6) 6.8(10.7) 5.8(8.6) 9.14(16.6) Sheet grown in-situ 1.24(2.4) 2(3.1) 1(1.5) 9.3(16.9) Total 51.84(100) 63.5(100) 67.5(100) 54.9(100)

Note: The values represent soils lost due to rills erosion (m3) and ﬁgures in the parenthesis represented percentages.

Figure-3. Rill formation due to insufficient maintenance of terraces. before the first rainstorms and occurrences of loose and bare soils the probability of rill prevalence on the cropping field is minimized. due to repeated tillage practices are most important factors for Unfortunately, large numbers of farmers often prepare to mitigate terrace damages and rills formations in the downhill parts of erosion which is often more vigorous on bare lands due to larger cropped fields. For instance, when the first terraces of the uphill erosive force run-off. This probably happens due to inadequate sections of cultivated fields are well maintained and stabilized, the knowledge about the source of initiation of rills along with topo- lengths of the rills on the middle and lower parts of the specific graphic characters of cropping areas or inadequate labour to cultivated fields are largely decreased. This perhaps could be owing excavate soils from terraces and diversion ditches. Moreover, bund to the substantial reduction of higher erosive force of the run-off by destruction by farmers, lack of adapting well-constructed terraces the first stabilized bunds. This implies that when terraces on (unable to drain surface water) and improper drainage ditches are cropping areas are timely and properly maintained and stabilized, important factors for the damage of terraces and rill occurrences in E. Debie et al. / International Soil and Water Conservation Research 7 (2019) 239e247 245

Fig. 4. Rill development due to entering of concentrated runoff from uphill fields. the terraced cultivated fields. Because of these interactive factors, fields is 1.35 times that of wheat fields. This further pointed out that the upper section of cropping areas between terraces is more tef crop fields are highly vulnerable to the erosive power of vulnerable to rill erosion than middle and lower parts. concentrated run-off that enters from uphill areas initiated by Rill numbers decrease when moving from upper to lower parts damaging terraces and ditches. of cropping areas between terraces due to the contribution of On the other hand, out of the total soil volume removed by rills indigenous drainage ditches or furrows. Even rills usually disappear in the wheat cropped fields, 16.6% and 16.9% accounted for from the middle parts of terraced areas towards the lower parts. improper constructed traditional ditches and strength of overland Drainage ditches or furrows play an important role in interrupting flow, respectively. However, of the total volume of soil eroded by sheet flow growing at its early stages. Moreover, in the fields where rills in the tef-cropped fields, 8.6% and 1.5% was accounted for terraces are well constructed, stabilized with vegetative practices improper constructed traditional ditches and overland flow like sesbania sesban, properly maintained and complemented with strengths, respectively. This indicates that the influence of sheet proper practices of indigenous drainage ditches; the occurrence of flow and improper constructed ditches on rill erosion in tef fields rills in the terraced cultivated fields often disappears. This perhaps was lower than in wheat fields. As observed in field observation, the indicates that the combined use of technically fitted indigenous up-down row-direction of wheat seed sowing could cause exces- drainage ditches and introduced terraces stabilized with sesbania sive erosion; rills were often generated in tillage paths as the result sesban are more likely to curb soil loss than a single practice. of strengthened sheet erosion in-situ (Figure-5). So, under the In the non-terraced fields, entering of concentrated run-off from conventional tillage, growing of sheet-wash into rills within crop- uphill areas and damage of ditches were observed as the principal ping horizons (between traditional ditches or terraces) was iden- factors accelerating the formation of rills and development of rill tified. Alternatively, when seed sowing was in contour row, erosion (Figure-4), whereas they had insignificant effect on rill growing of sheet-wash into rills disappeared. Because of consid- erosion in the terraced fields. For example, of the total soil lost due erable proportion of bare tilled surface, particularly in the early to rills in the non-terraced areas, about 43% was because of entering stage of crop growth, row crops on sloping fields give rise to more of concentrated run-off from uphill fields (Table-3). This indicates rill and inter-rill erosion problems (Arnhold et al., 2014; Morgan, that terracing has substantial contribution to control erosive power 2005). run-off entering from uphill areas. From the total volume of soil Similarly, poorly designed traditional ditches with significantly eroded by rills, damage of traditional ditches accounted for 43% gradient enough by themselves often accelerated rill erosion (Table-3). Based on frequent observations, absence of accompanied through bed erosion on channels of ditches and cut-off drains control practices (like terraces); long length of ditches along with (Figure-5). The volume of removed parts that are not covered by shallow depths, erosive power run-offs and unsafe gradients of crop seedlings could be an indication of soil loss from bed surface of ditches are contributing factors for breakage of traditional drainage drainage ditches. This occurrence was largely observed in the ditches and for the prevalence of rills on the downhill cropped wheat-cropped fields probably due to the flow of high erosive force fields. Sometimes, rills are also occurring due to the damage of cut- of run-off (which generates from relatively wider affected cropping off drains (locally named as Tekebikeb) because of inadequate horizons) on the channels of ditches and cut-off drains. However, excavation of tilled soils from channels during construction and according to the informal discussion with the owners of surveyed maintenance time. cultivated fields, a large number of farmers do not recognize The degree of influence of the factors on rill erosion also differs negative effects of improper constructed ditches and cut-off drains in tef and wheat cropped fields. On the one hand, out of the total for causing soil loss due to rills. eroded volume of soil due to rills, the combined effect of entering run-off from the uphill fields, and the damage of terraces and 4. Conclusions ditches accounted for 89.9% in tef and 66.5% in the wheat-cropped fields. This implies that because of the combined effect of these This study assessed the effect of conservation structures on accelerating factors, the proportion of soil loss due to rills in tef reducing the magnitude of soil loss by using a rill survey 246 E. Debie et al. / International Soil and Water Conservation Research 7 (2019) 239e247

Fig. 5. Rills formation in tef and wheat cropped fields due to ill-designed drainage ditches. methodology on 24 cultivated fields (with respect to crop type and of soil erosion and important for pragmatic assessment of the location in hill slope direction of the fields) and the factors accel- effectiveness of structural practices. erating rill erosion in the Beribere micro-watershed, in the northwest highlands of Ethiopia. Based on the results of the rill survey Acknowledgements study, the total volume of soil removed by rills in the study 3 1 catchment was about 22.8 m ha in the non-terraced fields and We thank Dr. Arega Bazezew for his help in reading the final 3 -1 10.6m ha in the terraced fields. This indicates that rill erosion was reviewed version of this manuscript. The field work of the study reduced by 53.5% in the terraced fields compared to the non- was sponsored by the Graduate School of Addis Ababa University, terraced fields. In the terraced area, terrace damage is the most Ethiopia. important contributing factor for soil loss due to rills. In the non- fi terraced elds, entering of concentrated run-off from uphill area Appendix A. Supplementary data and ditch damages were observed as principal accelerating factors for formation and development of rill erosion, whereas these fac- Supplementary data to this article can be found online at fi fi tors had insigni cant effects on rill erosion in the terraced elds. https://doi.org/10.1016/j.iswcr.2019.06.001. Moreover, soil erosion demonstrated spatial difference in terms fi of crop cover and relative location in hill slope direction of the elds References in the Beribere catchment. For instance, tef cropped fields were less vulnerable to rill erosion as compared to the wheat cropped fields. Amare, T., Derebe, A., Yitaferu, B., Steenhuis, T. S., Hurni, H., & Zeleke, G. (2014). This is probably due to the higher contribution of sheet erosion and Combined effect of soil bund with biological soil and water conservation erosion on beds of ditches to the magnitude of rill erosion in wheat measures in the northwest highlands of Ethiopia. Echohydrology and Hydrobi- ology, 14,192e199. fields than tef. Sheet flow grown in-situ and ditches bed erosion Amdihun, A., Gebremariam, E., Rebelo, L., & Zeleke, G. (2014). Modelling soil erosion mainly occurred as a result of inappropriate gradient of row crop dynamics in Blue Nile basin: A landscape approach. Journal of Environmental e sowing and widely spaced improper constructed ditches, respec- Sciences, 8(5), 243 258. Arnhold, S., Lindner, S., Lee, B., Martin, E., Kettering, J., Nguyen, T. T., et al. (2014). tively. All surveyed fields from middle-hill of the valley sides were Conventional and organic farming: Soil erosion and conservation potential for the most susceptible to rill erosion than other relative location in row crop cultivation, 219e220 Geoderma,89e105. Retrieved from https://doi. hill slope direction. org/10.1016/j.geoderma.2013.12.023. Belay, M., & Bewket, W. (2012). A participatory assessment of soil erosion and farm It is concluded that in spite of their advantages on considerable management practices in northwest Ethiopia. In Paper presented at the 8th in- soil loss reduction, terracing structures are inadequate to curb soil ternational symposium of AgroEnviron 2012, 1-4 may 2012, wageningen, The erosion by water effectively. Erosion control practices (especially Netherlands. Retrieved from http://library.wur.nl/ojs/index.php/AE2012/article/ viewFile/12449/12539 Accessed June 2012. terracing and traditional ditches) often cause rill formation unless Belay, M., & Bewket, W. (2015). Enhancing rural livelihoods through sustainable they properly constructed, maintained and stabilized. This aspect land and water management in northwest Ethiopia. Geography Environment makes them site specific erosion control practices for addressing Sustainability, 8(2), 79e100. ’ Bewket, W. (2003). Land degradation and farmers' acceptance and adoption of the problem. Improving farmers skills of designing and practicing conservation technologies in the Digil Watershed, northwestern highlands of indigenous and introduced structural practices through effective Ethiopia. Social Science Research Report Series/OSSREA, 29, 65. Retrieved from: communication between farmers and extension staffs should http://library.wur.nl/WebQuery/wurpubs/326000. fi accentuate. Furthermore, it is suggested that since erosion control Bewket, W., & Sterk, G. (2003). Assessment of soil erosion in cultivated elds using a survey methodology for rills in the Chemoga watershed, Ethiopia. Agriculture, practices can do little to prevent detachment of soil particles, they Ecosystems & Environment, 97,81e93. need to be complemented with considerable agronomic SWC Boix-Fayos, C., Martínez-Mena, M., Arnau-Rosalen, E., CalvoCases, A., Castillo, V., & fi practices to reduce soil, water and nutrient losses at acceptable Albaladejo, J. (2006). Measuring soil erosion by eld plots: Understanding the sources of variation. Earth-Science Reviews, 78, 267e285. levels and thus enable the cultivated fields to sustain crop yields. Casalı, J., Loizu, J., Campo, M. A., De Santisteban, L. M., & lvarez-Mozos, J. A. (2006). Eventually, the study corroborates that rill surveying method pro- Accuracy of methods for field assessment of rill and ephemeral gully erosion. e vides significant semi-quantitative information on the magnitude Catena, 67,128 138. Evans, R. (2002). An alternative way to assess water erosion of cultivated land e E. Debie et al. / International Soil and Water Conservation Research 7 (2019) 239e247 247

field-based measurements and analysis of some results. Applied Geography, 22, Publishing. 187e208. Nyssen, J., Clymans, W., Poesen, J., Vandecasteele, I., De Baets, S., Haregeweyn, N., Fisseha, G., Gebrekidan, H., Kibret, K., Bedadi, B., & Yitaferu, B. (2011). Participatory et al. (2009). How soil conservation affects the catchment sediment budget-a development of soil conservation measures at the Debre-Mewi watershed in comprehensive study in the north Ethiopian highlands. Earth Surface Pro- the upper catchment of the Blue Nile Basin, northwest Ethiopia. Biodiversity and cesses and Landforms, 34, 1216e1233. Environmental Sciences, 1(6), 199e213. Nyssen, J., Poesen, J., Gebremichael, D., Vancampenhout, K., D’aes, M., Yihdego, G., Gebremichael, D., Nyssen, J., Poesen, J., Deckers, J., Haile, M., Govers, G., et al. (2005). et al. (2007). Interdisciplinary on-site evaluation of stone bunds to control soil Effectiveness of stone bunds in controlling soil erosion on cropland in the erosion on cropland in northern Ethiopia. Soil and Tillage Research, 94,151e163. Tigray Highlands, northern Ethiopia. Soil Use & Management, 21, 287e297. Okoba, B. O., & De Graaff, J. (2005). Farmers' knowledge and perceptions of soil Govers, G., & Poesen, J. (1988). Assessment of the inter-rills and rills contributions to erosion and conservation measures in the central highlands, Kenya. Land total soil loss from an upland field plots. Geomorphology, 1(4), 343e354. Degradation And Development, 16,475e487. Gyssels, G., & Poesen, J. (2003). The importance of plant root characteristics in Pimentel, D. (2006). Soil erosion: A food and environmental threat. Environment, controlling concentrated flow erosion rates. Earth Surface Processes and Land- Development And Sustainability, 8,119e137. forms, 28, 371e384. Pimentel, D., Harvey, C., Resosudarmo, P., Sinclair, K., Kurz, D., McNair, M., et al. Haile, M., Herweg, K., & Stillhardt, B. (2006). Sustainable land management: A new (1995). Environmental and economic cost of soil erosion and conservation approach to soil and water conservation in Ethiopia. Centre for Development and benefits. Science, 267,1117e1123. Environment (CDE), University of Bern. Stocking, M. (2001). Field assessment of erosion and soil productivity from the Haregeweyn, N., Berhe, A., Tsunekawa, A., Tsubo, M., & Tsegaye, D. (2012). Inte- perspective of the land user. London: Earth Scan. grated watershed management as an effective approach to curb land degra- Stroosnijder, L. (2005). Measurement of erosion: Is it possible? Catena, 64,162e173. dation: A case study of the enabered watershed in northern Ethiopia. Subhatu, A., lemann, T., Hurni, K., Portner, B., Kasswmar, T., Zeleke, G., et al. (2017). Environmental Management, 50, 1219e1233. Deposition of eroded soil on terraced croplands in Minchet catchment, Ethio- Hargeweyen, N., Tsunekawa, A., Poesen, J., Tsubo, M., Tsegaye, D., Almaw, A., et al. pian highlands. International Soil and Water Conservation Research, 5, 212e220. (2017). Comprehensive assessment of soil erosion for better land use planning Tadesse, L., Suryabhagvan, K. V., Sridhar, G., & Legesse, G. (2017). Land use and land in river basins: Case study of upper Blue Nile River. The Science of the Total cover changes and soil erosion in Yezat watershed, northwest Ethiopia. Inter- Environment, 574,95e108. national Soil and Water Conservation Research, 5,85e94. Herweg, K. (1996). Field manual for assessment of current erosion damage. Soil con- Taye, G., Poesen, J., Van Wesemael, B., Vanmaercke, M., Teka, D., Deckers, J., et al. servation Research programme (SCRP) -Ethiopia and centre for development and (2013). Effects of land use, slope gradient, and soil and water conservation environment (CDE). Switzerland: University of Berne. structures on runoff and soil loss in semi-arid Northern Ethiopia. Physical Ge- Herweg, K., & Ludi, E. (1999). The performance of selected soil and water conser- ography, 34(3), 236e259. vation measures-case studies from Ethiopia and Eritrea. Catena, 36,99e114. Tefera, B., & Sterk, G. (2010). Land management, erosion problems and soil and Herweg, K., & Ostrowski, M. W. (1997). The influence of errors on erosion process water conservation in Fincha’ watershed, western Ethiopia. Land Use Policy, 27, analysis . Centre for development and environment (CDE), Ethiopia soil conserva- 1027e1037. tion Research programme. Research report 33. Tegene, B. (2000). Processes and causes of accelerated soil erosion on cultivated Hudson, N. W. (1981). Soil Conservation (second ed.). London: Batsford, 324p. fields of South Welo, Ethiopia. Eastern Africa Social Science Research Review, Hurni, H. (1993). Land degradation, famine, and land resource scenarios in Ethiopia. 16(1), 1e22. In D. Pimentel (Ed.), World soil erosion and conservation (pp. 27e62). UK: Vandaele, K., & Poesen, J. (1995). Spatial and temporal patterns of soil erosion rates Cambridge University Press. in an agricultural catchment, central Belgium. Catena, 25,213e226. Hurni, H. (1998). Degradation and conservation of the resources in the Ethiopian Zeleke, G., Kassie, M., Pender, J., & Yesuf, M. (2006). Stakeholder analysis for sus- highlands. Mountain Research and Development, 8,123e130. tainable land management (SLM) in Ethiopia: Assessment of opportunities, Lal, R. (2001). Soil degradation by erosion. Land Degradation & Development, 12, strategic constraints, information needs, and knowledge gaps. In Environmental 519e539. economics policy forum for Ethiopia (EEPFE). Addis Ababa. Morgan, R. P. (2005). Soil erosion and conservation (3rd ed.). USA: Blackwell International Soil and Water Conservation Research 7 (2019) 248e257

Contents lists available at ScienceDirect

International Soil and Water Conservation Research

journal homepage: www.elsevier.com/locate/iswcr

Original Research Article Assessment of the determinants that inﬂuence the adoption of sustainable soil and water conservation practices in Techiman Municipality of Ghana

* Kwasi Adjepong Darkwah a, , Joana Deladem Kwawu b, Frank Agyire-Tettey b, Daniel Bruce Sarpong c a Department of Statistics, University of Ghana, Ghana b Department of Economics, University of Ghana, Ghana c Department of Economics and Agribusiness, University of Ghana, Ghana article info abstract

Article history: This paper assesses the relationship between farmer characteristics and the degree to which nine soil Received 25 July 2017 and water conservation practices (SWCPs) are adopted by 300 maize farmers in Techiman Municipality, Received in revised form Ghana. Farmers were surveyed for their adoption of nine SWCPs, and 24 other characteristics including 27 October 2018 demographics, socio-economic factors, risk factors and costs of production. Sustainable soil and water Accepted 28 April 2019 conservation practices (SWCPs) in sub-Saharan Africa such as Ghana are important because they have Available online 24 May 2019 positive effects on yield, increase sustainability of farming, stop degradation and reduce soil erosion. The adoption of sustainable soil and water conservation practices in the agricultural industry of Ghana has Keywords: Sustainable soil and water conservation been variable. This study aims to explain differences in farmer adoption of SWCPs, by assessing factors practices that vary with the number of different SWCPs used by farmers. Hence, the Poisson model was used. The Adoption results show that farmer's household size, farm size, access to credit services and formal training of Correlation maize farmer have a positive significant association with the number of soil and water conservation Poisson regression practices adopted by maize farmers while distance to nearest output market, distance to input center, access to extension services, and risk of pest and diseases have a negative significant association with the number of soil and water conservation practices adopted by maize farmers at 5% significance level. The study concludes that any further research in Techiman Municipality on soil and water conservation practices should acknowledge the mixture of personal and demographic, institutional, socio-economic and risk factors. This suggests that agricultural policies formulated by the government should be aimed at supporting maize farmers to have access to extension service contact for frequent disseminating of agricultural technology information which is likely to increase the rate of adoption of soil and water conservation practices. © 2019 International Research and Training Center on Erosion and Sedimentation and China Water and Power Press. Production and Hosting by Elsevier B.V. This is an open access article under the CC BY-NC- ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction that maize has received more recognition than cassava as the main food crop based on the number of calories consumed in Africa. Maize has been identiﬁed over the years as one of the main According to Xu, Guan, Jayne, and Black (2009), the selling of maize essential cereal crops used as food to feed humans and animals in have dominated the agricultural markets in the world. Due to this, both developed and developing countries. It is mostly cultivated by the Ghanaian government has made several efforts to enhance small scale farming households. Webb and Highly (2000) stated maize productivity in Ghana. Recently, there has been an increase in the production of food crops by smallholders but this is still characterized by low productivity (FAO, 2015). The Ministry of Food “ * Corresponding author. and Agriculture's (MOFA's) stated vision is modernized agriculture E-mail addresses: [email protected] (K.A. Darkwah), [email protected]. culminating in a structurally transformed economy and evident in edu.gh (J.D. Kwawu), [email protected] (F. Agyire-Tettey), [email protected] food security, employment opportunities and reduced poverty”. (D.B. Sarpong). https://doi.org/10.1016/j.iswcr.2019.04.003 2095-6339/© 2019 International Research and Training Center on Erosion and Sedimentation and China Water and Power Press. Production and Hosting by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). K.A. Darkwah et al. / International Soil and Water Conservation Research 7 (2019) 248e257 249

One of the key goals in the Food and Agriculture Sector Develop- been done to assess the factors that influence the adoption of some ment Policy (FASDEP II, 2007) for Ghanaians is the transformation of these sustainable agricultural practices such as soil and water of agriculture and improved the productivity of farmers. conservation practices. In this study, the soil and water conserva- The revenues generated from the agricultural sector directly tion practices used comprises of row planting, intercropping, influences economic growth, food security, social welfare and improved seed use, composting, cover crops, crop rotation, inter- reduction in poverty (Mango, Makate, Tamene, Mponela, & cropping, mulching, agrochemical use, zero tillage, fertilizer use, Ndengu, 2017). Tweneboah (2000) declared that one of the de- and manure. terminants of a household food security is the availability of maize The purpose of this study was to assess the factors that have an in stock irrespective of other foods available in the house. One way association with the adoption of sustainable soil and water con- of attaining food security without foreign food aids in Ghana is servation practices for maize production in Ghana. The hypothesis through the expansion of agricultural productivity such as maize of the study was that the number of new technology for soil and production using improved agricultural technologies (Alhassan, water conservation practices adopted by farmers is associated with Salifu, & Adebanji, 2016). Even though the Government of Ghana the combined impact of personal and demographic, socio- and the Ministry of Food and Agriculture have made several efforts economic, institutional, risk, cost of production and psychological to introduce new varieties of maize, there is low productivity of factors. This is important in promoting the adoption process since maize by farmers (Alhassan et al., 2016). The low levels of maize pursuing those factors can enhance the adoption of soil and water productivity in Ghana could be ascribed to the low level of adoption conservation practices. The adoption process was considered based of sustainable soil and water conservation practices (MoFA, 2011). on the innovation-decision theory. Aikins, Afuakwa, & Owusu-Akuoko (2012) in their studies stated This study adds to an existing body of literature on farmer poor agronomic practices such as traditional seeds usage, low usage adoption of new SWCPs in Ghana, for example, Nkegbe and of chemical fertilizer, zero tillage as the main contributing factors Shankar (2014) and Abdul-Hanan (2017). Assessing vital factors that affect maize production in Ghana. that motivate maize farmers to adopt soil and water conservation Mucavele (2013) identified that most of the smallholder farmers practices is also an important contribution to literature since the in sub-Saharan African nations have not stopped practicing tradi- findings can be used to check further glitches of resource degra- tional methods of farming. Majority of the smallholder farmers dation that is a menace to efforts to minimize poverty reduction in employ no or least developed inputs to their farming activities Techiman Municipality. The practice of soil and water conservation instead of applying modern soil and water conservation technolo- such as no-tillage also prevents soil erosion which threatens the gies. Mango et al. (2017) stated that the shortage of variation in economic viability and survival of the agricultural industry (Busari, farming systems, with maize constituting 80e90% of the crop, also Kukal, Kaur, Bhatt, & Dulazi, 2015). Furthermore, this study fills the intensifies glitches of degradation of land, water and soil resources gap in understanding the adoption of modern technology such as which results in soil fertility problems and therefore have a nega- soil and water conservation practices that increase agricultural tive effect on agriculture. productivity in the Techiman Municipality of Ghana. Degradation of the soil results in the weakening of soil quality which lessens the productivity of the soil. This, therefore, jeopar- 2. Literature review dizes the sustainability of agriculture, the quality, and stability of the environment and further has an adverse effect on the economic Adoption is defined by Abdul-Hanan (2017) as the extent to and social development of the nation. Without taking the necessary which modern technology is utilized. Several fields have described soil and water conservation practices, the cost of tackling soil adoption based on their own perception. Sociologists referred to degradation will increase and productivity will continually reduce, adoption as the nature of communication channels, Economists hence influencing revenues generated from exporting agricultural described adoption in terms of profitability while Anthropologist products as well as causing a rise in food insecurity for most people and Geographers referred to adoption as the compatibility of in sub-Saharan Africa. innovation (Boahene, Snijders, & Folmer, 1999). This paper used the According to Ajayi (2007), adoption of land, soil and water adoption definition of Abdul-Hanan (2017) based on the extent to conservation practices has become a key concern in the develop- which farmers in Techiman Municipality adopts the nine soil and ment policy agenda for sub-Saharan Africa in their quest to curb the water conservation practices. Factors affecting the adoption of issue. Soil and water conservation practices are mostly adopted by modern technology can be grouped into personal and demographic farmers in Techiman Municipality to reduce soil erosion, conserve characteristics, socio-economic, institutional and psychological soil and water, reduce on-farm costs, increase soil fertility, reduce factors, cost of production and risk factors (Feder, Just, & Zilberman, toxic contamination of surface water and groundwater, reduce air 1985; Kaliba, Verkuijl and Mwangi, 2000; Rogers, 2003; Mazvimavi pollution as a result of soil tillage machinery and safeguard the soil & Twomlow, 2009; Mango, Makate, Tamene, Mponela, & Ndengu, long-term productivity. The adoption of land, soil, and water con- 2017). servation practices that comprise mulching, intercropping, etc. has Kaliba, Verkuijl, and Mwangi (2000) used the Tobit and Probit been considered as part of the agenda of agricultural development models to assess the factors of adoption of improved technologies policy (Mango et al., 2017). Aside the profits farmers derive from such as improved maize seeds and use of inorganic fertilizer in the agriculture as well as the state and international initiatives to intermediate and lowland areas of Tanzania. They identified that inspire farmers to invest in them, there is still low adoption rate of variety characteristics, access to extension services, rainfall, and on- some of these sustainable agricultural practices in the rural regions farm field trials were discovered to have a significant effect on the of developing countries (Kassie, Zikhali, Manjur, & Edwards, 2009). adoption of improved seed and use of inorganic fertilizer for maize Ghana is also experiencing low adoption rate regardless of production. hastened soil erosion, continuous soil degradation that curb yield Mango et al. (2017) studied the factors that affect the awareness and several attempts by the government to encourage maize and adoption of land, soil, and water conservation practices by farmers to adopt various soil and water conservation technologies farmers in Mozambique. They sampled 312 households in Tete (Abdul-Hanan, 2017). Several factors have been identified to have province of Mozambique, Southern regions of Malawi and Eastern an influence on the adoption of modern technology for soil and Province of Zambia and used the t-test to classify adopters and non- water conservation practices but relatively little empirical work has adopters of soil, land, and water conservation techniques and 250 K.A. Darkwah et al. / International Soil and Water Conservation Research 7 (2019) 248e257 binomial logit models to identify the determinants that affect 3. Materials and methods farmers' knowledge of conservation measures and adoption of land productivity practices. Their findings revealed that the household 3.1. Study area head's age, education, agricultural advice reception, farmer group membership and pieces of land owned affect the decision of The study was carried out in the Techiman Municipality of farmers to adopt modern technology. Ghana. It lies between latitude 8000 North and 7035 South and Bayard, Jolly, and Shannon (2006) examined the factors that longitude 1049 East and 2030 west. According to Ghana's 2010 affect the adoption and management of soil conservation practices population census, Techiman Municipality has a total population of such as rock walls in Haiti. They sampled 115 farmers comprising of 147,788 which represent 6.4% of the total population of Brong Ahafo 68 rock wall adopters and 47 non-adopters in Fort-Jacques using region. The Municipality covers a total land area of about probit models to identify the significant factors affecting the 669.7sq.km and lies within the semi-equatorial climate zone management of land improvement technology. Their findings (Ghana Statistical Service, 2014) which occurs widely in the tropics revealed that education, age, group membership, and per capita and experiences a maxima pattern of rainfall with a mean annual income affect the ability to manage the rock walls negatively while rainfall ranging between 1260 mm and 1600 mm in April to July age squared and the interaction between age and per capita income Basically, the Municipality falls within the three vegetation zones affect the ability to manage the rock walls positively. which are semi-deciduous zone located in the South, the guinea- Boz and Akbay (2005) examined the impact of certain socio- savanna woodland in the Northwest and the transitional zone economic features and communication behaviour on the adoption located in the South-East, West, and North of the Municipality. The of maize among Kahramanmaras farm workers in Turkey. They main economic activity in the Municipality is agriculture which sampled 264 operators using stratified sampling technique and employs about 36.3% of its economically active labour force. The compared early adopters, late adopters and nonadopters of maize farming systems practiced is rotational bush fallow and shifting by employing chi-square test of independence and the ordered cultivation (Ghana Statistical Service, 2014). Fig. 1 shows the map of probit method to find the extent to which certain socioeconomic Techiman Municipality. features and communication behaviour affect these adopter fi groups. Their ndings showed that adoption of maize in Kahra- 3.2. Variable Description manmaras, Turkey is affected by farm size, mechanization level, opinion leadership, credit use, agricultural investment, educational The variables used in this article are consistent with Kaliba, level of farmers; travel to nearest town, income, use of extension Verkuijl, and Mwangi (2000), Rogers (2003), Mazvimavi and personnel and credit use. Twomlow (2009), and Mango, Makate, Tamene, Mponela, & In Ghana, Abdul-Hanan, Ayamga and Donkoh (2014) estimated Ndengu, 2017. The count dependent variable used in the adoption the factors of adoption of SWC technology and examined the model indicates the number of technology adoption for soil and fi impact of SWC technology adoption on technical ef ciency using water conservation practices implemented by farmers. The tech- the Poisson model and stochastic frontier model respectively. The nology adoption for soil conservation practices adopted in the study employed data from the Ghana Agriculture Production Sur- study area was developed to reduce soil erosion, to conserve soil vey (GAPS) and sampled 1530 farm households chosen from 20 and water and to safeguard the soil long-term productivity. Nine fi districts across Ghana and identi ed that credit, farm size, group main technologies employed for soil and water management membership and proximity to input sale points positively affect practices were included in the study. These technology adoptions SWC adoption while credit, education and extension services for soil and water conservation comprise of agronomic methods fi ’ fi signi cantly affect farmers technical ef ciency. They concluded such as row planting, intercropping, improved seed use, compost- that keen approaches such as access to education, extension ser- ing, cover crops, crop rotation, mulching, agrochemical use, zero vices and credit to assist farmers should be encouraged. tillage and fertilizer use and manure. Table 1 shows the explanatory Also, Abdul-Hanan (2017) investigated the number of soil and variables used in the study. water conservation practices used by maize farmers in the three Northern Regions of Ghana using the Poisson model. The findings of the study show that male farmers, farmers who are closer to 3.3. Data collection instruments the input store, access to credit, and ownership of livestock employ the highest number of soil and water conservation The study employed a pre-tested semi-structured questionnaire practices. Their study supports the essence of the road network for data collection. The structure of questions in the data collection and agricultural credit provision to aid production technology instrument was a combination of close-ended and open-ended adoption by farmers. questions. The study collected data on the personal and de- Nkegbe and Shankar (2014) investigated the factors of the in- mographic, socioeconomic, institutional, risk, cost of production of tensity of adoption of soil and water conservation practices the farmer, farm and soil characteristics. The survey was conducted employed by smallholder producers in Northern Ghana. The study from February 2017 to June 2017. used count data models for the analysis and found that social capital, wealth, access to information and per capita landholding 3.4. Sampling technique and sample size affect the decision of smallholder producers to intensively adopt soil and water conservation practices. A total of 300 maize farmers were selected for the study using In this study, the analysis was extended to comprise of risk the multi-stage sampling technique. At the first stage, the Techiman factors such as risk of pest and diseases, risk of flood, risk of wild Municipality of Ghana was purposively selected due to the mu- animals and risk of drought; psychological factors such as percep- nicipality recording the highest average yield of 2.24 metric tonnes tion of farmers on the attributes of a technology, perception on the in 2014. At the second stage, 34 major maize producing commu- price of the technology, perception on compatibility of the tech- nities were sampled from six maize operational areas which were nology on their soil and cost of production such as cost of labour chosen from a total of nine operational areas namely Mangoase, and cost of Fertilizer. Forikrom, Tanoso, Nsonkonee, Nsuta, and Bamiri/Aworopataa Operational area. At the third stage, the stratified sampling method K.A. Darkwah et al. / International Soil and Water Conservation Research 7 (2019) 248e257 251

Fig. 1. Map of Techiman Municipality.

Table 1 Explanatory variables used in the study.

Variable Description Expected Sign

Category 1 Personal and Demographic Factors GNR Gender of Household Head (Dummy, 1 ¼ Male, 0 ¼ Female) Negative/Positive AGE Household Head's Age (Years) Negative EDU Household Head's Educational Level Positive EXP Experience of the farmer (Years) Positive HHSIZE Household size (Count) Negative/Positive FTM Formal Training on Maize (Dummy, 1 ¼ Yes, 0 ¼ No) Positive Category 2 Socio-Economic Factors FSIZE Farm Size (Hectares) Positive/Negative OFI Off-farm income (Amount in Ghana Cedi) Positive TL Total number of labour (Count) Positive Category 3 Institutional Factors ACS Access to Credit Services (Dummy, 1 ¼ Yes, 0 ¼ No) Positive AEXT Access to Extension Services (Dummy, 1 ¼ Yes, 0 ¼ No) Positive MAA Membership in an agricultural Association (1 ¼ Yes, 0 ¼ No) Positive DOM Distance to nearest output Market (Hours) Negative/Positive AMM Access to Mass Media (Dummy, 1 ¼ Yes, 0 ¼ No) Positive DTI Distance to input center (Hours) Positive Category 4 Psychological Factors PAT Perception of farmers on the attributes of a technology (Average) Negative/Positive PPT Perception of farmers on the price of the technology (Average) Negative PCT Perception of farmers on the compatibility of the technology on their soil (Average) Negative Category 5 Risk Factor ROD Risk of drought (Dummy, 1 ¼ Yes, 0 ¼ No) Negative ROF Risk of the ﬂood (Dummy, 1 ¼ Yes, 0 ¼ No) Negative RWA Risk of Wild Animals (Dummy, 1 ¼ Yes, 0 ¼ No) Negative RPD Risk of Pest and Diseases (Dummy, 1 ¼ Yes, 0 ¼ No) Negative Category 6 Cost of Production CL Cost of labour (Amount in Ghana Cedi) Negative CF Cost of Fertilizer (Amount in Ghana Cedi) Negative 252 K.A. Darkwah et al. / International Soil and Water Conservation Research 7 (2019) 248e257 was applied to collect data from 34 communities in the 6 opera- Cost of Fertilizer were observed to have a coefﬁcient of variation tional areas. In the last stage, a total of 300 maize farmers were (CV) of 2.89 and 3.01 respectively showing large variability. randomly sampled from the 34 communities in the 6 operational areas. All the respondents signed the informed consent forms as an 4.1. Pairwise correlation analysis and multicollinearity test of agreement to partake in the research study. explanatory variables

3.5. Poisson regression model The study tested the presence of multicollinearity using the correlation analysis and variable inflation factor. Table 3 presents fi The Poisson regression model which is an appropriate model for the Pairwise correlation coef cients for the explanatory variables fl estimation of count data (Greene, 1997) is used to estimate the that in uence the adoption of modern technologies for soil and number of soil and water conservation practices adopted by water conservation in Ghana while Table 4 presents the variable fl fi farmers. Before a Poisson regression analysis is performed, there is in ation factor. Table 3 shows that all the correlation coef cients fi the need to test over dispersion because it causes the standard are less than 0.5. The pairwise correlation coef cient was con- deviation to exceed the mean. Coxe, West, and Aiken (2009) stated ducted between the predictor variables to identify any cases of that both deviance and Pearson Chi-square statistic are employed multicollinearity. From Table 3, the pairwise correlation shows that to examine overdispersion. An index of dispersion is given when there is no multicollinearity. the total Chi-square is divided by the corresponding degrees of From Table 4, the VIF value of each explanatory variable is less freedom. An equidispersion gives an index of 1 for the ratio of mean than 10 which shows that all the explanatory variables can be to standard deviation. included in the Poisson regression model. We assumed that the number of independent soil and water conservation practices i adopted by farmers follows a Poisson dis- 4.2. Soil and water conservation practices tribution with rate n l . Here n denotes the number of soil and i i i Nine main technologies used for soil and water conservation water conservations practices and logðl Þ is the linear predictor. If Y i practices were incorporated in the study. Table 5 shows the list of has a Poisson distribution, then the effects of the covariates the 9 main technologies used for soil and water conservation considered can be modeled using a log-linear model of the form; practices by maize farmers in Ghana and their rate of adoption by b 300 farmers. From Table 5, the highest and lowest soil and water InY ¼ a þ b X þ b X þ ::: þ b X : (1) 1 1 2 2 k k conservation practice adopted by maize farmers was chemical Taking exponents of both sides give the predicted value of Y in fertilizer and mulching respectively. Mulching was identified to be counts. This gives the equation the least practice because of its arduous nature when using it. b ðaþb þb þ:::þb Þ Y ¼ e 1X1 2X2 kXk (2) 4.3. Number of soil and water conservation practices adopted by farmers Where Y is the dependent variable measured as the number of sustainable soil and water conservation practices each farmer Table 6 shows the number of soil and water conservation adopts and Xk is the independent variables which are categorized practices adopted by farmers. The mean number of soil and water into personal and demographic factors, socio-economic factors, conservation practices adopted by farmers is 4.2967 with a stan- institutional factors, risk factors, psychological factors and cost of dard deviation of 2.1781. From Table 6, the highest (modal) number production factors. The parameters in the Poisson regression model of soil and water conservation practices adopted by farmers was 6 were estimated using a maximum likelihood method which com- accounting for 16.67% of total adopters. 7 of the maize farmers out putes deviance that epitomizes variability. Deviance is a relative of 300 signifying 2.33 percent practice none of the soil and water and not an absolute metric. Small deviance shows a better fit conservation. whereas large deviance shows a poor fit. According to Coxe et al. (2009), the deviance is employed to compute a pseudo R2. This 4.4. Pairwise correlation analysis of soil and water conservation pseudo R2 falls within the range 0 and 1 and increases when several practices explanatory variables are incorporated into the model (Coxe et al., 2009). Table 6 presents the Pairwise correlation coefficients for the soil and water conservation practices adopted by the sampled maize devianceðfitted modelÞ farmers. Table 7 shows that all the correlation coefficients are less R2 ¼ 1 (3) deviance devianceðintercept aloneÞ than 0.5 with the highest correlation between the application of chemical fertilizer and zero tillage. Sharma, Bailey, and Fraser The study used STATA version 14 for the data analysis. (2010) interpreted the correlation coefficient greater than 0.5 as high. Therefore, the estimated correlations among maize farmers’ 4. Results selection of soil and water conservation practices are not high. Sharma et al. (2010) interpreted high correlation coefficients Table 2 presents the description and summary statistics of the among agricultural technologies as signifying that technologies are variables employed in the Poisson regression estimations. The chosen and used at the same time. The pairwise correlation analysis result from the table shows that the typical farmer in the Techiman of soil and water conservation practices showed that the maize Municipality is a youth with an average age of approximately 45 farmers employed a mixture of soil and water conservation prac- years. Also, the majority (81.33%) of the farmers in the Techiman tices on maize farms but at different times. Municipality are males with 48.67% of the farmers having no formal training. On the other hand, 78% of the farmers have access to credit 4.5. Poisson regression results facilities, 53.33% are members of agricultural associations, 80% have access to extension services and 87.67% have experienced the risk Twenty-four variables were used as the explanatory variables to of pest and diseases. The variables such as Off-farm Income and identify the determinants that influence the adoption of modern K.A. Darkwah et al. / International Soil and Water Conservation Research 7 (2019) 248e257 253

Table 2 Descriptive Statistics of explanatory variables.

Variable Item Frequency Percentage

Personal and Demographic Factors Gender of Household Head (GNR) Male 244 81.33 Female 56 18.67 Household Head's Educational Level (EDU) No Schooling 146 48.67 Primary School 40 13.33 MSCL/JHS 97 32.33 SHS/VOC/TECH 15 5.00 Tertiary 2 0.67 Formal Training on Maize (FTM) Yes 227 75.67 No 73 24.33 Institutional Factors Access to Credit Yes 234 78.00 No 66 22.00 Access to Extension Services Yes 240 80.00 No 60 20.00 Membership of an agricultural association Yes 160 53.33 No 140 46.67 Access to Mass Media Yes 208 69.33 Risk Factor No 92 30.67 Risk of drought Yes 142 47.33 No 158 52.67 Risk of ﬂood Yes 19 6.33 No 281 93.67 Risk of Wild Animals Yes 32 10.67 No 268 89.33 Yes 263 87.67 Risk of Pest and Diseases No 37 12.33

Variable (unit) Mean Std. Dev. Min. Max. CV

Personal and Demographic Factors Household Head's Age (in years) 45.2367 11.8214 19 80 0.26 Experience of the farmer (in years) 17.59 11.4554 0 60 0.65 Household Size (count) 5.4 2.8284 1 19 0.52 Socio-Economic Factors Farm Size (hectares) 4.0033 3.2285 1 25 0.81 Off-Farm Income (amount in Cedi) 69.7833 201.4173 0 2000 2.89 Total number of Labour (amount in Cedi) 223.63 124.0142 28 630 0.55 Institutional Factors Distance to nearest Output Market (hours) 1.3826 1.0582 0 5 0.77 Distance to input center (hours) 1.3900 1.0306 0 5 0.74 Psychological Factors Perception of farmers on the Attributes of a Technology (Average) 4.688 0.8813 1.8 6 0.19 Perception of farmers on the Price of the Technology (Average) 2.7587 0.5238 1.2 4.8 0.19 Perception of farmers on the Compatibility of the Technology on their soil (Average) 4.3133 0.9634 1.2 6 0.22 Cost of Production Cost of Labour (amount in Cedi) 15.8767 10.8939 10 200 0.69 Cost of Fertilizer (Amount in Cedi) 501.69 1510.38 0 10200 3.01

technology for maize production in Ghana. Before performing 3.51% while the Wald Chi-squared value is 66.15 with a P-value of Poisson regression analysis, the researcher analyzed the correlation 0.000. This shows the overall significance of the Poisson model between the predictor variables and performed the variable infla- used to explain the variables that influence the adoption of modern tion factor test to identify any cases of multicollinearity as well as technology. The low Pseudo R-squared value of 3.51% is as a result observing the p-values of Pearson Chi-squares and Deviance to of most of the explanatory variables were categorical. identify the existence of overdispersion on the data. The Multi- Several variables were identified to have an association with the collinearity test is an initial assumption for estimating parameters. number of soil and water conservation practices adopted by maize Tables 3 and 4 showed no multicollinearity, 7, hence all the farmers. The findings from Table 8 show that farmer's household explanatory variables can be included in the Poisson regression size, farm size, access to credit services and formal training of maize model. farmer have a positive significant association with the number of The value of Deviance goodness-of-fit is 68.85035 with soil and water conservation practices adopted by maize farmers Prob > chi2 (275) ¼ 1.0000 while the value of Pearson goodness-of- while distance to nearest output market, distance to input center, fit is 67.69374 with Prob > chi2 (275) ¼ 1.0000. This shows that the access to extension services, and risk of pest and diseases have a data is not overdispersed and does not have an excessive number of negative significant association with the number of soil and water zeros. Hence the Poisson regression model is appropriate. conservation practices adopted by maize farmers at 5% significance Table 8 shows the Poisson regression analysis of the factors level. influencing the adoption of soil and water conservation practices. The unexpected number of soil and water conservation practices The response variable is the number of soil and water conservation adopted by maize farmers was found to increase with household practices utilized by the farmers as their adoption of modern size, farm size, access to credit and formal training of maize farmer. technology. From Table 8, the estimated Pseudo R-squared value is Specifically, an increase in household size by one person is 254 K.A. Darkwah et al. / International Soil and Water Conservation Research 7 (2019) 248e257

Table 4 Multicollinearity test.

Variable VIF

GNR 1.13 AGE 1.01 EDUC 1.09 FTM 1.28 HHSIZE 1.50 EXP 1.30 0.01 0.01 0.01 1.00 0.02 1.00 0.12 0.02 1.00 FSIZE 1.08 OFI 0.07 TL 1.18 0.03

AMM 1.28 ACS 1.36 DOM 1.05 0.02 0.04 1.00 0.09 0.10 0.04 DTI 1.81 AEXT 1.03 MAA 1.47 0.07 0.00 1.00 0.01 1.00 PPT 1.51 0.08 0.04 0.01 PAT 1.03 PCT 1.78

0.05 ROD 1.23 ROF 1.07 RWA 1.15 0.07 0.00 0.12 0.04 0.04 0.10 0.02 0.04 0.07 RPD 1.01 CL 1.01 CF 1.01 0.01 0.03 0.02 0.06 0.02 0.09 1.00 0.02 0.10 0.25 0.01 Mean VIF 1.19 0.01 0.19 0.22 0.12 0.19 0.19 0.58 1.00 0.01 0.01 associated with a 1.89% increase in the number of soil and water conservation practices adopted by maize farmers in Techiman 0.08 0.09 1.00 0.14 0.09 1.00 0.22 0.04 0.07 0.07 0.05 0.06 Municipality. Also, the number of soil and water conservation 1.00 practices adopted by maize farmers in Techiman Municipality was found to be increased by 18.17% with an increase in access to credit 0.02 0.00 0.02 0.06 0.31 0.29 0.08 0.24 .01 0 services. A farmer who received formal training on maize will increase its adoption of soil and water conservation practices by 15%. 0.00 0.10 0.50 0.14 1.00 0.05 0.07

0.11 0.06 Finally, an increase in farm size by 1 ha is associated with a 10.84% 0.03 0.02 0.07 0.02 increase in the number of soil and water conservation practices adopted by maize farmers in Techiman Municipality. 0.06 1.00 0.01 0.05 0.01 0.01 0.05 0.06 On the other hand, a number of soil and water conservation practices adopted by maize farmers was found to decrease with distance to input center, access to extension services, distance to 0.06 1.00 0.05 0.07 0.04 0.20 0.10 0.01 0.04 0.08 0.17 0.23 0.04 0.15 0.43 1.00 0.04 0.15 0.44 0.04 output market, risk of pest and diseases and cost of labour. An hour increase in time to reach the input center is associated with a 3.36% 0.01 1.00 0.08 0.09 0.06 0.01 0.12 0.02 0.01 0.18 0.04 0.18 0.12 0.01 0.01 0.01 0.02 0.00 0.02 0.04 0.13 0.02 decrease in the number of soil and water conservation practices adopted by maize farmers in Techiman Municipality. Also, an hour increase in time to reach the output market is associated with a 0.01 0.13 0.08 0.04 0.07 0.09 0.06 0.19 0.08 0.08 0.06 0.03 0.09 0.04 0.03 0.04 0.06 0.04 0.02 0.07 0.01 7.13% decrease in the number of soil and water conservation practices adopted by maize farmers in Techiman Municipality. A farmer who had contact with extension ofﬁcers will decrease his or her 0.06 0.12 0.08 0.09 0.06 0.09 0.010.17 1.00 0.06 0.09 0.12 0.01 0.05 0.02 0.02 adoption of soil and water conservation practices by 14.93%. An increase in the risk of pest and diseases attack n maize farm is associated with a 15.52% decrease in the number of soil and water 0.16 0.03 0.08 0.07 0.07 0.06 0.02 0.02 0.15 0.01 0.04 conservation practices adopted by maize farmers in Techiman Municipality. 0.08 0.07 0.20 1.00 0.04 0.13 0.21 0.01 0.06 0.06 0.29 0.02 0.04 0.08 0.23 0.01 0.05

5. Discussion 0.16 1.00 0.15 0.14 1.00 0.20 0.10 0.03 0.21 0.15 1.00 0.05 0.17 0.03 0.05 0.15 0.02 0.05 0.22 0.05 0.01 0.04 0.30 0.12 0.02 0.13 0.03 0.07 0.04 0.04 0.07 0.07 The size of a farmers’ household was identiﬁed to have a positive cient analysis of explanatory variables result.

fi and significant association with the number of soil and water 0.03 1.00 0.01 0.08 0.06 0.06 0.06 0.16 0.01 0.11 0.06 0.12 0.00 0.07 0.36 0.17 0.03 0.02 0.11 0.05 0.02 0.09 0.03 0.01 0.05 0.14 0.05 0.07 0.01 conservation practices adopted by the farmers (b ¼ 0.0188649; p- value ¼ 0.073). This finding is consistent with a study carried out by 0.19 0.16 0.02 0.03 0.05 0.21 0.07 0.01 0.07 0.03 0.04 0.22 0.23 0.07 0.02 0.03 0.02 0.00 0.02 Abdul-Hanan, Ayamga, and Donkoh (2014) on smallholder adop- GNR AGE EDUC FTM HHSIZE EXP FSIZE OFI TL AMM ACS DOM DTI AEXT MAA PPT PAT PCT ROD ROF RWA RPD CL CF tion of soil and water conservation techniques in Ghana who concluded in support of this finding. Their findings found that farmers with larger household size are likely to adopt technology FTM 0.07 GNRAGEEDUC 1.00 0.06 1.00 HHSIZE EXP AEXT FSIZE OFI 0.07 RWA 0.17 0.06 MAA TL AMM 0.12 PPT ACS DOM 0.03 DTI RPD 0.08 PAT PCT ROD 0.01 0.05 CL ROF 0.09 0.11 CF 0.04

Table 3 Pairwise correlation coef techniques more than those with smaller household size. K.A. Darkwah et al. / International Soil and Water Conservation Research 7 (2019) 248e257 255

Table 5 Soil and water conservation practices.

Practice Number of Adopters Percentage of Adopters

Application of Chemical Fertilizer (ACF) 208 69.3 Zero Tillage (ZT) 161 53.7 Crop Rotation (CR) 153 51.0 Intercropping (IC) 140 46.7 Row Planting (RP) 138 46.0 Cover Cropping (CC) 136 45.3 Application of Manure (AOM) 123 41.0 Composting (COM) 119 39.7 Mulching (MUL) 111 37.0

Table 6 Table 8 Number of Soil and Water Conservation Practices Adopted by farmers results. Poisson regression analysis results. ﬁ > Number of SWCP Adopted Number of Adopters Percentage of Adopters Variables Coef cient Robust Std. Err. z P z

0 7 2.33 GNR 0.0070480 0.0804514 0.09 0.930 1 30 10.00 AGE -.00318750 0.0024610 1.30 0.195 2 37 12.33 EDU 0.0357630 0.0312720 1.14 0.253 3 35 11.67 EXP 0.0021953 0.0029768 0.74 0.461 4 46 15.33 HHSIZE 0.0188649 0.0105124 1.79 0.073* 5 49 16.33 FTM 0.1500303 0.0684921 2.19 0.028** 6 50 16.67 FSIZE 0.1084088 0.0483515 2.24 0.007*** 7 24 8.00 OFI 0.0001347 0.000149 0.90 0.366 8 15 5.00 TL 0.0000628 0.0002312 0.27 0.786 9 7 2.33 ACS 0.181661 0.0742221 2.45 0.014** Total 300 100.0 AEXT 0.1493431 0.0631944 2.36 0.018** DTI 0.0335981 0.0197819 1.70 0.089* DOM 0.0712955 0.0343411 2.08 0.038** AMM 0.0669587 0.0853463 0.78 0.433 Also, the findings from this study revealed a positive and sig- MAA 0.0633175 0.0743707 0.85 0.395 nificant association between farm size and the adoption of soil and PAT 0.0290061 0.0454942 0.64 0.524 b ¼ PPT 0.024532 0.0596977 0.41 0.681 water conservation practices by farmers ( 0.1084088; p- PCT 0.0458169 0.0404236 1.13 0.257 value ¼ 0.007). This shows that farmers with larger farm sizes have ROD 0.0787128 0.0623624 1.26 0.207 high financial resources and additional lands to apportion to ROF 0.1191406 0.1134500 1.05 0.294 improving technology adoption. They also have the capability to RWA 0.02194880 0.0833716 0.26 0.792 RPD 0.1551671 0.0756185 2.05 0.040** buy better and modern technologies as well as have the ability to fi CL 0.0024952 0.0041676 0.60 0.56 endure risk if the technology fails to function. This nding is CF 8.22e-060 0.0000157 0.52 0.600 consistent with the study conducted by Alene, Poonyth, and Hassan fi fi fi *, **and *** are 10%, 5% and 1% signi cance level respectively. (2000) who concluded in support of this nding. They identi ed Number of Observations ¼ 300 Wald Chisquare(24) ¼ 66.15. that farmers, with larger farm size, mostly adopt more modern Prob>chisquare ¼ 0.0000 Pseudo R-Squared ¼ 0.0351. technologies than their farmers with small farm size. This is, Log Pseudo likelihood ¼638.02391. however, consistent with our a priori expectation. Abdul-Hanan Iteration 0: log pseudolikelihood ¼638.02393. ¼ et al. (2014) and Tadesse and Belay (2004) had a similar finding. Iteration 1: log pseudolikelihood 638.02391. Formal training of maize farmers was found to have a positive fi and signi cant association with the number of soil and water association with the number of soil and water conservation prac- b ¼ conservation practices adopted by the farmers ( 0.1,500,303; p- tices adopted by the farmers (b ¼ 0.181661; p-value ¼ 0.014). This ¼ fi value 0.028). This nding is consistent with a study carried out by finding is consistent with a study carried out by Foltz (2001) who fi fi Fikru (2009) who concluded in support of this nding. He identi ed concluded in support of this finding. Foltz (2001) concluded that that training has a positive association with the adoption of soil or farmers who have access to credit are mostly motivated to adopt stone bund and tree plantations. Hence, training of farmers is a any modern technology as compared to farmers who are capital- means of creating awareness and encouragement for the adoption constrained. Abdul-Hanan et al. (2014) had a similar finding. of modern technologies. Farmer's access to extension services has a negative and Access to credit was found to have a positive and significant

Table 7 Pairwise correlation analysis of soil and water conservation practices results.

ACF ZT CC COM CR MUL AOM RP IC

ACF 1.00 ZT 0.4983 1.00 CC 0.0188 0.0539 1.00 COM 0.0812 0.0839 0.1257 1.00 CR 0.0278 0.0386 0.3434 0.1133 1.00 MUL 0.0605 0.1583 0.1239 0.4230 0.0330 1.00 AOM 0.0482 0.0542 0.2483 0.3770 0.1257 0.1753 1.00 1.00 RP 0.0627 0.1583 0.0865 0.1950 0.1153 0.3178 0.0873 IC 0.1584 0.2126 0.1414 0.1566 0.2753 0.0581 0.1848 0.1019 1.00 256 K.A. Darkwah et al. / International Soil and Water Conservation Research 7 (2019) 248e257 significant association with the number of soil and water conser- frequent disseminating of agricultural technology information vation practices adopted by the farmers (b ¼0.1493431; p- which is likely to increase the rate of adoption of soil and water value ¼ 0.018). This finding is consistent with a study carried out by conservation practices. The limitation of the study lies in the fact Agbamu (1995) on the analysis of farmers' characteristics in rela- that most of the variables were categorical which resulted in the tion to adoption of soil management practices in the Ikorodu area of low Pseudo R-squared. Also, the study was based only in Techiman Nigeria who concluded in support of this finding. His study found Municipality and focused only on the determinants, but not the that farmers' contact alone with extension officers will not enhance effects of adoption. We hope that future studies would address the adoption of modern technology if services of extension officers these limitations. such as information dissemination on modern technology are ineffective and inaccurate. According to Bamire, Fabiyi, and Acknowledgement Manyong (2002), farmers who do not receive frequent visits from extension officers have lower likelihoods of adapting to modern The authors wish to acknowledge the support of the USAID- technology as compared to farmers who are visited frequently. Ghana/ University of Ghana: Institutional Capacity Building Agri- Risk of pest and diseases were identified to have a negative and culture Productivity for funding this research work. significant association with the number of soil and water conservation practices adopted by the farmers (b ¼0.1551671; p- References value ¼ 0.040). This is consistent with our priori expectation because when farmers have a high risk of pest and diseases Abdul-Hanan, A. (2017). Determinants of adoption of soil and water and conservation techniques: Evidence from Northern Ghana. International Journal of affecting farms, they reduce their adoption in modern technology Sustainable Agricultural Management and Informatics, 3(1), 31e43. on soil and water conservation practices in order to save time and Abdul-Hanan, A., Ayamga, M., & Donkoh, S. A. (2014). Smallholder adoption of soil money. and water conservation techniques in Ghana. African Journal of Agricultural Research, (5), 539e546. Distance to input center was found to have a negative and sig- Agbamu, J. U. (1995). Analysis of farmers' characteristics in relation to adoption of nificant association with the number of soil and water conservation soil management practices in the Ikorodu area of Nigeria. Japanese Journal of practices adopted by the farmers (b ¼0.0335981; p- Tropical Agriculture, 39(4), 213e222. value ¼ 0.089). This finding is consistent with a study carried out by Aikins, S., Afuakwa, J., & Owusu-Akuoko, O. (2012). Effect of four different tillage practices on maize performance under rainfed conditions. Carbon, 1,1e51. Abdul-Hanan et al. (2014) on smallholder adoption of soil and Ajayi, O. C. (2007). User acceptability of sustainable soil fertility technologies: water conservation techniques in Ghana who concluded in support Lessons from farmers' knowledge, attitude and practice in southern Africa. e of this finding. Their findings found that the closer a farmer's farm Journal of Sustainable Agriculture, 30(3), 21 40. Alene, A. D., Poonyth, D., & Hassan, R. M. (2000). Determinants of adoption and was to the input store, the more the adoption of Soil and Water intensity of use of improved maize varieties in the Central highlands of Conservation Practices. The nearness of a farm to an input store Ethiopia: A Tobit analysis. Agrekon, 39(4), 633e643. does not only motivate the farmer to purchase the input, but it Alhassan, A., Salifu, H., & Adebanji, A. O. (2016). Discriminant analysis of farmers' adoption of improved maize varieties in Wa Municipality, Upper West Region of minimizes the transportation cost of conveying the input to the Ghana. SpringerPlus, 5(1), 1514. farmer's house. Bamire, A. S., Fabiyi, Y. L., & Manyong, V. M. (2002). Adoption pattern of fertiliser Distance to output market was found to have a negative and technology among farmers in the ecological zones of south-western Nigeria: A e fi Tobit analysis. Australian Journal of Agricultural Research, 53(8), 901 910. signi cant association with the number of soil and water conser- Bayard, B., Jolly, C. M., & Shannon, D. A. (2006). The adoption and management of vation practices adopted by the farmers (b ¼0.0712955; p- soil conservation practices in Haiti: The case of rock walls. Agricultural Eco- value ¼ 0.038). This finding is consistent with a study carried out by nomics Review, 7(2), 28. Boahene, K., Snijders, T. A., & Folmer, H. (1999). An integrated socioeconomic Teklewold, Kassie, and Shiferaw (2013) who concluded in support analysis of innovation adoption: The case of hybrid cocoa in Ghana. Journal of of this finding. Their findings identified that the means of trans- Policy Modeling, 21(2), 167e184. portation to output market is associated with the likelihood of Boz, I., & Akbay, C. (2005). Factors influencing the adoption of maize in Kahra- e adoption of improved seed and conservation tillage. According to manmaras province of Turkey. Agricultural Economics, 33(s3), 431 440. Busari, M. A., Kukal, S. S., Kaur, A., Bhatt, R., & Dulazi, A. A. (2015). Conservation them, a farmer who employs the services of public transport, bi- tillage impacts on soil, crop and the environment. International Soil and Water cycle or donkey/horse cart is more probable to adopt improved Conservation Research, 3(2), 119e129. seed and conservation tillage and suggested that improved road Coxe, S., West, S. G., & Aiken, L. S. (2009). The analysis of count data: A gentle introduction to Poisson regression and its alternatives. Journal of Personality infrastructure and easy access to public transportation system is Assessment, 91(2), 121e136. vital in expediting adoption. FAO, I. (2015). WFP (2015). The state of food insecurity in the world 2015. Meeting the. Feder, G., Just, R. E., & Zilberman, D. (1985). Adoption of agricultural innovations in developing countries: A survey. Economic Development and Cultural Change, 6. Conclusion 33(2), 255e298. Fikru, A. (2009). Assessment of adoption behaviour of soil and water conservation The study assessed the factors that have an association with practices in the Koga Watershed, Highlands of Ethiopia. MPS Thesis. United States: Cornell University. modern technology for soil and water conservation practices by Foltz, J. D. (2001). The economics of water-conserving technology adoption in maize farmers in the Techiman Municipality of Ghana. Empirical Tunisia: An empirical estimation of farmer technology choice. Economic results from the Poisson regression model shows that personal and Development and Cultural Change, 51(2), 359e373. Greene, W. H. (1997). FIML estimation of sample selection models for count data. demographic factors such as household size and formal training on Kaliba, A. R., Verkuijl, H., & Mwangi, W. (2000). Factors affecting adoption of maize, institutional factor such as access to credit services, access to improved maize seeds and use of inorganic fertilizer for maize production in extension services, distance to output market and distance to input the intermediate and lowland zones of Tanzania. Journal of Agricultural & e center, socio-economic factor such as farm size as well as risk factor Applied Economics, 32(1), 35 47. Kassie, M., Zikhali, P., Manjur, K., & Edwards, S. (2009 August). Adoption of sus- such as risk of pest and diseases were identified to have an asso- tainable agriculture practices: Evidence from a semi-arid region of Ethiopia. In ciation with soil and water conservation practices adopted by Natural Resources Forum (Vol. 33, No. 3, pp. 189e198). Oxford, UK: Blackwell farmers. The study suggests that any further research in the study Publishing Ltd. Mango, N., Makate, C., Tamene, L., Mponela, P., & Ndengu, G. (2017). Awareness and area on soil and water conservation practices should acknowledge adoption of land, soil and water conservation practices in the Chinyanja Tri- the mixture of personal and demographic, institutional, socio- angle, Southern Africa. International Soil and Water Conservation Research, 5(2), economic and risk factors. Also, some of the agricultural policies 122e129. Mazvimavi, K., & Twomlow, S. (2009). Socioeconomic and institutional factors formulated by the government should be aimed at supporting influencing adoption of conservation farming by vulnerable households in maize farmers to have access to extension service contact for Zimbabwe. Agricultural Systems, 101(1), 20e29. K.A. Darkwah et al. / International Soil and Water Conservation Research 7 (2019) 248e257 257

MoFA, S. R. I. D. (2011). Agriculture in Ghana-Facts and figures (2010). Ghana: Min- Statistical Service. istry of Food and Agriculture (MoFA)-Statistics, Research and Information Tadesse, M., & Belay, K. (2004). Factors influencing adoption of soil conservation Directorate (SRID). measures in southern Ethiopia: The case of Gununo area. Journal of Agriculture Mucavele, F. G. (2013). True contribution of agriculture to economic growth and and Rural Development in the Tropics and Subtropics (JARTS), 105(1), 49e62. poverty reduction. A synthesis report. Malawi, Mozambique and Zambia. Teklewold, H., Kassie, M., & Shiferaw, B. (2013). Adoption of multiple sustainable Nkegbe, P. K., & Shankar, B. (2014). Adoption intensity of soil and water conserva- agricultural practices in rural Ethiopia. Journal of Agricultural Economics, 64(3), tion practices by smallholders: evidence from Northern Ghana. Bio-based and 597e623. Applied Economics Journal, 3(1050-2016-85757), 159. Tweneboah, C. K. (2000). Modern agriculture in the tropics. Food Crops. Co. wood Rogers, E. M. (2003). Elements of diffusion. Diffusion of Innovations, 5,1e38. publishers. Sharma, A., Bailey, A., & Fraser, I. (2010). Technology adoption and pest control Webb, E., & Highly, D. (2000). Managing maize stocks in developing countries. strategies among UK cereal farmers: Evidence from parametric and nonpara- Postharvest Newsletter, 55. metric count data models. Journal of Agricultural Economics. https://doi.org/ Xu, Z., Guan, Z., Jayne, T. S., & Black, R. (2009). Factors influencing the profitability of 10.1111/j.1477-9552.2010.00272. fertilizer use on maize in Zambia. Agricultural Economics, 40(4), 437e446. Statistical Service. (2014). 2010 population and housing census report. Ghana International Soil and Water Conservation Research 7 (2019) 258e265

Contents lists available at ScienceDirect

International Soil and Water Conservation Research

journal homepage: www.elsevier.com/locate/iswcr

Original Research Article Effects of rice husk biochar on selected soil properties and nitrate leaching in loamy sand and clay soil

* Mohammad Ghorbani a, , Hossein Asadi b, Sepideh Abrishamkesh a a Department of Soil Science, University of Guilan, Rasht, 4818168984, Iran b Department of Soil Science, University of Tehran, Karaj, 77871-31587, Iran article info abstract

Article history: Biochar is a product of pyrolysis of biomass in the absence of oxygen and has a high potential to Received 3 February 2019 sequester carbon into more stable soil organic carbon (OC). Despite the large number of studies on Received in revised form biochar and soil properties, few studies have investigated the effects of biochar in contrasting soils. The 17 May 2019 current research was conducted to evaluate the effects of different biochar levels (0 (as control), 1% and Accepted 29 May 2019 3%) on several soil physiochemical properties and nitrate leaching in two soil types (loamy sand and clay) Available online 3 June 2019 under greenhouse conditions and wet-dry cycles. The experiment was performed using a randomized design with three levels of biochar produced from rice husks at 500 C in three replications. Cation Keywords: ﬁ Soil incubation study exchange capacity increased signi cantly, by 20% and 30% in 1% and 3% biochar-amended loamy sand Carbon sequestration soil, respectively, and increases were 9% and 19% in 1% and 3% biochar-amended clay soil, respectively. Biochar rates Loamy sand soil did not show improvement in aggregate indices, including mean weight diameter, Soil texture geometric mean diameter, water stable aggregates and fractal dimension, which was contrary to the Water stable aggregate results for the clay soil. Rice husk biochar application at the both rates decreased nitrate leaching in the clay soil more than in the loamy sand. Our study highlights the importance of soil type in determining the value of biochar as a soil amendment to improve soil properties, particularly soil aggregation and reduced nitrate leaching. The beneﬁts of the biochar in the clay soil were greater than in the loamy sand soil. © 2019 International Research and Training Center on Erosion and Sedimentation and China Water and Power Press. Production and Hosting by Elsevier B.V. This is an open access article under the CC BY-NC- ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction improvement of soil properties (Pandian, Subramaniayan, Gnasekaran, & Chitraputhirapillai, 2016) and plant growth Currently, environmental concerns about human induced CO2 enhancement (Rajkovich et al., 2012) can be achieved as a result of emission and global warming, high nutrient levels in surface and biochar addition to soil. groundwater, and eutrophication have necessitated identification Despite the many researches on influence of biochar on soil of efficient strategies to mitigate them. An option to carbon fertility and plant growth (Manolikaki & Diamadopoulos, 2017; sequestration, reduce methane and nitrous oxide emissions (Juriga Mehmood, Baquy, & Xu, 2018), and its carbon sequestration po- et al., 2018; Pratiwi & Shinogi, 2016), and reduce nutrient leaching tential (Van Trinh et al., 2017), there are a limited knowledge on the (Liu et al., 2017) could be the addition of biochar to soils. Biochar is effects of biochar on soil physical properties such as soil structure produced by pyrolysis of different carbon rich residues of agricul- characteristics in contrasting soils. Aggregate stability is one of the ture and forestry (Van Trinh et al., 2017), some grasses, and most important factor affecting soil environmental functions e.g. municipal and industrial wastes. Biochar can act as a long-term soil capacity to store and stabilize organic carbon (Wang et al., 2018), carbon sequestration when buried in soil, i.e. remaining for hun- and resistance against erosive agents (Chaplot & Cooper, 2015). dreds of years (Ghorbani & Amirahmadi, 2018). Furthermore, Aggregate stability can be improved through the formation of biochar-soil minerals complexes (Juriga et al., 2018). However, present reports about the effects of biochar on soil

* aggregate stability are not always in agreement with each other and Corresponding author. fi E-mail addresses: [email protected] (M. Ghorbani), [email protected] are biochar type, application rate and soil type-speci c. Decrease in (H. Asadi), [email protected] (S. Abrishamkesh). aggregation in a loamy sand soil amended by biochar after 96 days https://doi.org/10.1016/j.iswcr.2019.05.005 2095-6339/© 2019 International Research and Training Center on Erosion and Sedimentation and China Water and Power Press. Production and Hosting by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). M. Ghorbani et al. / International Soil and Water Conservation Research 7 (2019) 258e265 259 of incubation were observed (Busscher, Novak, & Ahmedna, 2011). soil were collected from two different sites: Lahijan (loamy sand, In a study, biochar application increased mean weight and geo- 3713002.000N5000043.900E) and Rasht (clay, 3711 051.000N metric mean diameter in a silty loam soil, but it has no significant 4939006.200E) in Guilan province, North of Iran. The two soils are influence on soil aggregation and stability in a sandy loam soil classified as Typic Hapludepts and Typic Hapludalfs, respectively (Xiang-Hong, Feng-Peng, & Xing-Chang, 2012). Also in the other (Wilson et al., 2008). Rice is the main agricultural crop in Guilan study, researchers found that the biochar addition to an Ultisol had province. The rice husk was used to produce biochar. The biochar no significant effect on mean weight diameter of aggregates (Peng, was produced by an electrical muffle furnace at peak temperature Ye, Wang, Zhou, & Sun, 2011). Furthermore, it is reported that the of 500 C and 30e45 min pyrolysis process. use of biochar can delay nitrate leaching which should be beneficial both to the environment and to plants because of nitrogen holding 2.2. Soil and biochar analysis in the soil for longer periods (Knowles, Robinson, Contangelo, & Clucas, 2011). There are few studies reported on the ability of bio- Soil samples were air dried and ground to pass through a 2-mm char to hold nitrate. sieve. Physical and chemical properties of the soil samples (Table 1) For example, significant reduction of leaching of fertilizer N were determined measured as follows: Soil pH and electrical con- from soil has been reported as a result of amendment with biochar ductivity (EC) were measured in 1:1 soil to water solution and in produced from forest residues (Manolikaki & Diamadopoulos, saturated extract, respectively. The soil texture was determined by 2017). Cation exchange by bamboo biochar led to sorption of the hydrometer method (Gee & Or, 2002). Organic carbon (OC) was ammonium ions and retardation of vertical translocation of measured by wet oxidation method (Nelson & Sommers, 1996). ammonium into sub surface soil layers during a 70-day incubation Nitrogen (N) was determined by the Kjeldahl method (Page, Miller, (Ding et al., 2010). Reduction of nitrate leaching from soil amended & Keeney, 1982). Bulk density (BD) was determined by the clod by biochar produced from pecan shells has been demonstrated over method (Radcliffe & Simunek, 2010). Cation exchange capacity 25 and 67 days (Chaplot & Cooper, 2015). (CEC) was determined by using ammonium acetate extraction at Rice is the staple food in Iran and the single most important pH 7 (Sumner & Miller, 1996). crop. Rice husk is the main byproduct of the rice-processing in- Rice husk biochar was analysed for its basic properties (Table 1). dustry and is produced in large quantities in north of Iran. The The pH and EC of the biochar were measured in deionized water at advantages and drawbacks of burning vs incorporation of rice straw 1:20 (biochar: water) weight ratios (Rajkovich et al., 2012). BD was in rice growing have been discussed (Varela Milla, Rivera, Huang, determined by the clod method (Radcliffe & Simunek, 2010). CEC Chien, & Wang, 2013). Rice husk ash for the improvement of rice was determined by using 0.05 M hydrochloric acid at pH 7 growth and yield in the peat soil of Sumatra were used (Masulili, (Munera-Echeverri et al., 2018). Carbon and hydrogen content of Utomo, & Syechfani, 2010). Also has been shown the ability of biochar was determined by dry combustion analysis using an rice husk and rice husk ash to remove heavy metals from the sys- Elemental analyser (Perkin Elmer 2400 II, Massachusetts, USA). tem (Ahmaruzzaman & Gupta, 2011). Application of biochar in Scanning electron microscopy (Fig. 1) was also performed on bio- tsunami-affected paddy fields in Sri Lanka were investigated, and char by mean of digital scanning electron microscope (KYKY- the experimental results showed that the application of 2 t rice- EM3200). husk-charcoal ha 1 increased the grain yield from less than 4tha 1 for the control treatment to more than 5 t ha 1 for the 2.3. Incubation experiment biochar treatment (Reichenauer, Panamulla, Subasinghe, & Wimmer, 2009). Now, the rice husks have been used as a bio-fuel Incubation experiments were carried out in two parts to study for electricity generation (Chaplot & Cooper, 2015). Rice husk bio- the effects of biochar on (i) selected soil properties and (ii) nitrate char, a byproduct of the thermochemical conversion from the rice leaching. husk to biofuel, has been proposed to be a new soil amendment. Rice husk biochar can be recycled easily in the riceewheat system 2.3.1. Incubation experiment of selected soil properties without adverse effect on the soil's health (Shackley et al., 2012). In incubation experiments of selected soil properties, one kg The potential of rice husk biochar as a soil amendment to improve samples of the soils (loamy sand and clay) were placed in plastic soil physicochemical properties has been reported in several pots (11 cm width and 15 cm depth) and mixed with treatments. studies (Li et al., 2016; Masulili et al., 2010). Rice husk conversion The three treatments were unamended soil (control) and amended into biochar may result in secondary carbon advantages via soils with 1 and 3% by weight of biochar, called here after B1 and B3, avoiding burning it in field and bio-resource recycling, which have respectively. Soils and biochar were mixed thoroughly, and then been considered as a matter of concern with air pollution of Iranian the pots were subjected to wetting (field capacity plus 20 percent) agriculture and climate change mitigation. On the other hand, ni- and drying (permanent wilting point) cycles. The unamended trate leaching losses and its health hazard is a serious problem in control was also mixed uniformly. The incubated pots were placed most northern paddy fields of Iran (Darzi-Naftchali, Shahnazari, & in a room at 25 C and weighed every 7 days to control moisture Karandish, 2017). In this context, we hypothesized that rice-husk content. All treatments were prepared in triplicate. The incubation biochar application in two soil types could improve physicochem- time was 10 months and soils were analysed at 10th month to ical properties in general and also we expect contrasting im- determine the selected soil properties including BD, porosity, OC, provements due to the effect of application rate and soil type. pH, EC, CEC and soil aggregation indices. Soil BD was determined by Under these contexts the objectives of this study were to evaluate the clod method (Radcliffe & Simunek, 2010). Porosity was calcu- the effects of rice husk biochar on several soil physiochemical lated from bulk density and particle density (assuming properties and nitrate leaching in a loamy sand soil and a clay soil. 2.65 g cm 3). The OC content was determined by the modified WalkleyeBlack method (Chan et al., 1995). The air-dried, disturbed 2. Materials and methods soil samples were passed through a 2-mm sieve and analysed for pH and EC using 1:2.5 and 1:5 (soil: water) suspensions, respec- 2.1. Soil and biochar tively. CEC was determined by using ammonium acetate extraction at pH 7 (Sumner & Miller, 1996). In this study soil aggregation was Soil samples (0e25 cm depth) including a loamy sand and a clay determined by two methods of wet aggregate size distribution and 260 M. Ghorbani et al. / International Soil and Water Conservation Research 7 (2019) 258e265

Table 1 Physiochemical characteristics of the soil samples, rice husk biochar and compost.

Property Loamy sand soil Clay soil Rice husk biochar Compost

pH 4.28 6.87 9.18 6.11 EC (dS m 1) 0.33 0.28 0.347 2.31 CEC (cmol 1) 12.26 5.67 17.57 10.3 BD (g cm 1) 1.61 1.31 0.84 0.73 Total carbon (g kg 1) 5.3 11.0 478 28.9 H(gkg 1) ee24.3 e N(gkg 1) 0.4 0.5 0 1.1 C/N eee 9.63 Sand (%) 66.4 17.6 ee Silt (%) 15.3 36.6 ee Clay (%) 17.3 45.8 ee

EC: electrical conductivity, CEC: cation exchangeable capacity, BD: bulk density, H: hydrogen, N: nitrogen, C/N: carbon/nitrogen.

Fig. 1. Scanning electron micrographs of biochar scale bar of 250 (left) and 500 (right).

Fig. 2. Regression relationships between pH v. cation exchange capacity for Loamy sand (a) and Clay (b) soils. wet aggregate stability. distribution: 2.3.2. Wet aggregate size distribution ¼ WA WC WSA (1) Wet-sieving method was used to assess the wet aggregate size WO WC distribution. After the air-dried soils prewetted for about 24 h with Where D is the fractal dimension, M(x) is the mass of aggregates tap water, the soil immersed in a container of tap water on a set of i left on the sieve of a nest of sieves, x is the size of aggregates in mm, sieves which had openings of 4, 2, 1, 0.5, 0.25, 0.125, 0.074 and i K is a constant, relative to the number of fragments in one volume 0.053 mm in diameters, and the sieving took place at 35 oscillations unit. per minute (along 38.1 mm amplitude) for 10 min. The remained material after wet-shaking in each sieve was carefully removed and dried at 105 C. The mean weight diameter (MWD) and geometric 2.3.3. Wet aggregate stability mean diameter (GMD) of soil aggregates were determined using In a single-sieve wet-sieving method, 4 g of 1e2 mm air-dried wet sieving data. aggregates was weighted into 0.26 mm (60 Mesh) sieve, pre- Aggregate mass is a scale dependent soil property which can be wetted by water for 24 h, and then shacked vertically about quantified by fractal geometry (Salako, 2006). Assuming scale- 1.5 cm, 35 times min 1, for 3 min. After drying, the weight of un- invariant density and shape of aggregates, the following equation stable and stable aggregates was determined. Dividing the weight is used for the estimation of fractal dimension (D) from mass-size of the stable aggregates by the total aggregate weight (without M. Ghorbani et al. / International Soil and Water Conservation Research 7 (2019) 258e265 261 sand particles > 0.05 mm) gives the index of water stable aggre- Increases were 9% and 19% in 1% and 3% biochar-amended clay soils, gates (WSA): respectively (Table 2). BD is an indirect measure of the soil compaction. In this study, W W fi < WSA ¼ A C 100 (2) BD signi cantly (P 0.05) decreased with increasing biochar WO WC application rate in both soils (Table 2). For example, BD of 3% biochar-treated soil was decreased from 1.31 to 1.21 g cme3 and where WSA is the water stable aggregate, WA is the weight of ma- 1.61 to 1.26 g cme3 in loamy sand and clay, respectively (Table 2). terial on the sieve after wet sieving, WC is the weight of sand material, WO is the weight of aggregates placed on the sieve prior to 3.2. Soil aggregation wet sieving size i. The mean comparison of interaction of biochar and soil type on 2.3.4. Incubation experiment of nitrate leaching the indices of soil aggregation are given in Fig. 3. No significant In another part of the study, the impact of biochar on nitrate differences in all aggregation indices were observed among the leaching was studied. Compost was used as a source of nitrogen for biochar application rates in loamy sand soil, whereas application of studying nitrate leaching. The compost used in this study was 3% biochar led to a significant increase in MWD, GMD, WSA and azocompost, which was used to create a nitrogen source for decrease in D in the clay soil (P < 0.05). Results showed that MWD nitration and also to compare with the biochar effect on aggregate increased significantly from 1.19 mm (control) to 1.50 mm (B3), stability. Azocompost is a mixture of organic matter including GWD increased significantly from 0.93 mm (control) to 1.02 mm azolla and rice straw, which is related to microorganisms in a warm (B3), WSA increased significantly from 68% (control) to 85% (B3) environment, and is treated with a suitable ventilation and D increased significantly from 4.42 (control) to 3.9 (B3) in Clay (Yousefzadeh et al., 2013). The azocompost was prepared from the soil (Fig. 3a, b, c and d). Rice Research Institute of Iran, which according to the tests carried out by the institute (Table 1), the percentage of nitrogen contained 3.3. Nitrate leaching in the compost was 11%. One kg samples of the two types of soil (loamy sand and clay) were placed in plastic pots (11 cm width and The effect of treatments and soil type on the amount of nitrate in 15 cm depth) and mixed with treatments. The 6 treatments were; leachates was significant at P < 0.05. The highest amount of nitrate unamended soil (control), and amended soils with 1 and 3% by leaching was recorded for C (soil only treated with compost) and it weight of biochars (called here after B1, B3), and amended soil with was significantly higher in all treatments containing compost 1% by weight of compost (called here after C), and combinations of compared with ones without compost (Fig. 4). Treating the soil 1 and 3% by weight of biochars with 1% compost (called here after with biochar at both 1% and 3% application rates decreased B1þC and B3þC). During the experiment, the pots were placed significantly nitrate leaching in both soils, i.e. in the presence or under wetting and drying cycles. In each cycle, outflow drained absence of compost. For example, nitrate leachate in the absence of water from the pots was collected after irrigation, and nitrate compost was decreased from 5.98 (control) to 4.83 (B1) and 3.56 concentration was measured by spectrophotometry according to (B3), and 5.98 (control) to 4.83 (B1) and 3.56 (B3), in loamy sand (Clesceri, Greenberg, Eaton, Rice, & Franson, 2005) for a 9-months and clay, respectively. Also nitrate leachate in the presence of period. compost was decreased from 9.37 (C) to 6.01 (B1þC) and 5.21 (B3þC), and 9.12 (C) to 5.57 (B1þC) and 4.76 (B3þC), in loamy sand 2.4. Data analysis and clay, respectively. Significant differences in nitrate leachate were observed in loamy sand and clay soils among treatments The statistical analysis of the effect of biochar on soil properties without compost (Fig. 4). in two type of soils was performed in factorial arrangement in a Mean comparison of time effect on the nitrate leaching content completely randomized design with three replications. The tripli- is presented in Fig. 5. During the experiment, nitrate leaching cate data of selected soil properties and amount of leachate nitrate reduced significantly and almost continuously, reached to a steady were subjected to mean separation analysis using the 2-way rate on the last three months. ANOVA and 1-way ANOVA test, respectively, using SAS 9.4 software (Delwiche & Slaughter, 2012). Treatment means were sepa- 4. Discussion rated using the least significant difference (l.s.d.) test. Least-square means were used to test for significant differences among the 4.1. Physicochemical soil properties treatments at P ¼ 0.05. Linear regression analysis was performed to investigate the relationships among pH and CEC using the PROC The pH of soil receiving 3% biochar increased by 0.69 and 0.49 REG program in SAS. units in loamy sand and clay soils, respectively, over the control. A higher absolute increase in pH in loamy sand than clay soil may be 3. Results due to its initial low pH allowing a greater increase. An increase in pH in biochar-amended soils has been reported in other long-term 3.1. Physicochemical soil properties incubation studies (Ghorbani & Amirahmadi, 2018; Manolikaki & Diamadopoulos, 2017). The pH increase in the soil may be due to Both 1% and 3% application rates showed significant increases in the high pH of rice-husk biochar (9.18), which is mainly attributed soil pH for both loamy sand (pH 4.76 and 5.06, respectively) and to the ash being dominated by alkaline carbonates, alkali earth clay (pH 6.84 and 7.20, respectively) soils (Table 2). metals and organic anions (Li et al., 2016). Another reason for an No significant differences in soil EC were observed among the increase in soil pH could be the high CEC of biochar-amended soils. biochar application rates in both soils compared with the control Exchangeable Al and soluble Fe tend to decrease as a result of (Table 2). increasing exchangeable sites leading to low potential acidity OC content significantly increased in 1% and 3% biochar- (Masulili et al., 2010). Significantly high coefficients of determina- amended soils. CEC increased significantly (P < 0.05) by 20% and tion (R2 ¼ 0.67 and 0.34 for loamy sand and clay soils, respectively; 30% in 1% and 3% biochar-amended loamy sand soils, respectively. P < 0.05; Fig. 2 a, b) confirmed the effect of CEC on pH. This showed 262 M. Ghorbani et al. / International Soil and Water Conservation Research 7 (2019) 258e265

Table 2 Physiochemical properties (mean ± s.d.) of biochar amended soils.

1 1 3 Treatments pH EC (dS m ) OC (%) CEC (cmolc kg )BD(gcm) Porosity (%) Clay Control 6.71 ± 0.02b 0.28 ± 0.003a 1.10 ± 0.03b 12.17 ± 1.07b 1.31 ± 0.05b 50.56 ± 1.17b B1 6.84 ± 0.03 ab 0.25 ± 0.002a 1.35 ± 0.03a 13.28 ± 1.25 ab 1.28 ± 0.03b 51.69 ± 1.12 ab B3 7.20 ± 0.01a 0.22 ± 0.006a 1.87 ± 0.04a 14.44 ± 0.57a 1.21 ± 0.04c 54.33 ± 1.11a Loamy Sand Control 4.36 ± 0.02d 0.33 ± 0.001a 0.53 ± 0.02b 5.71 ± 1.13d 1.61 ± 0.02a 39.24 ± 1.12c B1 4.76 ± 0.03c 0.28 ± 0.002a 1.18 ± 0.01 ab 6.87 ± 1.19cd 1.42 ± 0.04a 46.41 ± 1.15b B3 5.06 ± 0.05c 0.24 ± 0.002a 1.42 ± 0.03a 7.40 ± 0.94c 1.26 ± 0.06b 52.45 ± 1.03a

Biochar treatments: B1, 1%; B3, 3%. EC, electrical conductivity; OC, organic carbon; CEC: cation exchangeable capacity; BD, bulk density. The means with the same letters are not signiﬁcantly different at P > 0.05.

Fig. 3. The Interaction effects of biochar and soil type on (a) MWD: Mean weight diameter; (b) GMD: Geometric mean diameter, (c) WSA: Water stable aggregates, and (d) D: Fractal dimension. The means with the same letters are not significantly different at P > 0.05. the potential of rice-husk biochar as a soil amendment to correct which was related to the content of OC by offering more available medium acidity, especially in loamy sand soils. Although the pH of substrates for microbial growth. Nevertheless, when expressed as clay soil also increased, it stayed within the neutral pH range, with the mineralization of biochar per unit native OC in different soils, no potential detrimental impact on nutrient availability for crops. the revised degradation of biochar had a negative relationship with Contradictory results have been found in literature; some the soil pH (Chaplot & Cooper, 2015). studies reported a significant increase (Ghorbani & Amirahmadi, The increase in CEC may be due to the high CEC of biochar 2018; Masulili et al., 2010) and others reported no significant in- (19.21 cmolc kge1) from high specific surface area and high crease in soil EC with biochar application (Jien & Wang, 2013; Singh porosity (Jien & Wang, 2013). In addition, oxidation of biochar and Mavi et al., 2018). the development of surface negative charge over time may increase The highest OC content was observed for loamy sand (1.42%) and CEC (Mehmood et al., 2018). The percentage increase in CEC was clay (1.87%) soils amended with 3% biochar (Table 2). Increases in higher in loamy sand than clay even at similar application rates. The OC content were reported with the application of biochar in general CEC of the two soils responded differently to biochar application. (Jien & Wang, 2013; Li et al., 2016) as well as rice-husk biochar even Biochar oxidation could decrease because of interaction of biochar after 30 days of incubation (Masulili et al., 2010). The carbon with soil minerals (Chaplot & Cooper, 2015). The loamy sand (which sequestration potential of biochar is related to the properties of include 66.4% sand) with low active mineral content due to a biochar and soil (e.g., pH, OC, and clay content) (Wang et al., 2018). smaller fraction of finer particles (clay and silt). Therefore, surface- More biochar C was mineralized in the high-pH soil (clay) than in charge development of biochar could be higher in loamy sand soil the low-pH soil (loamy sand). Higher microbial biomass was than clay soil. However, improvement of CEC in both soils could attributed to higher biochar mineralization in the high-pH soil, provide agronomic and environmental benefits. M. Ghorbani et al. / International Soil and Water Conservation Research 7 (2019) 258e265 263

Fig. 4. Effects of treatments on nitrate leaching. B1: 1% biochar; B3: 3% biochar; C: compost; B1þC: 1% biochar þ compost; B3þC: 3% biochar þ compost. The means with the same letters are not signiﬁcantly different at P > 0.05.

Fig. 5. Effects of time on nitrate leaching. The means with the same letters are not signiﬁcantly different at P > 0.05.

Any reduction in BD could be due to the increase in the volume are the low organic matter content and the low particles of less of soil biochar mixtures over time. Further, total porosity in both than 0.05 mm in the loamy sand soil (Herath et al., 2013). The role soils significantly increased (Table 2). This indicated an increase in of organic matter in aggregate stability can be attributed to the the volume of biochar amended soils, which may be due to the hydrophobicity of these materials. So that, in a single study results rearrangement of soil and biochar particles. These results are showed that increasing organic matter in the soil from 2.3 to 3.5 consistent with other soilebiochar incubation studies (Herath, percent increased the aggregate stability (in particular, MWD) and Camps-Arbestain, & Hedley, 2013; Jien & Wang, 2013). Particle reduced soil loss (Lado, Ben-Hur, & Shainberg, 2004). This role of rearrangement from release of applied pressure by soileorganic organic matter is complete when the percentage of clay in the soil is particles could reduce the BD in soils with organic amendments high. The particles of the clay have a high level of surface and, as a added (Pandian et al., 2016). result, are a lot of free energy that, in order to stabilize, this energy must reach its lowest level. The flocculation of clay particles and the 4.2. Soil aggregation formation of aggregates, which have a lower surface in volume unit, helps to reduce this energy. The greater the organic matter, the Soil aggregate stability is a key soil quality parameter controlling greater the aggregation (Van Trinh et al., 2017). sensitivity to crusting and erosion (Abrishamkesh, Gorji, Asadi, Biochar significantly enhanced the aggregate stability indices of Bagheri-Marandi, & Pourbabaee, 2015). High aggregate stability the clay soil mainly consisting of silt and clay particles but had no indicates the high resistance to soil erosion and predominance of effect on the loamy sand with more sand content (Herath et al., macroaggregates (Pandian et al., 2016). The passage of time will 2013). The difference may be explained by the clay role as an certainly increase the stability of aggregates in clay soils, but it does aggregator in binding organic molecules by bivalent and polyvalent þ þ þ not have a significant effect on the loamy sand soil. The reason for cations (such as Ca2 ,Al3 , and Fe3 ) which result in formation of this can be found in the soil components. Probably reasons for this macro aggregates and making up a protecting layer to protect soil 264 M. Ghorbani et al. / International Soil and Water Conservation Research 7 (2019) 258e265 organic carbon from microbial decomposition and promote ag- particles) is less than loamy sand (with 17% clay particles). How- gregation (Jien & Wang, 2013). This can also be due to the high ever, treatment C (compost), despite a significant increase in percentage of particles smaller than 0.05 mm (clay and silt) in clay nitrification, is not capable of preserving nitrate, in contrast to texture compared to the loamy sand (Xiang-Hong et al., 2012). It is treatments containing biochar 3%, and much of it is leached. The not clear how the effect of this mechanism is. But some researchers low level of leaching in treatments B3, as well as B3þC compared to have argued that this is due to the high concentrations of activated treatment C can be attributed to the high level of specific surface in carbon on particle surfaces smaller than 0.05 mm in soils that biochar particles (Ghorbani & Amirahmadi, 2018). combine with biochar (Liu et al., 2017). This means that particles Higher nitrate leaching on the first month can be attributed to smaller than 0.05 mm will prefer black carbon particles to other the presence of more nitrate at the beginning of the experiment. organic compounds (Li et al., 2016). Some studies have shown that Increasing nitrate leaching in the fourth and fifth months shows an the addition of humic acid to soil by increasing acidic ligands, such increase in nitrification in these two months. This could be due to as eCOOH, improves soil properties, including the absorption of the fourth and fifth readings in the wet months of the year. different elements (Singh Mavi et al., 2018). The existence of such Increased moisture content is one of the conditions that increases ligands has been confirmed in a variety of different biochars. nitrate production (Metwally & Pollard, 1959). The moisture Therefore, it can be assumed that the active agent groups in the absorbing of biochar and the high porosity in its structure make it particles of biochar are the same reason for the flocculation of soil possible for microorganisms and other nitrate synthesis processes particles and the formation of aggregates. This may also be the (Darzi-Naftchali et al., 2017). reason why D was reduction in the 3% biochar-amended soil. The D provides a combined measure of irregularities and aggregates 5. Conclusions rutting (Gregory, Bird, Watts, & Whitmore, 2012). Studies have shown that D of aggregates is an important factor in the distribu- We examined the effect of rice husk biochar on selected soil tion of aggregate size. The soil aggregates are more stable, resulting properties and nitrate leaching in two types of soil (loamy sand and in less D. In contrast, unstable aggregates have greater fragmenta- clay) in Iran. Enhancing soil structure and nutrient status are an tion and higher D, which reflects the disadvantages of the soil urgent need for sustainable soil management. Results from this structure (Caruso, Barto, Siddiky, Smigelski, & Rillig, 2011). In the study showed that rice husk biochar application had a positive case of water stable aggregates (WSA) (Liu et al., 2017), also re- effect on soil properties, aggregation and nitrate retention of soil. In ported an increase in WSA in biochar amended soils having clay particular, 3% biochar application improved soil aggregate stability content >19%. In this study, loamy sand soil with low clay content indices and nitrate retention in the clay soil. Rice husk biochar has a (17.3%) showed little improvement in MWD, GMD, WSA and D even potential for ameliorating the moderate level of acidity in Loamy with the highest rate (3%) of biochar application. Therefore, high sand soil. However, the application of biochar even at the highest clay content in clay soils may have facilitated the formation of soil rate (3%) in this study did not show any threat of developing macro aggregates. alkalinity or salinity in any soil. Although in treatments of biochar Although biochar is a recalcitrant organic carbon compound, it is without compost, addition of 1% biochar was enough to signifi- not inert. Biochar could gradually change into stable humus after cantly decrease the amount of nitrate leaching. The different effects incorporation into soil (Wang et al., 2018) which its interaction with of rice husk biochar on CEC, nitrate leaching and soil aggregation in clay may led to improvement of aggregate stability. Similarly, an 11- two soils can be related to contrasting characteristics of soils, month incubation study showed that sawdust biochar application particularly soil texture. However, the Loamy sand soil is likely to had not any significant effect on the aggregate stability of the loamy require higher biochar application rates and incubation time than sand, but silty loam (Xiang-Hong et al., 2012). In the other study tested in this study to improve soil aggregation. Our study high- results showed that 2.5 and 5% of biochar application increased lights the importance of soil type in determining the value of rice mean weight diameter of clay soils after 63-day incubation (Jien & husk biochar as a soil amendment. Any improvement in aggrega- Wang, 2013). Biochar oxidation can produce acidic function groups tion can enhance carbon sequestration in soil and consequently and humic materials (Li et al., 2016). In spite of slow oxidation of mitigate CO2 emission to atmosphere. The positive changes in ni- biochar, the promotion of the biochar oxidation has been reported trate retention of soil may increase and regulate the nitrogen in presence of organic matter addition (Munera-Echeverri et al., availability to the crops in addition to alleviating the concerns 2018). The more organic carbon in the clay compared to the about not point pollution of water bodies by nitrate. loamy sand could led to formation of more humic material and increase of aggregate stability. References

4.3. Nitrate leaching Abrishamkesh, S., Gorji, M., Asadi, H., Bagheri-Marandi, G., & Pourbabaee, A. (2015). Effects of rice husk biochar application on the properties of alkaline soil and lentil growth. Plant Soil and Environment, 61(11), 475e482. This suggests the role of the compost as a source of nitrate Ahmaruzzaman, M., & Gupta, V. K. (2011). Rice husk and its ash as low-cost ad- (Hepperly, Lotter, Ulsh, Seidel, & Reider, 2009). Results showed that sorbents in water and wastewater treatment. Industrial & Engineering Chemistry fi Research, 50(24), 13589e13613. an increase in biochar consumption signi cantly reduces the Busscher, W. J., Novak, J. M., & Ahmedna, M. (2011). Physical effects of organic amount of leachate. The anionic adsorption potential and high su- matter amendment of a southeastern US coastal loamy sand. Soil Science, perficial levels of biochar are among the possible reasons for the 176(12), 661e667. adsorption of nitrate ions and their maintenance in the soil (Li et al., Caruso, T., Barto, E. K., Siddiky, M. R. K., Smigelski, J., & Rillig, M. C. (2011). Are power laws that estimate fractal dimension a good descriptor of soil structure and its 2016). Results of a study by (Xiang-Hong et al., 2012) showed that link to soil biological properties? Soil Biology and Biochemistry, 43(2), 359e366. biochar can be effectively used to improve the fertilizer nitrogen Chan, Y., Vowles, P., McTainsh, G., Simpson, R., Cohen, D., & Bailey, G. (1995). Use of a fi use efficiency in sandy soil and reduce nitrate loss from loamy soil. modi ed Walkley-Black method to determine the organic and elemental carbon content of urban aerosols collected on glass fibre filters. Chemosphere, As reviewed by (Liu et al., 2017), high cation exchange capacity, 31(11e12), 4403e4411. enhancement of soil water holding capacity and microbial immo- Chaplot, V., & Cooper, M. (2015). Soil aggregate stability to predict organic carbon bilization of nitrogen as results of biochar application are probable outputs from soils. Geoderma, 243, 205e213. Clesceri, L. S., Greenberg, A. E., Eaton, A. D., Rice, E. W., & Franson, M. A. H. (2005). drivers of nitrogen uptake and retention in soil. Perhaps this jus- Standard methods for the examination of water and wastewater. tifies why nitrate leaching in clay soil treatments (with 47% clay Darzi-Naftchali, A., Shahnazari, A., & Karandish, F. (2017). Nitrogen loss and its M. Ghorbani et al. / International Soil and Water Conservation Research 7 (2019) 258e265 265

health risk in paddy fields under different drainage managements. Paddy and 961e1010. Water Environment, 15(1), 145e157. Page, A., Miller, R., & Keeney, D. (1982). Methods of soil analysis. Part 2. Chemical and Delwiche, L. D., & Slaughter, S. J. (2012). The little SAS book: A primer. SAS Institute. Microbiological properties (2nd ed.) (Madison, Wisconsin, USA). Ding, Y., Liu, Y.-X., Wu, W.-X., Shi, D.-Z., Yang, M., & Zhong, Z.-K. (2010). Evaluation Pandian, K., Subramaniayan, P., Gnasekaran, P., & Chitraputhirapillai, S. (2016). Effect of biochar effects on nitrogen retention and leaching in multi-layered soil of biochar amendment on soil physical, chemical and biological properties and columns. Water, Air, & Soil Pollution, 213(1e4), 47e55. groundnut yield in rainfed Alfisol of semi-arid tropics. Archives of Agronomy and Gee, G. W., & Or, D. (2002). 2.4 Particle-size analysis. Methods of Soil Analysis, 4(598), Soil Science, 62(9), 1293e1310. 255e293. Peng, X., Ye, L., Wang, C., Zhou, H., & Sun, B. (2011). Temperature-and duration- Ghorbani, M., & Amirahmadi, E. (2018). Effect of rice husk Biochar (RHB) on some of dependent rice straw-derived biochar: Characteristics and its effects on soil chemical properties of an acidic soil and the absorption of some nutrients. properties of an ultisol in southern China. Soil and Tillage Research, 112(2), Journal of Applied Sciences & Environmental Management, 22(3), 313e317. 159e166. Gregory, A., Bird, N., Watts, C., & Whitmore, A. (2012). An assessment of a new Pratiwi, E. P. A., & Shinogi, Y. (2016). Rice husk biochar application to paddy soil and model of dynamic fragmentation of soil with test data. Soil and Tillage Research, its effects on soil physical properties, plant growth, and methane emission. 120,61e68. Paddy and Water Environment, 14(4), 521e532. Hepperly, P., Lotter, D., Ulsh, C. Z., Seidel, R., & Reider, C. (2009). Compost, manure Radcliffe, D. E., & Simunek, J. (2010). Soil physics with HYDRUS: Modeling and ap- and synthetic fertilizer influences crop yields, soil properties, nitrate leaching plications. CRC press. and crop nutrient content. Compost Science & Utilization, 17(2), 117e126. Rajkovich, S., Enders, A., Hanley, K., Hyland, C., Zimmerman, A. R., & Lehmann, J. Herath, H., Camps-Arbestain, M., & Hedley, M. (2013). Effect of biochar on soil (2012). Corn growth and nitrogen nutrition after additions of biochars with physical properties in two contrasting soils: An alfisol and an andisol. Geo- varying properties to a temperate soil. Biology and Fertility of Soils, 48(3), derma, 209,188e197. 271e284. Jien, S.-H., & Wang, C.-S. (2013). Effects of biochar on soil properties and erosion Reichenauer, T. G., Panamulla, S., Subasinghe, S., & Wimmer, B. (2009). Soil potential in a highly weathered soil. Catena, 110, 225e233. amendments and cultivar selection can improve rice yield in salt-influenced Juriga, M., Simanský, V., Horak, J., Kondrlova, E., Igaz, D., Pollakov a, N., et al. (2018). (tsunami-affected) paddy fields in Sri Lanka. Environmental Geochemistry and The effect of different rates of biochar and biochar in combination with N fer- Health, 31(5), 573e579. tilizer on the parameters of soil organic matter and soil structure. Journal of Salako, F. (2006). Fractal scaling of soil particles in agricultural landscapes of Ecological Engineering, 19(6), 153e161. Nigerian savannas. International Agrophysics, 20(4), 337. Knowles, O., Robinson, B., Contangelo, A., & Clucas, L. (2011). Biochar for the miti- Shackley, S., Carter, S., Knowles, T., Middelink, E., Haefele, S., Sohi, S., et al. (2012). gation of nitrate leaching from soil amended with biosolids. The Science of the Sustainable gasificationebiochar systems? A case-study of rice-husk gasifica- Total Environment, 409(17), 3206e3210. tion in Cambodia, Part I: Context, chemical properties, environmental and Lado, M., Ben-Hur, M., & Shainberg, I. (2004). Soil wetting and texture effects on health and safety issues. Energy Policy, 42,49e58. aggregate stability, seal formation, and erosion. Soil Science Society of America Singh Mavi, M., Singh, G., Singh, B. P., Singh Sekhon, B., Choudhary, O. P., Sagi, S., Journal, 68(6), 1992e1999. et al. (2018). Interactive effects of rice-residue biochar and N-fertilizer on soil Li, J.-h., Lv, G.-h., Bai, W.-b., Liu, Q., Zhang, Y.-c., & Song, J.-q. (2016). Modification and functions and crop biomass in contrasting soils. Journal of Soil Science and Plant use of biochar from wheat straw (Triticum aestivum L.) for nitrate and phos- Nutrition, 18(1), 41e59. phate removal from water. Desalination and Water Treatment, 57(10), Sumner, M. E., & Miller, W. P. (1996). Cation exchange capacity and exchange co- 4681e4693. efficients. Methods of Soil Analysis Part 3dChemical Methods(methodsofsoilan3), Liu, Z., He, T., Cao, T., Yang, T., Meng, J., & Chen, W. (2017). Effects of biochar 1201e1229. application on nitrogen leaching, ammonia volatilization and nitrogen use ef- Van Trinh, M., Tesfai, M., Borrell, A., Nagothu, U. S., Bui, T. P. L., & Quynh, V. D. (2017). ficiency in two distinct soils. Journal of Soil Science and Plant Nutrition, 17(2), Effect of organic, inorganic and slow-release urea fertilisers on CH 4 and N 2 O 515e528. emissions from rice paddy fields. Paddy and Water Environment, 15(2), 317e330. Manolikaki, I., & Diamadopoulos, E. (2017). Ryegrass yield and nutrient status after Varela Milla, O., Rivera, E. B., Huang, W.-J., Chien, C., & Wang, Y.-M. (2013). Agro- biochar application in two Mediterranean soils. Archives of Agronomy and Soil nomic properties and characterization of rice husk and wood biochars and their Science, 63(8), 1093e1107. effect on the growth of water spinach in a field test. Journal of Soil Science and Masulili, A., Utomo, W. H., & Syechfani, M. (2010). Rice husk biochar for rice based Plant Nutrition, 13(2), 251e266. cropping system in acid soil 1. The characteristics of rice husk biochar and its Wang, Y., Wang, Z.-L., Zhang, Q., Hu, N., Li, Z., Lou, Y., et al. (2018). Long-term effects influence on the properties of acid sulfate soils and rice growth in West Kali- of nitrogen fertilization on aggregation and localization of carbon, nitrogen and mantan, Indonesia. Journal of Agricultural Science, 2(1), 39. microbial activities in soil. The Science of the Total Environment, 624,1131e1139. Mehmood, K., Baquy, M. A.-A., & Xu, R.-k. (2018). Influence of nitrogen fertilizer Wilson, M., Burt, R., Indorante, S., Jenkins, A., Chiaretti, J., Ulmer, M., et al. (2008). forms and crop straw biochars on soil exchange properties and maize growth Geochemistry in the modern soil survey program. Environmental Monitoring on an acidic ultisol. Archives of Agronomy and Soil Science, 64(6), 834e849. and Assessment, 139(1e3), 151e171. Metwally, S., & Pollard, A. (1959). Effects of soil moisture conditions on the uptake Xiang-Hong, L., Feng-Peng, H., & Xing-Chang, Z. (2012). Effect of biochar on soil of plant nutrients by barley and on the nutrient content of the soil solution. aggregates in the loess plateau: Results from incubation experiments. Interna- Journal of the Science of Food and Agriculture, 10(11), 632e636. tional Journal of Agriculture and Biology, 14(6). Munera-Echeverri, J. L., Martinsen, V., Strand, L. T., Zivanovic, V., Cornelissen, G., & Yousefzadeh, S., Modarres-Sanavy, S. A. M., Sefidkon, F., Asgarzadeh, A., Mulder, J. (2018). Cation exchange capacity of biochar: An urgent method Ghalavand, A., & Sadat-Asilan, K. (2013). Effects of Azocompost and urea on the modification. The Science of the Total Environment, 642,190e197. herbage yield and contents and compositions of essential oils from two geno- Nelson, D. W., & Sommers, L. E. (1996). Total carbon, organic carbon, and organic types of dragonhead (Dracocephalum moldavica L.) in two regions of Iran. Food matter. Methods of Soil Analysis Part 3dChemical Methods (Methodsofsoilan 3), Chemistry, 138(2e3), 1407e1413. International Soil and Water Conservation Research 7 (2019) 266e274

Contents lists available at ScienceDirect

International Soil and Water Conservation Research

journal homepage: www.elsevier.com/locate/iswcr

Review Article Review of remote sensing and geospatial technologies in estimating rooftop rainwater harvesting (RRWH) quality

* Masayu Norman a, c, Helmi Z.M. Shafri a, b, , Shattri B. Mansor a, b, Badronnisa Yusuf a a Department of Civil Engineering, Faculty of Engineering, Universiti Putra Malaysia (UPM), 43400 Serdang, Selangor, Malaysia b Geospatial Information Science Research Centre (GISRC), Faculty of Engineering, Universiti Putra Malaysia (UPM), 43400 Serdang, Selangor, Malaysia c Centre of Studies for Surveying Science & Geomatics, Faculty of Architecture, Planning & Surveying, Universiti Teknologi MARA (Perlis), 02600 Arau, Perlis, Malaysia article info abstract

Article history: Rooftop rainwater harvesting (RRWH) systems can provide low-cost decentralized water to urban and Received 27 February 2019 rural households that have no access to treated water. These systems are considered the key strategic Received in revised form adoption measures for communities affected by climate change. Roofing materials, roofing conditions, 3 May 2019 roofing geometry, weather conditions, and land use/land cover (LULC) conditions can significantly affect Accepted 13 May 2019 the quality of RRWH. Therefore, the effects of these factors on RRWH quality must be analyzed carefully. Available online 26 May 2019 Remote sensing and Geographical Information System (GIS) have been widely used in urban environmental analysis. However, these technologies have never been used to analyze, map, and model the Keywords: Roofing materials effect of various factors on RRWH quality. This review determines the research gaps in the use of geo- fi Roofing conditions spatial technologies in estimating RRWH quality and simulate the implications of roo ng materials and Spatial modeling roofing surface conditions towards the urban environment. An approach for the integrated use of remote Environmental impact sensing and GIS to assess the quality of RRWH is also proposed. Remote sensing © 2019 International Research and Training Center on Erosion and Sedimentation and China Water and GIS Power Press. Production and Hosting by Elsevier B.V. This is an open access article under the CC BY-NC- ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction use of rainwater has been widely introduced as a reliable alternative. Accordingly, studies on the Rainwater Harvesting System Freshwater scarcity is a worldwide problem due to the (RWHS), have increased significantly in many countries, particu- increasing world population, improving living standards, changing larly on the techniques and treatment system (Matos, Santos, consumption patterns, and the expansion of irrigated agriculture Pereira, Bentes, & Imteaz, 2014; Fonseca, Hidalgo, Díaz-Delgado, (Ercin & Hoekstra, 2014; de Fraiture & Wichelns, 2010). It was re- Vilchis-Frances, & Gallego, 2017) evaluation of potential water ported that two third of the world population will be exposed to saving (Khastagir & Jayasuriya, 2010; Belmeziti, Coutard, & De water scarcity by 2025 (Fig. 1), and among the countries which will Gouvello, 2014), financial cost (Sample & Liu, 2014; Morales- be affected are some African countries, Asian countries, United Pinzon, Rieradevall, Gasol, & Gabarrell, 2015), and also quality States of America, Sudan and China (UN-HABITAT, 2003). At pre- and environmental analysis (Evans, Coombes, & Dunstan, 2006; sent, freshwater sources are becoming limited and polluted. In Tito et al., 2014). Rainwater can be used either directly from the addition, drinking water is becoming contaminated and these surface water or groundwater that has been recharged (Fig. 2). This phenomena have become major public concerns. Therefore, several system is the most significant approach for providing sustainable studies have been carried out to assess global water scarcity in water cycles, particularly in urban development (Ajim et al., 2017; terms of physical, social, and economical aspects (Mekonnen & Campisano & Lupia, 2017; Lee, Bak, & Han, 2012; Lye, 2009; Zhang Hoekstra, 2016; Oki et al., 2001; Wolfe & Brooks, 2003). et al., 2014). To reduce and minimize the consequences of water scarcity, the Rooftop rainwater harvesting (RRWH) system becomes popular due to its capability to provide more opportunity for collecting water from a larger surface area since building rooftops is the most important land cover type and almost half of the entire surface in * Corresponding authorDepartment of Civil Engineering, Faculty of Engineering, Universiti Putra Malaysia (UPM), 43400 Serdang, Selangor, Malaysia. the cities are covered with roofs (Ugai, 2016). Although RRWH is E-mail address: [email protected] (H.Z.M. Shafri). assumed to be a relatively safe source of non-potable and potable https://doi.org/10.1016/j.iswcr.2019.05.002 2095-6339/© 2019 International Research and Training Center on Erosion and Sedimentation and China Water and Power Press. Production and Hosting by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). M. Norman et al. / International Soil and Water Conservation Research 7 (2019) 266e274 267

Previous studies have applied remote sensing and GIS techniques approach broadly in rooftop rainwater harvesting studies. Yet, these technologies focused on their applications for RWH potential area (Bakir & Xingnan, 2008; Gaikwad, 2015; Hari, Reddy, Vikas, Srinivas, & Vikas, 2018). For example, a geo-information system comprising digital data sets of satellite imagery, topography, soil, vegetation, hydrology, and meteorology was established and this information was used to identify the areas which are generally suitable for water harvesting in the Syrian Desert (Bakir & Xingnan, 2008). Hence, the present study reviews global literature drawn from journal articles, conference proceedings and research reports. The specific objectives were also determined: (1) to evaluate the determinants of RRWH quality and their effects, (2) to discuss several examples of remote sensing applications on the water quality, (3) to assess the capabilities of remote sensing and Geographical Infor- mation System (GIS) in RWH application, and (4) to evaluate the remote sensing approaches used in the detection of roofing mate- Fig. 1. Predicted water scarcity by 2025 (Source: UN-HABITAT, 2003). rials and conditions, and (5) to determine gaps in existing research that can be filled by future research. Furthermore, this review proposes the development of advanced roofing materials and conditions and a novel spatial model to map, simulate, and predict the quality of harvested rainwater based on several factors.

2. Issues of RRWH quality and its determinants

The absorptions of contaminants in roof runoff are highly dependent on the sources, anthropogenic activities, flow pathways, and interactions with the environment. The quality of harvested rainwater is dependent on roofing materials and environmental circumstances such as local temperature and atmospheric pollution (Lee et al., 2012). Zhang et al. (2014) categorized the determinants of roof runoff quality into three namely roof characteristics (such as type, age, and roughness), rainfall characteristics (such as precipitation, antecedent dry weather period, and rainfall intensity), and environmental characteristics (such as seasonal variation, atmospheric pollution, and roof surrounding environment). Chang, Fig. 2. The usage of harvested rainwater. McBroom, and Beasley (2004) suggested that the composites of roofing materials can leach into the runoff. Moreover, aging roofs contain large quantities of suspended solids. water and is regarded as non-polluted storm water, yet the quality Catchment in rainwater harvesting systems refers to areas that of the collected rainwater represents another challenge whereby directly contribute to the harvested runoff, such as roof and pave- roof runoff can be contaminated from the rainwater source to the ment. However, the research on the quality and public health risks point of consumption due to several factors that have been iden- posed by water harvested from roofs is relatively limited (Gwenzi tified from previous studies (Boller, 1997; Gwenzi, Dunjana, Pisa, et al., 2015a, 2015b). Although there are many factors that affect Tauro, & Nyamadzawo, 2015a, 2015b; Zhang et al., 2014). the quality of harvested rainwater, but this study comprehensively Most of the existing RRWH are used directly from the collection reviews the five main sources of contaminants namely roofing tank for non-potable uses, such as for toilet flushing, landscape materials, roofing conditions, roofing geometry, weather condi- irrigation, and car washing, in which the water does not require tions, and land use/land cover conditions. treatment. However, minimum treatment is still needed before it can be utilized for potable uses such as for drinking, cooking, 2.1. Roofing materials showering, and cloth washing even though rainwater is relatively clean from major contaminants (Md Lani, Yusop, & Syafiuddin, Roofing is an example of a pollutant source that originates from 2018). Consequently, RRWH quality must be analyzed carefully materials collected on the roof and the roofing physical itself before it can be used for potable or non-potable purposes and the (Tobiszewski, Polkowska, Konieczka, & Namiesnik, 2010). Boller quality should depend strongly on its use and water quality (1997) discovered that roof runoff does not only contain heavy standard. metals, but also organic halogens and polycyclic aromatic hydro- Many studies of RRWH focused on water quality assessment, carbons (PAHs). Wallinder, Bahar, Leygraf, & Tidblad, (2007) re- and findings identified that many confounding factors contribute to ported that roofing can be an important contaminant source in water quality, for instance, roofing materials, roofing surface con- urban environments, where most of the landscapes are covered by ditions, roofing geometry, weather conditions, and land use/land roofs. cover (LULC) conditions. The examples of such studies are listed in The complex role of roofing material involves serving as the Table 1. Nevertheless, the results obtained are mostly presented in a basis for pollutants, a provisional basin where contaminants are form of tabular data, in which it makes it difficult to convey the adsorbed on its surface and a re-emitter once runoffs flush the information regarding water quality quickly and effectively. adsorbed pollutants. Chang et al. (2004) compared the quality of 268 M. Norman et al. / International Soil and Water Conservation Research 7 (2019) 266e274

Table 1 Example of studies in assessing the factors affecting the quality of RRWH.

Factor Reference Finding

Roof material Zhang et al. (2014) ⁃conventional roof materials such as concrete, asphalt and ceramic tile have significant effect on RRWH quality Lye (2009) ⁃levels of heavy metal contamination specific to runoff from rooftop areas containing exposed metal surfaces Roof surface Clark et al. (2008) ⁃examined various roofing materials to explore the influence of aging and weathered roofs on runoff conditions Gwenzi et al. (2015a, ⁃aging metal roofing sheets are consistently associated with high concentrations of heavy metals in roof runoff 2015b) Roof geometry Farreny et al. (2011) ⁃runoff can be quickly shed from the steep roof and that contamination in such a roof can be treated easily Ugai (2016) ⁃A large catchment area increases the concentration of pollutants in roof runoff Weather Sazakli et al. (2007) ⁃antecedent and prevailing weather conditions, such as wind and rainfall characteristics, influence the physic-chemical and conditions Evans et al. (2006) microbiological qualities of roof runoff ⁃weather patterns can also significantly influence the bacterial load of roof runoff Surrounding LULC Huston et al. (2011) ⁃poor rainwater quality in industrialized catchments is attributed to air pollution and atmospheric deposition from traffic and industrial emissions and aerosols

harvested rainwater in Nacogdoches, Texas from four roofing ma- can negatively influence harvested rainwater quality. The assess- terials namely painted aluminium, galvanized iron, composite ment of roof deterioration will become increasingly difficult shingle, and wood shingle. They determined that roofing material because old roofs are possibly weathered (Cilia et al., 2015). Forster significantly influences pH, electrical conductivity, and zinc levels (1996a, 1996b) and Simmons, Hope, Lewis, Whitmore, and Gao and that wood shingles have the worst runoff quality compared (2001) reported that corrosion and weathering of zinc sheets with those of other investigated roofing materials. According to result in high concentrations of zinc in rainwater. Meanwhile, Clark Despins, Farahbakhsh, and Leidl (2009), the association of total et al. (2008) examined various roofing materials to explore the organic carbon, color, and turbidity of harvested rainwater quality influence of aging and weathered roofs on runoff. Their results from steel roofs is safer than that from asphalt shingle roofs. indicated that nutrient release exists in traditional galvanized Unlike conventional roofing materials (i.e. metal, concrete tiles, metal roofing and wood products as a consequence of natural and alternative roofing materials), cool roofs are appropriate for degradation. rainwater harvesting purposes. However, roofing materials, such as The comparison of runoff quality from concrete, tile, asbestos green and shingle roofs, produce high dissolved organic carbon and galvanized iron roofs shows high variability in water quality concentrations (Mendez et al., 2011). Lee et al. (2012) compared the parameters among different materials. This result is attributed to quality of harvested rainwater in Seoul, South Korea from four the dissimilar ages of roofing materials. Similarly, aging metal different roofing materials such as concrete tiles, clay tiles, galva- roofing sheets are consistently associated with high concentrations nized steel, and wooden shingles. They determined that galvanized of heavy metals in roof runoff (Gwenzi et al., 2015a, 2015b). In steel is mainly appropriate for rainwater harvesting purposes after conclusion, the ratio of physical-chemical parameters in roof runoff the first flush followed by clay tiles. Zhang et al. (2014) evaluated to that in free-falling water is also higher for most parameters in the impact of conventional roofing materials, for examples, asphalt, aging roofs than that in new roofs (Meera & Ahammed, 2011, pp. ceramic tile, and concrete, on the quality of harvested rainwater in 686e691). Chongqing, China and compared them with other roofing material Quek and Forster (1993) examined surface roughness and (i.e. green roof). They revealed that the ceramic tile roof is the most showed that low roughness results in maximum suspended parti- appropriate material for rainwater harvesting purposes among cle concentrations in runoff and the comparatively high inclination other investigated roofing materials because of the low mean of zinc sheet roof. They also compared tar felt and pantile roofs and concentration of water quality parameters. However, the mean revealed that high concentrations of suspended solid on the two values of total nitrogen (TN) and chemical oxygen demand for roofs are caused by high levels of weathering on the AC roof surface. ceramic tile roof exceed the standard for drinking water. In addition, the roughness of the AC roof becomes high due to Several studies confirmed that asbestos dust severely affects trapped particles, except when rain intensities are high. human health. People who are exposed to asbestos fibers poten- The wash off of dryly deposited substances, such as ammonia tially have a long-term risk of developing lung cancer, larynx can- and Polycyclic Aromatic Hydrocarbons (PAHs), increased the wet cer, pleural mesothelioma, and asbestosis. Consequently, Frassy deposition load of organic micro pollutants from roof surfaces; et al. (2014) decided to map the coverage area of meanwhile, leaching from the roofing material increased the con- asbestosecement (AC) using Multispectral Infrared & Visible Im- centrations of inorganic ions, such as calcium. These processes aging Spectrometer sensor. They managed to detect the AC surface show the influence of roof surface on generalizing the level of roof with 80% accuracy. runoff pollution, particularly within a small area (Forster, 1998). Roofing materials, such as concrete and metal roofs, significantly affect the quality of harvested rainwater because of their smooth 2.2. Roofing conditions surfaces. Metal roofs differ from other roofs because of the low degree of bacteria; these roofs are highly recommended for rain- Roofing conditions refer to the degradation level of roofs. This water harvesting purposes (Mendez et al., 2011). Correspondingly, review paper divides roofing conditions based on two elements: (1) Yaziz, Gunting, Sapari, and Ghazali (1989) proved that the bacterial roofing materials to extract information on roofing age and (2) roof concentrations in harvested rainwater from metal roofs are lower surface to extract information on roughness or texture. than those from concrete tile roofs. Aging roofs have high potential to give harmful effect to the The International Agency for Research on Cancer categorized quality of harvested rainwater, and the inconsistent effects of asbestos as carcinogenic to humans. Due to that, Tobiszewski et al. roofing material on roof runoff quality can be attributed to the (2010) investigated the effect of roofing material on the concen- existence of confounding factors. In this regard, the age of roofing trations of contaminants in roof runoff waters and the variations of materials and the frequency of maintenance also influence roof contamination concentration with time. Four types of roofing runoff quality. Chang et al. (2004) demonstrated that aging roofs M. Norman et al. / International Soil and Water Conservation Research 7 (2019) 266e274 269 materials, namely asbestos tile, zinc sheeting, ceramic tile, and This result may be attributed to the fact that clay tiles are prone to bituminous membrane, were compared and examined. These ma- physical and chemical degradation. Sloping smooth roofs are con- terials are known sources of pollutants. Thus, the concentrations of tradicted to improve the quality of its harvested rainwater. This pollutants increased in runoff water collected from asbestos tile valuable finding can help city policy planners establish guidelines and tar paper, perhaps because of the preferable sorption of pes- for roofing slope by redesigning existing buildings. ticides on the surfaces of these materials. Harvested rainwater quality based on roofing slope and rough- Lee et al. (2012) examined the effect of different roofing mate- ness is important for local administrations and town planners in rials, including concrete tiles, galvanized steel, wooden shingles, redesigning buildings and towns from the viewpoint of sustainable and clay tiles, on the microbiological and chemical qualities for rainwater management. Sloping smooth roofs are preferred for domestic application. They also assessed the quality of harvested rainwater harvesting because of their superior and proven rainwater according to the existence of mosses and lichens on the performance. roof surface. The microbiological, chemical and physical qualities of harvested rainwater were badly affected by mosses and lichens. The 2.4. Weather conditions level of porosity of three roofing materials, namely concrete tiles, clay tiles, and wooden shingle tiles, is high. Thus, these roofing Weather is the most influential factor of the microbiological materials trap pollutants easily. Mosses and lichens are commonly quality of harvested rainwater. Apart from other factors, such as found in clay tiles, wooden shingle tiles, and concrete tile roofs. The relative source location, weather patterns can also significantly highest degree of colonization is observed in wooden shingle tile influence the bacterial load of roof runoff (Evans et al., 2006). roofs. Sazakli, Alexopoulos, and Leotsinidis (2007) found that antecedent Cilia et al. (2015) analyzed the weathering status of and prevailing weather conditions, such as wind and rainfall asbestosecement (AC) roof by introducing a roof deterioration in- characteristics, influence the physic-chemical and microbiological dicator based on a spectral index (i.e. visible and near-infrared qualities of roof runoff. bands). This device facilitated the analysis of roofing conditions According to Yaziz et al. (1989), rainfall amounts and timing are based on mosses and lichens that are usually developed on aging critical factors that influence roof runoff pollution. Thus, low rain- and weathered roofs. The authors assumed that the color of fall events resulted in high concentrations of contamination. The weathered is commonly dark roofs because of the vegetation that quality of runoff from roof catchment systems is significantly settled. The influence of roof surface on determining the quality of influenced by the deposition of numerous contaminants from the harvested rainwater is another important aspect in evaluating the atmosphere on roof surfaces, particularly throughout a dry period. impact of roofs on the surrounding environment. The volume of contaminants deposited on the roof surfaces increases when the dry period between rainfall events is extended. 2.3. Roofing geometry According to Radaideh, Al-Zboon, Al-Harahsheh, and Al-Adamat (2009), lead concentration is high when the rainfall amount is The use of potable water can be reduced by designing the less than 20 mm/day. roofing for rainwater collection. Roof design characteristics, such as Similarly, Jiries, Hussein, and Halaseh (2001) discovered the size, slope, and length, also influence the quality of roof runoff. The correlation between the intensity of rainfall and the contamination potential volume of rainwater collected for harvested rainwater level of the quality of harvested rainwater. This relationship is systems is dependent on the size of the catchment area of the roof shown in the water quality of Zarqa Governorate which is low when (Ugai, 2016). The catchment region is constructed from the roof the rainfall volume is reduced to less than 140 mm/year. “footprint” and can be calculated by determining the region of the Despins et al. (2009) concluded that color is another water building and the region of the overhang of the roof. A large quality parameter because of its sensitivity to rainfall or drought catchment area increases the concentration of pollutants in roof conditions. Color can be affected by rainfall when its amount is less runoff. than 0.05 and the antecedent dry period is less than 0.05. This Roofing slope largely contributes to the level of runoff. Param- finding is attributed to the increase of detected fecal coliforms in eters, such as roofing materials and roofing geometry, significantly summer and fall seasons. Such an increase is due to cold air tem- affect the contamination level in harvested rainwater (Yaziz et al., peratures inhibiting the growth of bacteria within the cistern itself. 1989). Roofing slope is significant for the promotion of rainwater as an alternative water supply to prevent flood water and water 2.5. Surrounding land use/land cover (LULC) scarcity. The quickness of water runoff during rainfall will be affected by Evans et al. (2006) mentioned that the location of the catchment the roofing slope. Farreny et al. (2011) mentioned that runoff can be area of harvested rainwater can significantly influence the bacterial quickly shed from the steep roof and that contamination in such a load of roof runoff, particularly when this factor integrates with roof can be treated easily. Meanwhile, a flat and less steep roof other factors such as weather patterns. Similarly, environmental causes water to flow slowly, thereby increasing the probability of conditions, such as proximity of roofs to major roads or industri- pollutants to endure on the catchment surface. Accordingly, the alized areas, can reduce the quality of harvested rainwater (Forster, authors developed a model for estimating the runoff and initial 1996a, 1996b), and the existence of birds can cause the same abstraction of each roof and assessed the physicochemical con- problem. It is further emphasized that, the chemical composition of taminations of roof runoff. They concluded that a flat gravel roof collected rainwater is significantly affected by the following factors: produces the most polluted runoff. This effect is likely due to weather conditions; location of the sampling points; and industrial, mosses and lichens that grow on weathered roofs. urban, or agricultural activities (Meera & Ahammed, 2006). In contrast, Quek and Forster (1993) mentioned that a flat gravel Harvested rainwater quality depends on the location of har- roof is likely to become a sink, particularly for particles, and vested rainwater systems. Despins et al. (2009) argued that the site allowed the constituents of sedimentation because of its hydraulic environment has a negative impact on the quality of harvested properties. Ammonia from flat gravel roofs does not exhibit the rainwater. They compared urban and rural sites and observed the lowest level. However, clay roof shows a high level of contamina- statistically significant effect of site environment. The high usage of tion, whereas plastic and metal roofs exhibit a similar water quality. nitrogen that contains agricultural inputs also influences rainwater 270 M. Norman et al. / International Soil and Water Conservation Research 7 (2019) 266e274 quality. Consequently, the quality of rainwater in rural areas is utilized for the estimation of chlorophyll. High accuracy results significantly better than that in urban areas. obtained from this study indicated the feasibility of applying the Radaideh et al. (2009) conducted a comparative study of rain- satellite images for water quality studies (Lim, MatJafri, Abdullah, water quality in two catchments with contrasting land use prac- Alias, & Mohd. Saleh, 2008). The potential applications of the tices; the results reveal that rainwater quality is poorer in Landsat 8 Operational Land Imager (OLI) for estimating the water industrialized areas than that in agricultural areas. Water quality quality in the Nakdong River, Korea, has been evaluated and the varies depending on the location. Specifically, industrial activities results demonstrate that the Landsat 8 OLI may provide a useful also contributed to the low quality of harvested rainwater in Zarqa tool for investigating the spatio-temporal variability of water Governorate. quality parameter in surface waters (J. Lim & Choi, 2015). Given the diffused nature of air pollution and wind transport, The latest technology in visible/near-infrared imaging spec- the effects of industrialization on water quality may extend beyond troscopy has proven that high-resolution remote sensing can also the spatial boundaries of industrial activities (Uygur, Karaca, & help to monitor and understand the more diffuse and complex Alagha, 2010). Huston, Chan, Chapman, and Shaw (2011) empha- effects of wetland restoration on water quality (Fichot et al., 2016). sized that poor rainwater quality in industrialized catchments is This technology can facilitate water quality monitoring by enabling attributed to air pollution and atmospheric deposition from traffic simultaneous depictions of several water quality indicators at a and industrial emissions and aerosols. very high spatial resolution. Furthermore, the critically affected Agricultural activity results in a higher concentration of phos- areas can be a remediation using contamination zones mapped by phamidon (a pesticide constitute) in rural agricultural catchments high-resolution (Ramadas & Samantaray, 2017), whereby the than in urban areas; such concentration results in the increase in mapping process is based on the spectral signatures collected by the toxicity of roof runoff in rural areas (Rouvalis, Karadima, Zioris, sensors. These spectral signatures provide information on con- Sakkas, & Albanis, 2009). Given the inconsistent findings on poor taminants and their concentration. Similarly, the water quality rainwater quality attributed to industrial and traffic emissions parameter was obtained using ASTER data and the spatial distri- (Gikas and Tsihrintzis, 2012), the effects of industrial activities on bution map for each parameter of water quality was then estab- roof runoff quality are proposed to be site-specific and may depend lished for Qaroun Lake in Egypt (Abdelmalik, 2018). on the rainfall characteristics, the magnitude of industrial emis- The application of remote sensing is also applied for monitoring sions, and other confounding factors (Gwenzi et al., 2015). water quality parameters such as suspended matter, phyto- Roofing materials and roofing age (surface conditions) can be plankton, turbidity, and dissolved organic matter (Usali & Hasmadi, categorized as factors that become priorities in RRWH quality 2010). This information is useful in providing a solution for future study, considering that almost every year the research is made on water resources planning and management for the government that factor. These factors have a high number of studies in 2012, especially in formulating policy related to water quality. Interest- 2014 and 2015 as shown in Fig. 3. ingly, with appropriate in situ validation, remotely sensed satellite images can provide high-accuracy interpolation, generating 3. Applications of remote sensing in water quality mapping spatially-explicit water quality maps more efficiently, even for places with limited in situ samplings (Haji Gholizadeh et al. (2016). The conventional methods of water quality monitoring have In summary, the rapid development of remote sensing tech- several limitations such as labor-intensive, time-consuming, and niques enables the application of remotely sensed data for moni- costly (Gholizadeh, Melesse, & Reddi, 2016). Consequently, the use toring large-scale and long-term water quality, especially the of remote sensing in water quality assessment can overcome these launch of hyper-resolution satellites. Moreover, it demonstrates limitations due to its capability to monitor water quality parame- that remote sensing can play an important role in water quality ters (i.e. suspended sediments (turbidity), chlorophyll, and tem- assessment by using high-resolution satellite image. perature). Several studies have successfully utilized remote sensing technology for water quality monitoring and management. For 4. Application of remote sensing and GIS in rainwater example, a new algorithm has been developed for chlorophyll harvesting mapping over Kedah, Malaysia, in which ALOS satellite data is Many researchers have developed and applied various meth- odologies and criteria to identify suitable sites and technique for RWH (Arshad, Shabbir, & Shakoor, 2013; Singh, 2014; Tiwari, Goyal, & Sarkar, 2018). Descriptions of environmental contamination in urban areas can be developed in a sophisticated approach using dynamic models, which can deal with spatialetemporal data. GIS manages the spatial description of urban environments. Modeling focuses on the functioning of environmental processes which include different forms of urban pollution. Remote sensing and GIS approach are applied to identify the potential RWH site. A GIS-based model, known as a decision support framework, was used to optimize and identify the locations to implement RWH management strategies effectively and efficiently (Abrefa, Eric, Eric, & James, 2013). A system which was developed by (Kumar & Viswanadh, 2012) enable them to determine the suitable sites for rainwater harvesting. Check dams, Percolation tanks, Contour trenches are the examples of structures identified using this system. Combination of Analytic Hierarchy Process (AHP) and Geospatial Hydrologic Modeling Extension manage to identify the potential RWH area which could minimize stagnant storm Fig. 3. Frequency of studies based on a factor which affected RRWH quality. water up to 26% through supplementing city water supply annually M. Norman et al. / International Soil and Water Conservation Research 7 (2019) 266e274 271 up to 20 L/person/day. Thus, this integration is believed to be an pollution (Clark et al., 2008), and building rooftops functions as the important guide to the decision supporting system for urban areas foremost role in the urban storm water runoff and yet covers almost (Aysha & Shoukat, 2015). half of the entire surface area in urban environments (Ugai, 2016). The roofing layer was developed as a database which was then Existing studies discovered that harvested rainwater for urban utilized for RRWH potential area evaluation and the selection of areas depends on several factors, such as the quality of roof runoff area is based on various urban roofing materials. WorldView-3 in terms of roofing materials, roof conditions, roof geometry, satellite image was used to obtain spectral and spatial informa- weather conditions, and land use, to retain urban sustainability. tion of buildings and roofs. Results from this study were then used However, roofing materials and roofing conditions were identified to develop RRWH model for an urban environment (Radzali, Shafri, as the most important determinants for RRWH quality. The Norman, & Saufi, 2018). Furthermore, the GIS examination tech- extraction of roof information is appreciated and the understand- nique was applied to measure the RRWH potential, whereby it is an ing of diverse materials and their conditions can assist in the quality efficient assessment of rooftop water collecting in the Wollert, assessment of rooftop harvested rainwater. which is a suburb in Melbourne, Victoria. With the use of GIS, it was The process of roof deterioration, in which metals released from conceivable to appraise the aggregate sum of water harvestable at aged roofs, for instance, contributes to harmful effects even after 60 the household level (Baby, Arrowsmith, & Al-Ansari, 2019). years (Clark et al., 2008). Thus, remote sensing is recommended to Roof runoff in urban areas shows high levels of metallic con- obtain the information of roofing materials and roofing conditions centration because metallic elements are generated by corrosion of due to its capability to discriminate the spectral reflectance from roofing materials before being swept away by rainwater. Therefore, similar materials that usually causes spectral confusion (Myint, Bris and Robert (2009) modeled this phenomenon and evaluated Gober, Brazel, Grossman, & Weng, 2011). metallic flows from roofs in rainwater at the roof scale. They Several studies on urban roof mapping were conducted using defined five classes of roofs, namely, zinc plate, slate, red tile, brown high-resolution satellite images. tile, and flat roofs. Remote sensing data such as land use and land cover was Roof runoff in urban areas shows high levels of metallic con- applied to identify the suitable rooftop area as potential sites for centration because metallic elements are generated by corrosion of water harvesting (Gaikwad, 2015). Taherzadeh and Shafri (2013) roofing materials before being swept away by rainwater. Bris and proposed a model that can predict and detect four roofing mate- Robert (2009) modeled this phenomenon and evaluated metallic rials which are widely used in Malaysia namely concrete tile, clay flows from roofs in rainwater at the roof scale. They defined five tile, asbestos, and metal. These roofing materials information were classes of roofs namely zinc plate, slate, red tile, brown tile, and flat then used for RRWH quality assessment. Interestingly, existing roofs. Their proposed method is powerful and relatively easy to use geospatial data and derived data from satellite image were used to for areas when orthophotos are available. In other study, it was estimate the requirements water to irrigate fruits and vegetable discovered that urban runoff is a significant resource of environ- gardens through the collection and use of harvestable rainwater mental contamination and roofing material (particularly metal) is from buildings’ rooftops in the urban area of Rome (Italy). The re- the main basis of urban runoff metal contaminations. Therefore, a sults show that RRWH also provides benefit to urban agriculture in new method to evaluate yearly pollutant flow release from dis- residential areas (Lupia, Baiocchi, Lelo, & Pulighe, 2017). similar roofing material elements (such as gutters and rooftops) Bassani et al. (2007) used remote sensing MIVIS data and also was developed (Emna, Bernard, Gromaire, & Chebbo, 2014). TIR and SWIR multispectral data to assess the effect of deterioration GIS modeling was applied to identify suitable sites for rainwater level of asbestos roof based on a number of mineral fibers exposed. harvesting systems based on different criteria. Determining the They revealed that asbestos fiber air dispersion acts as a pathogenic best method or guideline for site selection is difficult, therefore a agent that can cause lung cancer in human beings. Moreover, a general method reviewed by Adham, Riksen, Ouessar, and Ritsema mass balance-based method was employed to estimate RWH per- (2016) is for selecting the suitable rainwater harvesting sites based formance of four large metropolitan areas of the U.S. Interestingly, on the methods developed in the past three decades. They deter- they applied geospatial analysis to characterize the number of in- mined the most important criteria for selecting suitable sites for dividuals living in each building as well as the available rooftop harvested rainwater, which include slope, land use/land cover, soil areas for rainfall harvesting (Rostad, Foti, and Montalto (2016). type, rainfall, distance to settlements/streams, and cost. Recent As mentioned earlier, asbestos is known as a harmful material, methods for selecting rainwater harvesting sites using GIS com- in which exposure to asbestos results in serious risks of developing bined with hydrological models and multi-criteria analysis were mesothelioma and lung cancer. Therefore, hyperspectral MIVIS data developed. was utilized to discriminate asbestos-cement roofs from other The results from GIS models are significant components for materials. The weathering conditions and deterioration status of environmental impact evaluation and can provide the foundation roofs were also assessed by observing the exposure of asbestos fiber for the decision-making process of environmental management from each type of roofing material (Fiumi, 2012). Furthermore, both agencies. With the application of GIS, it is possible to assess the the European and the Italian legislations have banned the manu- total potential of water that can be harvested and to determine the facture, importation, processing and distribution in commerce of suitable location for HRW. The total of harvested rainwater can be asbestos-containing products and have recommended action plans estimated by utilizing remote sensing and GIS techniques. Inte- for the safe removal of asbestos from public and private buildings. gration of these technologies provides a reliable, accurate and up- Thus, Frassy et al. (2014) utilized the hyperspectral remote sensing to-date database on land and water resources, which is a pre- MIVIS sensor to obtain AC roofing map over a large mountainous requisite for an integrated approach in identifying potential area of the Italian Western Alps. Remote sensing analysis provides runoff zones and suitable sites for water harvesting structures such valuable information that can assist authorities in producing AC as check dams, farm ponds, percolation tanks and bandies. roofing map. Knowledge concerning the roofing materials or the different 5. Remote sensing in the detection of roofing materials and kinds of ground areas is required because several kinds of pollution conditions are generated by roofing materials. Consequently, a superspectral sensor was designed for urban materials classification, and mate- The urban surface is a contributor to long-term environmental rials maps are provided for modeling and monitoring environment 272 M. Norman et al. / International Soil and Water Conservation Research 7 (2019) 266e274 applications (Bris & Robert-Sainte, 2009; Bris, Chehata, Briottet, & effects on the environment. This combination is useful in mapping Paparoditis, 2016). enhancement, particularly for roofing conditions. The fusion of In conclusion, remote sensing technologies can assist the au- optical and active systems is valuable if applied in certain areas that thorities in mapping the index of RRWH quality and it can provide have cloud cover. In addition, the use of the OBIA technique in useful information for supporting environmental monitoring in obtaining significant information on roofing materials and its terms of RRWH quality. related characteristics for harvested rainwater quality assessment can be applied because this technique can reduce within-class 6. Conclusion spectral variation. Moreover, the GIS modeling approach can be adopted by This review shows that the quality of rooftop harvested rain- implementing intelligent deep learning which combines spectral water is affected by roofing materials, roofing conditions, roofing and spatial dimensions. In terms of weather conditions, monthly geometry, weather conditions, and surrounding land use condi- rainfall (in millimeter) during low and high seasons should be tions (Fig. 4). Generally, remote sensing is useful in extracting collected. Surrounding LULC information can be adopted using various determinants of harvested rainwater quality. Such infor- image identification and classification. mation can then be compiled, mapped, and modeled in a GIS to This review identifies a few research gaps. Several factors are generate a spatial model. The resulting map and model can help to identified as RWH quality determinants, but there are only a few identify the potential rainwater harvesting sites and model the that significantly affect the quality of RWH; roofing materials, predicted quality over large areas. roofing conditions, roofing geometry, weather conditions and land From the review, it confirms that harvested rainwater quality use land cover. Moreover, roofing materials and roofing conditions can be affected by the quality of collected rainwater itself and the play a major role in determining RRWH quality. Thus, it is impor- different storage media used to store rainwater. Thus, their effects tant to extract this information in order to analyze and evaluate the against determinants parameters should be evaluated. Future study environmental impact. should be conducted to analyze physico-chemical and microbio- Many studies utilized the remote sensing approaches for water logical samples using standard methods. These results can be quality parameters [(i.e. suspended sediments (turbidity), chloro- compared with the water quality guidelines of the World Health phyll, and temperature)] assessment. However, most of the studies Organization (WHO). This approach is purposely evaluated to focused on obtaining water quality parameter in surface waters ascertain its suitability for potable and non-potable uses. instead of RWH or RRWH. Meanwhile, GIS application is only used The use of very high resolution (VHR) optical remote sensing for identifying a potential area for RRWH. Yet, the GIS model for data and active systems, such as LiDAR or radar, is recommended in simulating RRWH quality was not developed and this application is extracting land use types, investigating the determining factors not yet used for RRWH quality evaluation. Moreover, the VHR sat- that contribute to harvested rainwater quality, and evaluating their ellite system can be explored further to obtain information on

Fig. 4. The framework of integrated use of remote sensing and GIS to assess the quality and suitability of RRWH. M. Norman et al. / International Soil and Water Conservation Research 7 (2019) 266e274 273 various characteristics of roofs. The determining factors of RRWH Chang, M., McBroom, M. W., & Beasley, R. S. (2004). Roofing as a source of nonpoint e quality were not combined to provide maps of harvested rainwater water pollution. Journal of Environmental Management, 73, 307 315. https://doi. org/10.1016/j.jenvman.2004.06.014. quality. These issues present an excellent foundation for further Cilia, C., Panigada, C., Rossini, M., Candiani, G., Pepe, M., & Colombo, R. (2015). research. Mapping of asbestos cement roofs and their weathering status using hyperspectral aaerial images. ISPRS International Journal of Geo-Information, 4, 928e941. https://doi.org/10.3390/ijgi4020928. Disclosure statement Clark, S. E., Steele, K. A., Spicher, J., Siu, C. Y. S., Lalor, M. M., Pitt, R., et al. (2008). Roofing materials' contributions to storm-water runoff pollution. Journal of e No potential conflict of interest was reported by the authors. Irrigation and Drainage Engineering, 134(5), 638 645. https://doi.org/10.1061/ (ASCE)0733-9437. Despins, C., Farahbakhsh, K., & Leidl, C. (2009). Assessment of rainwater quality Funding from rainwater harvesting systems in Ontario, Canada. Journal of Water Supply: Research & Technology - Aqua, 58(2), 117e134. https://doi.org/10.2166/aqua. fi 2009.013. The authors acknowledge the facilities and nancial supports Emna, S., Bernard, G., Gromaire, M., & Chebbo, G. (2014). A new method for provided by Universiti Putra Malaysia (UPM) via the research grant modelling roofing materials emissions on the city scale: Application for zinc in UPM/700-1/2/GPB/2017/9543100, Universiti Teknologi MARA the city of Creteil (France). Environmental Science and Pollution Research, 21(8), 5284e5296. (Perlis), Malaysia and Ministry of Education (MOE) Malaysia via the Ercin, A. E., & Hoekstra, A. Y. (2014). Water footprint scenarios for 2050: A global FRGS research grant 03-02-14-1529FR(Vot no: 5524613). analysis. Environment International, 64,71e82. https://doi.org/10.1016/j.envint. 2013.11.019. Evans, C. A., Coombes, P. J., & Dunstan, R. H. (2006). Wind, rain and bacteria: The Acknowledgments effect of weather on the microbial composition of roof-harvested rainwater. Water Research, 40(1), 37e44. https://doi.org/10.1016/j.watres.2005.10.034. The authors acknowledge the facilities and financial supports Farreny, R., Tito, M., Guisasola, A., Taya, C., Rieradevall, J., & Gabarrell, X. (2011). Roof selection for rainwater harvesting: quantity and quality assessments in Spain. provided by Universiti Putra Malaysia (UPM) via the research grant Water Research, 45,3245e3254. UPM/700-1/2/GPB/2017/9543100, Universiti Teknologi MARA Fichot, C. G., Downing, B. D., Bergamaschi, B. A., Windham-Myers, L., Marvin- (Perlis), Malaysia and Ministry of Education (MOE) Malaysia via the Dipasquale, M., Thompson, D. R., et al. (2016). High-resolution remote sensing FRGS research grant 03-02-14-1529FR(Vot no: 5524613). The of water quality in the san Francisco Bay-Delta Estuary. Environmental Science and Technology, 50(2), 573e583. https://doi.org/10.1021/acs.est.5b03518. comments from anonymous reviewers in improving this article are Fiumi, L. (2012). Surveying the roofs of Rome. Cultural Heritage, 13, 304e313. highly appreciated. Fonseca, C., Hidalgo, V., Díaz-Delgado, C., Vilchis-Frances, A., & Gallego, I. (2017). Design of optimal tank size for rainwater harvesting systems through use of a web application and geo-referenced rainfall patterns. Journal of Cleaner Pro- References duction, 145, 323e335. https://doi.org/10.1016/j.jclepro.2017.01.057. Forster, J. (1996a). Patterns of roof runoff contamination and their potential im- Abdelmalik, K. W. (2018). Role of statistical remote sensing for Inland water quality plications. Water Science and Technology, 33(6), 39e48. parameters prediction. Egyptian Journal of Remote Sensing and Space Science, Forster, J. (1996b). Patterns of roof runoff contamination and their potential im- 21(2), 193e200. https://doi.org/10.1016/j.ejrs.2016.12.002. plications on practice and regulation of treatment and local infiltration. Water Abrefa, K. N., Eric, K. F., Eric, A., & James, K. S. (2013). Using a GIS based model as an Science and Technology, 33(6), 39e48. https://doi.org/10.1016/0273-1223(96) decision support framework for identifying suitable rainwater harvesting sites. 00329-0. International Journal of Advanced Technology & Engineering Research, 3(4). Forster, J. (1998). The influence of location and season on the concentrations of Adham, A., Riksen, M., Ouessar, M., & Ritsema, C. (2016). Identification of suitable micro ions and organic trace pollutants in roof runoff. Water Science and sites for rainwater harvesting structures in arid and semi arid regions_A review. Technology, 38(10), 83e90. https://doi.org/10.1016/S0273. International Soil and water Conservation research. Retrieved from www.elsevier. de Fraiture, C., & Wichelns, D. (2010). Satisfying future water demands for agri- com/locate/iswcr. culture. Agricultural Water Management, 97(4), 502e511. https://doi.org/10. Ajim, S., Krishna, K., Amol, N., Rushikesh, T., Ashish, S., & Shailesh, K. (2017). Rain 1016/j.agwat.2009.08.008. water harvesting. International Research Journal of Engineering and Technology Frassy, F., Candiani, G., Rusmini, M., Maianti, P., Marchesi, A., Nodari, F. R., et al. (IRJET), 4(3), 1060e1062. https://www.irjet.net. (2014). Mapping asbestos-cement roofing with hyperspectral remote sensing Arshad, M., Shabbir, A., & Shakoor, A. (2013). Investigation of potential water har- over a large mountain region of the Italian western alps. Sensors, 14, vesting sites at potohar using modeling approach. 15900e15913. https://doi.org/10.3390/s140915900. Aysha, A., & Shoukat, A. (2015). Potentiality of rainwater harvesting for an urban Gaikwad, S. (2015). Application of remote sensing and GIS in rainwater harvesting: community in Bangladesh. Journal of Hydrology, (528), 84e93. A case from Goa, India/aplicacion de teledeteccion y SIG en la recoleccion de Baby, S., Arrowsmith, C., & Al-Ansari, N. (2019). Application of GIS for mapping agua de lluvia: Un caso de Goa, India. International Journal of Scientific Engi- rainwater-harvesting potential: Case study Wollert, Victoria. Engineering, 11(01), neering and Research, 6(1), 633e639. 14e21. https://doi.org/10.4236/eng.2019.111002. Gikas, G. D., & Tsihrintzis, V. A. (2012). Assessment of water quality of first-flush roof Bakir, M., & Xingnan, Z. (2008). GIS and remote sensing applications for rainwater runoff and harvested rainwater. Journal of Hydrology, 466-467,115e126. harvesting in the syrian desert (al-badia). 12th International Water Technology Gholizadeh, M., Melesse, A. M., & Reddi, L. (2016). Spaceborne and airborne sensors Conference,1e10. Retrieved from http://iwtc.info/wp-content/uploads/2010/09/ in water quality assessment. International Journal of Remote Sensing, 37(14), Gis-And-Remote-Sensing-Applications-For-Rainwater-Harvesting-In-The- 3143e3180. https://doi.org/10.1080/01431161.2016.1190477. Syrian-Desert-Al-Badia.pdf. Gwenzi, W., Dunjana, N., Pisa, C., Tauro, T., & Nyamadzawo, G. (2015a). Water quality Bassani, C., Cavalli, R. M., Cavalcante, F., Cuomo, V., Palombo, A., Pascucci, S., et al. and public health risks associated with roof rainwater harvesting systems for (2007). Deterioration status of asbestos-cement roofing sheets assessed by potable supply: Review and perspectives. Sustainability of Water Quality and analyzing hyperspectral data. Remote Sensing, 109,361e378. Retrieved from Ecology, 6,107e118. https://doi.org/10.1016/j.swaqe.2015.01.006. www.elsvier.com/locate/rse. Gwenzi, W., Dunjana, N., Pisa, C., Tauro, T., & Nyamadzawo, G. (2015b). Water Belmeziti, A., Coutard, O., & De Gouvello, B. (2014). How much drinking water can quality and public health risks associated with roof rainwater harvesting sys- be saved by using rainwater harvesting on a large urban area? Application to tems for potable supply: Review and perspectives. Sustainability of Water Paris agglomeration. Water Science and Technology, 70(11), 1782e1788. https:// Quality and Ecology, 6,107e118. https://doi.org/10.1016/j.swaqe.2015.01.006. doi.org/10.2166/wst.2014.269. Hari, D., Reddy, K., Vikas, K., Srinivas, N., & Vikas, G. (2018). Assessment of rainwater Boller, M. (1997). Tracking heavy metals reveals sustainability deficits of urban harvesting potential using GIS. IOP Conference Series: Materials Science and drainage systems. Water Science and Technology, 35(9), 77e87. Retrieved from Engineering, 330(1), 1e9. https://doi.org/10.1088/1757-899X/330/1/012119. http://www.china-sds.org/kcxfzbg/addinfomanage/lwwk/data/kcx1223.pdf. Huston, R., Chan, Y. C., Chapman, H. F., & Shaw, G. (2011). Source apportionment of Bris, A. L., & Robert, P. (2009). Classification of roof Materials for rainwater Pollution. heavy metals and ionic contaminants in rainwater tanks in a subtropical urban Hannover: ISPRS High-Resolution Earth Imaging for Geospatial Information. area in Australia. Water Research, 46(4), 1121e1132. Bris, A. Le, & Robert-Sainte, P. (2009). Classification of Roof Materials for Rainwater Jiries, A. G., Hussein, H. H., & Halaseh, Z. (2001). The quality of water and sediments Pollution Modelization. Hannover: ISPRS High-Resolution Earth Imaging for of street runoff in Amman, Jordan. Hydrological Processes, 15(5), 815e824. Geospatial Information. https://doi.org/10.1002/hyp.186. Bris, A. L., Chehata, N., Briottet, X., & Paparoditis, N. (2014). Spectral band selection for Kumar, P., & Viswanadh, G. (2012). Developing a GIS model and identification of urban material classification using hyperspectral libraries. rainwater harvesting structures using geomatic approach - a case study of Campisano, A., & Lupia, F. (2017). A dimensionless approach for the urban-scale Hyderabad. IJWREM, 3(1), 1e8. evaluation of domestic rainwater harvesting systems for toilet flushing and Lee, J., Bak, G., & Han, M. (2012). Quality of roof-harvested rainwater - comparison of garden irrigation. Urban Water Journal, 14(9), 883e891. https://doi.org/10.1080/ different roofing materials. Environmental Pollution, 162, 422e429. https://doi. 1573062X.2017.1279192. org/10.1016/j.envpol.2011.12.005. 274 M. Norman et al. / International Soil and Water Conservation Research 7 (2019) 266e274

Lim, J., & Choi, M. (2015). Assessment of water quality based on Landsat 8 opera- Ramadas, M., & Samantaray, A. (2017). Applications of remote sensing and GIS in tional land imager associated with human activities in Korea. Environmental water quality monitoring and remediation. A State-of-the-Art Review, 225e246. Monitoring and Assessment, 187(6), 1e17. https://doi.org/10.1007/s10661-015- https://doi.org/10.1007/978-981-10-7551-3_13. 4616-1. Rostad, N., Foti, R., & Montalto, F. (2016). Resources , Conservation and Recycling Lim, H. S., MatJafri, M. Z., Abdullah, K., Alias, A. N., & Mohd Saleh, N. (2008). Water Harvesting rooftop runoff to flush toilets : Drawing conclusions from four major quality mapping using remote sensing technique in Penang Straits, Malaysia. U . S . cities. Resources, Conservation and Recycling, 108,97e106. https://doi.org/ Remote Sensing of the Ocean, Sea Ice, and Large Water Regions, 7105, 71050A. 10.1016/j.resconrec.2016.01.009. https://doi.org/10.1117/12.800573. Rouvalis, A., Karadima, C., Zioris, I., Sakkas, V., & Albanis, T. (2009). Determination of Lupia, F., Baiocchi, V., Lelo, K., & Pulighe, G. (2017). Exploring rooftop rainwater pesticides and toxic potency of rainwater samples in western Greece. Exotox- Harvesting Potential for Food Production in Urban Areas, 1e17. https://doi.org/10. ilogy and Environmental Safety, 72(3), 828e833. 3390/agriculture7060046. Sample, D. J., & Liu, J. (2014). Optimizing rainwater harvesting systems for the dual Lye, D. J. (2009). Rooftop runoff as a source of contamination: A review. The Science purposes of water supply and runoff capture. Journal of Cleaner Production, 75, of the Total Environment, 407(21), 5429e5434. https://doi.org/10.1016/j. 174e194. https://doi.org/10.1016/j.jclepro.2014.03.075. scitotenv.2009.07.011. Sazakli, E., Alexopoulos, A., & Leotsinidis, M. (2007). Rainwater harvesting, quality Matos, C., Santos, C., Pereira, S., Bentes, I., & Imteaz, M. (2014). Rainwater storage assessment and utilization in Kefalonia Island, Greece. Water Research, 41(9). tank sizing: Case study of a commercial building. International Journal of Sus- Simmons, G., Hope, V., Lewis, G., Whitmore, J., & Gao, W. (2001). Contamination of tainable Built Environment, 2(2), 109e118. https://doi.org/10.1016/j.ijsbe.2014.04. potable roof-collected rainwater in Auckland, New Zealand. Water Research, 004. 35(6), 1518e1524. https://doi.org/10.1016/S0043-1354(00)00420-6. Md Lani, N., Yusop, Z., & Syafiuddin, A. (2018). A review of rainwater harvesting in Singh, J. P. (2014). Selection of Suitable Sites for water Harvesting Structures in Malaysia : Prospects and challenges. Water, 10(506), 1e21. https://doi.org/10. Soankhad Watershed , Punjab using remote Sensing and Geographical Information 3390/w10040506. System ( RS & GIS ) Approach d A Case Study (January) https://doi.org/10.1007/ Meera, V., & Ahammed, M. (2006). Water quality of rooftop rainwater harvesting s12524-009-0009-7. systems: A review. Aqua, 257e268. Retrieved from http://www.iwaponline.com/ Taherzadeh, A., & Shafri, H. Z. M. (2013). Development of a generic model for the jws/055/jws0550257.htm. detection of roof materials based on an object-based approach using Meera, V., & Ahammed, M. M. (2011). Factors affecting the quality of roof-harvested WorldView-2 satellite imagery. Advances in Remote Sensing, 2, 312e321. rainwater. 12th International Conference on Enviromental Science and Technology. Tiwari, K., Goyal, R., & Sarkar, A. (2018). GIS-based methodology for identification of Mekonnen, M. M., & Hoekstra, A. (2016). Four billion people facing severe water suitable locations for rainwater harvesting structures. Water Resources Man- scarcity. Science Advances, 2(2). https://doi.org/10.1126/sciadv.1500323. agement, 1811e1825. https://doi.org/10.1007/s11269-018-1905-9. Mendez, C. B., Klenzendorf, J. B., Afshar, B. R., Simmons, M. T., Barrett, M. E., Tobiszewski, M., Polkowska, Z., Konieczka, P., & Namiesnik, J. (2010). Roofing ma- Kinney, K. A., et al. (2011). The effect of roofing material on the quality of terials as pollution emitters -concentration changes during runoff. Journal of harvested rainwater. Water Research, 45, 2049e2059. https://doi.org/10.1016/j. Environmental Studies, 19(5), 1019e1028. watres.2010.12.015. Ugai, T. (2016). Evaluation of sustainable roof from various aspects and benefits of Morales-Pinzon, T., Rieradevall, J., Gasol, C. M., & Gabarrell, X. (2015). Modelling for agriculture roofing in urban core. Procedia - Social and Behavioral Sciences, 216, economic cost and environmental analysis of rainwater harvesting systems. 850e860. https://doi.org/10.1016/j.sbspro.2015.12.082. Journal of Cleaner Production, 87(1), 613e626. https://doi.org/10.1016/j.jclepro. UN-HABITAT. (2003). Rainwater harvesting and utilization. Blue drop series, united 2014.09.021. Nations human settlements programme (UN-HABITAT) book 2. Beneficiaries & Myint, S. W., Gober, P., Brazel, A., Grossman, C. S., & Weng, Q. (2011). Per-pixel vs. Capacity Builders. Retrieved from http://www.cabdirect.org/abstracts/ object-based classification of urban land cover extraction using high spatial 20033157591.html. resolution imagery. Remote Sensing of Environment, 115(5), 1145e1161. https:// Usali, N., & Hasmadi, M. (2010). Use of remote sensing and GIS in monitoring water doi.org/10.1016/j.rse.2010.12.017. quality. Sustainable Development, 3(3), 228e238. Oki, T., Agata, Y., Kanae, S., Saruhashi, T., Yang, D., & Musiake, K. (2001). Global Uygur, N., Karaca, F., & Alagha, O. (2010). Prediction of sources of metal pollution in assessment of current water resources using total runoff integrating pathways. rainwater in Istanbul, Turkey using factor analysis and long-range transport Hydrological Sciences Journal, 46(6), 983e995. https://doi.org/10.1080/ models. Atmospheric Research, 95,55e64. 02626660109492890. Wallinder, I.,O., Bahar, B., Leygraf, C., & Tidblad, J. (2007). Modelling and mapping of Quek, U., & Forster, J. (1993). Trace metals in roof runoff. Water. Air and Soil Pollution, copper runoff for Europe. Journal of Environmental Monitoring: JEM, 9,66e73. 68, 373e389. Retrieved from http://site.ebrary.com/lib/rit/reader.action? https://doi.org/10.1039/b612041e. docID¼10667396&ppg¼17. Wolfe, S., & Brooks, D. B. (2003). Water scarcity: An alternative view and its im- Radaideh, J., Al-Zboon, K., Al-Harahsheh, A., & Al-Adamat, R. (2009). Quality plications for policy and capacity building. Natural Resources Forum, 27(2), sssessment of harvested rainwater for domestic uses. Jordan Journal of Earth and 99e107. https://doi.org/10.1111/1477-8947.00045. Environmental Sciences, 2(1), 26e31. Retrieved from http://jjees.hu.edu.jo/files/ Yaziz, M. I., Gunting, H., Sapari, N., & Ghazali, A. W. (1989). Variations in rainwater v2n1/4.pdf. quality from roof catchments. Water Research, 23(6), 761e765. https://doi.org/ Radzali, N. A. W. M., Shafri, H. Z. M., Norman, M., & Saufi, S. (2018). Roofing 10.1016/0043-1354(89)90211-X. assessment for rooftop rainwater harvesting adoption using remote sensing Zhang, Q., Wang, X., Hou, P., Wan, W., Li, R., Ren, Y., et al. (2014). Quality and sea- and GIS approach. International Archives of the Photogrammetry, Remote Sensing sonal variation of rainwater harvested from concrete, asphalt, ceramic tile and and Spatial Information Sciences - ISPRS Archives, 42(4/W9), 129e132. https://doi. green roofs in Chongqing, China. `Journal of Environmental Management, 132, org/10.5194/isprs-archives-XLII-4-W9-129-2018. 178e187. https://doi.org/10.1016/j.jenvman.2013.11.009. International Soil and Water Conservation Research 7 (2019) 275e285

Contents lists available at ScienceDirect

International Soil and Water Conservation Research

journal homepage: www.elsevier.com/locate/iswcr

Original Research Article The effect of hydrogel particle size on water retention properties and availability under water stress

Ahmad M. Abdallah

Department of Natural Resources and Agricultural Engineering, Faculty of Agriculture, Damanhour University, 59 Damanhour, Egypt article info abstract

Article history: The use of superabsorbent polymers or hydrogels could increase the water holding capacity (WHC) of Received 21 June 2018 sandy soil and reduce water loss by deep percolation. However, hydrogels' retained water availability to Received in revised form plants might be overestimated without taking into consideration the hydrogel particles size. Therefore, 2 May 2019 the ultimate objective of this study was to address the impacts of hydrogel particles size on hydrogel's Accepted 13 May 2019 retained water availability (plant available water, PAW), daily water consumption (DWC) and survival of Available online 16 May 2019 Guava seedlings subjected to drought. Moreover, some soil physical properties, i.e., WHC, water retention properties, and hydraulic conductivity (Ksat) were investigated. Hydrogel (WaterSorb, “WS”) application, Keywords: fi Water holding capacity particularly the WS of small particles, signi cantly reduced Ksat, and increased WHC and PAW. Therefore, fi e e Swelling rate seedlings grown in soil amended with WS ne (0.8 1.0 mm), WS medium (1.0 2.0 mm) and WS large (2 Water consumption e4 mm) survived for 27.0 ± 1.3, 24.0 ± 1.1 and 17.0 ± 0.7 days, respectively, compared to 13.0 ± 1.0 days Available water capacity for the control. The water stored in the WS of large particles was less readily available for plant roots. Seedlings survival Interestingly, hydrogels, had no effect on the DWC of the seedlings. Utilizing hydrogels as a soil amendment increases WHC, PAW, growth and survival of Guava seedlings, while the effect was less pronounced for the large hydrogel particles which had lower specific surface area and swelling rate. © 2019 International Research and Training Center on Erosion and Sedimentation and China Water and Power Press. Production and Hosting by Elsevier B.V. This is an open access article under the CC BY-NC- ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction water capacity (AWC, the amount of water that a plant can uptake) of light soil is vital to ensure efficient water use and sustainability. The scarcity of rain and irrigation water is a critical problem in It is widely accepted that WHC and AWC of the soil are positively arid and semiarid regions, where water is the determining factor correlated to the soil organic matter content. However, Minasny for crop productivity and the cultivated area (Han, Benner, & Flores, and McBratney (2018) reported that increasing WHC and AWC by 2018). Moreover, sandy soils are characterized by low water hold- organic matter addition might be overestimated or uncertain. ing capacity (WHC) and excessive drainage of rain and irrigation Minasny and McBratney (2018) conducted a meta-analysis using 60 water below the root zone, resulting in inefficient water and fer- published datasets and analyzed about 50000 measurements to tilizers use (Fan et al., 2005; Wei & Durian, 2014). These problems study the impact of organic matter addition to soil on AWC. They are exacerbated when cultivating crops with shallow roots and found that organic matter had a very slight effect on AWC. Inter- with high-water requirements (Bhardwaj, Shainberg, Goldstein, estingly, the significant increase in AWC might be only obtained Warrington, & JLevy, 2007; Fan et al., 2005). Therefore, increasing when large rates of organic matter were applied, which might not WHC (the total amount of water that a soil can hold) and available be visible under commercial production conditions (Ankenbauer & Loheide, 2017; Minasny & McBratney, 2018). Therefore, the development of non-traditional practices to improve WHC and AWC of light soils is gaining more attention (Belaqziz et al., 2013; Abbreviations: AWC, available water capacity; CLP, cross linked polyacrylamides; Guilherme et al., 2015). It is well established that superabsorbent DWC, daily water consumption; DW, distilled water; Ksat, saturated hydraulic polymers (SAPs) or hydrogels such as cross-linked polyacrylamides conductivity; PAW, plant available water; RWC, relative water content; SAPs, & superabsorbent polymers; SC, swelling capacity; SR, swelling rate; WHC, water (CLP) increase the WHC of soil (Li, He, Hughes, Liu, Zheng, 2014; holding capacity; WS, WaterSorb; WSF, WaterSorb fine; WSM, WaterSorb medium; Xu et al., 2015; Yang, Yang, Chen, Guo, & Li, 2014). The CLP are WSL, WaterSorb large. polymers with high molecular weight and high negative charge E-mail address: [email protected]. https://doi.org/10.1016/j.iswcr.2019.05.001 2095-6339/© 2019 International Research and Training Center on Erosion and Sedimentation and China Water and Power Press. Production and Hosting by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). 276 A.M. Abdallah / International Soil and Water Conservation Research 7 (2019) 275e285 that absorb a significant amount of water, up to 400e2000 g water Table 1 g 1 CLP, (Guilherme et al., 2015; Yang et al., 2014). Main physical and chemical properties of a representative soil samples in the experimental site (0e30 cm depth). When the hydrogel-amended soil dries, the absorbed water by hydrogel particles (about 90e95% of the retained water) could Soil property The value gradually be released to the plants (Chen, Zhang, Luo, & Fang, 2004; Particle size distribution Han et al., 2013; Hüttermann, Zommorodi, & Reise, 1999). In this Sand (%) 90.6 way, hydrogels application might increase not only WHC but also Silt (%) 5.65 AWC of sandy soils and reduce water loss by deep percolation and Clay (%) 4.23 Water retention points fi & fertilizer leaching (Banedjscha e Durner, 2015; Montesano, Saturation (%) 38.8 Parente, Santamaria, Sannino, & Serio, 2015; Yu et al., 2017). Field capacity (%) 15.33 Therefore, hydrogels application might enable longer intervals PWP (%) 7.30 Available water (mm m 1) 83.33 between irrigations, provides a buffer against water stress and Bulk Density (Mg m 3) 1.63 reduce the risk of temporary drought stress (Beniwal, Langenfeld- Total porosity (%) 38.8 Heyser, & Polle, 2010; Han et al., 2013). Additionally, hydrogles Saturated Hydraulic Conductivity (mm h 1) 106.33 application found to enhance plant growth at the early stages of Ece (dS m 1) 2.2 production (El-Asmar, Jaafar, Bashour, Farran, & Saoud, 2017; Yang Ph 8.1 1 et al., 2014; Zhang, Liu, Li, & Wang, 2006), reduce the failure of CEC (c mol kg ) 4.21 OM (%) 0.07 certain crops at establishment or reduce yield losses of some crops CaCO3 (%) 4.21 at specific growth stages (Orikiriza et al., 2013; Wu & Liu, 2008) and PWP, OM and CEC presents permanent wilting point, organic matter and cation aides seed germination, emergence and increases the seedling exchange capacity, respectively. Electrical conductivity (Ece) measured in soil paste survival (Beniwal et al., 2010; Orikiriza et al., 2013; Wei & Durian, extract. Soil reaction (PH) measured in 1:2.5 soil suspension. Organic matter 2014; Yang et al., 2014). The effect of SAPs was extended to other (calculated by multiplying the organic carbon content by a conversion factor of soil physical properties. It has been reported that hydrogels appli- 1.724). cation, reduces soil bulk density, saturated hydraulic conductivity, and penetration resistance, while increases soil aggregation and porosity (Agaba et al., 2010; Andry et al., 2009; Li et al., 2014; Xu et al., 2015). sieved soil. WaterSorb (WS), a potassium-based cross-linked Besides the low cost, availability, and safety, the main required polyacrylamide, (WaterSorb-227 S Church Ave, Fayetteville, AR. properties of hydrogels for agricultural purposes are the high USA) with three different particle size ranges (0.8e1.0, 1.0e2.0 and swelling rate (SR), swelling capacity (SC) and re-swellability. These 2.0e4.0 mm) each was mixed with the air-dried sandy soil at 0.3% properties are mainly dependent on: (I) polymer's network struc- (w/w) and control (soil without WS). Hereafter, the sizes 0.8e1.0, ture, (ii) hydrogel concentration, (iii), soil salinity and water quality 1.0e2.0 and 2.0e4.0 mm will be referred to WSF, WSM and WSL, (Banedjschafie & Durner, 2015; Shahid, Qidwai, Anwar, Ullah, & respectively. The specific surface area (SSA) of the WS particles, Rashid, 2012; Yu et al., 2017), (iv) application method (El-Asmar assuming a spherical shape, was calculated using the following et al., 2017; Li et al., 2014; Wei & Durian, 2014), (v) crop species equation (Gao, Li, Song, & Wang, 2000): (Bai, Song, & Zhang, 2013; Langaroodi, Ashouri, Dorodian, & Azarpour, 2013) and (vi) soil texture (Agaba et al., 2010; Bai et al., 6 SSA ¼ *10 6 (1) 2013; Montesano et al., 2015; Yu et al., 2017). dr It is well known that Hydrogels or SAPs, e.g., CLP efficiently increase WHC of the soil, particularly the light soils. However, Where; SSA is the specific surface area of WS particles (m2 g 1),dis hydrogels' retained water availability to plants' root system might the diameter of the WS particle (m) and r is the density of the WS be overestimated without taking into consideration the Hydrogel particles (0.7 Mg m 3). particles size. Yet, little information is known about the impact of hydrogel particles size on soil WHC, PAW and the ratio between the two (PAW/WHC). This ratio indicates the percentage of available water to plants in relation to the retained water in the hydrogel. 2.2. Measurements of the swelling capacity of the hydrogels and Therefore, the main objective of this study was to address the im- their response to salinity pacts of hydrogel particle size on hydrogel's retained water availability to plant, daily water consumption (DWC) and survival of The swelling capacity (SC, the water absorption capacity) of the Guava seedlings under drought conditions. A further objective was WS was determined with the teabag method as described by to investigate the effect of hydrogel particles size on some soil (Buchholz, 1998). For measuring SC, one g of each WS particle size physical properties i.e., WHC, AWC and saturated hydraulic was packed in a teabag and submerged into distilled water (DW) conductivity. until the equilibrium swelling was reached (z80 min). The mass of the swollen absorbent was then measured after the removal of the 2. Materials and methods extra water using a filter paper. Similarly, five levels of saline water (0.5, 1, 2, 3 and 4 dS m 1) were used for measuring SC as well. The 2.1. Soils and hydrogels used in the experiments equilibrium water absorbency of swollen samples was calculated using the following equation (Spagnol, Rodrigues, & Neto et al., Experiments were conducted using samples of sandy soil (90.6, 2012, Spagnol, Rodrigues, & Pereira et al., 2012). 5.65 and 4.23% of sand, silt and clay, respectively). The soil was characterized by low organic matter and low water holding ca- Ws Wd SC ¼ (2) pacity. The main physical and chemical properties of the used soil Wd are presented in Table (1). The soil was air-dried, ground, and passed through a 2-mm sieve in the case of laboratory experiments, Where: SC is the swelling capacity (g water g 1 WS), Ws and Wd while in the pots experiments, the pots were packed using un- are mass of the swollen and dry samples (g), respectively. A.M. Abdallah / International Soil and Water Conservation Research 7 (2019) 275e285 277

2.3. Measurements of swelling rate of different sized hydrogel randomized design in the greenhouse under controlled conditions particles (ambient temperature of 28.0 ± 2.0 C and a relative humidity of 70.0 ± 5.0%). The pots were weekly rotated within the greenhouse Swelling rate (SR, the unit of water absorbed for each time unit) to ensure uniform distribution of light and air to the seedlings. For was measured by measuring the amount of water absorbed at the first eight weeks, the seedlings were irrigated three times a various times using DW (Isik & Kis, 2004). The swelling rate of the week with 2 L of tap water (0.5 dS m 1) per pot, to ensure that the WSF, WSM and WSL was calculated at 0, 5, 10, 20, 30, 40, 50, 60, 70 seedlings were successfully established (Hüttermann et al., 1999). and 80 min. The SR (g sec 1) was calculated by dividing the amount After that, irrigation was terminated. Immediately, after ceasing of water absorbed (g) by the submerging time (sec). The initial and irrigation, initial water content was calculated by weighting the final swelling (swelling equilibrium) was estimated after 5 and pots (the changes in plant weights were ignored). The pot weights 80 min, respectively. were recorded daily to obtain the DWC (Agaba et al., 2010; Wang & Boogher, 1987). This process was continued until seedlings became 2.4. Measurements of re-swelling capability of different sized wilt. The signs of wilting were the brown leaves and brittle hydrogel particles branches (Agaba et al., 2010; Yang et al., 2014). The total time to reach wilting stage was considered as the survival time. The plant One gram of each WS particle size (WSF, WSM and WSL) was available water (PAW) was determined by obtaining the difference immersed in a DW for 80 min to reach the swelling equilibrium. between initial soil water content after suspending irrigation and The swollen WS were determined and then filtered out with a mesh the final soil water content at the individual plant death (Agaba screen and dehydrated at 80 C. After complete drying, the recov- et al., 2010; Hüttermann et al., 1999). Shoot heights were ered WS particles were rewetted and dehydrated twice consisting measured immediately after irrigation terminating and at the in- three swelling/dehydration cycles (Shahid et al., 2012). Similarly, dividual plant death. The non-survived seedlings were harvested the WS re-swelling capacity was measured at varied water salin- and the whole seedling was dried at 70 C until a constant weight ities (2.0 and 4 dS m 1). was reached and a total dry weight was determined. The relative water content (RWC) of leaves was measured before ceasing irri- 2.5. Measurement of the swelling capacity in the water and soil gation and after ten days of terminating irrigation. Relative water mixture content of leaves was measured as described by (Yamasaki & Dillenburg, 1999). Fully expanded leaves were detached, and leaf fresh weight (LFW) was measured. The sampled leaves were floated To compare the SC of WS, of each particle size, in the free water and in the soil mixture; soil samples (100 g), WS samples (1 g) of on DW at 4 C in a dark chamber for 24 h, and leaf turgid weight each particle size, and WS-soil mixtures (1 g WS 100 1 g soil) were (LTW) was measured. Leaf dry weight (LDW) was determined after transported into permeable nylon bags (Isik & Kis, 2004). Three drying at 70 C till they reached a constant weight. RWC was bags for each treatment were immersed in a container filled with calculated using the following equation: three L of tap water (0.5 dS m 1) for 15 min (Yu et al., 2011). The SC ðLFWe LDWÞ of the WS of each particle size was measured in the free water, RWCð%Þ¼ *100 (4) moreover, the SC in the soil mixture was measured as follow. ðTW eLDWÞ

D b SC ¼ (3) C

Where; D is the mass of the wet WS-soil mixture (g), b is the mass 2.8. Measurement of soil-moisture characteristics curves of dry soil (g) with the water absorbed by it, and C is the dry weight of WS in the WS-soil mixture (g). Immediately before harvesting the seedlings, undisturbed soil cores (65 mm diameter and 30 mm height) were sampled at 10 cm 2.6. Measurement of saturated hydraulic conductivity (Ksat) depth (three replicates for each treatment) for measuring soil moisture characteristics curves. Samples were saturated overnight 1 Saturated hydraulic conductivity was measured according to with tap water (0.5 dS m ). The soil moisture characteristic curve (Moutier, Shainberg, & Levy, 2000). A plastic cylinder (20 cm height was measured for the soil using pressure plate extractors (Dane & with a diameter of 7 cm) with a fine metal screen at the bottom Hopmans, 2002), in a matric potential of 0, 33, 100, 300, and were used to determine the hydraulic conductivity of soil amended 1500 kPa. The desired tension was applied until outflow ceased and with 0.3% WS of the three sized particles in addition to the control, the soil water was considered to be in equilibrium with the applied using three replicates for each treatment. The amended soil and pressure. Water content at each pressure was calculated from the control soil was packed in the cylinder at a bulk density of volume of outflow between pressure steps, the final water content, 1.6 Mg m 3 and covered with a filter paper to reduce soil surface and the weight of oven-dried soil. Furthermore, the AWC (mm/m) disturbance. Soil cylinders were saturated by capillary rise using was calculated by the subtraction of water content at Field capacity tap water (0.5 dS m 1). After the saturation process from the bot- (FC) and permanent wilting point (PWP). tom, the flow direction was reversed, and a constant pressure head of 4 cm was applied in the top of soil cylinder. The leachate was collected, and hydraulic conductivity was calculated. 2.9. Statistical analysis

2.7. Guava seedlings survival experiment Analysis of Variance (ANOVA) of the obtained data was performed using a completely randomized design in Glmmix proced- A one-year-old Guava seedlings (71 ± 1.12 cm-high) (Psidium ure in SAS 9.4 (SAS, Inc., Cary, NC. USA). The statistical differences guajava L.) were transplanted to 30 L pots with either control soil or among treatments' means across traits were conducted using amended soil with WSF, WSM or WSL at 0.3% (w/w) and replicated Tukey's test. Mean differences were considered signiﬁcant at ﬁve times. The seedlings were then placed in a completely p ¼ 0.05. 278 A.M. Abdallah / International Soil and Water Conservation Research 7 (2019) 275e285

3. Results abruptly with salinity till 0.5 dS m 1. When water salinity increased only to 0.5 dS m 1, SC was markedly reduced by 37.15, 43.17 and 3.1. Swelling rate of tested hydrogels 45.5% for WSF, WSM and WSL. The SC decreased gradually and significantly with further increase in water salinity behind The results of swelling rate (Fig. (1a)) showed a non-linear 0.5 dS m 1. However, even under severe water salinity (4 dS m 1), negative relationship between the average particle size and the testes WS of different particle size were still functional; as a swelling rate for all treatment. The absorption behavior followed significant amount of water was absorbed (81.33 ± 0.57, three stages; the first stage (0e5 min) showed a rapid increase in 69.33 ± 0.57 and 66.6 ± 1.15 g water g 1 WS for WSF, WSM and the amount of absorbed water with time. Meanwhile, the second WSL). The percentages of absorbed water at 4 dS/m to that in the stage (5e40 min) showed moderately absorption. The last stage DW were 33.4, 28.6 and 33.3% for WSF, WSM and WSL, respectively. (40e80 min) showed almost steady-state absorption in which, no The re-swelling capacity of the WSF, WSM and WSL was tested increase in water absorbancy over time was observed. Since the for three swelling/dehydration cycles in varied water salinities (0.0, absorbance of water is a function of time, two swelling conditions 2.0 and 4.0 dS m 1). The results of re-swelling capacity are shown were described at 5 min (initial swelling) and 80 min (final in Figure (3). The three WS particles size found to have the ability to swelling). The two swelling conditions were shown in Figure (1b) re-swell under different water salinity. However, the SC was as a function of WS particle sizes. As regard to the initial swelling, significantly reduced in the second and the third swelling/dehy- the amount of water absorbed was significantly (P 0.05) affected dration cycles at any given water salinity. For example, when DW by the hydrogel particles size. The amount of water absorbed after was used, in the first swelling/dehydration cycle, maximum SC was 5 min was 195.0 ± 0.47, 172.0 ± 0.81 and 125.0 ± 1.24 g g 1 (g water 244.3 ± 3.05, 242.0 ± 2.64 and 238.6 ± 1.52 g g 1 for WSF, WSM and g 1 WS) for WSF, WSM and WSL, respectively, corresponding to a swelling rate of 0.65 ± 0.002, 0.57 ± 0.003 and 0.42 ± 0.005 g s 1. The amount of water absorbed by the WS of different particle size at final equilibrium swelling time (after 80 min) was not significantly affected by hydrogel particles size; in which the amount of water absorbed was 244.3 ± 3.05, 242.0 ± 2.64 and 238.6 ± 1.52 g g 1 for WSF, WSM and WSL, respectively. The corresponding swelling rate values were 0.05 ± 0.001, 0.050 ± 0.0004 and 0.050 ± 0.0002 g s 1 for WSF, WSM and WSL, respectively.

3.2. The effect of salinity on swelling and re-swelling capacity of tested hydrogels

The results of SC of varying WS particle size under several water salinity were presented in Fig. 2. At all water salinity levels, the SC of the WS decreased in the following order: WSF > WSM > WSL. The different grained-WS were profoundly affected by water salinity; where a non-linear negative relationship was observed between the amount of water absorbed (g water g 1 WS) and water 1 salinity (dS m ). Maximum SC was observed in the DW (0.0 dS/m). Fig. 2. Swelling capacity (water absorption capacity, g water g 1 WaterSorb) for Using DW, the WSF, WSM and WSL absorbed 244.3 ± 3.05, various sized WaterSorb particles (WSF, WSM and WSL) under several water salinity. ± ± 1 242.0 2.64 and 238.6 1.52 g water g WS, respectively. The SC WSF, WSM and WSL denote WaterSorb ﬁne (0.8e1.0 mm), WaterSorb medium decreases were in distinct two stages, in which SC decreased (1.0e2.0 mm) and WaterSorb large (2.0e4.0 mm), respectively.

Fig. 1. Swelling capacity (Water absorption capacity, g water g 1 WaterSorb) as a function of time and the average particle size of the WaterSorb (WSF, WSM and WSL) (Panel a). Initial and Final swelling of various sized WaterSorb particles (Panel b). The initial swelling rate analysis showed an inverse linear relationship between average particle size and initial swelling rate. No significant correlation between particle size and final equilibrium swelling ratio was observed. WSF, WSM and WSL denote WaterSorb fine (0.8e1.0 mm), WaterSorb medium (1.0e2.0 mm) and WaterSorb large (2.0e4.0 mm), respectively. A.M. Abdallah / International Soil and Water Conservation Research 7 (2019) 275e285 279

Fig. 3. Swelling capacity (Water absorption capacity, g water g 1 WaterSorb) of the different WaterSorb particle size (WSF, WSM and WSL) during ﬁrst, second and third swelling/ dehydration cycles under several water salinity (0.0, 2.0 dS m 1 and 4.0 dS m 1). WSF, WSM and WSL denote WaterSorb ﬁne (0.8e1.0 mm), WaterSorb medium (1.0e2.0 mm) and WaterSorb large (2.0e4.0 mm), respectively. Data are means ± standard deviation.

WSL, respectively. Meanwhile, the SC was reduced in the second swelling/dehydration cycle to be 227.6 ± 2.52, 216 ± 5.29 and 207 ± 1.7 g water g 1 WS for WSF, WSM and WSL, respectively. Under salinity conditions, signiﬁcantly reduction in WS re- swellability was observed, compared to the DW treatment, with no signiﬁcant differences between the 2 dS m 1 and 4 dS m 1 treatments. Whereas, when DW was used, SC in the third swelling/ dehydration cycle was reduced by 10.1, 13.36 and 10.04% for WSF, WSM and WSL, respectively. Meanwhile, when a saline water of 2dSm 1 was used, the corresponding reduction values were 15.46, 22.64 and 22.02% for WSF, WSM and WSL, respectively. Similarly, the corresponding reduction in SC in the third swelling/dehydration cycle, in the 4 dS m 1 treatments were 13.11, 21.63 and 18.8 for WSF, WSM and WSL, respectively.

3.3. Hydrogels swelling in the free water system and soil mixture

The amount of water absorbed by the different WS particle size Fig. 4. The quantity of water retained by various sized WaterSorb particles (WSF, WSM in the free water system and the WS-soil mixtures were presented and WSL) in the free water system and in the soil mixture. Different letters on the top in Figure. (4). One g of WSF, WSM and WSL were tested for SC in a of bars indicate significant differences between treatments with p-value 0.05 (Tukey's test). Significance letters apply only within each parameter. WSF, WSM and free water system and compared with their SC in soil mixture (one fi e e 1 WSL denote WaterSorb ne (0.8 1.0 mm), WaterSorb medium (1.0 2.0 mm) and gofWS100 g sandy soil). In both systems, the SC was inversely WaterSorb large (2.0e4.0 mm), respectively. Data are means ± standard deviation. related to the average particle size and significantly decreased in the following order: WSF > WSM > WSL. At any particle size, the WS-soil mixtures retained significantly less water (112.3 ± 2.5, 7.16 ± 0.29, 16.0 ± 1.00, 17.37 ± 0.57 and 20.33 ± 0.58%. In spite of 96.3 ± 1.15 and 84.3 ± 1.52 g g 1 for WSF, WSM and WSL, respec- the significant increase in soil water content at PWP, the AWC was tively) compared to the free water system (153.3 ± 1.53, increased by 5.04, 4.56 and 3.61 folds for WSF, WSM and WSL, 134.6 ± 1.53 and 125.6 ± 4.0 g g 1 for WSF, WSM and WSL, respectively, relative to the control (Fig. 7). The significant (P 0.05) respectively). increase in PWP indicates that a part of the water that retained by the WS particles might not be available to the seedlings when 3.4. Soil-moisture characteristics curves and available water needed. A portion of 15.5, 18.7 and 26.3% of hydrogels' retained capacity water for WSF, WSM and WSL, respectively, was apparently not available to seedlings. Soil-moisture characteristics curves that obtained for WS-soil mixtures presented in (Fig. 5). The results indicate that the water 3.5. Saturated hydraulic conductivity retention of the amended soil depends on WS particle size. Mixing the WS with soil significantly increased the retained water at all Results presented in Figure (6) indicate the values of Ksat for the studied suctions. The WSF possessed the highest retained water WS-soil mixtures and the control. The Ksat values decreased compared with the WSM and WSF. Considering the field capacity significantly by mixing 0.3% of the WS, for all particles sizes, in (FC) at soil suction of 33 kPa, the Fc values were 15.3 ± 1.1, comparison to control. The Ksat was reduced significantly in the 58.0 ± 1.2, 55.4 ± 0.56 and 50.4 ± 1.5% for control, WSF, WSM and following order; WSL > WSM > WSF; whereas, the Ksat was WSL, respectively. Similarly, if permanent wilting point (PWP) is significantly reduced by 68.8, 47.4 and 38.9% for WSF, WSM and estimated at 1500 kPa, the corresponding PWP values were WSL, respectively. 280 A.M. Abdallah / International Soil and Water Conservation Research 7 (2019) 275e285

Fig. 5. Soil-moisture characteristics curves of the soil amended with different Fig. 7. Effects of WaterSorb particle size (WSF, WSM and WSL) on water content at WaterSorb Particle size (WSF, WSM and WSL). WSF, WSM and WSL denote WaterSorb field capacity (FA, %), permanent wilting point (PWP, %) and available water capacity fine (0.8e1.0 mm), WaterSorb medium (1.0e2.0 mm) and WaterSorb large (AWC, mm/m). Different letters on the top of the bars indicate significant differences (2.0e4.0 mm), respectively. between treatments with p-value 0.05 (Tukey's test). Significance letters apply only within each parameter. WSF, WSM and WSL denote WaterSorb fine (0.8e1.0 mm), WaterSorb medium (1.0e2.0 mm) and WaterSorb large (2.0e4.0 mm), respectively. Data are means ± standard deviation.

case of the control to 7.09 ± 0.1, 6.23 ± 0.05 and 4.02 ± 0.05 kg/pot for WSF, WSM and WSL, respectively, which correspond to 38.2 ± 0.79, 78.3 ± 1.1, 68.6 ± 0.56 and 44.3 ± 0.56 mm for control, WSF, WSM and WSL, respectively. The ratio between PAW and initial water content (total water storage in the soil after stopping irrigation), was significantly lower in the WS-amended soil, particularly in the WSL (Table 2). Therefore, a significant portion of water retained in the WS particles was not readily available to the seedlings. In this study, 80.8 ± 1.1% of the water present in the control soil after ceasing irrigation, was consumed. The corresponding values in the WS-amended soils, were, 76.8 ± 0.58, 73.1 ± 1.1 and 59.0 ± 0.71% for WSF, WSM and WSL, respectively (Table 2). A significant increase in the seedlings survival time was observed for the WS-treated soil, which was dependent on the PAW Fig. 6. Saturated hydraulic conductivity (Ksat) of the WaterSorb-soil mixtures as a and WS particle size. The seedlings treated with WSF, WSL and function of polymer particle size (WSF, WSM and WSL). Different letters on the top of WSM survived for 27.0 ± 1.3, 24.0 ± 1.1 and 17.0 ± 0.7 days, respec- bars indicate significant differences between treatments with p-value 0.05 (Tukey's tively, instated of 13.0 ± 1.0 days for the control. fi e test). WSF, WSM and WSL denote WaterSorb ne (0.8 1.0 mm), WaterSorb medium The pattern of DWC of different grained-WS and the control is e e (1.0 2.0 mm) and WaterSorb large (2.0 4.0 mm), respectively. Data are fi means ± standard deviation. shown in Figure (8). In the rst seven days, DWC was almost similar for all treatments. The average DWC, in the first seven days, was 3.49, 3.44, 3.40 and 3.41 mm day 1 for control, WSF, WSM and WSL, 3.6. Guava seedlings survivability experiment respectively. After seven days of ceasing irrigation, the soils amended with WS, particularly the WSF, significantly showed a Soil water balances for the studied trials before and after ceasing considerable higher DWC relative to the control. Moreover, when irrigation are presented in Table (2). The total soil water storage DWC was calculated by dividing the PAW by the survival days differed among the treatments significantly at the beginning and at (Table 2), the WS-treated seedlings tend to have similar values. The plant death. The stored water was significantly dependent on the only exception is being for the WSL, in which the DWC was WS particles sizes, in which the WSF possessed the greatest amount significantly lower than the other treatments. The DWC values of water retained in comparison to the control or the other WS sizes were 2.93 ± 0.20, 2.88 ± 0.16, 2.91 ± 0.17 and 2.61 ± 0.09 for control, at the trials begin. At the beginning of the desiccation experiment, WSF, WSM and WSL, respectively. the initial soil water content was 14.4 ± 0.24, 30.7 ± 0.22, A considerable growth of the seedlings in terms of plant height 28.33 ± 0.17 and 22.7 ± 0.25% for control, WSF, WSM and WSL, and TDM was observed, which was highest in the seedlings respectively, which was reduced to 2.8 ± 0.2%, 7.07 ± 0.14, growing in WS-amended soil (Table 3). The plant heights before 7.57 ± 0.14 and 9.33 ± 0.24% for control, WSF, WSM and WSL, ceasing irrigation after eight weeks of enough watering were respectively, at which the seedlings died. Thereby, the PAW differed approximately similar and did not differ significantly. However, by significantly among the studied treatments (Table 2). The PAW for the end of the experiment, plant heights differed significantly the WS treatments was significantly increased by a factor of 2.11, among the treatments. Plant height increased to be 103.8 ± 1.1, 1.86 and 1.21 for WSF, WSM and WSL, respectively, compared to the 97.8 ± 2.0 and 93.8 ± 2.38 cm for WSF, WSM and WSL, respectively, control treatment. PAW was increased from 3.5 ± 0.07 kg/pot in instead of 87.6 ± 1.1 for control. The effect of WS application was A.M. Abdallah / International Soil and Water Conservation Research 7 (2019) 275e285 281

Table 2 Water budget, plant available water, daily water consumption and number of survival days of the water-stressed Guava seedlings after ceasing irrigation.

Treatments Total soil water content after Total soil water content at Plant available water (PAW) PAW/WHC Number of Daily water ceasing irrigation (WHC) seedlings death (% of the initial survival days consumption water content) (mm/day) (Kg) (%) (Kg) (%) (kg) (mm)

Control 4.3 ± 0.07d 14.3 ± 0.24 d 0.8 ± 0.04 d 2.8 ± 0.12 c 3.5 ± 0.1 d 38.2 ± 1.13 d 80.8 ± 1.01 a 13.0 ± 1.0 d 2.93 ± 0.20 a WSL 6.8 ± 0.08 c 22.8 ± 1.54 c 2.8 ± 0.07 a 9.3 ± 0.24 a 4.0 ± 0.05 c 44.4 ± 0.56 c 59.0 ± 0.76 d 17.0 ± 0.7 c 2.61 ± 0.09 b WSM 8.5 ± 0.05 b 28.3 ± 1.22 b 2.3 ± 0.04 b 7.6 ± 0.17 b 6.2 ± 0.05 b 68.5 ± 0.56 b 73.1 ± 1.10 c 23.6 ± 1.1 b 2.91 ± 0.17 a WSF 9.2 ± 0.07 a 30.7 ± 1.12 a 2.1 ± 0.04 c 7.1 ± 0.65 b 7.1 ± 0.10 a 78.0 ± 1.1 a 76.8 ± 0.58 b 27.2 ± 1.3 a 2.88 ± 0.16 a Tuckey's value 0.089 0.30 0.06 0.21 0.10 1.18 1.02 0.60 0.11

Fig. 8. The pattern of daily water consumption (evapotranspiration, mm day 1) after the last watering as affected by WaterSorb of varying particle size (WSF, WSM and WSL). WSF, WSM and WSL denote WaterSorb ﬁne (0.8e1.0 mm), WaterSorb medium (1.0e2.0 mm) and WaterSorb large (2.0e4.0 mm), respectively.

extended to the TDM; whereas, TDM (g seedling 1) was increased on soil WHC, PAW and the ratio between the two (PAW/WHC). This from 177.8 ± 3.0 g in case of the control to be 196.2 ± 1.7, 194.4 ± 2.8 ratio indicates the percentage of available water to plants in rela- and 187.2 ± 3.1 g for WSF, WSM and WSL, respectively (Table 3). The tion to the total retained water in the hydrogel particles. In this RWC before terminating irrigation behaved similarly to the plant study, the swelling behavior (SR, SC, and re-swellability) of heights (Table 3). Meanwhile, the RWC after 10 days of irrigation different WS sized particles using DW and varying water salinities termination differed among the treatments and its values for the was investigated. After which, we assessed the impact of different control was the lowest among the treatments of WS. RWC of leaves WS sized particles on Ksat, WHC, PAW, DWC, growth and the sur- increased from 60.06 ± 0.81% in case of control to 74.8 ± 1.9, vival of Guava seedlings under drought conditions. The water ab- 75.4 ± 1.1, and 68.4 ± 1.1% for the WSF, WSM and WSL. sorbency of hydrogels, observed in the current study, was also previously observed (Agaba et al., 2010; Montesano et al., 2015; Yu et al., 2017). The water absorbency of hydrogels might be attributed 4. Discussion to its high molecular weight, high negative charge and for its hydrophilic functional groups, i.e., carboxyl, hydroxyl, and sulfonic Classically, hydrogels application increase WHC, PAW of the soil, groups (Guilherme et al., 2015; Hüttermann, Orikiriza, & Agaba, thus prolong plant survival under water stress. The current study 2009). Several studies (Ahmed, 2015; Spagnol et al., 2012a,b) introduce information about the impacts of hydrogel particle size

Table 3 Effect of Watersorb particle size (WSL, WSM and WSF) on the height, relative water content of leaves and total dry matter of Guava seedling subjected to drought stress.

Treatments Plant high (cm) Relative water content of leaves (%) Total dry matter (g/seedling) Before ceasing irrigation At Plant death Before ceasing irrigation Ten days after ceasing irrigation

Control 84.6 ± 1.14 a 87.2 ± 1.30 c 79.12 ± 0.75 a 54.6 ± 0.81 c 177.8 ± 3.03 c WSL 84.4 ± 1.14 a 93.8 ± 1.48 b 80.00 ± 1.22 a 68.4 ± 1.14 b 187.2 ± 3.11 b WSM 85.0 ± 1.51 a 97.8 ± 2.04 b 78.72 ± 1.13 a 75.4 ± 1.14 a 194.4 ± 2.88 a WSF 85.4 ± 1.58 a 103.8 ± 2.3 a 79.40 ± 1.14 a 74.8 ± 1.92 a 196.2 ± 1.78 a Tuckey's value 1.82 2.50 1.44 1.76 3.69

WSF, WSM and WSL denote Watersorb fine (0.8e1.0 mm), Watersorb medium (1.0e2.0 mm) and Watersorb large (2.0e4.0 mm), respectively. Different superscript letters indicate significant difference at p ¼ 0.05. Means with same alphabets within the same column indicate non-significant differences (p > 0.05, Tuckey's test). Data are mean ± standard deviation. 282 A.M. Abdallah / International Soil and Water Conservation Research 7 (2019) 275e285 reported that water absorption capacity depends mainly on the 2009). This observation is in agreement with Bhardwaj et al. (2007), crosslinking density, particle size and ionic osmotic pressure of the where the Ksat of a sandy soil amended with Stockosorb 500 me- hydrogels. The maximum SC was recorded for the smaller particles dium or Stockosorb 500 micro reduced to be 51.1 and 37.3 mm h 1 as compared with the medium and large particles. Similar results instead of 71.0 mm h 1 for the control. The remarkable Ksat were reported by (Bhardwaj et al., 2007), in which they found a reduction of WS-amended soil, might not due to the clay swelling þ negative correlation between the particles size of the CLP and the or dispersion. Since the concentration of Ca2 electrolytes in tab SC. In terms of swelling rate, the initial SR, was dramatically water, used in this test, usually enough to prevent clay swelling and dependent on the particle size. Our results support the findings of dispersion (Shainberg & Letey, 1984). Additionally, in the current (Ahmed, 2015; Bhardwaj et al., 2007; Singh, 1997) in which they study, almost there were no aggregates to be disturbed in such indicated that the fast swelling (higher swelling rate) is mainly sandy soil with less than 5% clay (Kay & Angers, 1999). The Ksat based on the hydrogel particles size, and thus, the time needed to reduction might be explained by the increased volume of the WS- reach equilibrium swelling (final swelling) was less for the reduced soil mixtures, due to polymer swelling during saturation, that ul- hydrogel particles size. The higher SR and SC of the fine particles timately led not only a decrease in pore space between soil particles relative to medium and large particles might be attributed to the (Levy, Goldstein, & Mamedov, 2005) but also resulted in pore higher SSA of the fine particles. Since, the large particle possesses blockage by the swollen polymer grains (Bhardwaj et al., 2007; Han small SSA while the small ones possess large SSA (Hillel, 1971). The et al., 2013). The significant differences between the WS treatments high SSA enhanced the capability of the small particle to absorb and could be attributed to the difference in their swelling properties, retains water films, as the contact area with water molecules is e.g., SR and SC (Bhardwaj et al., 2007). This is in agreement with the much higher compared to that for the large particles (Hillel, 1971). results of Mohammadi, Sefti, Salehi, and Moghadam (2015) in All tested WS particle sizes were highly sensitive to water which they reported that polymers of high swelling rate reduce salinity. It is well established that the presence of electrolytes in the water permeability in porous media compared with that of low solution decreases the absorption capacity of CLP (Guilherme et al., swelling rate. 2015; Johnson, 1984). Our results are in agreement with Akhter Regarding Guava seedlings survivability test, the increase in et al. (2004) who found that maximum SC was 505, 212 and PAW found to be dependent on the SSA of the tested hydrogels, 140 g g 1 in DW, tap water (EC ¼ 0.759 dS m 1) and saline water where a correlation of 0.94 was observed between the SSA and the (EC ¼ 5.09 dS m 1). The reduction in the SC of hydrogels might amount PAW of the WS-amended soil. Similarly, the increase in occur due to the competition between the molecules of solute and seedlings' survival was proportional to the increase in PAW. water for adsorption to the WS molecules (Johnson, 1984; Shahid Moreover, the significant increase in soil water content at PWP (at et al., 2012; Spagnol et al., 2012a,b). Moreover, the SC of hydro- the individual seedlings death) of the WS-amended soil, indicates gels might not only affected by the dissolved salts concentrations, that seedlings' roots do not have full access to water absorbed by but also affected by the nature of the dissolved salts (i.e., valence the hydrogels. The water stored in the large particles was less and ionic radius). The presence of ions, particularly the divalent readily available for plant root abstraction when needed. The þ þ cations (Mg2 and Ca2 ), permanently attached to the carboxylic variation in PAW among WS treatments could be explained by the acid groups of acrylic acid and thus, block the negative active sites variation in the SSA of the WS particles. When the soil particles get of the hydrogel and decrease the absorbency of the hydrogel dry, the water inside the WS particles gradually moves to the sur- (Spagnol et al., 2012a, b; Zhou et al., 2012). Zhou et al. (2012) re- rounding dry soil particles. The lower SSA of large particles allows ported that with higher cation valence, concentration and larger for less contact area with the dry soil particles, therefore, less water ionic radius, polymers water-absorption remarkably reduced as can be released into the soil (Hillel, 1971). The results agreed with þ þ þ þ þ þ þ þ Na < K < Mg2 < Ca2 < Fe2 < Fe3 < Al3 < Cu2 , while anions those of (Hüttermann et al., 1999), in which they indicated that a had no effects. significant part of the water stored in hydrogel particles was The SC at equilibrium was decreased during the second and apparently not available for the plants. Since, the plants (Pinus third swelling/dehydration cycle. Similar results were reported by halepensis) in the control soils died after the water content was (Akhter et al., 2004; Han et al., 2013; Shahid et al., 2012; Spagnol reduced to 0.6%, while, the trees in the soils amended with 0.4% et al., 2012a,b) in which they reported that the SC of tested Stockosorb K-400 died at a water content of 10.7%. Additionally, hydrogels showed a downward trend with time. The reduction in Han et al. (2013) reported that about 89% of the water retained in þ SC might be due to the presence of polyvalent cations (Ca2 and SAP might be available to plant roots. However, the results were in þ Mg2 ) form the insoluble salts with acrylic acid (relatively stable disagreement with those of Li et al. (2014) in which they reported acrylic acid carboxylates) (Shahid et al., 2012). the water absorbed by the hydrogel particles was not available to In this study, the water absorption of the WS in the free system wheat roots. The inconsistency among results might be attributed was higher than in the WS-soil mixture. In the WS-soil mixture, WS to the variability in plant species, environmental conditions and particles were surrounded by soil particles and subjected to a SAPs type, concentration and particle size. pressure from all directions (Bhardwaj et al., 2007; Buchholz, 1998). The increase in water storage owing to WS addition in the pot As a result, the degree of swelling of the absorbents was limited experiments was corroborated by the laboratory results (water compared with that in the free water (Singh, 1997). Obtained re- retention curve-using pressure plate test), but with higher water sults are in agreements with previous studies (Bhardwaj et al., content values in the plate test. A possible reason for the increase in 2007; Buchholz, 1998; Han et al., 2013). The maximum swelling soil water content at FC in the plate test relative to that in the pots in the WS-soil mixture was recorded for the WSF, which might be might be, in the plate test, the WS-soil mixture was saturated attributed to the high SSA of the fine particles that allowed for overnight, and thus the WS particles had enough time to reach full higher SR. Moreover, logically the number of polymer granules in swelling. Meanwhile, the inconsistency in PWP values, in the pots the small grain polymers is higher compared to the medium and and the plate test, could be attributed to that we might be over- large polymer granules. Therefore, the granules ability to resist the estimated the PWP in the pots experiments, to ensure that seed- confining pressure of the soil matrix increased and thus, more lings have died. However, the water release of retained water by water was absorbed by the fine particles (Bhardwaj et al., 2007). hydrogels under different conditions needs to be investigated in The reduction in Ksat, observed in this study, due to hydrogels further studies. It is worthy to note that, the increase in soil WHC, in application was previously reported (Agaba et al., 2010; Andry et al., the pots experiment, might be attributed not only to the water A.M. Abdallah / International Soil and Water Conservation Research 7 (2019) 275e285 283 absorbed by the hydrogels, but also, could be due to the reduction establishment in soils with low WHC. However, older plants would in the Ksat, thus decreasing water drainage below the root zone require larger amounts. The plant roots are hydrotropic; therefore, (Bhardwaj et al., 2007; Han et al., 2013). the hydrogel amendment is expected to impact root distribution, The results of this study confirm preliminary studies which have growth and behavior (Hüttermann et al., 1999). The feasibility of compared the increase in PAW, due to the hydrogel amendments, hydrogels application in the field has been previously confirmed and the survival time of trees. In the current study, the seedlings' (Hüttermann et al., 2009; Wang, Sun, & Wang, 1993). Where, the survival time was positively correlated (r ¼ 0.95) with the PAW. application of hydrogels in afforestation increases the success of Obtained results are in agreement with the results of Agaba et al. planting and thus replanting was not necessary (Wang et al., 1993), (2010), in which they indicated that amending the sandy soil therefore, the final costs of afforestation was reduced. Moreover, with 0.4% hydrogel significantly increased PAW by about three-fold hydrogels application lead to better plant growth and higher yield in eight of the nine tested tree species and prolonged survival for all (Agaba et al., 2010; El-Asmar et al., 2017; Hüttermann et al., 2009, the nine tree species. Moreover, the results of seedlings survival are 1999). Additionally, the reduced water losses via percolation, due to in agreement with earlier publications on the survival of Pinus trees hydrogel application, leads to the avoidance of several environ- grown on sandy soils amended with hydrogel under water stress mental and economic problems, which is of a big importance in (Agaba et al., 2010; Andry et al., 2009; El-Asmar et al., 2017; regions where water resources are limited and expensive. Hüttermann et al., 1999) and with those reporting improved survival of seedlings for dryland afforestation (Chirino, Vilagrosa, & 5. Conclusion Vallejo, 2011; Orikiriza et al., 2013). Furthermore, the results agree with the improved survival time and the lower leaf water The fine-grained WS had a higher swelling rate and re- potential of young citrus seedlings grown in perlite and peat media swellability as compared with the medium and large-grained WS. treated with hydrogel (Arbona et al., 2005) and for the growth of The amendment of the sandy soil with 0.3% WS, at any particle size, ornamental trees (Dehganl, Yeagerl, & Almira, 1994). However, significantly reduced saturated hydraulic conductivity, and other studies reported negative effects on Aleppo pine grown on increased both water holding capacity and available water capacity, loam soil in Spain (Del Campo, Hermoso, Flors, Lidon, & Navarro- with higher values for the fine-grained WS. Therefore, the survival Cerrillo, 2011). of Guava seedlings increased from 13.0 ± 1.0 days for control to After eight days of ceasing irrigation, seedlings grown in WS- 27.0 ± 1.3, 24.0 ± 1.1 and 17.0 ± 0.7 days for WSF, WSM and WSL, amended soil had higher DWC compared to the control. Since, respectively. The more prolonged survival together with the the soils amended with the WS, especially, the WSF, retained much enhanced growth of the seedlings, under drought stress, was more more water during the irrigation period than the control soils, they pronounced for the small-grained WS. The increased water content could afford to lose more water and still retain more soil moisture at PWP in the laboratory and pots tests indicate that seedlings do than control soil. In control soil, 50% of the PAW was evapo- not have full access to the water stored in the hydrogel. Since the transpired after 5 days (8 days before the seedlings died), whereas, water stored in the large particles was less readily available for in the case of the seedlings growing on WS-treated soil, it took 12, plant root abstraction when needed. The application of WS had no 10 and 7 days for WSF, WSM and WSL respectively. The results effect on the daily water consumption, in which the values were disagreed with those of Agaba et al. (2010), in which they found almost similar for all treatments as long as the control soil still had that the addition of Luquasorb (0.4%, w/w) decreased the daily available water to be absorbed the roots. The inconsistency in water evapotranspiration of eight tested tree species. This conflict might release from the hydrogel, under different conditions (in the plant be due to the different plant species, hydrogel type, concentration system and the plate test), needs to be investigated in further and particle size. studies. Further studies are also necessary for understanding the During the water stress period, a significantly better growth of relationship between hydrogel particle size and their durability in the seedlings (in terms of plant high and TDM), together with a the soil. higher leaves RWC were observed for the WS-treated soil as compared to the control. The improved growth of WS-treated Funding seedlings could be attributed to the increase in PAW which increased the seedlings survival (r ¼ 0.96) and thus, more dry This research did not receive any specific grant from funding matter accumulated. Furthermore, the increase in seedlings growth agencies in the public, commercial, or not-for-profit sectors. parameters was closely dependent on the increase in PAW, i.e., RWC of leaves (r ¼ 0.89), which promote the photosynthesis resulting in Conflicts of interest the higher dry matter (r ¼ 0.88) and higher plant height (r ¼ 0.91). The results are in accordance with the findings of (Orikiriza et al., The author declare no conflict of interest. 2013), in which they indicated that soils treated by SAPs markedly alleviated the inhibition of plant growth increased that were Acknowledgments caused by drought stress. The study clearly showed that the use of hydrogel (Watersorb) The author sincerely thanks Rich McLaughlin, Professor of Urban fi signi cantly reduced the downward movement of irrigation water, Soil & Water Management, Crop and Soil Sciences department, thus reduced percolation. Additionally, amending the soil with NCSU, USA, for providing the hydrogels used in this study. A special hydrogels, particularly with small sized particles, markedly thanks to Ibrahim Nassar, Professor of Soil Physics, Faculty of increased WHC, PAW, induce better growth of water-stressed Agriculture, Damanhour University, Egypt, for reviewing the plants and significantly prolonged their survival. Since, the hydro- manuscript and valuable suggestions. We also thank Dr. Ibrahim gels mixed only in the root zone of the trees, the cost of hydrogels is Salah, Dr. Ahmed ELghanam and Dr. Atef Nassar for their comments mainly depending on the amount used per tree and number of trees on a draft of the manuscript. per ha (Hüttermann et al., 2009). In our case, if we used 20e30 g per seedling, with a 600 seedling per ha, an amount of References 12e18 kg ha 1 is needed. This calculation was based on seedling plants, which would be suitable for supporting seedling Agaba, H., Orikiriza, L. J. B., Esegu, J. F. O., Obua, J., Kabasa, J. D., & Hüttermann, A. 284 A.M. Abdallah / International Soil and Water Conservation Research 7 (2019) 275e285

(2010). Effects of hydrogel amendment to different soils on plant available Hüttermann, A., Orikiriza, L. J. B., & Agaba, H. (2009). Review application of water and survival of trees under drought conditions. Clean - Soil, Air, Water, superabsorbent polymers for improving the ecological chemistry of degraded 38(4), 328e335. https://doi.org/10.1002/clen.200900245. or polluted lands 1 desertification as a major problem for global sustainability. Ahmed, E. M. (2015). Hydrogel: Preparation, characterization, and applications: A Clean - Soil, Air, Water, 37(7), 517e526. https://doi.org/10.1002/clen.200900048. review. Journal of Advanced Research, 6(2), 105e121. https://doi.org/10.1016/j. Hüttermann, A., Zommorodi, M., & Reise, K. (1999). Addition of hydrogels to soil for jare.2013.07.006. prolonging the survival of Pinus halepensis seedlings subjected to drought. Soil Akhter, J., Mahmood, K., Malik, K. A., Mardan, A., Ahmad, M., & Iqbal, M. M. (2004). and Tillage Research, 50(3e4), 295e304. https://doi.org/10.1016/S0167-1987(99) Effects of hydrogel amendment on water storage of sandy loam and loam soils 00023-9. and seedling growth of barley, wheat and chickpea. Plant Soil and Environment, Isik, B., & Kis, M. (2004). Preparation and determination of swelling behavior of 50(10), 463e469. https://doi.org/10.17221/4059-PSE. poly(acrylamide-co-acrylic acid) hydrogels in water. Journal of Applied Polymer Andry, H., Yamamoto, T., Irie, T., Moritani, S., Inoue, M., & Fujiyama, H. (2009). Water Science, 94(4), 1526e1531. https://doi.org/10.1002/app.21074. retention, hydraulic conductivity of hydrophilic polymers in sandy soil as Johnson, M. S. (1984). The effects of gel-forming polyacrylamides on moisture affected by temperature and water quality. Journal of Hydrology, 373(1e2), storage in sandy soils. Journal of the Science of Food and Agriculture, 35(11), 177e183. https://doi.org/10.1016/j.jhydrol.2009.04.020. 1196e1200. https://doi.org/10.1002/jsfa.2740351110. Ankenbauer, K. J., & Loheide, S. P. (2017). The effects of soil organic matter on soil Kay, B. D., & Angers, D. A. (1999). Structure. In M. E. Sumner (Ed.), Handbook of soil water retention and plant water use in a meadow of the Sierra Nevada, CA. science (pp. A229eA276). https://doi.org/10.1046/j.1365-2389.2001.00373.x. Hydrological Processes, 31(4), 891e901. https://doi.org/10.1002/hyp.11070. Langaroodi, N. B. S., Ashouri, M., Dorodian, H. R., & Azarpour, E. (2013). Study effects Arbona, V., Iglesias, D. J., Jacas, J., Primo-Millo, E., Talon, M., & Gomez-Cadenas, A. of super absorbent application , saline water and irrigation management on (2005). Hydrogel substrate amendment alleviates drought effects on young yield and yield components of peanut ( Arachis hypogaea L .). Annals of Bio- citrus plants. Plant and Soil, 270(1), 73e82. https://doi.org/10.1007/s11104-004- logical Research, 4(1), 160e169. 1160-0. Levy, G. J., Goldstein, D., & Mamedov, A. I. (2005). Saturated hydraulic conductivity Bai, W., Song, J., & Zhang, H. (2013). Repeated water absorbency of super-absorbent of semiarid soils. Soil Science Society of America Journal, 69(3), 653. https://doi. polymers in agricultural field applications: A simulation study. Acta Agriculturae org/10.2136/sssaj2004.0232. Scandinavica Section B Soil and Plant Science, 63(5), 433e441. https://doi.org/10. Li, X., He, J. Z., Hughes, J. M., Liu, Y. R., & Zheng, Y. M. (2014). Effects of super- 1080/09064710.2013.797488. absorbent polymers on a soil-wheat (Triticum aestivum L.) system in the Banedjschafie, s, & Durner, W. (2015). Water retention properties of a sandy soil field. Applied Soil Ecology, 73,58e63. https://doi.org/10.1016/j.apsoil.2013.08. with superabsorbent polymers as affected by aging and water quality. Journal of 005. Plant Nutrition and Soil Science, 178, 798e806, 2002. Minasny, B., & McBratney, A. B. (2018). Limited effect of organic matter on soil Belaqziz, S., Khabba, S., Er-Raki, S., Jarlan, L., Le Page, M., Kharrou, M. H., et al. (2013). available water capacity. European Journal of Soil Science, 69(1), 39e47. https:// A new irrigation priority index based on remote sensing data for assessing the doi.org/10.1111/ejss.12475. networks irrigation scheduling. Agricultural Water Management, 119,1e9. Mohammadi, S., Sefti, M. V., Salehi, M. B., & Moghadam, A. M. (2015). Hydrogel https://doi.org/10.1016/j.agwat.2012.12.011. swelling properties : Comparison between conventional and nanocomposite Beniwal, R. S., Langenfeld-Heyser, R., & Polle, A. (2010). Ectomycorrhiza and hydrogels for water shutoff treatment. Asia-Pacific Journal of Chemical Engi- hydrogel protect hybrid poplar from water deficit and unravel plastic responses neering,743e753 (June) https://doi.org/10.1002/apj. of xylem anatomy. Environmental and Experimental Botany, 69(2), 189e197. Montesano, F. F., Parente, A., Santamaria, P., Sannino, A., & Serio, F. (2015). Biode- https://doi.org/10.1016/J.ENVEXPBOT.2010.02.005. gradable superabsorbent hydrogel Increases Water retention properties of Bhardwaj, A. K., Shainberg, I., Goldstein, D., Warrington, D. N., & J Levy, G. (2007). growing media and plant growth. Agriculture and Agricultural Science Procedia, Water retention and hydraulic conductivity of cross-linked polyacrylamides in 4,451e458. https://doi.org/10.1016/j.aaspro.2015.03.052. sandy soils. Soil Science Society of America Journal, 71(2), 406. https://doi.org/10. Moutier, M., Shainberg, I., & Levy, G. J. (2000). Hydraulic gradient and wetting rate 2136/sssaj2006.0138. effects on the hydraulic conductivity of two calcium vertisols. Soil Science So- Buchholz, F. L. (1998). The structure and properties of superabsorbents poly- ciety of America Journal, 64(4), 1211e1219. https://doi.org/10.2136/sssaj2000. acrylates. In Buchholz, & Graham (Eds.), Modern superabsorbent polymer tech- 6441211x. nology, John Wiley Sons Inc. Wiley-VCH. https://doi.org/10.1002/1097- Orikiriza, L. J. B., Agaba, H., Eilu, G., Kabasa, J. D., Worbes, M., & Hüttermann, A. 0126(200011)49. (2013). Effects of hydrogels on tree seedling performance in temperate soils Chen, P., Zhang, W., Luo, W., & Fang, Y. (2004). Synthesis of superabsorbent poly- before and after water stress. Journal of Environmental Protection, 04(07), mers by irradiation and their applications in agriculture. Journal of Applied 713e721. https://doi.org/10.4236/jep.2013.47082. Polymer Science, 93,1748e1755, 93, 1748e1755 https://doi.org/10.1002/app. Shahid, S. A., Qidwai, A. A., Anwar, F., Ullah, I., & Rashid, U. (2012). Improvement in 20612. the water retention characteristics of sandy loam soil using a newly synthesized Chirino, E., Vilagrosa, A., & Vallejo, V. R. (2011). Using hydrogel and clay to improve poly(acrylamide-co-acrylic acid)/AlZnFe 2O 4 superabsorbent hydrogel nano- the water status of seedlings for dryland restoration. Plant and Soil, 344(1), composite material. Molecules, 17(8), 9397e9412. https://doi.org/10.3390/ 99e110. https://doi.org/10.1007/s11104-011-0730-1. molecules17089397. Dane, J. H., & Hopmans, J. W. (2002). Pressure plate extractor. In J. H. Dane, & Shainberg, I., & Letey, J. (1984). Response of soils to sodic and saline conditions. G. C. Topp (Eds.), Methods of soil analysis: Physical methods, Part 4 (pp. Hilgardia, 52(2), 1e57. https://doi.org/10.3733/hilg.v52n02p057. 688e690). Madison: Soil Science Society of America. Singh, J. (1997). Physical behavior of superabsorbent hydrogels in sand. Montreal: Dehganl, B., Yeagerl, T. H., & Almira, F. C. (1994). HortScience : A publication of the McGill Univ.. Retrieved from https://www.researchgate.net/publication/ American society for horticultural science. HortScience, 29(6), 641e644. 30008260_Physical_behavior_of_superabsorbent_hydrogels_in_sand. Retrieved from http://hortsci.ashspublications.org/content/29/6/641.short. Spagnol, C., Rodrigues, F. H. A., Neto, A. G. V. C., Pereira, A. G. B., Fajardo, A. R., Del Campo, A. D., Hermoso, J., Flors, J., Lidon, A., & Navarro-Cerrillo, R. M. (2011). Radovanovic, E., et al. (2012). Nanocomposites based on poly(acrylamide-co- Nursery location and potassium enrichment in Aleppo pine stock 2. Perfor- acrylate) and cellulose nanowhiskers. European Polymer Journal, 48(3), mance under real and hydrogel-mediated drought conditions. Forestry, 84(3), 454e463. https://doi.org/10.1016/j.eurpolymj.2011.12.005. 235e245. https://doi.org/10.1093/forestry/cpr009. Spagnol, C., Rodrigues, F. H. A., Pereira, A. G. B., Fajardo, A. R., Rubira, A. F., & El-Asmar, J., Jaafar, H., Bashour, I., Farran, M. T., & Saoud, I. P. (2017). Hydrogel Muniz, E. C. (2012). Superabsorbent hydrogel composite made of cellulose banding improves plant growth, survival, and water use efficiency in two nanofibrils and chitosan-graft-poly(acrylic acid). Carbohydrate Polymers, 87(3), calcareous soils. Clean - Soil, Air, Water, 45(7), 1700251. https://doi.org/10.1002/ 2038e2045. https://doi.org/10.1016/j.carbpol.2011.10.017. clen.201700251. Wang, Y., & Boogher, C. A. (1987). Effect of a medium-incorporated hydrogel on Fan, T., Stewart, B. A., Payne, W. A., Yong, W., Luo, J., & Gao, Y. (2005). Long-term plant growth and water use of two foliage species 1. Journal of Environmental fertilizer and water availability effects on cereal yield and soil chemical prop- Horticulture, 5,125e127. September. erties in northwest China. Soil Science Society of America Journal, 69(3), 842. Wang, J., Sun, J., & Wang, Z. (1993). Studies on a series of techniques in applying https://doi.org/10.2136/sssaj2004.0150. superabsorbents in planting at south hillside in lower mountain area of Beijing Gao, L., Li, Q., Song, Z., & Wang, J. (2000). Preparation of nano-scale titania thick film (V): Analysis of cost for applying superabsorbents in afforestation. J. Beijing For. and its oxygen sensitivity. Sensors and Actuators B: Chemical, 71(3), 179e183. Univ., (Supp. 13) https://doi.org/10.1002/clen.200900048. https://doi.org/10.1016/S0925-4005(00)00612-2. Wei, Y., & Durian, D. J. (2014). Rain water transport and storage in a model sandy soil Guilherme, M. R., Aouada, F. A., Fajardo, A. R., Martins, A. F., Paulino, A. T., with hydrogel particle additives. The European Physical Journal E, 37(97), 1e11. Davi, M. F. T., et al. (2015). Superabsorbent hydrogels based on polysaccharides https://doi.org/10.1140/epje/i2014-14097-x. for application in agriculture as soil conditioner and nutrient carrier: A review. Wu, L., & Liu, M. (2008). Preparation and characterization of cellulose acetate- European Polymer Journal, 72, 365e385. https://doi.org/10.1016/j.eurpolymj. coated compound fertilizer with controlled-release and water-retention. Poly- 2015.04.017. mers for Advanced Technologies, 19(7), 785e792. https://doi.org/10.1002/pat. Han, B., Benner, S. G., & Flores, A. N. (2018). Evaluating impacts of climate change on 1034. future water scarcity in an intensively managed semi-arid region using a Xu, S., Zhang, L., McLaughlin, N. B., Mi, J., Chen, Q., & Liu, J. (2015). Effect of synthetic coupled model of biophysical processes and water rights. Hydrology and Earth and natural water absorbing soil amendment soil physical properties under System Sciences Discussions,1e53. https://doi.org/10.5194/hess-2018-140. potato production in a semi-arid region. Soil and Tillage Research, 148,31e39. Han, Y., Yu, X., Yang, P., Li, B., Xu, L., & Wang, C. (2013). Dynamic study on water https://doi.org/10.1016/j.still.2014.10.002. diffusivity of soil with super-absorbent polymer application. Environmental Yamasaki, S., & Dillenburg, L. (1999). Measurements of leaf relative water content in Earth Sciences, 69(1), 289e296. https://doi.org/10.1007/s12665-012-1956-9. Araucaria angustifolia. Revista Brasilleira de Fisiologia, 11(2), 69e75. Retrieved Hillel, D. (1971). Soil and water; physical principles and processes. Academic Press. from http://www.cnpdia.embrapa.br/rbfv/pdfs/v11n2p69.pdf. A.M. Abdallah / International Soil and Water Conservation Research 7 (2019) 275e285 285

Yang, L., Yang, Y., Chen, Z., Guo, C., & Li, S. (2014). Inﬂuence of super absorbent under drying conditions. Soil Science Society of America Journal, 81(4), 889. polymer on soil water retention, seed germination and plant survivals for rocky https://doi.org/10.2136/sssaj2016.07.0231. slopes eco-engineering. Ecological Engineering, 62,27e32. https://doi.org/10. Zhang, J., Liu, R., Li, A., & Wang, A. (2006). Preparation, swelling behaviors, and 1016/j.ecoleng.2013.10.019. slow-release properties of a poly(acrylic acid-co-acrylamide)/sodium humate Yu, J., Shainberg, I., Yan, Y. L., Shi, J. G., Levy, G. J., & Mamedov, a I. (2011). Super- superabsorbent composite. Industrial & Engineering Chemistry Research, 45(1), absorbents and semiarid soil properties affecting water absorption. Soil Science 48e53. https://doi.org/10.1021/ie050745j. Society of America Journal, 75(6), 2305e2313. https://doi.org/10.2136/sssaj2010. Zhou, B., Liao, R., Li, Y., Gu, T., Yang, P., Feng, J., et al. (2012). Water-absorption 0397. characteristics of organic-inorganic composite superabsorbent polymers and its Yu, J., Shi, J. G., Ma, X., Dang, P. F., Yan, Y. L., Mamedov, A. I., et al. (2017). Super- effect on summer maize root growth. Journal of Applied Polymer Science, 126(2), absorbent polymer properties and concentration effects on water retention 423e435. https://doi.org/10.1002/app.36652. International Soil and Water Conservation Research 7 (2019) 286e296

Contents lists available at ScienceDirect

International Soil and Water Conservation Research

journal homepage: www.elsevier.com/locate/iswcr

Original Research Article Performance of Phragmites Australis and Cyperus Papyrus in the treatment of municipal wastewater by vertical ﬂow subsurface constructed wetlands * Fernando García-Avila a, , Jhanina Patino-Ch~ avez a, Fanny Zhinín-Chimbo a, Silvana Donoso-Moscoso a, Lisveth Flores del Pino b, Alex Aviles-A nazco~ a a Facultad de Ciencias Químicas, Universidad de Cuenca, Cuenca, 010107, Ecuador b Facultad de Ciencias, Universidad Nacional Agraria La Molina, Lima, 12175, Peru article info abstract

Article history: The use of constructed wetlands to treat municipal wastewater reduces energy consumption and Received 29 October 2018 therefore economic costs, as well as reduces environmental pollution. The purpose of this study was to Received in revised form compare the purification capacity of domestic wastewater using two species of plants sown in subsurface 4 April 2019 constructed wetlands with vertical flow built on a small scale that received municipal wastewater with Accepted 15 April 2019 primary treatment. The species used were Phragmites Australis and Cyperus Papyrus. For this purpose, a Available online 25 April 2019 constant flow of 0.6 m3 day 1 was fed from the primary lagoon to each of the two wetlands built on a pilot scale with continuous flow. Each unit was filled with granite gravel in the lower part and with silicic Keywords: Constructed wetland sand in the upper part of different granulometry, the porosity of the medium was 0.34, with a retention 1 fi Fitorremediacion time of 1.12 days and a hydraulic load rate of 0.2 m day . To analyze the puri cation capacity of Macrophytes wastewater, physical, chemical and biological parameters were monitored during three months. Samples Nutrients were taken at the entrance and exit in each experimental unit. The results obtained in the experimental Wastewater treatment tests for the two species of plants, indicated that the Cyperus Papyrus presented a greater capacity of pollutants removal as biochemical oxygen demand (80.69%), chemical oxygen demand (69.87%), ammoniacal nitrogen (69.69%), total phosphorus (50%), total coliforms (98.08%) and fecal coliforms (95.61%). In the case of Phragmites Australis retains more solids. The species with greater efficiency in the treatment of municipal wastewater for this study was Cyperus Papyrus. © 2019 International Research and Training Center on Erosion and Sedimentation and China Water and Power Press. Production and Hosting by Elsevier B.V. This is an open access article under the CC BY-NC- ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction (CW) have been increasingly recognized throughout the world in recent years (Mango, Makate, Tamene, Mponela, & Ndengu, 2017; Researches on the treatment of domestic wastewater through Orimoloye, Kalumba, Mazinyo, & Nel, 2018; Talukdar & Pal, 2017; technologies that demand low costs, low energy and maintenance Vymazal, 2011). CW that try to mimic the layout and functions of has been a priority in most countries around the world (Binder, natural wetlands have been widely used to improve water quality Gottle,€ & Shuhuai, 2015; Laaffat, Aziz, Ouazzani, & Mandi, 2016). (Kyambadde, Gumaelius, & Dalhammar, 2004). CW have evolved For the treatment of domestic wastewater, conventional treatments over time adding the use of plants, assimilating them to natural can be used in large populations and unconventional treatments in wetlands, consists pollutants removal system through various nuclei of small populations or in rural areas, where there is no natural processes, involving the operation of the support material sewerage service or buildings are widely dispersed (Tahir, Yasmin, or ﬁlter and the plant species. They are used throughout the world & Hassan Khan, 2016; Zidan, El-Gamal, Rashed,& El-Hady Eid, as a natural and economically favorable alternative to energy 2015). Because it is a natural process, which has a low maintenance (Ansari & Golabi, 2018). CWs include biological, chemical and cost, and also avoids chemical treatment, constructed wetlands physical processes similar to the processes that occur in natural treatment wetlands (Stefanakis, Akratos, & Tsihrintzis, 2014). CWs can improve water quality by retaining or transforming

* Corresponding author. constituents and have been used successfully to mitigate environ- E-mail address: [email protected] (F. García-Avila). mental pollution by eliminating a wide variety of wastewater https://doi.org/10.1016/j.iswcr.2019.04.001 2095-6339/© 2019 International Research and Training Center on Erosion and Sedimentation and China Water and Power Press. Production and Hosting by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). F. García-Avila et al. / International Soil and Water Conservation Research 7 (2019) 286e296 287 pollutants such as organic compounds, suspended solids, patho- Treatment Plant “El Guabo” were implemented, located in the gens, metals, colorants, pesticides and nutrients (Hansen, Kraus, Santa Isabel city, Ecuador. The plant is located at the following Bachand, Horwath, & Bachand, 2018). Horizontal Subsurface Flow geographic coordinates: Longitude 79.313732 W and Latitude Constructed Wetlands (HSSFCW) have been used mainly to treat 3.298460 S. This plant treats municipal wastewater through the municipal wastewater (Zidan, El-Gamal, Rashed, & El-Hady Eid, lagooning system, which serves as the primary treatment. After the 2015). CWs have been used and improved over time, using several primary treatment, the water is conducted to the wetlands as a types of plant species that adapt to the systems, flow types: hori- secondary treatment. Two experimental units were built, one for zontal, vertical or mixed, all carrying the same purpose that is the each species, Phragmites Australis and Cyperus Papyrus. Each purification of wastewater (Sanmuga Priya & Senthamil Selvan, wetland with its respective vegetation species operated indepen- 2017; Stefanakis et al., 2014). Vegetation is an important factor dently. The flow was applied continuously in each wetland. affecting the effectiveness of Vertical Subsurface Flow Constructed Wetlands (VSSFCW) in the elimination of COD, BOD , TSS and NH 5 4 2.2. Design and constructed of vertical subsurface flow wetland in all conditions tested (Abdelhakeem, Aboulroos, & Kamel, 2016; Coleman et al., 2001). Several species of plants sown in wetlands 2.2.1. Sizing of the piloto wetland improve the elimination of COD throughout the year compared to The criteria for the SSFCW design were taken from Metcalf & the gravel only systems, especially at low temperatures Eddy (1991); Crites and Tchobanoglous (1998); Kadlec et al. (Abdelhakeem et al., 2016), works efficiently in tropical climates (2000); García Serrano & Corzo Hernandez (2008); Abdelhakeem (Bohorquez, Paredes, & Arias, 2017). et al. (2016). To determine the size of the biological filtration sys- It was found that the presence of vegetation causes small vari- tem, the average temperature of the water (24.6 C) was deter- ations in the efficiency of elimination of chemical oxygen demand mined, thereby calculating the reaction rate constant, KT (1.31 (BOD ), suspended solids (SS), ammonia nitrogen (NH3-N), total 5 day 1). Considering the initial concentration of BOD (100 mg L 1) phosphorus (TP) and total nitrogen (TN) of livestock wastewater 5 entering the system, and the desired BOD5 concentration in effluent (Effendi, Widyatmoko, Utomo, & Pratiwi, 2018; Harmel et al., 2018). (10 mg L 1), the residence time of the water resulted in 3.25 days For all the above mentioned, the application of a VSSFCW was when the system operates intermittently and 1.12 days of retention chosen on a pilot scale in the wastewater treatment plant (WWTP) when the system operates continuously. “El Guabo” of the city of Santa Isabel, located in the province of The depth of the substrate, which can typically be 0.7 m. The Azuay-Ecuador. The choice of this type of subsurface flow wetland deeper the substrate is, the greater the load that the system can is due to the fact that they work with high rates of organic load, also process, but if the substrate is too deep, the bottom conditions during the operation of the system there were no presence of odors become anaerobic and can result in reduced elimination of BOD . or pests. They have greater tolerance to cold, and possibly, a greater 5 The effective porosity of the substrate (0.346 for sand and gravel) potential for assimilation per unit area than in surface flow systems and the depth of the substrate (0.65 m taking as reference the (Perez Villar, Dominguez, Hernandez, Sanchez, & Arteaga, 2012; greater root depth of the Papyrus plant) giving us as a result a final Stefanakis et al., 2014). Common reed (Phragmites Australis) is the 2 area of the 3 m system necessary for the reduction of BOD5.In most widely used species in subsurface flow constructed wetlands Table 1, the design parameters of the pilot scale experimental units & & € (SSFCW) (Vymazal Kropfelova, 2005; Vymazal Kropfelova, are presented. fi 2011), it has an appropriate ef cacy in the elimination of microbi- The CWs operated with a HLR of 0.2 m d 1 throughout the & ological parameters (Andreo-Martínez, García, Quesada, Almela, experiment. The granulometry of the material arranged in layers & 2017; Dadban Shahamat, Asgharnia, Kalankesh, Hosanpour, with their respective depths is shown in Table 2. 2018; Shahi et al., 2013). The root structures of Cyperus Papyrus provided more microbial fixation sites, sufficient residence time of wastewater, entrapment 2.2.2. Construction of pilot-scale wetlands and settlement of suspended particles, surface area for adsorption For the construction of the wetlands, clearing work was initially of contaminants, absorption, assimilation in plant tissues and ox- carried out, that is, cleaning and extraction of the weeds from the ygen for the oxidation of organic matter and inorganic in the site. Later the excavation was executed, in which the necessary area rhizosphere. Which explains its high efficiency treatment, which is for the construction of the experimental units established in the why it is a potential plant used to treat domestic wastewater design was delimited. In the surroundings, security channels were (Kyambadde, Kansiime, Gumaelius, & Dalhammar, 2004; also built in order to prevent the entry of extra flows and materials Perbangkhem & Polprasert, 2010); Cyperus Papyrus has also been carried during the rainy season. Then the wetlands were covered shown to reduce the content of phenols in wastewater (Vymazal, with high-density polyethylene (HDPE), to avoid drainage to the 2014). The hypothesis of this study, was that the wetland with underground aquifer of 750 mm, which is 2.2 times cheaper than Phragmites Australis would have greater removal of contaminants, reinforced concrete (Stefanakis et al., 2014; Tsihrintzis, 2017). For compared to the Cyperus Papyrus wetland; since it has been re- the conduction of the treated water, PVC pipes of 50 mm diameter ported a greater efficiency of Phragmites Australis (Zhang, Jinadasa, Gersberg, Liu, Jern, & Tan, 2014). The objective of this paper was to fi Table 1 investigate the puri cation capacity of domestic wastewater using Data of the design parameters to implement the pilot-scale constructed wetlands. two species of vegetation (Phragmites Australis and Cyperus Papyrus) in pilot system of VSSFCW in the WWTP “El Guabo” and Parameter Value Unit determine the improvement of water quality of download. Flow 0,6 m3 d 1 Depth 0,65 m Free edge 0,1 m Area 3 m 2 2. Materials and methods Long-wide relationship 3:1 Long 3 m 2.1. Study site description Width 1 m Time 1.12 día Slope 1 % The experimental wetlands on the land next to the Wastewater 288 F. García-Avila et al. / International Soil and Water Conservation Research 7 (2019) 286e296

Table 2 allowing to form a closed circuit and in turn distribute the liquid Depth of each layer of granulometric material. equally over the entire surface of the wetland. Material Effective size (mm) Depth (cm)

Coarse sand 2 10 2.3. Operation and monitoring of constructed wetlands Fine sand 8 30 Fine gravel 16 7.5 2.3.1. Operation of the VSSFCW Medium gravel 32 7.5 After building the VSSFCW and once planted Phragmites Aus- Coarse gravel 128 10 tralis and Cyperus Papyrus, the WCs were filled for a week. The flow of income to each wetland was taken from the lagoon exit of the were used; the drainage system was also assembled with 50 mm primary treatment system. From the second week, it began with a PVC pipes and the necessary accessories such as elbows, tees and continuous water supply and the flow was controlled through liquid solders; thus forming a circuit for water collection. Pipes meters. In this growth stage of plants, periodic monitoring was were distributed throughout the base of the wetland, in the form of carried out to verify the entry of water into the wetlands, observing a fishbone (Stefanakis et al., 2014). In regard to aeration pipes, the development of the plants and the purified effluent. Two 2 recommend installing 1 pipe for every 4 m (Kadlec et al., 2000). months were expected in order to obtain a greater growth of the At the bottom of the bed a drainage system for perforated pipes of ø plants, after which, the monitoring stage of the influent and 50 mm and covered with coarse gravel was installed (Da˛browski, effluent of the wetland began for three months. After four months Karolinczak, Gajewska, & Wojciechowska, 2017). A system of six of planting, the plants reached a certain height. To avoid excessive vertical tubes connected to the drainage collector system ensured accumulation of Cyperus Papyrus remains in the wetland, pruning gravitational ventilation, with the objective of increasing the was carried out. In the case of Phragmites Australis dried leaves amount of oxygen to the wetland (Da˛browski et al., 2017). were extracted. For the filling of the filtering medium, gravel and silicic sand The rainfall record in the study area was 561.4 mm of annual available near the place of study were used. In the vegetation rainfall, distributed mostly in the months of January to May with sowing, 12 plants of each species investigated for each wetland 406.5 mm, which is equivalent to 72.4%. In this area there are two were used (Fig. 1). Both types of vegetation were planted with a periods, one rainy from January to May and another dry from June 2 density of 4 plants m (Abou-Elela & Hellal, 2012; Brix & Arias, to December. The experiment began in February and ended in July, 2005; Ramprasada, Smithb, Memonc, & Philipa, 2017). Young that is, it was carried out during the rainy season. plants developed of Phragmites Australis were extracted from the banks of the Minas River, close to the study site; while Young plants 2.3.2. Monitoring program developed of Cyperus Papyrus were obtained from nurseries in the The monitoring allowed to evaluate the operation of the same zone, in such a way that a previous adaptation to the climate experimental units, identifying possible abnormalities, evidencing was not necessary. the fulfillment of the established objectives and checking if the Polyethylene hoses of 16 mm in diameter, were used for the physical, chemical and biological parameters in the effluent of the continuous feeding of water to the wetlands. The same ones that wetlands were within the maximum permissible limits established fl were drilled in order to regulate the ow at each point of the in the Ecuadorian environmental regulation. A period of two wetland. The holes are located according to the surface area; months was expected so that the plants could adapt to the new

Fig. 1. Wetlands built to scale. (a) Coarse gravel placement, (b) Fine gravel placement, (c) Initial development of Cyperus Papyrus, (d) Initial development of Phragmites Australis, (e) Fourth month of Cyperus Papyrus development, (f) Fourth month of Phragmites Australis development. F. García-Avila et al. / International Soil and Water Conservation Research 7 (2019) 286e296 289 environment of the CWs, until the plants reached a height of 0.5 m. (WTW, Germany), COD was analyzed by the K2Cr2O7 method according to APHA (1998). NH4-N, NO3-N and TP were analyzed with e 2.3.3. Location of sampling points and parameters analyzed a GENESYS 10S UV Vis Spectrophotometer following standard fi Three sampling points were identified: one point at the methods (APHA, 1998). Faecal coliforms (by membrane lter pro- entrance of water to the wetlands (because the residual water that cedure), TSS and alkalinity as standard methods for water and enters is the same for the two wetlands) and one point at the exit of wastewater examination (APHA, 1998) was analyzed. each wetland (Fig. 2). The monitoring had a period of three months a fortnightly sampling was carried out. The following physical, 2.4. Statistic analysis chemical and microbiological parameters were analyzed: pH, Temperature, Total Alkalinity, Electric Conductivity, Total Sus- Simple ANOVA was used for statistical analysis in order to pended Solids, Total Dissolved Solids, Nitrates, Ammoniacal Nitro- determine significant statistical differences in the performance of gen, Total Phosphorus, Biochemical Oxygen Demand, Chemical wastewater treatment by wetlands between Phragmites Australis Oxygen Demand, Total Coliforms, Fecal Coliforms (E. Coli). and Cyperus Papyrus. The variance analysis (ANOVA) was carried out using the InfoStat software package. A significance level of ¼ 2.3.4. Data collection p 0.05 was used. The samples were kept at 4 C and transported the same day for analysis to the Engineering Faculty Laboratory of the Cuenca Uni- 3. Results and discussion versity. Samples were collected in 2.5-L plastic bottles for physicochemical parameters, and in 50-ml containers for microbiological 3.1. Results of applied wetlands parameters (García, Ramos, Pauta, & Quezada, 2018a; INEN 2176, 1998). Temperature, pH, and conductivity were measured in situ The average value of the parameters monitored biweekly for using a Hach Multiparameter model HQ40d. BOD5 was quantified three months in the influent and effluent of the wetlands is pre- after five days of incubation at 20 C with Oxytop head gas sensors sented in Table 3.

Fig. 2. Location of sampling points, water inlet and outlet pipes. 290 F. García-Avila et al. / International Soil and Water Conservation Research 7 (2019) 286e296

Table 3 Inﬂuent and Efﬂuent wastewater characterization.

Parameter Units Influent Effluent Cyperus Papyrus Effluent Phragmites Australis

pH 6.95 ± 0.03 6.31 ± 0.1 6.21 ± 0.11 Temperature C 24.62 ± 0.55 24.37 ± 0.59 24.23 ± 0.54 1 Alkalinity mg L CaCO3 175.87 ± 11.93 68.8 ± 18.29 62.6 ± 12.12 Conductivity mScm 1 643.25 ± 27.23 633.67 ± 76.22 608.5 ± 44.2 Suspended solids mg L 1 88.83 ± 14.74 60 ± 16.85 33 ± 9.54 1 BOD5 mg L 95.75 ± 13.18 18.49 ± 2.58 23.57 ± 5.25 COD mg L 1 222.44 ± 16.99 67.02 ± 5.50 78.35 ± 4.75 Nitrates mg L 1 0.84 ± 0.2 7.16 ± 1.41 8.35 ± 4.75 Ammoniacal Nitrogen mg L 1 29.52 ± 2.56 8.95 ± 2.53 8.65 ± 1.98 Total Phosphorus mg L 1 7.42 ± 1.09 3.71 ± 0.06 3.76 ± 0.19 Total Coliforms MPN/100 ml 5.4E10 ± 2.22E9 1.038E9±4.18E8 2.14E9±5.3E8 Fecal Coliforms MPN/100 ml 1.71E10 ± 4.72E9 7.52E8±3.04E8 1.07E9±4.1E8

3.2. Removal percentages obtained with the application of pilot the renewal of oxygen, allowing a better efficiency of the filter beds wetlands and long-term functioning (Molle, Lienard, Grasmick, & Iwema, 2006). According to the HRL values mentioned by these authors, The average value of the purification efficiency of wastewater the HRL value (0.2 m d 1) used in this study is consistent with the treated with Phragmites Australis and Cyperus Papyrus is pre- HLR used in the aforementioned studies. sented in Table 4. Table 5 shows the removal percentages of NH4-N, TP, COD and BOD5 for different HLR used in other studies. These HLR values are 3.3. Role of hydraulic loading rate similar or close to the 0.2 m d 1 value used in this study. It can be observed that the removal values of COD and BOD5 (69.87%, 80.69%) To achieve an efficient pollutant removal rate (PRR), CWs obtained in this study for both types of vegetation were superior to require low HLR and longer hydraulic retention times (HRT) (Zhang the results obtained by Chang et al. (2007) and Yadav, Chazarenc, et al., 2014). The European recommendations for HLR of VFCW for and Mutnuri (2018), who used HRL of 0.2 and 0.22 m d 1 respec- wastewater treatment are 50e60 mm d 1 (Brix & Arias, 2005). tively. While, the NH4-N removal values (69.69%) obtained in this Vegetation can increase nitrogen and phosphorus uptake by study were lower than those reported by Chang et al. (2007), but increasing HLR (eg, 18e68 mm d 1), but decreases nutrient uptake higher than those reported by Yadav et al. (2018). Weedon (2010) if loading rates are too high (e.g., 135 mm d 1) (Tripathi, Ssrivastava, using an HLR of 0.34 m d 1 obtained higher PRR compared to this & Misra, 1991). There is still discussion about the relationship be- study. The removal percentages of NH4-N, TP, COD and BOD5 in this tween HLR and PRR, because there are no criteria that define the study were greater than those obtained by Yadav et al. (2018) when low and high values of HLR (Chang et al., 2007). Thus, Lin, Jing, Lee, they tested an HLR of 0.15 m d 1. The removal values of TP (50%) of and Wang (2002) considered a high HLR value of 0.135 m d 1; while this study were lower than those reported by Chang et al. (2007); Lin et al. (2005) considered a high HLR value between 1.57 and but higher than those reported by Da˛browski et al. (2017) and 1.95 m d 1. However, Schulz, Gelbrecht, and Rennert (2003) re- Maina, Mutua, and Oduor (2011) who tested with HLR of 0.1 and ported values up to 5410 m d 1. Some variables such as the size and 0.174 m d 1, respectively. distribution of the medium; the growth rate of biofilm; the TSS load rate affect or determine the maximum HLR and therefore the ex- 3.4. Results analysis istence of floods (Cooper, 2005). Weedon (2010) operated without floods in the range of 0.033e1.027 m d 1, but was flooded at 3.4.1. On-site parameters 1.27 m d 1. On the other hand, Johansen, Brix, and Arias (2002) The pH values shown in (Fig. 3a) indicate a slight decrease of the tested with HRL in the range of 0.20e1.20 m d 1, but had floods pH in the effluent with respect to the tributary, which may be at 0.80 m d 1. Stefanakis and Tsihrintzis (2012) used three hydraulic related to the production of organic substances that acidify the load rates: 0.19, 0.26 and 0.44 m d 1 in the long term in ten vertical medium, generated inside the wetland during the growth and flow wetlands, to treat synthetic wastewater. The authors showed death of the plant, and by the mineralization of organic matter that for each load increase, the systems achieved higher nitrogen (Coleman et al., 2001). These pH variations could be due, to that removal performance. The high HRL and less retention times allow during the day (morning and afternoon) ammonia is produced as a product of the decomposition of nitrogen compounds, contributing Table 4 to the increase in pH, but during the night, due to the release of Comparison of percentages of purification. carbon dioxide, the pH decreases. Therefore, the pH of the effluent

Parameter Cyperus Papyrus (%) Phragmites Australis (%) can be acidic during the night and basic during the day (Vymazal, 2014); in this study all the samples were taken at times from 8 pH 9,25 10,69 a.m. to 9am, so there are remnants of water that have rested at Temperature 1,02 1,56 Alkalinity 60,87 64,40 night. These results indicate the ability of macrophytes such as Conductivity 1,49 5,40 Phragmites Australis and Cyperus Papyrus to slightly modify the pH Suspended solids 32,46 62,85 conditions in the rhizosphere, which are consistent with the results BOD5 80,69 75,39 of Hussein and Scholz (2017) and Kyambadde et al. (2004) COD 69,87 64,78 respectively. Nitriﬁcation can be considered as the main cause of Nitrates 88,67 90,29 þ Ammoniacal Nitrogen 69,69 70,70 the pH decrease in VFCW, because during this process H ions are Total Phosphorus 50,00 49,38 released (Vymazal, 2007) and alkalinity is consumed (Paredes et al., Total Coliforms 98,08 96,02 2007). The average value of the alkalinity in the inﬂuent was Fecal Coliforms 95,61 93,74 1 1 175.84 mg L CaCO3, reducing on average to 68.8 and 62.6 mg L F. García-Avila et al. / International Soil and Water Conservation Research 7 (2019) 286e296 291

Table 5 Percentages of contaminants removal for different HLR used in other studies.

1 HLR (m d )NH4-N (%) TP (%) COD (%) BOD5 (%) Author 0.2 91.77 71.79 36.46 47.5 Chang et al. (2007) 0.34 90.9 56.4 93.8 98.3 Weedon (2010) 0.012 95.1 85.1 92.4 99.6 Weedon (2010) 0.15 54 64 65 Yadav et al. (2018) 0.22 58 61 62 Yadav et al. (2018) 0.1 89.2 30.2 84.5 88.1 Da˛browski et al. (2017) 0.16e0.32 88 92 68 72 Dan, Quang, Chiem, and Brix (2011) 0.174 97.8 47.1 Maina et al. (2011)

Influent Influent Effluent C. Papyrus Effluent C. Papyrus (a) Effluent P. Australis 225 (b) )

3 Effluent P. Australis 7.0 200 6.8 175 150 6.6 125 6.4 pH 100 6.2 75 6.0 50

Alkalinity (mg/L CaCO (mg/L Alkalinity 25 5.8 0 l y y e e l y y e e ly ril ri a a un un uly ri a a un un Ju Ap Ap 3M 7M 0J 4J 8J Ap 3M 7M 0J 4J 8 15 29 1 2 1 2 15 1 2 1 2 Time Time 800 (c) Influent Influent Effluent C. Papyrus (d) Effluent C. Papyrus 160 750 Effluent P. Australis Effluent P. Australis 140 700 120 650 100

600 80 60 550 40 Conductivity (uS/cm) 500 20

Suspended solids (mg/L) 0 450 ril il ay ay e y il il y y e e ly p pr une Jun ul pr pr a n un Ju A 13M 27M 10J 24 8J A A 3Ma 7M 0Ju 4J 8 15 29A 15 29 1 2 1 2

Fig. 3. Concentration of pH (a), alkalinity (b), electrical conductivity (c) and suspended solids (d) in the inﬂuent and the VSSFCW planted with Cyperus Papyrus and Phragmites Australis.

CaCO3 for Cyperus Papyrus and Phragmites Australis respectively higher with respect to the depuration value of the WC with Cyperus (Fig. 3b). The reduction of alkalinity justifies the reduction of pH Papyrus of 32.46% (Fig. 3d). (García-Avila, Ramos-Fernandez, & Zhindon-Ar evalo, 2018b). In According to Crites and Tchobanoglous (1998), solids in sus- addition, reactions of degradation of organic matter where het- pension mainly by physical processes are eliminated, such as erotrophic bacteria produce acetic acid, butyric acid and lactic acid sedimentation and filtration. Filtration occurs by impaction of can also explain the decreases in pH (Paredes et al., 2007). particles in the roots and stems of macrophytes or in gravel parti- The effluent temperature was between 21.8 C and 26.7 C; this cles in subsurface flow systems (Sanmuga Priya & Senthamil temperature is suitable for the efficient removal of organic matter Selvan, 2017). The voids and media grain structure which is (Bohorquez et al., 2017; Perez Villar et al., 2012). clearly visible in gravel, has remarkable impact on suspended solids fl The conductivity is directly proportional to the concentration of and trapping during the ow path (Abdelhakeem et al., 2016; & dissolved solids, the conductivity values are slightly reduced both Knowles, Dotro, Nivala, García, 2011; Tsihrintzis, 2017; Zidan in the WC with Cyperus Papyrus and Phragmites Australis (Fig. 3c). et al., 2015). Confirming that vertical flow wetlands have a relatively low ca- In the elimination of TSS, according to the statistical analysis, fi pacity to eliminate conductivity (Bohorquez et al., 2017; Hussein & there is no signi cant difference between the primary treatment > Scholz, 2017). and wastewater treatment with wetlands (p 0.05).

3.4.2. Totals suspended solids 3.4.3. Organic matter abatement With reference to totals suspended solids, the removal in the For the BOD5 parameter, there was a reduction in the concen- wetland with Phragmites Australis had a value of 62.85%, it was tration in the two wetlands, the values of removal percentages in 292 F. García-Avila et al. / International Soil and Water Conservation Research 7 (2019) 286e296

BOD5 are consistent with that reported by Vera, Martel, and higher denitrification rates (Hernandez & Mitsch, 2007), which is Marquez (2013). However, it can be shown that Cyperus Papyrus why VFSSCW have a low denitrification process. The absence of the showed a removal of 80.69% compared to 75.39% of Phragmites denitrification process is the main obstacle in vertical flow wet- Australis, therefore Cyperus Papyrus has a higher efficiency than lands to achieve greater nitrogen elimination (Abdelhakeem et al., Phragmites Australis (Fig. 4a). Organic material biodegradation 2016). Negative efficiencies were registered in the elimination of takes place in the biofilm (attached microbial population) along the NO3 , this was also reported by Tsihrintzis (2017). The increase of plant roots and stems and the surface of the substrate grains, which the NO3 levels indicated the nitrification of ammonium to nitrate, allows to reduce the BOD5 concentration (Stefanakis et al., 2014). which could be favored by the lower concentrations of organic It is the same case, for chemical oxygen demand COD, Cyperus matter and the consequent greater growth of autotrophic bacteria Papyrus presents a removal of 69.87% and Phragmites Australis (nitrifying). The system did not remove nitrates, probably because 64.78% (Fig. 4b); therefore, with Cyperus Papyrus slightly higher anaerobic conditions predominated in some system area. removal yields are obtained. The results are a bit lower than the For removal of ammoniacal nitrogen (NH4-N), the average concentration reductions reported by Vymazal (2002) for HSSF- values of the percentages obtained in each wetland, provides the CW, and Abdelhakeem et al. (2016) for a VSSF-CW (88%). The following results: for Cyperus Papyrus 69.69% and for Phragmites transformation of the COD is essentially affected by the microor- Australis 70.70% (Fig. 5b). In this case there is no significant dif- ganisms whose presence and activity is carried out by the presence ference between the two plant species, these results are similar to and processes mediated by the wetland plants (Barco & Borin, those obtained by Bohorquez et al. (2017); Barco and Borin (2017). 2017; Bruch, Alewell, Hahn, & Hasselbach, 2014). These results confirm that the VFCW have been used mainly for the In the reduction of BOD5 and COD, according to the statistical treatment of municipal and domestic wastewater, due to their analysis, it was significantly different (p < 0.05) between the pri- greater capacity of nitrification, they have also been used for the mary treatment and the water treatment with wetlands. In as treatment of other types of wastewater, mainly those with a high much, that the treatment between Phragmites Australis and concentration of ammoniacal nitrogen (Stefanakis et al., 2014). Cyperus Papyrus, shows that there is no significant difference be- The Total phosphorus (TP) data obtained in the effluent show tween both (p > 0.05). that the reduction of phosphorus is effective using both plants, observing very similar results between Cyperus Papyrus 50% and Phragmites Australis 49.38% (Fig. 5c). 3.4.4. N and P concentrations and contents These values are somewhat lower than those obtained by The concentration of nitrites NO in the effluent was greater 3 (Kyambadde et al., 2004; Vymazal, 2010). Phosphate removal effi- than in the influent for both cases (Fig. 5a). These data show that ciencies were significantly higher in surface flow constructed the concentration of NO in the effluent of the VSSFCW seeded 3 wetland (SFCW) than subsurface flow constructed wetland with the two species of plants increased slightly, producing nega- (SSFCW) (Hernandez, Galindo-Zetina, & Hernandez Hernandez, tive removal percentages. This reflects that conditions existed for 2017). The phosphorus transformations in the wetlands are the nitrification. These low results for the elimination of NO reflect the 3 adsorption and absorption by the substrate, being the mechanism absence of favorable conditions for their elimination by the well that governs the removal of this pollutant, without ruling out the aerated VSSFCW wetland. Nitrates by reducing them to nitrogen chemical precipitation, fragmentation and leaching, mineralization gas due to the denitrification process are eliminated. This process and burial (Vymazal, 2011). occurs in the presence of an organic substance available only in In the reduction of NH -N and TP, according to the statistical anaerobic and anoxic conditions, where nitrogen is used as an 4 analysis, it was significantly different (p < 0.05) between the pri- electron acceptor instead of oxygen (Abdelhakeem et al., 2016; mary treatment and the water treatment with wetlands. In as Kyambadde et al., 2004). The anaerobic conditions required for the much, that the treatment between Phragmites Australis and initiation of NO reduction are not met in the VSSFCW. These re- 3 Cyperus Papyrus, shows that there is no significant difference be- sults are consistent with those obtained by Vymazal (2010) and tween both (p > 0.05). Stefanakis et al. (2014), who found that constructed vertical flow wetlands successfully eliminate ammonia N, but very limited denitrification occurs. The VSSFCW offers good oxygen re- 3.4.5. Microbiological analysis quirements for the nitrification of NH4 but unfavorable conditions In relation to total coliforms (TC), the percentages of removal of for the denitrification of NO3 . Permanently flooded wetlands have the wetland with Cyperus Papyrus were 98.08% and Phragmites

Fig. 4. Concentration of BOD5 (a) and COD (b) in the inﬂuent and the VSSFCW planted with Cyperus Papyrus and Phragmites Australis. F. García-Avila et al. / International Soil and Water Conservation Research 7 (2019) 286e296 293

Influent Influent Effluent C. Papyrus 12 (a) Effluent C. Papyrus (b) Effluent P. Australis 11 Effluent P. Australis 40 10 9 35 8 30 7 25 6 5 20 4 15

Nitrates (mg/L) 3 10 2 1 5 Ammoniacal nitrogen (mg/L) Ammoniacal nitrogen 0 0 y y e e ril ay ay ne ne ly ril ril un un uly p M M Ju Ju Ju Ap Ap 3Ma 7Ma 0J 4J 8J 5A 13 27 10 24 8 15 29 1 2 1 2 1 Time Time Influent 12 (c) Effluent C. Papyrus Effluent P. Australis 11 10 9 8 7 6 5

Total Phosphorus (mg/L) 4 3 il ay ay ne ne ly pr M M Ju Ju Ju 5A 13 27 10 24 8 1 Time

Fig. 5. Concentration of Nitrates (a), ammonia nitrogen (b) and total phosphorus (c) in the inﬂuent and the VSSFCW planted with Cyperus Papyrus and Phragmites Australis.

Australis 96.02%. Of the two species, the greatest removal of total Fig. 7. The values obtained for all parameters are according to the coliforms was slightly greater with Cyperus Papyrus (Fig. 6a). The Ecuadorian regulations for the disposal. decline of the coliform population is due to filtration, sedimenta- In this study, native species Phragmites Australis and Cyperus tion, adsorption, absorption of vegetation. In reference to fecal Papyrus tolerated very well the treatment conditions in the con- coliforms (FC), Cyperus Papyrus presents 95.61% and Phragmites structed wetlands, showing high growth and biomass, allowing the Australis 93.74% removal (Fig. 6b), results consistent with that ob- plants to reach their potential to improve the treatment. tained by Kivaisi (2001). The CW are efficient and profitable sys- In relation to BOD5 and COD, Cyperus Papyrus showed greater tems to eliminate pathogens, the extraction mechanisms are not removal compared to Phragmites Australis; discarding the hy- well understood, but may include: physical (filtration, sedimenta- pothesis of this study, that the wetlands with Phragmites Australis tion); chemical (oxidation, UV irradiation, adsorption to OM, etc.); would have greater efficiency than the huedas with Cyperus and biological (depredation by other microorganisms, retention in Papyrus. This greater removal is due to the structure of the stems, biofilm, natural death, etc.) (Tsihrintzis, 2017). the Cyperus Papyrus possesses the porous structure; while in the The elimination efficiency of TC and FC was significantly Phragmites Australis, the greater stems do not have parenchyma. different (p < 0.05) between the primary treatment and the treat- Due to the type of root, Cyperus Papyrus does not develop deep ment with wetlands. There is no significant difference between roots and forms a kind of networks that helps a greater coverage of both wetlands (p > 0.05). the root area, which in turn allows a greater oxygenation of the For compliance with Ecuadorian regulations, in the case of TC system. In the Phragmites Australis case, the roots develop verti- and FC does not comply with the regulations, due to the high mi- cally, as they do not occupy a greater horizontal area, increasing the crobial load in the tributary. However removal percentages greater hydraulic conductivity compared to Cyperus Papyrus; what makes than 90% are evident (Fig. 7). It is observed that for parameters pH, the purification in deeper layers therefore does not guarantee conductivity, total dissolved and suspended solids, BOD5, COD, ni- removal of contaminants due to the size of the granular material. trates and total phosphorus if it complies with the Ecuadorian regulations. 4. Conclusions It was obtained a significant reduction of pollutants in the effluent wastewater after eight weeks of operation of the vertical The elimination efficiency of BOD5, COD, total coliforms, fecal subsurface wetland. The percentages of reduction of all the coliforms, ammonia nitrogen and phosphates was at 80.69, physical-chemical and microbiological parameters are presented in 69.87, 98.08, 95.61, 69.69 and 50.0% for Cyperus Papyrus 294 F. García-Avila et al. / International Soil and Water Conservation Research 7 (2019) 286e296

Influent Influent )

L Effluent C. Papyrus 11.5 (a) Effluent C. Papyrus Effluent P. Australis Effluent P. Australis 11.0 10.5 (b) 10.5 10.0 MPN/100 mL) MPN/100 10.0 100 m MPN/ 10 10 9.5 g 9.5 9.0 9.0 8.5 8.5 8.0 8.0 7.5 7.5

l y e e y (Lo Coliforms Fecal pri ay a n n ul il y y e e y A M Ju Ju J pr a a un un ul

Total Coliforms (Log Total Coliforms 8 15 13 27 M 10 24 A M M J J 8 J 15 13 27 10 24 Time Time

Fig. 6. Logarithm (base-10) of the most probable number (MPN/100 mL) of coliforms in the VSSFCW planted with Cyperus Papyrus and Phragmites Australis, determined in the inﬂuent and efﬂuent for each treatment: (a) total coliforms, (b) fecal coliforms.

100 (a) C. Papyrus (b) C. Papyrus 100 90 P. Australis P. Australis 80 80 60 70 40 60 20 - 50 NO3 0 40 -20 COD NH4-N TP TC FC 30 -40 20 -60

10 removal rate (%) Average Average removal rate (%) -80 0 pH Temp. Alkalinity EC TSS BOD520 -100

Fig. 7. Treatment performance of VSSFCW a) Percentages of pH removal, Temperature, Alkalinity, Conductivity, Suspended solids, BOD5 b) Percentages of COD removal, Nitrates, Ammoniacal Nitrogen, Total Phosphorus, Total Coliforms, Fecal Coliforms.

elimination and 75.39, 64.78, 96.02, 93.74, 70.70 and 49.38% for application in the treatment of a variety of contaminants pro- Phragmites Australis respectively. duced through various human activities. Cyperus Papyrus was slightly more efficient in eliminating these parameters, while Phragmites Australis was more efficient in Acknowledgments removing suspended solids. These results allow us to infer that Cyperus Papyrus could be a species of macrophyte ideal for The authors acknowledge to the Municipal GAD of Santa Isabel wetlands built on a large scale, due to its high elimination effi- city for allowing us to carry out the research in the wastewater ciencies. The system did not remove nitrates, probably because treatment plant. anaerobic conditions predominated in some system area. According to the results obtained, the construction of VFCW References planted with Cyperus Papyrus is recommended because it achieves high yields in the elimination of both physicochemical Abdelhakeem, S. G., Aboulroos, S. A., & Kamel, M. M. (2016). Performance of a vertical subsurface flow constructed wetland under different operational con- and biological pollutants present in urban/domestic wastewater ditions. Journal of Advanced Research, 7(5), 803e814. https://doi.org/10.1016/j. especially for small communities. jare.2015.12.002. Another advantage of this plant species is that it can withstand Abou-Elela, S. I., & Hellal, M. S. (2012). Water. Municipal wastewater treatment fl temperatures between 20 and 33 C, that is, it can adapt to warm using vertical ow constructed wetlands planted with Canna, Phragmites and Cyprus. Ecological Engineering, 47, 209e213. https://doi.org/10.1016/j.ecoleng. zones; additionally by its capacity of reproduction, it does it in a 2012.06.044. shorter time in comparison to the Phragmites Australis, presents Andreo-Martínez, P., García-Martínez, N., Quesada-Medina, J., & Almela, L. (2017). facility to carry out the pruning, and an important contribution Domestic wastewaters reuse reclaimed by an improved horizontal subsurface- flow constructed wetland: A case study in the southeast of Spain. Bioresource for its landscape value. Under the conditions of this study, a Technology, 233, 236e246. https://doi.org/10.1016/j.biortech.2017.02.123. slight disinfection is required to eliminate the residual pathogen Ansari, A., & Golabi, M. H. (2018). Prediction of spatial land use changes based on LCM in the treated effluent. No odors or insects were detected during in a GIS environment for Desert Wetlands e a Case study: Meighan Wetland, Iran. International Soil and Water Conservation Research. https://doi.org/10.1016/j. the entire VSSFCW operating time. iswcr.2018.10.001. Effective solutions for wastewater treatment have been pre- APHA. (1998). Standard methods for the examination of water and wastewater. sented using promising VSSFCW technology planted with Washington, DC: American Public Health Association. Barco, A., & Borin, M. (2017). Treatment performance and macrophytes growth in a Phragmites Australis and Cyperus Papyrus that have low con- restored hybrid constructed wetland for municipal wastewater treatment. struction cost and low operating requirements, and has Ecological Engineering, 107,160e171. https://doi.org/10.1016/j.ecoleng.2017.07. F. García-Avila et al. / International Soil and Water Conservation Research 7 (2019) 286e296 295

004. treatment and reuse in developing countries: A review. Ecological Engineering, Binder, W., Gottle,€ A., & Shuhuai, D. (2015). Ecological restoration of small water 16(4), 545e560. https://doi.org/10.1016/S0925-8574(00)00113-0. courses, experiences from Germany and from projects in Beijing. International Knowles, P., Dotro, G., Nivala, J., & García, J. (2011). Clogging in subsurface-flow Soil and Water Conservation Research, 3,141e153. https://doi.org/10.1016/j.iswcr. treatment wetlands: Occurrence and contributing factors. Ecological Engineer- 2015.04.004. ing, 37(2), 99e112. https://doi.org/10.1016/j.ecoleng.2010.08.005. Bohorquez, E., Paredes, D., & Arias, C. A. (2017). Vertical flow-constructed wetlands Kyambadde, J., Kansiime, F., Gumaelius, L., & Dalhammar, G. (2004). A comparative for domestic wastewater treatment under tropical conditions: Effect of different study of Cyperus papyrus and Miscanthidium violaceum-based constructed design and operational parameters. Environmental Technology, 38(2), 199e208. wetlands for wastewater treatment in a tropical climate. Water Research, 38(2), https://doi.org/10.1080/09593330.2016.1230650. 475e485. https://doi.org/10.1016/j.watres.2003.10.008. Brix, H., & Arias, C. A. (2005). The use of vertical flow constructed wetlands for on- Laaffat, J., Aziz, F., Ouazzani, N., & Mandi, L. (2016). Biotechnological approach of site treatment of domestic wastewater: New Danish guidelines. Ecological En- greywater treatment and reuse for landscape irrigation in small communities. gineering, 25(5), 491e500. https://doi.org/10.1016/j.ecoleng.2005.07.009. Saudi Journal of Biological Sciences. https://doi.org/10.1016/j.sjbs.2017.01.006. Bruch, I., Alewell, U., Hahn, A., Hasselbach, R., & Alewell, C. (2014). Influence of soil Lin, Y. F., Jing, S. R., Lee, D. Y., Chang, Y. F., Chen, Y. M., & Shih, K. C. (2005). Per- physical parameters on removal efficiency and hydraulic conductivity of ver- formance of a constructed wetland treating intensive shrimp aquaculture tical flow constructed wetlands. Ecological Engineering, 68,124e132. https://doi. wastewater under high hydraulic loading rate. Environmental Pollution, 134, org/10.1016/j.ecoleng.2014.03.069. 411e421. https://doi.org/10.1016/j.envpol.2004.09.015. Chang, J., Zhang, X., Perfler, R., Xu, Q. S., Niu, X. Y., & Ge, Y. (2007). Effect of hydraulic Lin, Y. F., Jing, S. R., Lee, D. Y., & Wang, T. W. (2002). Nutrient removal from aqua- loading rate on the removal efficiency in a constructed wetland in subtropical culture wastewater using a constructed wetlands system. Aquaculture, 209, China. Fresenius Environmental Bulletin, 16(9), 1082e1086. 169e184. https://doi.org/10.1016/S0044-8486(01)00801-8. Coleman, J., Hench, K., Garbutt, K., Sextone, A., Bissonnette, G., & Skousen, J. (2001). Maina, C. W., Mutua, B. M., & Oduor, S. O. (2011). Evaluating performance of vertical Treatment of domestic wastewater by three plant species in constructed wet- flow constructed wetland under various hydraulic loading rates in effluent lands. Water, Air, and Soil Pollution, 128, 283e295. https://doi.org/10.1023/A: polishing. Journal of Water, Sanitation and Hygiene for Development, 1(2), 1010336703606. 144e151. https://doi.org/10.2166/washdev.2011.025. Cooper, P. (2005). The performance of vertical flow constructed wetland systems Mango, N., Makate, C., Tamene, L., Mponela, P., & Ndengu, G. (2017). Awareness and with special reference to the significance of oxygen transfer and hydraulic adoption of land, soil and water conservation practices in the Chinyanja Tri- loading rates. Water Science and Technology : A Journal of the International As- angle, Southern Africa. International Soil and Water Conservation Research, 5, sociation on Water Pollution Research, 51(9), 81e90. https://doi.org/10.2166/wst. 122e129. https://doi.org/10.1016/j.iswcr.2017.04.003. 2005.0293. Metcalf, & Eddy. (1991). Wastewater engineering: Treatment, disposal and reuse (3rd Crites, R., & Tchobanoglous, G. (1998). Small and decentralized wastewater man- Edition). McGraw Hill. agement systems. Boston, MA, USA: McGraw Hill Inc. Molle, P., Lienard, A., Grasmick, A., & Iwema, A. (2006). Effect of reeds and loading Dadban Shahamat, Y., Asgharnia, H., Kalankesh, L., & Hosanpour, M. (2018). Data on operations on hydraulic behavior of vertical flow constructed wetlands under wastewater treatment plant by using wetland method, Babol, Iran. Data in Brief, hydraulic overloads. Water Research, 40, 606e612. https://doi.org/10.1016/j. 16, 1056e1061. https://doi.org/10.1016/j.dib.2017.12.034. watres.2005.11.026. Dan, T. H., Quang, L., Chiem, N. H., & Brix, H. (2011). Treatment of high-strength Orimoloye, I. R., Kalumba, A. M., Mazinyo, S. P., & Nel, W. (2018). Geospatial analysis wastewater in tropical constructed wetlands planted with Sesbania sesban: of wetland dynamics : Wetland depletion and biodiversity conservation of Isi- Horizontal subsurface flow versus vertical downflow. Ecological Engineering, mangaliso Wetland, South Africa. Journal of King Saud University Science. https:// 37(5), 711e720. https://doi.org/10.1016/j.ecoleng.2010.07.030. doi.org/10.1016/j.jksus.2018.03.004. Da˛browski, W., Karolinczak, B., Gajewska, M., & Wojciechowska, E. (2017). Appli- Paredes, D., Kuschk, P., Mbwette, T. S., Stange, F., Müller, R. A., & Koser,€ H. (2007). cation of subsurface vertical flow constructed wetlands to reject water treat- New aspects of microbial nitrogen transformations in the context of wastewater ment in dairy wastewater treatment plant. Environmental Technology, 38(2), treatment - a review. Engineering in Life Science, 7(1), 13e25. https://doi.org/10. 175e182. https://doi.org/10.1080/09593330.2016.1262459. 1002/elsc.200620170. Effendi, H., Widyatmoko, Utomo, B. A., & Pratiwi, N. T. M. (2018). Ammonia and Perbangkhem, T., & Polprasert, C. (2010). Biomass production of Papyrus (Cyperus orthophosphate removal of tilapia cultivation wastewater with Vetiveria ziza- Papyrus) in constructed wetland treating low-strength domestic wastewater. nioides. Journal of King Saud University Science. https://doi.org/10.1016/j.jksus. Bioresource Technology, 101(2), 833e835. https://doi.org/10.1016/j.biortech. 2018.04.018. 2009.08.062. García-Avila, F., Ramos-Fernandez, L., & Zhindon-Ar evalo, C. (2018b). Estimation of Perez Villar, M. M., Domínguez, E. R., Tack, F., Hernandez Ruiz, J. M., Sanchez corrosive and scaling trend in drinking water systems in the city of Azogues, Morales, R., & Arteaga, L. E. (2012). Vertical subsurface wetlands for wastewater Ecuador. Ambiente & Agua - An Interdisciplinary Journal of Applied Science, 13(5), purification. Procedia Engineering, 42, 1960e1968. https://doi.org/10.1016/j. 1e14. https://doi.org/10.4136/ambi-agua.2237. proeng.2012.07.592. García, J., & Corzo, A. (2008). Depuracion con humedales construidos. Guía de practica Ramprasada, C., Smithb, C. S., Memonc, F. A., & Philipa, L. (2017). Removal of de Diseno,~ Construccion y Explotacion de Sistemas de Humedales de Flujo Sub- chemical and microbial contaminants from greywater using a novel con- superficial. Barcelona, Espana:~ Universidad Politecnica de Catalunya. structed wetland: GROW. Ecological Engineering, 106,55e65. https://doi.org/10. García, F., Ramos, L., Pauta, D., & Quezada, D. (2018a). Evaluation of water quality 1016/j.ecoleng.2017.05.022. and stability in the drinking water distribution network in the Azogues city, Sanmuga Priya, E., & Senthamil Selvan, P. (2017). Water hyacinth (Eichhornia Ecuador. Data in Brief, 18,111e123. https://doi.org/10.1016/j.dib.2018.03.007. crassipes) e an efficient and economic adsorbent for textile effluent treatment Hansen, A. M., Kraus, T. E., Bachand, S. M., Horwath, W. R., & Bachand, P. A. (2018). e a review. Arabian Journal of Chemistry, 10, S3548eS3558. https://doi.org/10. Wetlands receiving water treated with coagulants improve water quality by 1016/j.arabjc.2014.03.002. removing dissolved organic carbon and disinfection byproduct precursors. The Schulz, C., Gelbrecht, J., & Rennert, B. (2003). Treatment of rainbow trout farm ef- Science of the Total Environment, 622e623, 603e613. https://doi.org/10.1016/j. fluents in constructed wetland with emergent plants and subsurface horizontal scitotenv.2017.11.205. water flow. Aquaculture, 217(1e4), 207e221. https://doi.org/10.1016/S0044- Harmel, R. D., Pampell, R., Gentry, T., Smith, D. R., Hajda, C., Wagner, K., et al. (2018). 8486(02)00204-1. Vegetated treatment area (VTAs) efficiencies for E. coli and nutrient removal on Shahi, D. H., Eslami, H., Ehrampoosh, M. H., Ebrahimi, A., Ghaneian, M. T., small-scale swine operations. International Soil and Water Conservation Ayatollah, S., et al. (2013). Comparing the efficiency of Cyperus alternifolius and Research, 6,153e164. https://doi.org/10.1016/j.iswcr.2018.02.002. Phragmites Australis in municipal wastewater treatment by subsurface con- Hernandez, M. E., Galindo-Zetina, M., & Hernandez Hernandez, J. C. (2017). structed wetland. Pakistan Journal of Biological Sciences, 16(8), 379e384. https:// Greenhouse gas emissions and pollutant removal in treatment wetlands with scialert.net/abstract/?doi¼pjbs.2013.379.384/. ornamental plants under subtropical conditions. Ecological Engineering, 114, Stefanakis, A., Akratos, C., & Tsihrintzis, V. A. (2014). Vertical flow constructed wet- 88e95. https://doi.org/10.1016/j.ecoleng.2017.06.001. lands: Eco-engineering systems for wastewater and sludge treatment. Elsevier. Hernandez, M. E., & Mitsch, W. J. (2007). Denitrification in created riverine wet- Stefanakis, A. I., & Tsihrintzis, V. A. (2012). Effects of loading, resting period, tem- lands: Influence of hydrology and season. Ecological Engineering, 30(1), 78e88. perature, porous media, vegetation and aeration on performance of pilot-scale https://doi.org/10.1016/j.ecoleng.2007.01.015. vertical flow constructed wetlands. Chemical Engineering Journal, 181e182, Hussein, A., & Scholz, M. (2017). Dye wastewater treatment by vertical-flow con- 416e430. https://doi.org/10.1016/j.cej.2011.11.108. structed wetlands. Ecological Engineering, 101,28e38. https://doi.org/10.1016/j. Tahir, U., Yasmin, A., & Hassan Khan, U. (2016). Phytoremediation : Potential flora for ecoleng.2017.01.016. synthetic dyestuff metabolism. Journal of King Saud University Science, 28, INEN 2176. (1998). Norma Tecnica Ecuatoriana. Agua. Calidad del agua. Muestreo. 119e130. https://doi.org/10.1016/j.jksus.2015.05.009. Tecnicas de muestreo. https://archive.org/details/ec.nte.2169.1998. Talukdar, S., & Pal, S. (2017). Impact of dam on inundation regime of flood plain Johansen, N. H., Brix, H., & Arias, C. A. (2002). Design and characterisation of a wetland of punarbhaba river basin of barind tract of Indo-Bangladesh. Inter- compact constructed wetland removing BOD, nitrogen and phosphorus for national Soil and Water Conservation Research, 5,109e121. https://doi.org/10. single household sewage. In Paper presented at the 8th international conference 1016/j.iswcr.2017.05.003. on wetland treatment for water pollution control, Arusha, Tanzania,September Tripathi, B. D., Ssrivastava, J., & Misra, K. (1991). Nitrogen and phosphorus removal- 2002. capacity of four chosen aquatic macrophytes in tropical freshwater ponds. Kadlec, R. H., Knight, R. L., Vymazal, J., Brix, H., Cooper, P. F., & Haberl, R. (2000). Environmental Conservation, 18,143e147. https://doi.org/10.1017/ Constructed wetlands for water pollution control: Processes, performance, design S037689290002172X. and operation. Scientific and Technical Report No. 8. London: IWA. Tsihrintzis, V. A. (2017). The use of vertical flow constructed wetlands in wastewater Kivaisi, A. K. (2001). The potential for constructed wetlands for wastewater treatment. Water Resources Management, 31, 3245e3270. https://doi.org/10. 296 F. García-Avila et al. / International Soil and Water Conservation Research 7 (2019) 286e296

1007/s11269-017-1710-x. ecoleng.2005.07.005. Vera, L., Martel, G., & Marquez, M. (2013). Two years monitoring of the natural Vymazal, J., & Kropfelov€ a, L. (2011). A three-stage experimental constructed system for wastewater reclamation in Santa Lucia, Gran Canaria Island. wetland for treatment of domestic sewage: First 2 years of operation. Ecological Ecological Engineering, 50,21e30. https://doi.org/10.1016/j.ecoleng.2012.08.001. Engineering, 37(1), 90e98. https://doi.org/10.1016/j.ecoleng.2010.03.004. Vymazal, J. (2002). The use of sub-surface constructed wetlands for wastewater Weedon, C. (2010). A decade of compact vertical flow constructed wetlands. Water treatment in the Czech Republic : 10 years experience. Ecological Engineering, 18, Science and Technology: A Journal of the International Association on Water 633e646. https://doi.org/10.1016/S0925-8574(02)00025-3. Pollution Research, 62, 2790e2800. https://doi.org/10.2166/wst.2010.041. Vymazal, J. (2007). Removal of nutrients in various types of constructed wetlands. Yadav, A., Chazarenc, F., & Mutnuri, S. (2018). Development of the «French system» The Science of the Total Environment, 380(1e3), 48e65. https://doi.org/10.1016/j. vertical flow constructed wetland to treat raw domestic wastewater in India. scitotenv.2006.09.014. Ecological Engineering, 113,88e93. https://doi.org/10.1016/j.ecoleng.2018.01. Vymazal, J. (2010). Constructed wetlands for wastewater treatment. Water, 2, 001. 530e549. https://doi.org/10.3390/w2030530. Zhang, D. Q., Jinadasa, K., Gersberg, R. M., Liu, Y., Jern Ng, W., & Tan, S. K. (2014). Vymazal, J. (2011). Constructed wetlands for wastewater treatment: Five decades of Application of constructed wetlands for wastewater treatment in developing experience. Environmental Science & Technology, 45,61e69. https://doi.org/10. countries - a review of recent developments (2000-2013). Journal of Environ- 1021/es101403. mental Management, 141,116e131. https://doi.org/10.1016/j.jenvman.2014.03. Vymazal, J. (2014). Constructed wetlands for treatment of industrial wastewaters: A 015. review. Ecological Engineering, 73,724e751. https://doi.org/10.1016/j.ecoleng. Zidan, A. R., El-Gamal, M. M., Rashed, A. A., & El-Hady Eid, M. A. (2015). Wastewater 2014.09.034. treatment in horizontal subsurface flow constructed wetlands using different Vymazal, J., & Kropfelova, L. (2005). Growth of Phragmites Australis and phalaris media (setup stage). Water Science, 29(1), 26e35. https://doi.org/10.1016/j.wsj. arundinacea in constructed wetlands for wastewater treatment in the Czech 2015.02.003. republic. Ecological Engineering, 25(5), 606e621. https://doi.org/10.1016/j. International Soil and Water Conservation Research 7 (2019) 297e307

Contents lists available at ScienceDirect

International Soil and Water Conservation Research

journal homepage: www.elsevier.com/locate/iswcr

Original Research Article Technosols on mining wastes in the subarctic: Efﬁciency of remediation under Cu-Ni atmospheric pollution

* Marina V. Slukovskaya a, , Viacheslav I. Vasenev b, Kristina V. Ivashchenko b, c, Dmitry V. Morev d, Svetlana V. Drogobuzhskaya a, Liubov A. Ivanova e, Irina P. Kremenetskaya a a I.V. Tananaev Institute of Chemistry and Technology of Rare Elements and Mineral Raw Materials, Kola Science Centre, Russian Academy of Sciences, 184209, Academgorodok 26a, Apatity, Murmansk Region, Russia b Department of Landscape Design and Sustainable Ecosystems, RUDN University, 117198, Miklykho-Maklaya str. 8, Moscow, Russia c Institute of Physicochemical and Biological Problems in Soil Science, Russian Academy of Sciences, 142290, Institutskaya str., 2, Pushchino, Moscow Region, Russia d Department of Ecology, Russian State Agrarian University e Moscow Timiryazev Agricultural Academy, 127550, Timiryazevskaya str. 49, Moscow, Russia e N.A. Avrorin Polar-Alpine Botanical Garden-Institute, Kola Science Centre, Russian Academy of Sciences, 184209, Academgorodok 18a, Apatity, Murmansk Region, Russia article info abstract

Article history: The copper-nickel factory's emissions in the Murmansk region, Russia, led to the degradation of plant Received 24 August 2018 cover and topsoil with the subsequent formation of industrial barrens. In this study, the industrial Received in revised form barrens were remediated by means of Technosol engineering, when grasses were sown on the two 12 April 2019 different types of mining wastes (carbonatite and serpentinite-magnesite) covered by hydroponic Accepted 28 April 2019 vermiculite. The serpentinite-magnesite waste was significantly different from the carbonatite waste in Available online 30 May 2019 the content of silicon (Si) and manganese (Mn), pH, and texture. Both wastes had an alkaline pH level and high content of calcium (Ca) and magnesium (Mg). The vegetation and Technosol properties at the Keywords: Industrial barrens remediated sites were analyzed in 2017 and compared to the initial state (2010 year) to assess the effi Metals ciency of the long-term remediation. The quality and sustainability of Technosols based on the Phytostabilization serpentinite-magnesite wastes were substantially higher compared to the carbonatite-based Technosol. Soil organic carbon Biomass and a projective cover of the grass community depended on Si content in the original mining Basal respiration waste and were found to be higher in the serpentinite-magnesite Technosol. The content of organic Pedogenesis carbon and its fractions, microbial biomass and basal respiration after seven years of Technosol evolution was comparable to natural values. These parameters were directly related to plant cover state and were inversely proportional to copper (Cu) content in Technosol. The Technosol development led to the reduction of nickel (Ni) and Cu migration in soil-plant ecosystems due to neutralization and adsorption properties of mining wastes and phytostabilization by underground parts of grass communities. The Technosol development in its early stage of pedogenesis indicates the efficiency of applied remediation technology to the degraded acidic soil under the conditions of industrial atmospheric pollution. © 2019 International Research and Training Center on Erosion and Sedimentation and China Water and Power Press. Production and Hosting by Elsevier B.V. This is an open access article under the CC BY-NC- ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction

Environmental pollution caused by emissions from mining and producing industries is a common ecological problem for industrial * Corresponding author. regions around the globe (Ajmone-Marsan & Biasioli, 2010; Pacyna, E-mail addresses: [email protected] (M.V. Slukovskaya), vasenyov@ Pacyna, & Aas, 2009; Peuelas & Filella, 2002). The main pollutants mail.ru (V.I. Vasenev), [email protected] (K.V. Ivashchenko), dmorev@ rgau-msha.ru (D.V. Morev), [email protected] (S.V. Drogobuzhskaya), typically are copper (Cu), nickel (Ni), zinc (Zn), lead (Pb) and some [email protected] (L.A. Ivanova), [email protected] other metals, whose high contents could pose risks for human and (I.P. Kremenetskaya). https://doi.org/10.1016/j.iswcr.2019.04.002 2095-6339/© 2019 International Research and Training Center on Erosion and Sedimentation and China Water and Power Press. Production and Hosting by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). 298 M.V. Slukovskaya et al. / International Soil and Water Conservation Research 7 (2019) 297e307 environment (Morel, Chenu, & Lorenz, 2015; Aguilar et al., 2018; (Archegova & Likhanova, 2012; Huot, Simonnot, & Morel, 2015). Kumar, Gogoi, Kumari, & Borah, 2016). Due to high toxicity and This study aimed to investigate changes in the properties of persistence, long-term sedimentation and accumulation of these Technosols engineered on the mining wastes after seven years of metals in soils result in ecosystem degradation (Ali, Malik, development. The following hypotheses were tested: 1) engineer- Shinwari, & Qadir, 2015; Jiang et al., 2017; Laumbach & Kipen, ing Technosols on the mineral mining wastes is an effective 2012). The main factors causing ecosystem degradation in the approach for metal pollutants' phytostabilization in the Subarctic metal-polluted lands are soil toxicity, degradation of organic mat- zone; 2) the efficiency of phytostabilization differs depending on ter, and disruption of macronutrient cycles (Antoniadis et al., 2017; the mining wastes’ mineral composition, and 3) soil-forming pro- Kabata-Pendias, 2000; Manninen, Zverev, Bergman, & Kozlov, cesses in the engineered Technosols occurring during seven years 2015). Functions and ecosystem services of such soils are highly of the experiment result in improved soil quality and show signs of degraded and partly lost (Ding et al., 2018; Morel et al., 2015). sustainable ecosystem development. Heavily polluted soils lose the capacity to support sustainable vegetation cover leading to the development of industrial barrens, 2. Materials and methods which is a final digression stage of a vegetative succession where plant cover is less than 10% (Ganicheva, Lukina, Kostina, & Nikonov, 2.1. Research area 2004; Kozlov & Zvereva, 2007). Industrial barrens are not rare in high altitudes and are abundant in the Russian Subarctic (Kozlov & The study was conducted near the Cu-Ni factory of ‘Kola Mining Zvereva, 2007; Manninen et al., 2015). and Metallurgical Company’ (Kola MMC, Monchegorsk production Industrial development of the Russian Subarctic zone has star- site), the Murmansk region, Russia. Town of Monchegorsk locates in ted in the first half of the 20th century and has resulted in air, water the Subarctic zone (forest tundra) where the growing season lasts and soil pollution in the region (Abakumov, Lupachev, & Andreev, from June to August (AMAP, 1998). The mean temperature 2017; Moskovchenko, Kurchatova, Fefilov, & Yurtaev, 2017). The is 13.8 C in January and þ14.1 C in July, the annual precipitation formation of new industrial barrens is mainly driven by intense is 561 mm; the frost-free period ranges from 50 to 100 days aerial emission of metal dust and aerosols with further deposition (Manninen et al., 2015). Ferric Podzol is a dominant soil type in the on soils. The extremely high content of metals in soil lead to region (Kashulina, Kubrak, & Korobeinikova, 2015). degradation of vegetation, resulting in water and wind erosion and ‘Kola MMC’ of Monchegorsk has been operating since 1938. depletion of organic carbon (C) in soils. High vulnerability of sub- Today, it is the largest non-ferrous enterprise in the North-West of arctic soils and ecosystems to such kind of pollution raises public Russia, which produces Ni powder and pellets, Ni combines and full awareness on the ecological sustainability of the region and stim- plate cathodes, Cu cathodes, electrolytic cobalt (Co) and other ulates development and implementation of appropriate technolo- metal-based products (Norilsky Nickel, 2018). In this regard, ‘Kola gies for soil remediation at industrial barrens (Moiseenko et al., MMC’ is the main source of pollution in the region (Report …, 2006; Koptsik et al., 2014; 2016). 2017). Processing of Ni-Cu ores has a permanent impact on soils Phytoremediation is a principal approach to restore eroded and through atmospheric emission of gases and dust produced by polluted soils of industrial barrens. Cereals with fast biomass smelters (Kashulina, 2017; Slukovskaya et al., 2017). Continuous growth are widely used for phytoremediation purposes because of emission and deposition of Cu and Ni for eighty years resulted in their high capacity to accumulate metals in the plant tissues the accumulation of more than 6000 mg kg 1 Cu and 9000 mg kg 1 (Salinas, Villavicencio, Bustillos, & Aragon, 2012; Koptsik, 2014). Ni in the upper organic soil horizons (Kashulina, 2017). The previ- Among different phytoremediation approaches, phytostabilization ous investigations report a strong degradation of soils in the area, has been reported as highly effective for the long run (Shutcha et al., including metal contamination, transformation of soil profiles, 2010; Koptsik, 2014; Slukovskaya et al., 2017; 2018). In the condi- changes in water regime and С and N cycles, and secondary tion of high pollution, engineering of new soils, referred as Tech- pollution of terrestrial and water ecosystems (Evdokimova, nosols (WRB 2015; Morel et al., 2015), is necessary to support Mozgova, & Kalabin, 2011; Kashulina et al., 2015; Lyanguzova, phytostabilization and to set up a new plant succession for soil Goldvirt, & Fadeeva, 2015; Manninen et al., 2015). restoration. The Technosols are substantially different from the natural soils of the region by parent material; both plant cover and 2.2. Site description and experimental design designed soil properties of the Technosols are expected to sustain the current and continuous pollution. Materials used for Tech- The experimental site locates 1.5 km from the Cu-Ni factory on nosols’ construction should buffer and absorb metal compounds, the top of the hill (6756.4030N, 3250.2870E; 143 m above sea reducing their toxicity for vegetation (Koptsik, 2014; Mahar et al., level). In 2010, the remediation plots were established (Fig. 1). The 2016; Sharma, Tiwari, Hasan, Saxena, & Pandey, 2018). Wastes of Ferric Podzol at the experimental site at that time was heavily mining industries usually possess the required properties, i.e., an eroded, and plant cover was degraded. In result, organic (O) and alkaline medium, high sorption capacity for the majority of metals podzolic (E) horizons in the profile were almost lost, and the and high contents of macronutrients. Therefore, such wastes can be remaining topsoil could be defined as an AleFe-humus horizon used for Technosol construction as a part of aided phytor- with gravel and rocks. On the top of the eroded Ferric Podzols, two emediation (Gruszecka-Kosowska, Baran, Wdowin, & Franus, 2017; types of Technosols were engineered, each containing two layers: Radziemska, 2018; Slukovskaya et al., 2017). the top layer from hydroponic vermiculite (1 cm depth) and the The efficiency of remediation of the metal-polluted soils is sublayer from mining wastes (5 cm depth) (Ivanova et al., 2013). traditionally assessed by reduction of metals' labile content, Two different types of mining wastes were studied as a material for reduction of toxicity and migration of metals in comparison to the sublayer construction such as carbonatite waste (CW) and initial state (Sharma et al., 2018). The development of soil-forming serpentine-magnesite waste (SM). The hydroponic vermiculite was processes (e.g., an increase of macronutrients’ stocks, C accumula- used for the topsoil engineering of both Technosols in a similar tion and shifting pH from acid to neutral) in the Technosols engi- amount to increase the production of grass cover at the early stages neered for the remediation purposes can be used as an important of ontogenesis, which is an essential approach for phytor- integral indicator of the efficient remediation, which can be emediation in the subarctic conditions where the vegetation period considered as the first step of sustainable ecosystem development is quite short (Ivanova, 2011; Ivanova, Gorbacheva, Slukovskaya, M.V. Slukovskaya et al. / International Soil and Water Conservation Research 7 (2019) 297e307 299

Fig. 1. Location and view of the experimental plots (near the JSC Kola MMC factory in Monchegorsk, the Murmansk Region, Russia) with constructed Technosols (CW, carbonatite waste; SM serpentine-magnesite waste) after 7 years of the soil development.

Kremenetskaya, & Inozemtseva, 2014). The eroded Ferric Podzol was performed by DAK 100 autoclaves (Berghof, Germany) after covered by the 1-cm layer of hydroponic vermiculite (without microwave digestion in an SW4 system. Hydrochloric and HNO3 mining wastes in the profile) was as a control plot. acids in the ration of 3:1 were applied for preliminary soil Five 1 m2 experimental plots (spatial replicates) were estab- decomposition. Soil samples were extracted with ammonium- lished for each type of Technosol and the control plots (total of all acetate buffer solution (pH ¼ 4.65) to measure phytoavailable the plots was 15) on the flat homogeneous area. The plots located at forms of elements (Carter & Gregorich, 2007). In soil samples, total a 0.5 m-distance from each other. The seed mixture of Festuca rubra and phytoavailable forms of magnesium (Mg), calcium (Ca), silicon L., F. pratensis Huds., Bromus inermis Leyss., x Festulolium F. Aschers. (Si), potassium (K), manganese (Mn), iron (Fe), Ni, Cu, Co, and zinc et Graebn. (4:3:1:1) was sown at each plot. Complex NPK fertilizers (Zn) contents were determined. In roots, spikes and leaves of grass, (16:16:16, 60 g m 2) were applied once per year (2010e2016) the content of Ca, Mg, Si, Cu, Ni, and Co was analyzed by a Perki- except for the last year (2017). nElmer ELAN 9000 DRC-e Inductively Coupled Plasma Mass Spec- trometer (ICP-MS). Multielement calibration solutions Inorganic 2.3. Sampling and analysis Venrures (USA) (IV-STOCK-29, IV-STOCK-21, IV-STOCK-28) were used for the instrumental calibration. For quality control, standard Sampling campaign was carried out in August 2017. The Tech- reference materials were used: CRM-SOIL-A (High purity standards, nosols were sampled by the monolith method, and 1000 cm3 of soil USA), GSO 7126-84 Bil-1, GSO 902-76 SP-2, GSO 903-76 SP-3, SDPS, was collected (Bohm,1979).€ Monoliths were transported to the Kola SKR-1 (Russia). Science Centre analytical lab, separated into two layers (0e5 and Total soil content of C and N were determined using an element 5e10 cm) and air-dried at 22 C prior analysis. The reference Ferric analyzer (Leco CHN-932). Organic C was determined by dichromate Podzol was sampled by the envelope method (corners and center) oxidation. Fulvic and humic acids content was determined by a 1 from 5 cm depth, and composited samples were prepared. The sodium pyrophosphate extraction (0.1 mol L Na4P2O7 10 H2O height of grass sampled from monolith was measured and above- mixed with 0.1 mol L 1 NaOH, pH ¼ 13). Free and bounded with Ca ground biomass was collected from one-quarter of plots by cutting and moveable sesquioxideshumic substances completely dissolve with further measurements of air-dry weight. in pyrophosphate solution. Separation of these two fractions provides an additional determination of humic substances extracted 2.3.1. Mineralogical and radionuclides analysis from undecalcified soil with 0.1 mol L 1 NaOH. X-ray phase analysis to determine primary minerals in the mining wastes was done on a Thermo Scientific ARL X’TRA Powder 2.3.3. Microbiological analysis X-ray Diffraction System diffractometer at irradiation (Cu The microbial biomass C (MBC) was measured by the substrate- Ka ¼ 1.54 Å) in Bregg-Brentano geometry. The activity of radionu- induced respiration (SIR) method (Ananyeva, Susyan, Chernova, & clides was measured using gamma-ray scintillation spectrometer Wirth, 2008; Anderson & Domsch, 1978). The method is based on with PROGRESS software. Radiation-hygienic assessment of min- the maximal initial response (CO2 production) to glucose addition eral materials was conducted according to the radiation safety (10 mg g 1). Soil samples with added glucose were incubated for standards (SanPiN, 2009). 3e5 h at 22 C. Carbon dioxide production was measured by gas chromatography (KrystaLLyuks4000 M, ‘Meta-Chrom’ manufac- 2.3.2. Soil and biomass physicochemical analysis turer, Russia) equipped with a thermal conductivity detector. The Soil texture was analyzed by dry sieving through meshes with soil SIR rate was calculated considering CO2 content, volume gas sizes of 2.5, 1.0, 0.6, 0.3 and 0.2 and < 0.2 mm. The preparation of phase in a flask, and incubation time. The value of the SIR rate was soil and plant samples for determination of total elements content used to estimate soil MBC (Anderson & Domsch, 1978). Basal 300 M.V. Slukovskaya et al. / International Soil and Water Conservation Research 7 (2019) 297e307 respiration (BR) was measured in soil samples after 24 h incubation of smaller particles (<1 mm). More than 50% of the SM's particles at 22 C with 0.1 mL g 1 soil water added. The ratio of BR to MBC or originated from coarse sand, whereas fine sand dominated in CW. microbial metabolic quotient (qCO2) and MBC/SOC ratio were Serpentine-magnesite contained insignificant amounts of natural calculated. radionuclides Ra-226, Th-232, and K-40, while CW included natural Prior to the microbiological analysis, air-dried soil samples were radionuclides U-238, Th-232, and K-40. Technogenic radionuclide moistened up to 55% of water holding capacity and pre-incubated Cs-137 was observed in both mining wastes. The specific activity of 1 (soil ~200 g, 22 С, 7 d, air exchange) to avoid an excess CO2 pro- radionuclides in wastes was on average 330 Bq$kg (CW) and 15 duction by disturbance during the preparation procedures Bq$kg 1 (SM), which allows handling these substrates without (Ananyeva et al., 2008; Creamer et al., 2014). restrictions according to the existing regulations (SanPiN, 2009).

2.3.4. Statistical analysis 3.2. Efficiency of the seven-year remediation experiment: a The measurements were performed in 3e5 replicates per comparative assessment of engineered technosols on two types of sample to determine chemical and microbial properties. The results the mining wastes were calculated for dry soil (105 C, 8 h). Welch's t-test was used to check statistically significant differences in the properties between 3.2.1. Physical and chemical properties of the technosols the two Technosol groups. The relationship between Technosol Implementation of the investigated waste materials for Tech- properties was analyzed through Pearson's correlation coefficient nosols construction had a different effect after seven years of and estimated by multiple regressions. The significance of differ- remediation. The pH of the Technosol’ top layer decreased from 9.2 ence was reported based on the 0.05 p-level. The experimental data to 8.0 for CW and from 8.4 to 7.9 for SM. The pH in the sublayer of were statistically processed and visualized using R 3.2.4 software. Technosol (degraded topsoil) increased to 6.8 for CW and 7.4 for SM Principal component analysis (PCA) was applied using the PC-ORD compared to the initial pH of 4.6 in the Al-Fe horizon in 2010 software. (Table 1). The reduction potential of CW Technosol was significantly higher (on average 173 mV) compared to SM (on average 152 mV). 3. Results The content of total and phytoavailable forms of the studied chemical elements also differed between the Technosols and 3.1. Initial conditions of the wastes and eroded topsoil depended on the mineralogical composition of the wastes (Table 2). In the top layer of CW Technosol, the total contents of Si, K, Ca, Mn, Mineral wastes were chosen for the field experiment due to Cu and Zn were on average higher than in SM, however, the content their high alkalizing properties and nutrient content. The wastes of Mg, Fe, and Ni was significantly less (p < 0.05). Phytoavailable differed by the origin and chemical properties. Both wastes were content of Mg, K, Ni, and Si was on average 2.3, 2.7, 2.2 and 1.6 times strongly alkaline, and pH of CW was almost one unit higher less in the top layer of the CW Technosol compared to SM, whereas compared to SM. The content of phytoavailable Cu in CW the contents of phytoavailable Ca, Mn, Fe and Cu were significantly (7.2 mg kg 1) was twenty times higher compared to SM higher (p < 0.05). (0.2 mg kg 1), whereas a similar content of phytoavailable Ni was After 7 years, a slight increase for the total content of Co (by the found in both wastes (Table 1). Total and organic C content in the factor of 2) and Fe (by the factor of 4e5) was observed, whereas the initial CW waste was 100 and 23%, respectively, higher than in SM, content of Zn and Mn remained unchanged. The content of phy- whereas N was on average six times lower. Percent of organic C in toavailable Cu was an order of magnitude higher than that of Ni in total C content was 21 and 38% for CW and SM, respectively. The the top layer and 2.5 times higher in the sublayer. Percentage of sum of humic and fulvic acids gave 60% of the organic C stocks in phytoavailable forms of Cu and Ni in the top layer of the CW the SM waste and only 42 and 48% in the CW waste and in the Al-Fe Technosol was twice higher than in the sublayer, whereas no dif- horizon of the degraded soil, respectively. The physical properties ference between the layers of the SM Technosol was observed. of the applied wastes also differed (Fig. 2). Serpentine-magnesite Total Cu and Ni content increased on average 1.3e2.2 and included mainly particles >1 mm in size, whereas CW consisted 6.2e7.6 times for the engineered Technosols after the seven-year

Table 1

Selective chemical properties of the mining wastes (carbonatite waste (CW) and serpentine-magnesite (SM)) and the eroded topsoil (0e5 cm). Note: organic carbon (Corg), carbon of fulvic (CFA) and humic (CHA) acids.

Properties Carbonatite waste (CW) Serpentine-magnesite (SM) Eroded topsoil

Reference to the aGroup of artifabricats, the sub-group of artiindustrats aGroup of naturfabricats, the subgroup of b Ferric Podzol classiﬁcation lithostrats Origin Wastes of the secondary ore processing Overburden rock Industrially polluted and eroded soil Location The Kovdor Mining and Processing Factory JSC, the The Khalilovskoye magnesite deposit, the Industrial barren, the Murmansk Murmansk region, Russia Orenburg region, Russia region, Russia Primary minerals Calcite, Forsterite, Dolomite Serpentine, Magnesite e

pHw 9.2 8.4 4.6 Cu phytoavailability, % 2.0 0.1 61 Ni phytoavailability, % 6.0 5.0 1.5 C, % 3.46 1.59 1.36 N, % 0.01 0.08 0.07 C/N 293.0 30.0 21.3

Corg,% 0.74 0.60 1.34 CHAþFA,% 0.31 0.36 0.65 CHA,% ND ND 0.35 a Classiﬁcation of Technogenic surface formations according to Shishov, Tonkonogov, Lebedeva, & Gerasimova, 2004; Prokofeva & Vasenev, 2017. b Classiﬁcation of Soils according to WRB, 2015. M.V. Slukovskaya et al. / International Soil and Water Conservation Research 7 (2019) 297e307 301

Fig. 2. The granulometric composition of the serpentine-magnesite (SM) and carbonatite wastes (CW), residue on the sieves with different mesh size (% of weight).

Table 2 Mean content of elements (total/phytoavailable forms, mg$kg 1) in Technosols on carbonatite waste (CW) and serpentine-magnesite waste (SM).

Elements CW SM CW SM

0e5cm 5e10 cm

Mg 90,608/2460 152,801/5736 55,539/1842 137,233/5277 Ca 131,889/80,240 6207/2103 85,753/49,947 7097/1556 Si 5766/214 582/351 5616/283 840/344 K 3757/369 1353/995 3109/463 1347/773 Mn 1126/190 527/54 803/132 538/50 Fe 39,184/606 44,109/70 35,329/471 42,506/82 Ni 1127/15 1697/33 1400/40 1527/32 Cu 721/133 317/80 915/281 350/88 Co 73/5 84/3 70/3 73/3 Zn 55/7 43/5 57/5 47/0

formation period compared with its content after the first year of the experiment (Slukovskaya et al., 2018), indicating on-going pollution of the sites. At the same time, phytoavailable content of Cu in both layers decreased on average by two times. Percent of phytoavailable Ni increased slightly for the seven-year period but did not reach 3% of total content (Fig. 3). The principal component analysis (PCA) was able to explain 71 and 16% of the total variance in the chemical elements of the upper layer of Technosols (Fig. 4). Calcium, Mn, and Fe contributed strongly to the main component Axis 1, while Axis 2 had strong positive loadings for S and Si. PCA allowed observing a substantial difference between CW and SM Technosols regarding Ca, Mn, and Fe content. Also, differences in copper and nickel accumulation by Fig. 3. Phytoavailable of Cu and Ni (in %) in relation to the total content of the elements in Technosols on carbonatite waste (CW) and serpentine-magnesite (SM) waste in the Technosols are shown. second (2011) and seventh (2017) years of the experiment. A considerable increase in C content of the sum of humic and fulvic acids (on average 1.7 and 1.6 times, respectively) and its increase in organic C stocks (up to 52 and 39%) in the top layers were efficiency of soil C-use (qCO2 decreasing) and its availability to found for the CW and SM Technosols. The CHA/CFA ratio was above microorganisms (MBC/Corg increasing) were higher in the SM 1.0 for both Technosols (Table 3). In the sublayer of CW and SM Technosol compared to CW (Fig. 5). A similar tendency was found Technosol (the Al-Fe horizon of degraded soil), organic C also for the sublayers of Technosols. The values of MBC, BR and MBC/Corg increased up to 40 and 25% compared to the initial state. However, for SM Technosol were high and not significantly different between an accumulation of humic and fulvic acids was less intensive the top- and sublayers (p > 0.05). The available Cu content in e compared to the topsoil with the increase of CHA only by 18% in Technosols explained up to 78 95% of the total variance of the seven years of the remediation experiment. Nitrogen content the microbial properties (Fig. 6). The high content of available Cu 1 SM Technosol (0e5 cm) was 1.5 times higher than in the CW. The C/ (>132 mg kg ) in Technosol decreased MBC, BR, and MBC/Corg on N ratio of 19 and 17 estimated for the top- and sublayers of the SM average by 1.4e2.5 times, and increased qCO2 on average by 1.8 Technosol, respectively, indicated more favorable conditions for times. However, the influence of Ni on the microbial properties was plant growth compared to the very high values obtained for the CW not observed. layer (51 and 34% for top- and sublayers, respectively).

3.2.3. Plant cover status 3.2.2. Microbial properties The quantity and quality of biomass substantially increased after In the SM Technosol, the content of MBC and BR rate was on seven years of the remediation efforts. The SM Technosols provided average 1.7 and 1.4 times higher compared to CW. Moreover, the better conditions for plant growth, compared to CW: total grass 302 M.V. Slukovskaya et al. / International Soil and Water Conservation Research 7 (2019) 297e307

Fig. 4. Principal component analysis (PCA) for chemical elements in the upper layer of the Technosols on the carbonatite waste (CW) and serpentine-magnesite (SM) waste.

Table 3 Copper content was higher on average by the factor of 1.5 in the Total (C) and organic carbon (Corg), fulvic (CFA), humic (CHA) acids, total nitrogen (N) plant roots of the CW Technosol compared to SM. Nickel content in e e contents in the top (0 5 cm) and sub (5 10 cm) layers of the Technosols on car- different parts of plants was similar for CW and SM Technosols. For bonatite waste (CW) and serpentine-magnesite (SM) waste. both Technosols, the majority of pollutants (46e85%) was accu- Properties CW Technosol SM Technosol mulated in the roots. top sub top sub 3.3. Soil and ecosystem development after remediation C 4.22 2.70 2.31 2.02 N 0.08 0.08 0.12 0.11 C/N 50.9 33.5 19.3 17.0 Although the remediation effect of the implemented waste Corg 1.00 1.88 1.50 1.46 materials was different, the efficiency of both materials for soil and CHAþFA 0.52 0.70 0.59 0.69 ecosystem development was demonstrated by comparison of soil C 0.33 0.39 0.39 0.39 HA and biomass properties in the constructed Technosols to the initial wastes and the reference (eroded) soil. Vegetation covered 100% of the sites in both cases (Fig. 7), which is the most widely reported biomass, the density of aerial parts, and height and thickness of the indicator of efficient remediation. Intensive growth of above- and sod on the SM Technosol were 15e30% higher (Table 4). underground biomass triggered the humification process, which Calcium content in the plant tissues on the CW Technosol was resulted in the accumulation of organic C as well as N, Mg and Ca in 2.5 times higher in the aboveground parts and 3.4 times in the the Technosols. Moreover, CHAþFA in the top layers almost doubled compared to the SM Technosol. On the contrary, Mg content in the compared to the eroded soil. Nitrogen content increased by 6.4 and plant tissues of the SM Technosol was 2.0e2.8 times higher 1.5 times compared to the initial content in the top layers and 1.2 compared to the CW Technosol. Silicon content in the aboveground and 1.7 times for the sublayers of eroded soil. The content of Mg and biomass on the SM Technosol was higher by the factor of 2e3 than Ca increased in the sublayers by the factor of 100 and 290 for the on CW. Therefore, Ca, Mg and Si content in all parts of the plants CW Technosols and by the factor of 250 and 16 for the SM Tech- were affected by the chemical properties of the mineral wastes. nosols, respectively. Both BMC and BR in Technosols increased 3e12

Fig. 5. Microbial metabolic quotient (qCO2) and percentage of microbial biomass C to organic C (MBC/Corg) in the top- and sublayers of Technosols on carbonatite waste (CW) and serpentine-magnesite (SM) wastes (circle represents a mean, bars e 1 standard deviation). Dotted lines are the mean values for the eroded Ferric Podzol in the AleFe-humus horizon. M.V. Slukovskaya et al. / International Soil and Water Conservation Research 7 (2019) 297e307 303

alkaline conditions in the formed Technosols. Both CW and SM Technosols accumulated Cu and Ni; therefore, they performed as alkaline geochemical barriers for metals deposited from atmospheric emissions and thus hampered their vertical and lateral migration (Koptsik, 2014; Vodyanitskii, 2008). The content of plant-available forms of Cu in CW Technosol was 2e4 times higher than in SM. Manganese oxides are the primary carriers of Cu in soils (Kropacheva, Didik, & Kornev, 2013; Vodyanitskii, 2008, 2009). The high content of Mn in the reducing environment in CW Technosol led to the accumulation of Cu on the alkaline geochemical barrier, formed by carbonatite mining waste. Also, the total Cu strongly correlated with total phosphorus (P) content (r ¼ 0.9, p < 0.05). Phosphorus is a component of organic matter; thus, the formation of metal-organic complexes in both Technosols could be possible (Kabata-Pendias, 2000; Rinklebe & Shaheen, 2017). The accumulation of Ni in the top layer of SM Technosol occurred 1.5e2 times more intensively than in CW. Geochemical speciﬁcity of Ni deposition slightly differed from Cu; although Mn is the leading carrier of Ni, Si-compounds can take an active part in Ni precipitation on chemical alkaline barriers (Vodyanitskii, 2008, 2009; Rinklebe & Shaheen, 2017). Silicon mobile content in the SM Technosol was 1.6 times higher than in CW and predominantly occurred in the form of magnesium silicate, i.e. Si is a part of the compounds actively interacting with metals (Sharma et al., 2018; Slukovskaya et al., 2017). Nickel can partly replace Mg in serpentine minerals through the ion exchange reaction. Lower values of Eh and pH in the SM Technosol also contributed to the decrease in Ni mobility (Zelano et al., 2016; Slukovskaya et al., 2017; Rinklebe & Shaheen, 2017). Thus, the main factors of the metals’ accumula- Fig. 6. Relationship between available Cu content and microbial biomass carbon tion in Technosols were the initial mineralogical composition of the (MBC), basal respiration (BR), the microbial metabolic quotient (qCO2) and the portion wastes (especially Mn and Si content), oxidation-reduction condi- of microbial biomass C to organic C (MBC/Corg) of Technosols. tions, and accumulation of soil organic matter. times compared to the eroded soils and initial waste materials, 4.2. Phytotoxicity of the engineered technosols indicating gradual development of a soil microbial community. This outcome is conﬁrmed by the decrease of the microbial metabolic The ecosystems in the impact zone of the Cu-Ni factory in the quotient in Technosols compared to the eroded reference soil. Low Subarctic is mainly affected by soil macronutrients and metal pol- values of qCO have traditionally considered a relevant indicator of 2 lutants (Evdokimova et al., 2011). The “toxicity module” index (Mt) less stressful conditions for microbiota and overall ecosystem sus- was previously developed to assess soil conditions in remediation tainability (Bastida, Zsolnay, Hernandez, & García, 2008; Vasenev experiments in the industrial barren (Evdokimova et al., 2011; et al., 2018, pp. 18e35). Slukovskaya et al., 2017). Based on the experimental data, Mt was calculated as follows (expressed in molar units): Mt ¼ 4. Discussion (nNi þ nCu) 100/(nCa þ nMg). It was shown earlier that a hun- dredfold surplus of the sum of molar macronutrient content rela- 4.1. Immobilization of Cu and Ni tive to the sum of metals (Mt ¼ 1) neutralizes toxic effect of Cu and Ni on the photosynthetic apparatus of plants, whereas a twofold The high content of Ca and Mg in the mineral wastes promoted increase in this index (Mt ¼ 2) leads to the inhibition of

Table 4 Properties of grass cover on the carbonatite waste (CW) and serpentine-magnesite (SM) waste Technosols.

Properties CW SM

Biometric index Biomass, kg m 2 0.7 ± 0.1 1.1 ± 0.2 Height of aboveground parts, cm 54.8 ± 26.6 67.2 ± 11.6 Thickness of the sod, cm 10.0 ± 0.5 11.5 ± 0.6 The density of plant cover, pcs·m 2 1144 ± 128 1610 ± 141 The content of chemical elements, mg$kg 1 Leaves Spikes Roots Leaves Spikes Roots Ca 2978 3286 13,046 1202 1340 3808 Mg 1288 1814 6136 2757 3371 17,042 Si 3374 2137 3132 6238 6433 2195 Cu 200 692 957 204 582 632 Ni 75 341 353 66 309 354 Co 3 13 94 3 11 52 304 M.V. Slukovskaya et al. / International Soil and Water Conservation Research 7 (2019) 297e307

Fig. 7. Grass cover at the experimental sites immediately after plots development (a), after 1 year (b) and after 7 years (c and d). photosynthesis (Slukovskaya et al., 2017b). The Mt index of the artiindustrats subgroup (Prokofeva & Vasenev, 2017), and mostly initial soil was 75, i.e., this soil was extremely toxic, while Mt for presented by the particles < 1 mm in size. Serpentinite-magnesite Technosols was <1 indicating the absence of a significant toxic ef- materials were mostly overburdened rocks (lithostrats). The grain fect of metals on plants ontogenesis. size of the SM waste was related to its origin; only one-half of the The SM Technosol provided more favorable conditions for particles was <1 mm, while another half was 1e2.5 mm in size. biomass development compared to the CW Technosol based on These differences in the origin and granulometric composition of biomass productivity and other biometric indices. The accumula- the mining wastes could lead to unequal distribution of available tion of Si by plants increases their resistance to abiotic and biotic water and air in Technosols. Serpentinite-magnesite with a high stress factors (Imtiaz et al., 2016; Liang, Sun, Zhu, & Christie, 2007; water saturation rate could provide better conditions for an organic Matychenkov, 2008; Rizwan, Meunier, Miche, & Keller, 2012). The matter accumulation (Huot et al., 2015). mix of serpentine-containing mining waste with podzol soil increased the productivity of six grass species in the Subarctic due 4.3. Changes in C quantity and quality to high Si content in this waste (Slukovskaya et al., 2018). Therefore, an active accumulation of Si in the aboveground parts of the plants Carbon accumulation is considered a principal function per- in the SM Technosol may influence higher biometric indices in this formed by soils, which can be lost due to degradation and pollution case. Besides, high accumulation of Cu in the CW Technosol may (Ding et al., 2018; Lal, 2004). Consequently, restoration of C stocks slightly inhibit the growth of root systems. by C accumulation is an important indicator of soil remediation. In spite of the surface contamination of leaves, the translocation Total C content of the SM Technosol was similar to the reference factor calculated as a ratio of metals’ content in leaves and roots was Ferric Podzols under non-disturbed northern taiga ecosystems, below one for both Technosols, that allows referring the grass whereas CW Technosol contained twice more total C. Total C con- species as “excluders” of Cu and Ni (Lange et al., 2017). Thus, the tent in the CW Technosol was similar to C content in Retisols under formed plant cover had a stabilizing function for metal pollutants, south taiga (Dobrovolsky & Urusevskaya, 2004). Likely, this in- i.e., promoted their fixation in the soil and prevented migration crease in C content in Technosols was partly inherited from the through the food chains (Antoniadis et al., 2017; Mahar et al., 2016). original waste materials and partly accumulated due to the humus- The high content of plant-available Ca and Mg was the most accumulation processes. A predominant relict origin of total C beneficial property of the mining wastes, which supported plant stocks in the CW Technosols was clearly illustrated by doubling C growth on the engineered Technosols. The CW material was content in the original carbonate-containing CW compared to the dominated by Ca, whereas Mg was a dominating element in the SM SM waste. This assumption was further confirmed by a comparison wastes. This difference in macronutrients has a clear relation to of topsoil C , MBC, and aboveground biomass in the CW and SM waste origin (Echevarria, 2018). Also, the presence of phytoavail- org Technosols with all the indicators being higher in the SM able Si in the SM mining waste had a positive effect on grass pro- Technosols. ductivity. The increased plant cover contributed to metals’ The quality of C stocks was also different between Technosols. phytostabilization, prevented soil erosion and contamination of the The C/N ratio of CW Technosols indicates an N impoverishment of surrounding areas. organic matter and an unfavorable condition for plants. The C /C Mining wastes used as parent materials for Technosol formation HA FA and MBC/C ratios were higher for the SM Technosol compared to differed in the predominant grain size. Carbonatite waste was the org the CW Technosols suggesting the overall lower quality of organic iron ores' secondary beneficiation waste and regarded to matter in the CW Technosol. For both substrates, the CHA/CFA in the M.V. Slukovskaya et al. / International Soil and Water Conservation Research 7 (2019) 297e307 305 top layer was considerably higher than in the sublayer, indicating being catabolized to CO2, and the only low amount is available for the humification of roots and aboveground biomass as a domi- microbial biomass accumulation (Chander, Dyckmans, Joergensen, nating soil forming process in the engineered topsoils. This process Meyer, & Raubuch, 2001; Joergensen & Emmerling, 2006). contributed to a 10e100% increase of Corg in the SM Technosols compared to the initial eroded soil and waste materials. Thus, the 4.5. Soil formation and ecosystem development at the early stages SM Technosol provided more favorable conditions for plants after remediation growth regarding organic matter quality than CW with an optimal C: N ratio and accumulation of organic C, including C . HA Restoration of ecosystems to the zonal type of vegetation in the A rate of C accumulation of 0.5e1.0% during seven years was org Subarctic conditions takes at least 30e35 years (Archegova & much faster than expected for the natural soils in the region and Likhanova, 2012). The first (intensive) recovery phase takes about comparable to the C accumulation rate reported for the urban 5e7 years and the primary functions of this stage are the formation constructed Technosols in a temperate climate (Vasenev & of an organic-accumulative layer on the basis of a mineral horizon Kyzyakov, 2018). Similarly, the adapted RothC model predicted and prevention of erosion (Archegova & Likhanova, 2012; that amount of organic C stored in the whole 1-m profile of the Archegova & Likhanova, 2012; Leguedois et al., 2016; Ser e et al., constructed Technosol under wastes over the 12-year evolution 2017). The target of this phase is a rapid establishment of a vege- would be 5 times higher than the corresponding value observed in tation cover by sowing adapted mixtures of perennial grasses and a natural soils in similar climatic conditions and plant cover (Rees fertilizer’ application. At the second (assimilation) stage, a zonal et al., 2017). Such a prompt C accumulation, i.e., C sequestration type of vegetation gradually replaces the initial herbal community. at the early stages of soil formation and development, is a very Initial soil conditions created at the first stage of remediation positive result of the remediation on the key soil function. Some largely determine the success of the whole restoration process previous studies (e.g., Shchepeleva et al., 2017) show the opposite (Slukovskaya et al., 2018). trend, where constructed Technosols at the golf course in Moscow The first (intensive) stage must fix the upper layer of the city were significant C sources during the first three years after the polluted soil with the root systems of perennial grasses to prevent construction. Apparently, in high altitudes where low temperatures further erosion and its duration is usually about 3e5 years (Ivanova, hamper the decomposition of C, young Technosols can act as C Kostina, Kremenetskaya, & Inozemtseva, 2010; Archegova & sinks, which is essential for climate mitigation perspectives. Likhanova, 2012). In the conditions of air and soil pollution, this stage can take much longer due to the need to reduce soil toxicity 4.4. Microbial properties’ assessment for plants and microorganisms (Mahar et al., 2016; Salinas et al., 2012). Microbial biomass C, BR, and eco-physiological ratios are widely Despite the high level of pollution of soil and air by metals in the used as indices for assessing soil microbial community in a variety impact zone of the Cu-Ni factory, the sustainable development of a of terrestrial ecosystems (Nielsen & Winding, 2002; Joergensen & plant-soil system in our experiment was clearly demonstrated. The Emmerling, 2006; Murugan, Loges, Taube, Sradnick, & increase of N and organic C stocks, microbiological biomass, and Joergensen, 2014). Microbial biomass C together with N are activity, as well as productivity and the projective cover of plant important properties directly contributing to the increase in rich- communities observed at the experimental sites indicates the ness and density of plants for Technosols under wastes in aban- continuing sustainable development of the ecosystem. Although doned pyritic tailings (Zornoza, Gomez-Garrido, Martínez- the application of the NPK complex fertilizer stopped after six years Martínez, Dolores Gomez-L opez, & Faz, 2017). In our study, MBC of the experiment, positive dynamics in soil and biomass properties content and BR rate in the SM Technosols reached up to 319 mgC remained, marking the end of the intensive stage and the transition g 1 and 0.50 mg C-CO g 1 h 1, corresponding to the values in the 2 to the assimilation stage of ecosystem recovery. upper 10-cm layer of the 120 years old Ferric Podzol of pine forest 1 1 1 257 mgC g and 0.74 mg C-CO2 g h , respectively (Stolnikova, Ananyeva, & Chernova, 2011). This indicates that microbial 5. Conclusion biomass and its activity in SM Technosol are close to achieving the natural soil levels. Likely, the indices will increase with the subse- Mining wastes were implemented to construct Technosols for quent Technosol evolution due to the organic matter accumulation remediation of metal polluted and eroded acidic soils in the impact linked to inputs from plant residues. zone of the factory processing non-ferrous metals in the Subarctic Creation of Technosols with the application of waste can be fast zone. After seven years of the experiment, sustainable positive and efficient. In four years of the experiment, the diversity of changes in soil properties and ecosystem state were demonstrated. Bacteria and Archaea, richness of earthworms and Collembola The Technosols construction remediated the territory by storing Cu abundance were similar compared to the natural soil of a meadow and Ni and reducing their toxicity. This affirms the first hypothesis (Ser e et al., 2017). However, the efficiency of remediation largely of the study and shows that engineering Technosol is an effective depends on the types of applied wastes. In our study, the SM remediation approach in the Subarctic. Further comparison be- Technosol had more favorable conditions for microbial commu- tween the waste materials affirms the second research hypothesis nities’ growth. The heavily degraded lower layer under the SM regarding the different efficiency of the wastes and demonstrates waste had a similar content of MBC, BR rate, and MBC/Corg ratio for the advantages of the serpentine-magnesite material for imple- the top- and sublayers of Technosols suggesting a successful rees- mentation in Subarctic conditions. tablishment of microbial properties. The differences in microbial Based on results obtained in this study, we suggest that mining properties of Technosols could be explained by the characteristics wastes used for Technosol constructions must have the following of the applied wastes associated with immobilization of metals, in main properties and functions: particular, Сu. The high amount of available Cu in the CW Technosol was the limiting factor decreasing microbial biomass content, mi- o Weekly-alkaline plant-friendly medium; crobial respiration, C-use efficiency and its availability to microor- o Low dissolution rate for the long-term remediation effect; ganisms. Copper content in the CW Technosol is quite toxic for o The inclusion of minerals capable of adsorbing metal pollutants microorganisms and leads to the higher amount of organic matter to reduce their exchangeable forms; 306 M.V. Slukovskaya et al. / International Soil and Water Conservation Research 7 (2019) 297e307

o Coarse to fine sand grain size to allow the evolution of under- Bastida, F., Zsolnay, A., Hernandez, T., & García, C. (2008). Past, present and future of e e ground parts of plants; soil quality indices. A Biological Perspective Geoderma, 147(3 4), 159 171. https://doi.org/10.1016/j.geoderma.2008.08.007. o Plant available forms of macronutrients and microelements to Bohm,€ W. (1979). Methods of studying root systems. Ecological Studies, 33. Springer: ensure an independent existence of plant communities; Berlin. nd o Presence of labile Si to increase stress-resistance of plants in Carter, M. R., & Gregorich, E. G. (2007). Soil sampling and methods of analysis (2 ed.). CRC Press. severe climatic and contamination conditions. Chander, K., Dyckmans, J., Joergensen, R. G., Meyer, B., & Raubuch, M. (2001). Different sources of heavy metals and their long-term effects on soil microbial A combination of these properties allows reducing metals properties. Biology and Fertility of Soils, 34,241e247. Creamer, R. E., Schulte, R. P. O., Stone, D., et al. (2014). Measuring basal soil respi- migration in ecosystems through such interrelated processes as ration across Europe: Do incubation temperature and incubation period mat- neutralization, adsorption, and phytostabilization, necessary for ter? Ecological Indicators, 36, 409e418. the sustainable development and functioning of vegetation and Ding, K., Wu, Q., Wei, H., Yang, W., Ser e, G., Wang, S., et al. (2018). Ecosystem ser- soils under the continuing pollution load. vices provided by heavy metal-contaminated soils in China. Journal of Soils and Sediments, 18(2), 380e390. We observed the intensive development of soil-forming pro- Dobrovolsky, G. V., & Urusevskaya, I. S. (2004). Soil geography (2nd ed.). Moscow: cesses, e.g. humus accumulation, in the engineered Technosols. The MSU (In Russian). fi plant growth and the increase of microbial biomass led to the Echevarria, G. (2018). Genesis and behaviour of ultrama c soils and consequences for nickel biogeochemistry. In A. Van der Ent, G. Echevarria, A. J. M. Baker, & accumulation of soil organic matter including humic and fulvic J. L. Morel (Eds.), Agromining: Farming for metals: Extracting unconventional re- acids. A comparison between the initial and remediated soils also sources using plants (pp. 135e156). Cham: Springer International Publishing. confirms positive changes in pH levels and macronutrient cycles. Evdokimova, G. A., Mozgova, N. P., & Kalabin, G. V. (2011). Contents and toxicity of fi heavy metals in soils of the zone affected by aerial emissions from the Sever- The obtained evidence of soil formation con rms our third research onikel enterprise. Eurasian Soil Science, 44(2), 237e244. hypothesis and demonstrates that the soil-forming processes are Ganicheva, S. N., Lukina, N. V., Kostina, V. A., & Nikonov, V. V. (2004). Technogenic essential indicators of remediation efficiency. Thus, sustainable degradation and demutation succession in coniferous forests of the Kola Peninsula. Lesovedenie, 3,57e67 (In Russian). development and evolution of a plant-soil complex can prevent Gruszecka-Kosowska, A., Baran, P., Wdowin, M., & Franus, W. (2017). Waste dolo- further erosion processes, soil degradation, and environmental mite powder as an adsorbent of Cd, Pb(II), and Zn from aqueous solutions. pollution in highly vulnerable Subarctic ecosystems. Environment and Earth Science, 76, 521. https://doi.org/10.1007/s12665-017- 6854-8. Huot, H., Simonnot, M.-O., & Morel, J.-L. (2015). Pedogenetic trends in soils formed Acknowledgments in technogenic parent materials. Soil Science, 180(4e5), 182e192. Imtiaz, M., Rizwan, M. S., Mushtaq, M. A., Ashraf, M., Shahzad, S. M., Yousaf, B. S., et al. (2016). Silicon occurrence, uptake, transport and mechanisms of heavy This work was supported by the Russian Foundation for Basic metals, minerals and salinity enhanced tolerance in plants with future pros- Research Project 16-35-60022. Design of the experiment and multi- pects: A review. Journal of Environmental Management, 183,521e529. https:// year monitoring have been supported by the Joint-stock company doi.org/10.1016/j.jenvman.2016.09.009. & ‘ ’ Ivanova, L.A., Kremenetskaya, M.V., Gorbacheva, T.T., Inozemtseva, E.S., Korytnaya, Kola Mining and Metallurgical Company . Analysis of C quantity O.P. Method of creation of vegetative ground cover in recultivation of disturbed and quality was supported by the Russian Science Foundation soils. Patent RF, N 2484613. 20.06.2013. Bulletin N 17. Project 17-77-20046. Data analysis and processing were supported Ivanova, L.A. (2011) Method for biologically recultivating industrial wastelands. Int. “ e ” Appl. WO/2011/084079, Int.Cl.6 A 01 B 79/02, A 01 G 1/00, 31/00. No. PCT/ by the RUDN University 5 100 Project. The authors would like to RU2010/000001. Applied: 11.01.10. Published: 14.07.11. thank Prof. Ivan Vasenev, Dr.Sc. Nadezhda Ananyeva, Ph.D. candi- Ivanova, L. A., Gorbacheva, T. T., Slukovskaya, M. V., Kremenetskaya, I. P., & date Andrey Novikov, Ph.D. candidate Irina Mosendz and Olga Inozemtseva, E. S. (2014). Innovation technologies of reclamation of disturbed e Korytnaya for conducting plant and soil analyses, and Ph.D. lands. Ekologiya proizvodstva, 2,58 68 (in Russian). Ivanova, L. A., Kostina, V. A., Kremenetskaya, M. V., & Inozemtseva, Y.e. S. (2010). candidate Anna Paltseva for editing English and valuable comments Accelerated formation of anti-erosion grass in the technogenically disturbed on the article. areas: The arctic circle, 4(2) Vestnik MSTU, 13, 977e983 (In Russian). Jiang, Y., Shi, L., Guang, A.-L., Mu, Z., Zhan, H., & Wu, Y. (2017). Contamination levels and human health risk assessment of toxic heavy metals in street dust in an References industrial city in Northwest China. Environmental Geochemistry and Health, 1e14. https://doi.org/10.1007/s10653-017-0028-1. Abakumov, E., Lupachev, A., & Andreev, M. (2017). Trace element content in soils of Joergensen, R. G., & Emmerling, C. (2006). Methods for evaluating human impact on the King George and Elephant islands, maritime Antarctica. Chemistry and soil microorganisms based on their activity, biomass, and diversity in agricul- Ecology, 33(9), 856e868. https://doi.org/10.1080/02757540.2017.1384821. tural soils. Journal of Soil Science and Plant Nutrition, 169, 295e309. rd Aguilar, M., Mondaca, P., Ginocchio, R., Vidal, K., Sauve, S., & Neaman, A. (2018). Kabata-Pendias, A. (2000). Trace elements in soils and plants (3 ed.). Boca Raton, Comparison of exposure to trace elements through vegetable consumption Florida, US: CRC Press. between a mining area and an agricultural area in central Chile. Environmental Kashulina, G. M., Kubrak, A. N., & Korobeinikova, N. M. (2015). Soil acidity status in Science & Pollution Research, 25(19), 19114e19121. https://doi.org/10.1007/ the vicinity of the Severonikel copperenickel industrial complex, Kola penin- s11356-018-2116-x. sula. Eurasian Soil Science, 48(4), 432e444. Ajmone-Marsan, F., & Biasioli, M. (2010). Trace elements in soils of urban areas. Kashulina, G. M. (2017). Extreme pollution of soils by emissions of the copper-nickel Water, Air, and Soil Pollution, 213, 121e143. https://doi.org/10.1007/s11270-010- industrial complex in the Kola Peninsula. Eurasian Soil Science, 50(7), 837e849. 0372-6. Koptsik, G. N. (2014). Modern approaches to remediation of heavy metal polluted Ali, Z., Malik, R. N., Shinwari, Z. K., & Qadir, A. (2015). Enrichment, risk assessment, soils: A review. Eurasian Soil Science, 47(7), 707e722. and statistical apportionment of heavy metals in tannery-affected areas. Inter- Koptsik, G. N., Koptsik, S. V., & Smirnova, I. E. (2016). Alternative technologies for national journal of Environmental Science and Technology, 12(2), 537e550. remediation of technogenic barrens in the Kola Subarctic. Eurasian Soil Science, https://doi.org/10.1007/s13762-013-0428-4. 49(11), 1294e1309. https://doi.org/10.1134/S1064229316090088. AMAP assessment report: Arctic pollution issues. (1998). Oslo, Norway: Arctic Moni- Kozlov, M., & Zvereva, E. (2007). Industrial barrens: Extreme habitats created by toring and Assessment Programme. non-ferrous metallurgy. Reviews in Environmental Science and Biotechnology, 6, Ananyeva, N. D., Susyan, E. A., Chernova, O. V., & Wirth, S. (2008). Microbial 231e259. respiration activities of soils from different climatic regions of European Russia. Kropacheva, T. N., Didik, M. V., & Kornev, V. I. (2013). Modeling sorption of cations of European Journal of Soil Biology, 44,147e157. heavy metals by hydroxides in the presence of EDTA. Sorption and Chromato- Anderson, J. P. E., & Domsch, K. H. (1978). A physiological method for the quanti- graphic Processes, 13(3), 360e368 (in Russian). tative measurement of microbial biomass in soils. Soil Biology and Biochemistry, Kumar, M., Gogoi, A., Kumari, D., & Borah, R. (2016). Review of perspective, prob- 10(3), 215e221. lems, challenges, and future scenario of metal contamination in the urban Antoniadis, V., Levizoua, E., Shaheen, S. M., Ok, Y. S., Sebastian, A., Baum, C., et al. environment. Journal of Hazardous, Toxic, and Radioactive Waste, 21(4), (2017). Trace elements in the soil-plant interface: Phytoavailability, trans- 04017007. https://doi.org/10.1061/(ASCE)HZ.2153-5515.0000351. location, and phytoremediation e a review. Earth-Science Reviews, 171, Lal, R. (2004). Agricultural activities and the global carbon cycle. Nutrient Cycling in 621e645. Agroecosystems, 70,103e116. Archegova, I. B., & Likhanova, I. A. (2012). Biological recultivation problem and its Lange, B., van der Ent, A., Baker, A. J. M., Echevarria, G., Mahy, G., Malaisse, F., et al. solution in the European northeast, the Komi Republic as an example. Izvestiya (2017). Copper and cobalt accumulation in plants: A critical assessment of the Komi nauchnogo tsentra UrO RAN, 1(9), 29e34 (In Russian). current state of knowledge. New Phytologist, 213,537e551. M.V. Slukovskaya et al. / International Soil and Water Conservation Research 7 (2019) 297e307 307

Laumbach, R., & Kipen, H. (2012). Respiratory health effects of air pollution: Update Rizwan, M., Meunier, J.-D., Miche, H., & Keller, C. (2012). Effect of silicon on reducing on biomass smoke and traffic pollution. The Journal of Allergy and Clinical cadmium toxicity in durum wheat (Triticum turgidum L. cv. Claudio W.) grown Immunology, 129,3e11. in a soil with aged contamination. Journal of Hazardous Materials, 209e210, Leguedois, S., Ser e, G., Auclerc, A., Cortet, J., Huot, H., Ouvrard, S., et al. (2016). 326e334. Modelling pedogenesis of Technosols. Geoderma, 262,199e212. Salinas, M. Z., Villavicencio, M. V., Bustillos, L. G. T., & Aragon, A. G. (2012). Liang, Y., Sun, W., Zhu, Y. G., & Christie, P. (2007). Mechanisms of silicon-mediated Assessment of in situ and ex situ phytorestoration with grass mixtures in soils alleviation of abiotic stresses in higher plants: A review. Environmental Pollu- polluted with nickel, copper, and arsenic. Physical Chemistry Earth Journal, tion, 147, 422e428. https://doi.org/10.1016/j.envpol.2006.06.008. 37e39,52e57. Lyanguzova, I. V., Goldvirt, D. K., & Fadeeva, I. K. (2015). Transformation of poly- SanPiN 2.6.1.2523-09. (2009). Radiation safety standards NRB-99/2009 (In Russian). metallic dust in the organic horizon of AleFe-humus podzol (field experiment). Ser e, G., Schwartz, C., Cortet, J., Guimont, S., Watteau, F., Simmonnot, M.-O., et al. Eurasian Soil Science, 48(7), 701e711. (2017). Pedological engineering for brownfield reclamation. Soils within Cities: Mahar, A., Wang, P., Ali, A., Awasthi, M. K., Lahori, A. H., Wang, Q., et al. (2016). Global Approaches to their Sustainable Management e Composition, Properties, Challenges and opportunities in the phytoremediation of heavy metals and Functions of Soils of the Urban Environment,152e158. contaminated soils: A review. Ecotoxicology and Environmental Safety, 126, Sharma, S., Tiwari, S., Hasan, A., Saxena, V., & Pandey, L. M. (2018). Recent advances 111e121. in conventional and contemporary methods for remediation of heavy metal- Manninen, S., Zverev, V., Bergman, I., & Kozlov, M. (2015). Consequences of long- contaminated soils. Biotec, 8(216), 18. term severe industrial pollution for above ground carbon and nitrogen pools Shchepeleva, A. S., Vasenev, V. I., Mazirov, I. M., Vasenev, I. I., Prokhorov, I. S., & in northern taiga forests at local and regional scales. The Science of the Total Gosse, D. D. (2017). Changes of soil organic carbon stocks and CO2 emissions at Environment, 536,616e624. the early stages of urban turf grasses' development. Urban Ecosyst, 20(2), Matychenkov, V. V. (2008). The role of mobile silicon compounds in plants and the 309e321. soil-plant system. Pushchino. Synopsis of Ph.D. dissertation. Pushchino 2008, Shishov, L. L., Tonkonogov, V. D., Lebedeva, I. I., & Gerasimova, M. I. (2004). Classi- 34 p. (In Russian). fication and diagnostics of Russian soils. Oikumena, Smolensk, 342 (in Russian). Moiseenko, T. I., Voinov, A. A., Megorsky, V. V., Gashkina, N. A., Kudriavtseva, L. P., Shutcha, M. N., Mubemba, M. M., Faucon, M.-P., Luhembwe, M. N., Visser, M., Vandish, O. I., et al. (2006). Ecosystem and human health assessment to define Colinet, G., et al. (2010). Phytostabilization of copper-contaminated soil in environmental management strategies: The case of long-term human impacts katanga: An experiment with three native grasses and two amendments. In- on an Arctic lake. The Science of the Total Environment, 369(1e3), 1e20. https:// ternational Journal of Phytoremediation, 12,616e632. doi.org/10.1016/j.scitotenv.2006.06.009. Slukovskaya, M. V., Ivanova, L. A., Kremenetskaya, I. P., Gorbacheva, T. T., Morel, J. L., Chenu, C., & Lorenz, K. (2015). Ecosystem services provided by soils of Drogobuzhskaya, S. V., Lashchuk, V. V., et al. (2018). Rehabilitation of industrial urban, industrial, traffic, mining, and military areas (SUITMAs). Journal of Soils barren in Arctic region using mining wastes. The Open Ecology Journal, 11,1e13. and Sediments, 15(8), 1659e1666. Slukovskaya, M. V., Kremenetskaya, I. P., Ivanova, L. A., & Vasilieva, T. N. (2017). Moskovchenko, D. V., Kurchatova, A. N., Fefilov, N. N., & Yurtaev, A. A. (2017). Remediation in conditions of an operating copper-nickel plant: Results of Concentrations of trace elements and iron in the arctic soils of Belyi island (the perennial experiment. Nonferrous Metals, 2,20e26. Kara sea, Russia): Patterns of variation across landscapes. Environmental Slukovskaya, M. V., Novichonok, E. V., Kremenetskaya, I. P., Mosendz, I. A., Monitoring and Assessment, 189(5), 210. https://doi.org/10.1007/s10661-017- Drogobuzhskaya, S. V., & Markovskaya, E. F. (2017). Using Festuca rubra L. In 5928-0. phytoremediation: Complex assessment of influence of technogenic grounds on Murugan, R., Loges, R., Taube, F., Sradnick, A., & Joergensen, R. G. (2014). Changes in plant. Uchenye Zapiski Petrozavodskogo Gosudarstvennogo Universiteta, 4(165), soil microbial biomass and residual indices as ecological indicators of land use 70e80 (in Russian). change in temperate permanent grassland. Microbial Ecology, 67, 907e918. Stolnikova, E. V., Ananyeva, N. D., & Chernova, V. O. (2011). The microbial biomass Nielsen, M. N., & Winding, A. (2002). Microorganisms as indicators of soil health. and its activity and structure in the soils of old forests in the European Russia. National environmental research institute. Denmark. Technical report N. 388. Eurasian Soil Science, 44,479e494. Norilsky Nickel. (2018). The official site of the Nornickel Group. Retrieved from Vasenev, V. I., Castaldi, S., Vizirskaya, M. M., Ananyeva, N. D., Shchepeleva, A. S., https://www.nornickel.com/. Mazirov, I. M., et al. (2018). Urban soil respiration and its autotrophic and het- Pacyna, J. M., Pacyna, E. G., & Aas, W. (2009). Changes of emissions and atmospheric erotrophic components compared to adjacent forest and cropland within the deposition of mercury, lead, and cadmium. Atmospheric Environment, 43(1), Moscow megapolis. Springer Geography. https://doi.org/10.1007/978-3-319- 117e127. https://doi.org/10.1016/j.atmosenv.2008.09.066. 70557-6_4. Peuelas, J., & Filella, I. (2002). Metal pollution in Spanish terrestrial ecosystems Vasenev, V., & Kyzyakov, Y. (2018). Urban soils as hotspots of anthropogenic carbon during the twentieth century. Chemosphere, 46(4), 501e505. https://doi.org/ accumulation: Review of stocks, mechanisms and driving factors. Land Degra- 10.1016/S0045-6535(01)00171-0. dation & Development, 29(6), 1607e1622. https://doi.org/10.1002/ldr.2944. Prokofeva, T. V., & Vasenev, V. I. (2017). Technogenic surface formations of classifi- Vodyanitskii, Y.u. N. (2008). Heavy metals and metalloids in soils. Moscow, GNU cation and diagnostic system of Russian Soils. Soils within Cities (pp. 98e99). Pochvennyy institute imeni V.V. Dokuchaeva RASKhN, 84 (in Russian). Catena-Schweizerbart. Vodyanitskii, Y.u. N. (2009). Mineralogy and geochemistry of manganese: A review Radziemska, M. (2018). Study of applying naturally occurring mineral sorbents of of publications. Eurasian Soil Science, 42(10), 1170e1178. Poland (dolomite halloysite, chalcedonite) for aided phytostabilization of soil World reference base for soil resources (3rd ed.). (2015) Retrieved from http://www. polluted with heavy metals. Catena, 163,123e129. fao.org/soils-portal/soil-survey/soil-classification/world-reference-base/en/. Rees, F., Dagois, R., Derrien, D., Schwartz, C., Simonnot, M. O., Morel, J. L., et al. Zelano, I., Sivry, Y., Quantin, C., Gelabert, A., Tharaud, M., Nowak, S., et al. (2016). (2017). Storage of carbon in constructed Technosols: In situ monitoring and Study of Ni exchangeable pool speciation in ultramafic and mining environ- modeling. In Abstract book of 9th international congress soils of urban industrial ments with isotopic exchange kinetic data and models. Applied Geochemistry, traffic mining and military areas (pp. 152e153). 64,146e156. Report on the status and environmental protection of the Murmansk region in Zornoza, R., Gomez-Garrido, M., Martínez-Martínez, S., Dolores Gomez-L opez, M., & 2016. (2017). Murmansk region government. Retrieved from http://mpr.gov- Faz, A. (2017). Bioaugmentation in Technosols created in abandoned pyritic murman.ru (In Russian). tailings can contribute to enhance soil C sequestration and plant colonization. Rinklebe, J., & Shaheen, S. M. (2017). Redox chemistry of nickel in soils and sedi- The Science of the Total Environment, 593e594,357e367. ments: A review. Chemosphere, 179, 265e278. International Soil and Water Conservation Research 7 (2019) 308–315

Contents lists available at ScienceDirect

International Soil and Water Conservation Research

journal homepage: www.elsevier.com/locate/iswcr

Original Research Article Spatial distribution and source identiﬁcation of heavy metals (As, Cr, Cu and Ni) at sub-watershed scale using geographically weighted regression

Maziar Mohammadi a, Abdulvahed Khaledi Darvishan a,n, Nader Bahramifar b a Department of Watershed Management Engineering, Faculty of Natural Resources, Tarbiat Modares University, 4641776489, Mazandaran Province, Iran b Department of Environmental Science, Faculty of Natural Resources, Tarbiat Modares University, 4641776489, Mazandaran Province, Iran article info abstract

Article history: Heavy metals are among the most important sources of water and soil pollution. These elements ac- Received 26 October 2018 cumulate in the agricultural soil through using contaminated water for irrigation, fertilizers, pesticide Received in revised form and enter to the river systems by water erosion. Therefore, land use plays a serious role in water and 22 January 2019 sediment pollution. In this regards, geographically weighted regression was used to investigate the Accepted 31 January 2019 spatial correlation between sediment heavy metals and land uses in a highland watershed. The landuse Available online 16 February 2019 map was used to calculate the area percentage of landuse types in sub-watersheds, followed by geo- Keywords: graphically weighted regression method to investigate the spatial correlation of As, Cr, Cu and Ni versus Land use three types of land uses. The highest correlation was observed for irrigated plots versus As and Ni in Sediment pollution upstream and for rainfed plots versus AS, Ni and Cr in the downstream. The relationship between heavy GWR metals and developed lands was more complicated and the highest correlation was found for Ni and As Talar watershed at outlet (R2 ¼ 0.52–0.89), for Cr in the upstream (R2 ¼ 0.50–0.76), and for Cu in the upstream and downstream (R2 ¼ 0.36–0.60). The results indicated that there is a positive correlation between heavy metals and land uses which varies with the level of agricultural and urbanization development at sub- watershed. Based on the ﬁndings, appropriate policies and decisions should be taken on agricultural land to prevent the transfer of heavy metals by sediment to aquatic environments. & 2019 International Research and Training Center on Erosion and Sedimentation and China Water and Power Press. Production and Hosting by Elsevier B.V. This is an open access article under the CC BY-NC- ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction pollution load index (PLI), and multivariate statistical analyses have been used to assess the role of natural and human sources Sediment transportation by rivers has been one of the critical on sediment heavy metal pollution (Chen, Lu, & Yang, 2012; issues for its role in enhancing nutrients and various Yongming, Peixuan, Junji, & Posmentier, 2006), they ignore spatial contaminants, including heavy metals that cause environmental relationship between the pollutants and the land use. Variables in problems due to their toxic and bio-accumulative and persistence watersheds are non-stationary, which means that the structure or properties (Govind & Madhuri, 2014). Sources of heavy metals are relationship between the variables will change in different categorized in two natural processes (erosion-weathering cycle of locations (McMillen, 2004). There is a constant regression rocks) and anthropogenic input (Li et al., 2013). Human activities coefﬁcient between dependent and independent variables in can affect the input-output ratio of natural cycle (Micó, Recatalá, traditional statistical methods such as ordinary least square, which Peris, & Sánchez, 2006). Land use and land cover pattern are means the spatial change of variables are not considered. among the most important driving forces that affect sediment To overcome the disadvantages of traditional regression pollution. Although such quality indices as contamination factor models, geographically weighted regression (GWR) has been (CF), enrichment factor (EF), geo-accumulation index (I-geo), developed to investigate the local correlation between dependent versus independent variables, which is widely used in different ﬁelds (Clement, Orange, Williams, Mulley, & Epprecht, 2009; n Corresponding author. Gao & Li, 2011; Ogneva-Himmelberger, Pearsall, & Rakshit, 2009; E-mail address: [email protected] (A. Khaledi Darvishan). Peer review under responsibility of International Research and Training Center Tu, 2011). This approach has been widely applied in environmental on Erosion and Sedimentation and China Water and Power Press. studies following topics of water pollution and land use https://doi.org/10.1016/j.iswcr.2019.01.005 2095-6339/& 2019 International Research and Training Center on Erosion and Sedimentation and China Water and Power Press. Production and Hosting by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). M. Mohammadi et al. / International Soil and Water Conservation Research 7 (2019) 308–315 309

(Huang, Huang, Pontius, & Zhang, 2015; Tu, 2011), while few 30% and more than 60% cover study area of 47.8%, 38.0% and 14.2%, studies have been conducted on the relationship between respectively. In terms of lithology, the watershed is covered by ig- sediment contamination and land use. For example, GWR and neous, metamorphic and sedimentary rocks. The average and the traditional multivariate statistical analyses have been combined to highest recorded discharges were 7.95 m3 s1 and 93.46 m3 s1,re- evaluate the effect of land use and human impacts on heavy metal spectively (Kavian, Mohammadi, Gholami, & Rodrigo-Comino, 2018). pollution in sediments of Sheyang River (Wu, Yang, Guo, & In Talar watershed, the most important land uses are rainfed agri- Han, 2017). Similarly, the relationships between land-use types culture, irrigated agriculture, forest, rangeland and residential area. and heavy metal concentrations and their corresponding chemical fractionation have been assessed using GWR to consider their 2.2. Land use and sediment quality data spatial variation (Xia et al., 2018). Talar is one of the highland watersheds in Mazandaran (Iran) that like the other watersheds in The digital elevation model (DEM) of Talar watershed was this province has experienced severe land use changes due to downloaded from the United States Geological Survey (USGS) population growth and lack of proper land management since the website with resolution of 30 m (https://gdex.cr.usgs.gov/gdex/). past few decades. Considering the importance of the Talar river Using the DEM, the watershed was divided into 12 sub-watersheds in supplying agricultural and drinking water, it is necessary to using "Watershed Delineator" placed in the main menu of evaluate the temporal-spatial changes of water and sediment ArcSWAT extension version 2012.10.19 in ArcGIS version 10.3.1 quality. The main objective of present study is to apply GWR to software. In order to obtain a land use map, the Landsat satellite assess the spatial distribution and the correlation of heavy metals image (2001) was downloaded from USGS (https://earthexplorer. (As, Cr, Cu and Ni) with different land uses at sub-watershed scale. usgs.gov/) and then classified into land uses, including rainfed agriculture, irrigated agriculture, forest, range and developed area. The above mentioned process was done by a supervised classifi- 2. Methods cation, followed by assessing accuracy using Kappa coefficient and overall accuracy in ENVI 5.1 software. Finally, the percentage of 2.1. Study area land use classes for each sub-watershed was calculated (Table 1). Sediment quality data were obtained from Iran's Geological Talar watershed covers a total area of 2905.38 ha to the south of Survey and Exploration. Surface sediment samples (top 5 cm) were the Caspian Sea, Mazandaran province, Iran, between 35° 44′ 23.6″ to collected in July 2007. As, Cr, Cu and Ni have been used to assess 36° 19′ 1.6″ Nand52° 35′ 22.2″ to 53° 23′ 34.19″ E(Fig.1).TheTalar the sediment pollution in the Talar basin, the statistical properties river and its main branches which originate from the Alborz mountain of which are shown in Table 2. and flow through the highlands and Ghaemshahr plain and finally reach the Caspian Sea, are important sources for agricultural and 2.3. Geographically weighted regression (GWR) drinking water supplies. With a Mediterranean rainfall regime, the annual average precipitation is 552.7 mm with annual average max- Geographically weighted regression (GWR) is a statistical method imum and minimum temperature of 21.1 and 7.7 °C, respectively. developed to assess spatial correlation between dependent and in- Three slope classes including less than 12% (north plain), more than dependent variables, which varies in different regions (Fotheringham,

Fig. 1. Location of the study area in Mazandaran province. 310 M. Mohammadi et al. / International Soil and Water Conservation Research 7 (2019) 308–315

Table 1 Percentage of land uses in the sub-watersheds (SWs) and the total area.

Forest (%) Irrigated Agriculture (%) Rainfed Agriculture (%) Rangeland (%) Developed area (%) Area (km2)

SW1 95.91 2.96 0.35 0.01 0.76 40.62 SW2 87.79 4.88 3.06 0.40 3.86 122.98 SW3 74.22 2.45 12.62 8.32 2.38 49.46 SW4 94.71 0.04 3.69 0.98 0.58 17.11 SW5 64.44 2.59 19.48 8.22 5.26 57.79 SW6 68.37 1.29 13.95 12.87 3.53 28.47 SW7 17.03 0.91 32.14 46.80 3.11 26.05 SW8 31.14 0.07 16.48 49.62 2.69 21.60 SW9 38.37 1.53 18.68 38.62 2.80 33.61 SW10 14.42 0.55 19.29 65.54 0.20 32.52 SW11 13.88 0.04 31.72 51.91 2.45 37.14 SW12 13.81 0.13 16.46 68.11 1.50 32.95

Table 2 In this study, we deﬁned three land use classes (irrigated Descriptive statistics of As, Cr, Cu and Ni. agricultural land, rainfed agricultural land and developed areas) and four heavy metals (As, Cr, Cu and Ni) as independent and ﬁ Metals Mean Standard Variance Coef cient Minimum Maximum (ppm) deviation of variation dependent variables, respectively. As a result, 12 (3 4) GWR models were established for 12 sub-watersheds for better As 6.98 2.368 5.60 33.9 1.40 12.40 comparison of spatial relationships. Cu 22.27 14.58 212.61 65.46 6.60 65.50 Cr 108.99 65.54 4295.91 60.14 49.50 409.58 Ni 36.89 34.59 1196.74 93.78 12.60 198.28 3. Results and discussion

Charlton, & Brunsdon, 1998). One of the main drawbacks of the tra- 3.1. Accuracy assessment of land use classification ditional regression models (such as ordinary least squares) is that they assume constant relationships among spatial variables without con- The accuracy assessment plays an important role in remote sidering their spatial dependency. The evaluation ability of sensing and classification of land use changes. The most common non-stationarity is the major advantage of GWR over traditional accuracy assessment elements are overall accuracy, producer's regression models (Attorre, De Sanctis, Francesconi, & Bruno, 2007). accuracy and Kappa coefficient (Luo et al., 2018). In this study, Non-stationarity indicates the changing of variable by varying Kappa coefficient and overall accuracy were 0.94% and 96.69%, locations (Mennis, 2006) that can be calculated by GWR. The above respectively (Table 3). mentioned method establishes separate equations with considering the independent and dependent variables inside a strip (the distance 3.2. Heavy metal pollution assessment from each phenomenon). The GWR method can be expressed as Eq. (1): The sediment quality guidelines (SQGs) are valuable tools to p represent concentration limits of contaminants in sediments and yuvuvxj =+ββε0()jj,,∑ j ()jjijj + their effects on aquatic system. In this study, indices of effects ()1 i=1 range low (ERL) and effects range medium (ERM) introduced by β Long et al. (1995) were used to determine sediment contamination where (uvj, j) the geographical coordinates for j, i(uvj, j)isthelocal regression coefficient for independent variables χiatlocationj,and level in sub-watersheds. These indices are established based on ( )andεj is the intercept and error term, respectively. Βi(uj,νj) empirical methods and their values are determined by field β0 uvj, j was estimated by Eq. (2) to determine a minimum: sampling and laboratory analysis, and statistical methods. ERL indicates a probability of minimal negative biological impact ⎛ ⎞2 n p whereas ERM represent frequent negative biological effects on βββuv,,,=−− w⎜ y uv uvx⎟ 0()jj ∑∑jk⎝ k 0 ()jj j ()jjij⎠ aquatic environment exposed to polluted sediment. k==1 i 1 ()2 As presented in Table 4, the concentration of As exceeded the where wjk is the distance decay function for location j and k, with the ERL limits for SWs of 1, 4, 5, 6, 7, 9 and 11, of which SW9 had the basic assumption that The local coefficient is higher when the ob- highest level of contamination. The lowest standard deviation and servations are closer to the sample at point j (Brunsdon, Fothering- coefficient of variation were also calculated for As, indicating the ham, & Charlton, 2002). lowest spatial variability (Table 2). Cu sediment pollution was

Table 3 The accuracy for land use map.

Land use Forest Irrigated agriculture Rainfed agriculture Residential area Range land Total Producer accuracy (%)

Forest 472 0 0 0 0 472 100 Irrigated agriculture 0 72 0 0 0 72 100 Rainfed agriculture 0 0 23 0 4 27 85.2 Developed area 0 0 0 42 27 69 60.9 Rangeland 0 0 0 0 297 297 100 Total 472 72 23 42 328 937 –

Kappa index: 0.94. Overall accuracy: 96.69%. M. Mohammadi et al. / International Soil and Water Conservation Research 7 (2019) 308–315 311

Table 4 Comparison of heavy metals concentrations in sub-basins with ERL and ERM values (Long, Macdonald, Smith, & Calder, 1995).

Heavy metal concentrations in this study

Sub-watersheds As (ppm) Cu (ppm) Ni (ppm) Cr (ppm)

1 8.65a 15.16 23.94a 93.18a 2 6.30 13.90 22.10a 75.90 3 6.10 17.00 19.20 100a 4 9.60a 34.30a 53.11b 122.85a 5 9.50a 14.80 23.50a 99.20a 6 8.40a 13.10 13.60 81.30a 7 10.40a 56.10a 88.51b 191a 8 1.40 65a 198b 384b 9 12.40a 44a 64.97b 141a 10 4.40 57a 128.15b 280a 11 10.80a 42.23a 65.97b 132.57a 12 8.10 32.62 62.28b 135.20a Proposed concentration (Long et al., 1995) ERL (ppm) 8.2 34.00 20.90 81.00 ERM (ppm) 70.00 270.00 51.60 370.00

a Exceeded the ERL. b Exceeded the ERM.

Fig. 2. Spatial distribution of heavy metals concentration in sub-watersheds. 312 M. Mohammadi et al. / International Soil and Water Conservation Research 7 (2019) 308–315

Fig. 3. Local R2 coefficient between irrigated agricultural land and heavy metals Local R2 values for the GWR models of irrigated agricultural land and heavy metal pollution. observed in SWs of 4, 7, 8, 9, 10 and 11. For Ni, all SWs except 3 and biological effects of Cr sediment pollution for all SWs. The amount 7 exceeded the permissible limit, the highest concentration of Cr contamination in the SW 8 was significantly higher than all of which was measured for SW10. Cr sediment contamination other SWs and exceeded ERM value. Cr is an abundant metal in the is observed in all SWs. According to Fig. 2, contamination of earth's crust, but industries such as smelting, tanning and metal sediments by metals (As, Cr, Cu and Ni) is more intensive in the finishing are also the main anthropogenic sources in the upstream than in the downstream. environment (Burbridge, Koch, Zhang, & Reimer, 2012). The spatial distribution of heavy metals concentrations showed There is a low probability of adverse biological effect for some that As exceeded the ERL limit in some SWs, indicating a low SWs caused by Cu, while the concentration of Cu in all SWs did not probability of adverse biological effects at the studied sites (Fig. 2). exceed the ERM. Sever pollution of Ni was observed in southern Due to the presence of coal mining activities, the hot spots of As SWs and the concentration of this element exceeded ERM for concentration were observed in the central and southern many of SWs. In other words, adverse biological impact occurred sub-watersheds, where it could have been transferred to the river in these regions. The natural sources of Ni include water erosion, by water erosion process. Mining activities are important origin of weathering and forest fires while discharge from industrial As pollution in many parts of the world, and several studies have activities, such as refining, melting and electroplating are addressed this challenge (Gomez-Gonzalez et al., 2016; Jamieson, considered as anthropogenic origins (Attig et al., 2010). 2014; Posada-Ayala, Murillo-Jiménez, Shumilin, Marmolejo- Rodriguez, & Nava-Sánchez, 2016). In terms of accumulation, it has 3.3. Spatial correlation between heavy metals versus land uses been stated that As can be accumulated in sediments, particularly obtained by GWR in river systems near to mining areas (Drewniak & Sklodowska, 2013). Cr, Cu and Ni showed homogenous spatial distribution and Land use types is highly related to human activities that their hot spot areas were located in the southern SWs, especially influence the sediment pollution in river system. In present the southeast. Totally, the concentration of Cr, Cu and Ni decreased research, GWR was used to assess the correlations between land from south to north and only the concentration of Cr exceeded the uses and heavy metals pollution in Talar watershed at the ERL threshold in SW 9. Also, there is a low probability of adverse sub-watershed scale. Figs. 3–5 show the local R2 between land M. Mohammadi et al. / International Soil and Water Conservation Research 7 (2019) 308–315 313

Fig. 4. Local R2 coefﬁcient between rainfed agriculture land and heavy metals.

uses (rain-fed agriculture, irrigated agriculture and developed Eroded soil from agricultural land in highland watersheds is area) and heavy metals (As, Cr, Cu and Ni). one of the main sources of water and sediment pollution in rivers There is a clearly positive correlation between the percentage (Niazi, Obropta, & Miskewitz, 2015). In fact, soil erosion in of irrigated land and metals for all of the sites, especially for As and agricultural lands are a pathway for transferring heavy metals Ni for SWs of 1 and 2. The lowest correlation of metals versus to rivers (Jiao, Ouyang, Hao, & Lin, 2015). About 99% of heavy irrigated land was found for Cu, among which only SW8 reach to metals could be accumulated in sediments after entering the 0.2–0.48. Moreover, the lowest local R2 was found in SWs of 3, 5, aquatic environment, so sediments are known as a sink for these 8 and 9 for Cu which measured 0–0.04 (Fig. 3). pollutants (Sun, Wang, & Hu, 2011). When sediment contamina- GWR results showed that the percentage of rainfed agricultural tion level is higher than the permissible limits, it is considered as land had positive relationships with As, Cr, Ni and Cu for all sampling contaminated (Yu et al., 2017). Various works have revealed point. Similar to the irrigated land, the percentage of rainfed relationship between agriculture land use types and heavy metals agricultural land also showed strong positive relationships with As (Tu, 2013; Wu et al., 2017). The concentration of Cu, Cd and Cr and Ni in downstream and Cr and Cu in upstream. Also, the weakest were reportedly controlled by reasonable application of fertilizer local R2 for Cu and Ni were obtained for SWs of 3, 9, 10 and 12. and pesticide (Gu et al., 2014; Xia et al., 2018). Jiao et al. (2015) Totally, the results of spatial correlation showed a positive and high highlighted the importance of land conversion in release of heavy coefﬁcient of local R2 for the two agricultural land use types with metals and proposed the identiﬁcation of erosion processes at heavy metals. In general, the changes in the percentage of watershed scale to better controlling of heavy metals pollution. agricultural land in SWs affect their spatial correlation with heavy Based on Fig. 5 and Table 1, there is a positive correlation metals. It can be concluded that the agricultural lands are the main between developed area and heavy metals. The highest correlation sources of sediment heavy metal pollution due to extensive between SWs versus heavy metals was obtained for As, Ni, Cr and application of fertilizers, pesticide and other anthropogenic inputs. Cu, respectively. The highest local R2 was obtained at the outlet 314 M. Mohammadi et al. / International Soil and Water Conservation Research 7 (2019) 308–315

Fig. 5. Local R2 coefﬁcient between developed area and heavy metals. point (SWs 1 and 2) for Ni and As, in the upstream regions, The highest correlation between developed areas and As was ob- indicating Cr and upstream and downstream for Cu. Several tained at the outlet because of highest level of urbanization. This studies have shown high relationship between water pollution study demonstrated the GWR as a good technique in under- parameters, including heavy metals and the land development standing the spatial variation of heavy metals and its relationship scales, such as urban, residential and commercial areas due to with land use at sub-scale. wastewater and pollutant discharges (Medeiros et al., 2017; Schoonover, Lockaby, & Pan, 2005; Woli, Nagumo, Kuramochi, & Hatano, 2004). Acknowledgements

The authors are grateful to Iran geological survey and mineral 4. Conclusions explorations for providing necessary data and also to Dr. Jafar Sey- fabadi for his great comments to improve the English of the paper. In order to assess heavy metal pollution risk in Talar river watershed, sediment quality data combined with GWR were used to identify spatial correlation between heavy metal pollution and References different land uses. Heavy metal contaminations was traced in the northern region of the watershed due to the intensive agriculture Attig, H., Dagnino, A., Negri, A., Jebali, J., Boussetta, H., Viarengo, A., & Banni, M. and human activities. A positive correlation was observed between (2010). Uptake and biochemical responses of mussels Mytilus galloprovincialis As and Ni and irrigated agriculture, specially for outlet point. In exposed to sublethal nickel concentrations. Ecotoxicology and Environmental – downstream, rainfed agriculture showed a positive relationships Safety, 73(7), 1712 1719. Attorre, F., De Sanctis, M., Francesconi, F., Bruno, F., et al. (2007). Comparison of with As and Ni but Cr in upstream. Also, the percentage of de- interpolation methods for mapping climatic and bioclimatic variables at re- veloped areas had the positive correlation with metals in all SWs. gional scale. International Journal of Climatology, 27(13), 1825–1843. M. Mohammadi et al. / International Soil and Water Conservation Research 7 (2019) 308–315 315

Brunsdon, C., Fotheringham, A. S., & Charlton, M. (2002). Geographically weighted Impact of forest maintenance on water shortages: Hydrologic modeling and summary statistics-a framework for localized exploratory data analysis. Com- effects of climate change. Science of The Total Environment, 615, 1355–1363. puters, Environment and Urban Systems, 26(6), 501–524. McMillen, D. P. (2004). Geographically weighted regression: The analysis of spatially Burbridge, D. J., Koch, I., Zhang, J., & Reimer, K. J. (2012). Chromium speciation in varying relationships. Oxford University Press. river sediment pore water contaminated by tannery effluent. Chemosphere, 89 Medeiros, A. C., Faial, K. R. F., do Carmo Freitas Faial, K., da Silva Lopes, I. D., de (7), 838–843. Oliveira Lima, M., Guimarães, R. M., & Mendonça, N. M. (2017). Quality index of Chen, X., Lu, X., & Yang, G. (2012). Sources identification of heavy metals in urban the surface water of Amazonian rivers in industrial areas in Pará, Brazil. Marine topsoil from inside the Xi’an Second Ringroad, NW China using multivariate Pollution Bulletin (pp. 156–164), 156–164. statistical methods. Catena, 98,73–78. Mennis, J. (2006). Mapping the results of geographically weighted regression. The Clement, F., Orange, D., Williams, M., Mulley, C., & Epprecht, M. (2009). Drivers of Cartographic Journal, 43(2), 171–179. afforestation in Northern Vietnam: Assessing local variations using geo- Micó, C., Recatalá, L., Peris, M., & Sánchez, J. (2006). Assessing heavy metal sources graphically weighted regression. Applied Geography, 29(4), 561–576. in agricultural soils of an European Mediterranean area by multivariate ana- Drewniak, L., & Sklodowska, A. (2013). Arsenic-transforming microbes and their lysis. Chemosphere, 65(5), 863–872. role in biomining processes. Environmental Science and Pollution Research, 20 Niazi, M., Obropta, C., & Miskewitz, R. (2015). Pathogen transport and fate modeling – (11), 7728 7739. in the Upper Salem River Watershed using SWAT model. Journal of Environ- Fotheringham, A. S., Charlton, M. E., & Brunsdon, C. (1998). Geographically mental Management, 151,167–177. weighted regression: A natural evolution of the expansion method for spatial Ogneva-Himmelberger, Y., Pearsall, H., & Rakshit, R. (2009). Concrete evidence & – data analysis. Environment and Planning A, 30(11), 1905 1927. geographically weighted regression: A regional analysis of wealth and the land Gao, J., & Li, S. (2011). Detecting spatially non-stationary and scale-dependent re- cover in Massachusetts. Applied Geography, 29(4), 478–487. lationships between urban landscape fragmentation and related factors using Posada-Ayala, I. H., Murillo-Jiménez, J. M., Shumilin, E., Marmolejo-Rodriguez, A. J., – geographically weighted regression. Applied Geography, 31(1), 292 302. & Nava-Sánchez, E. H. (2016). Arsenic from gold mining in marine and stream Gomez-Gonzalez, M. A., Voegelin, A., Garcia-Guinea, J., Bolea, E., Laborda, F., & sediments in Baja California Sur, Mexico. Environmental Earth Sciences, 75(11), Garrido, F. (2016). Colloidal mobilization of arsenic from mining-affected soils 996. by surface runoff. Chemosphere, 144,1123–1131. Schoonover, J. E., Lockaby, B. G., & Pan, S. (2005). Changes in chemical and physical Govind, P., & Madhuri, S. (2014). Heavy metals causing toxicity in animals and properties of stream water across an urban-rural gradient in western Georgia. fishes. Research Journal of Animal, Veterinary and Fishery Sciences, 2(2), 17–23. Urban Ecosystems, 8(1), 107–124. Gu, Y. G., Li, Q. S., Fang, J. H., He, B. Y., Fu, H. B., & Tong, Z. J. (2014). Identification of Sun, G.-X., Wang, X.-J., & Hu, Q.-H. (2011). Using stable lead isotopes to trace heavy heavy metal sources in the reclaimed farmland soils of the pearl river estuary in metal contamination sources in sediments of Xiangjiang and Lishui Rivers in China using a multivariate geostatistical approach. Ecotoxicology and Environ- China. Environmental Pollution, 159(12), 3406–3410. mental Safety, 105,7–12. Tu, J. (2011). Spatially varying relationships between land use and water quality Huang, J., Huang, Y., Pontius, R. G., & Zhang, Z. (2015). Geographically weighted across an urbanization gradient explored by geographically weighted regres- regression to measure spatial variations in correlations between water pollu- – tion versus land use in a coastal watershed. Ocean and Coastal Management, sion. Applied Geography, 31(1), 376 392. 103,14–24. Tu, J. (2013). Spatial variations in the relationships between land use and water Jamieson, H. E. (2014). The legacy of arsenic contamination from mining and pro- quality across an urbanization gradient in the watersheds of northern Georgia, – cessing refractory gold ore at Giant Mine, Yellowknife, Northwest Territories, USA. Environmental Management, 51(1), 1 17. Canada. Reviews in Mineralogy and Geochemistry, 79(1), 533–551. Woli, K. P., Nagumo, T., Kuramochi, K., & Hatano, R. (2004). Evaluating river water Jiao, W., Ouyang, W., Hao, F., & Lin, C. (2015). Anthropogenic impact on diffuse trace quality through land use analysis and N budget approaches in livestock farming – – metal accumulation in river sediments from agricultural reclamation areas areas. Science of the Total Environment, 329(1 3), 61 74. with geochemical and isotopic approaches. Science of the Total Environment, Wu, S.-S., Yang, H., Guo, F., & Han, R.-M. (2017). Spatial patterns and origins of heavy 536, 609–615. metals in Sheyang River catchment in Jiangsu, China based on geographically Kavian, A., Mohammadi, M., Gholami, L., & Rodrigo-Comino, J. (2018). Assessment weighted regression. Science of the Total Environment, 580,1518–1529. of the spatiotemporal effects of land use changes on runoff and nitrate loads in Xia, F., Qu, L., Wang, T., Luo, L., Chen, H., Dahlgren, R. A., & Huang, H. (2018). Dis- the Talar River. Water, 10(4), 445. tribution and source analysis of heavy metal pollutants in sediments of a rapid Li, X., Liu, L., Wang, Y., Luo, G., Chen, X., & Yang, X. (2013). Heavy metal con- developing urban river system. Chemosphere, 207,218–228. tamination of urban soil in an old industrial city (Shenyang) in Northeast China. Yongming, H., Peixuan, D., Junji, C., & Posmentier, E. S. (2006). Multivariate analysis Geoderma, 192,50–58. of heavy metal contamination in urban dusts of Xi’an, Central China. Science of Long, E. R., Macdonald, D. D., Smith, S. L., & Calder, F. D. (1995). Incidence of adverse the Total Environment, 355(1–3), 176–186. biological effects within ranges of chemical concentrations in marine and es- Yu, R., Hu, G., Lin, C., Yang, Q., Zhang, C., & Wang, X. (2017). Contamination of heavy tuarine sediments. Environmental Management, 19(1), 81–97. metals and isotopic tracing of Pb in intertidal surface sediments of Jinjiang Luo, P., Zhou, M., Deng, H., Lyu, J., Cao, W., Takara, K., & Schladow, S. G. (2018). River Estuary, SE China. Applied Geochemistry, 83,41–49.

GUIDE FOR AUTHORS

INTRODUCTION Permission of the International Research and Training Center on Erosion and Sedimentation or the China Water and Power Press is Types of paper required for resale or distribution outside the institution and for all Contributions falling into the following categories will be considered other derivative works, including compilations and translations (please for publication: Original research papers, reviews, letters to the editor. consult [email protected] or [email protected]). If excerpts Please ensure that you select the appropriate article type from the list from other copyrighted works are included, the author(s) must obtain of options when making your submission. Authors contributing to written permission from the copyright owners and credit the source(s) special issues should ensure that they select the special issue article in the article.] type from this list. Role of the funding source You are requested to identify who provided f nancial support for BEFORE YOU BEGIN the conduct of the research and/or preparation of the article and to Ethics in Publishing brief y describe the role of the sponsor(s), if any, in study design; in For information on Ethics in Publishing and Ethical guidelines for the collection, analysis and interpretation of data; in the writing of the journal publication see http://www.elsevier.com/publishingethics and report; and in the decision to submit the paper for publication. If the http://www.elsevier.com/ethicalguidelines. funding source(s) had no such involvement then this should be stated. Please see http://www.elsevier.com/funding. Conﬂ ict of interest Language and language services All authors are requested to disclose any actual or potential conf ict of interest including any f nancial, personal or other Please write your text in good English (American or British usage is relationships with other people or organizations within three accepted, but not a mixture of these). Authors who require information years of beginning the submitted work that could inappropriately about language editing and copyediting services pre- and post- inf uence, or be perceived to inf uence, their work. See also submission please visit http://webshop.elsevier.com/languageediting http://www.elsevier.com/conf ictsof nterest. or our customer support site at http://support.elsevier.com for more information. Submission declaration Submission Submission of an article implies that the work described has not been published previously (except in the form of an abstract or as part of a Submission to this journal proceeds totally online. Use the following published lecture or academic thesis), that it is not under consideration guidelines to prepare your article. Via the homepage of this journal for publication elsewhere, that its publication is approved by all authors (http://www.evise.com/evise/faces/pages/navigation/NavController. and tacitly or explicitly by the responsible authorities where the work jspx?JRNL_ACR=ISWCR) you will be guided stepwise through the was carried out, and that, if accepted, it will not be published elsewhere creation and uploading of the various f les. The system automatically including electronically in the same form, in English or in any other converts source f les to a single Adobe Acrobat PDF version of the language, without the written consent of the copyright-holder. article, which is used in the peer-review process. Please note that even though manuscript source f les are converted to PDF at submission for Changes to authorship the review process, these source f les are needed for further processing This policy concerns the addition, deletion, or rearrangement of author after acceptance. All correspondence, including notif cation of the names in the authorship of accepted manuscripts: Editor’s decision and requests for revision, takes place by e-mail and via the author’s homepage, removing the need for a hard-copy paper Before the accepted manuscript is published in an online issue: trail. If you are unable to provide an electronic version, please contact Requests to add or remove an author, or to rearrange the author the editorial off ce prior to submission [example: e-mail: paige.chyu@ names, must be sent to the Journal Manager from the corresponding gmail.com or [email protected]; telephone: +86 10 68786416; or author of the accepted manuscript and must include: (a) the reason the fax: 86 10 68411174]. name should be added or removed, or the author names rearranged and (b) written conf rmation (e-mail, fax, letter) from all authors that Additional Information they agree with the addition, removal or rearrangement. In the case Tables and f gures may be presented with captions within the main of addition or removal of authors, this includes conf rmation from body of the manuscript; if so, f gures should additionally be uploaded the author being added or removed. Requests that are not sent by the as high resolution f les. corresponding author will be forwarded by the Journal Manager to the corresponding author, who must follow the procedure as described PREPARATION above. Note that: (1) Journal Managers will inform the Journal Editors of any such requests and (2) publication of the accepted Use of wordprocessing software manuscript in an online issue is suspended until authorship has been It is important that the f le be saved in the native format of the agreed. wordprocessor used. The text should be in single-column format. Keep the layout of the text as simple as possible. Most formatting codes After the accepted manuscript is published in an online issue: Any will be removed and replaced on processing the article. In particular, requests to add, delete, or rearrange author names in an article do not use the wordprocessor’s options to justify text or to hyphenate published in an online issue will follow the same policies as noted words. However, do use bold face, italics, subscripts, superscripts etc. above and result in a corrigendum. When preparing tables, if you are using a table grid, use only one grid Copyright for each individual table and not a grid for each row. If no grid is [example: Upon acceptance of an article, authors will be asked used, use tabs, not spaces, to align columns. The electronic text should to complete a ‘Journal Publishing Agreement’. Acceptance of the be prepared in a way very similar to that of conventional manuscripts agreement will ensure the widest possible dissemination of information. (see also the Guide to Publishing with Elsevier: http://www.elsevier. An e-mail will be sent to the corresponding author conf rming receipt com/guidepublication). Note that source f les of f gures, tables and of the manuscript together with a ‘Journal Publishing Agreement’ text graphics will be required whether or not you embed your form or a link to the online version of this agreement. f gures in the text. See also the section on Electronic illustrations. To avoid unnecessary errors you are strongly advised to use the and in front of the appropriate address. Provide the full postal “spell-check” and “grammar-check” functions of your wordprocessor. address of each aff liation, including the country name, and, if available, the e-mail address of each author. LaTeX • Corresponding author. Clearly indicate who will handle If the LaTeX f le is suitable, proofs will be produced without rekeying correspondence at all stages of refereeing and publication, also the text. The article should preferably be written using Elsevier’s post-publication. Ensure that telephone and fax numbers (with document class “elsarticle”, or alternatively any of the other recognized country and area code) are provided in addition to the e-mail classes and formats supported in Elsevier’s electronic submissions address and the complete postal address. Contact details must system, for further information see http://www.elsevier.com/wps/f nd/ be kept up to date by the corresponding author. authorsview.authors/latex-ees-supported. • Present/permanent address. If an author has moved since the The Elsevier “elsarticle” LaTeX style f le package (including detailed work described in the article was done, or was visiting at the time, instructions for LaTeX preparation) can be obtained from the a “Present address” (or “Permanent address”) may be indicated Quickguide: http://www.elsevier.com/latex. It consists of the as a footnote to that author’s name. The address at which the f le: elsarticle.cls, complete user documentation for the class f le, author actually did the work must be retained as the main, bibliographic style f les in various styles, and template f les for a quick aff liation address. Superscript Arabic numerals are used for such start. footnotes. Article structure Abstract Subdivision - numbered sections A concise and factual abstract is required. The abstract should state brief y the purpose of the research, the principal results and major Divide your article into clearly def ned and numbered sections. conclusions. An abstract is often presented separately from the article, Subsections should be numbered 1.1 (then 1.1.1, 1.1.2, ...), 1.2, etc. so it must be able to stand alone. For this reason, References should (the abstract is not included in section numbering). Use this numbering be avoided, but if essential, then cite the author(s) and year(s). Also, also for internal cross-referencing: do not just refer to “the text”. Any non-standard or uncommon abbreviations should be avoided, but if subsection may be given a brief heading. Each heading should appear essential they must be def ned at their f rst mention in the abstract on its own separate line. itself. Introduction Keywords State the objectives of the work and provide an adequate background, avoiding a detailed literature survey or a summary of the results. Authors are invited to submit keywords associated with their paper. Material and methods Abbreviations Provide suff cient detail to allow the work to be reproduced. Methods Def ne abbreviations that are not standard in this f eld in a footnote already published should be indicated by a reference: only relevant to be placed on the f rst page of the article. Such abbreviations that modif cations should be described. are unavoidable in the abstract must be def ned at their f rst mention there, as well as in the footnote. Ensure consistency of abbreviations Theory/calculation throughout the article. A Theory section should extend, not repeat, the background to the Acknowledgements article already dealt with in the Introduction and lay the foundation for further work. In contrast, a Calculation section represents a practical Collate acknowledgements in a separate section at the end of the article development from a theoretical basis. before the references and do not, therefore, include them on the title page, as a footnote to the title or otherwise. List here those individuals Results who provided help during the research (e.g., providing language help, Results should be clear and concise. writing assistance or proof reading the article, etc.). Discussion Nomenclature and units This should explore the signif cance of the results of the work, not Follow internationally accepted rules and conventions: use the repeat them. A combined Results and Discussion section is often international system of units (SI). If other quantities are mentioned, appropriate. Avoid extensive citations and discussion of published give their equivalent in SI. Authors wishing to present a table literature. of nomenclature should do so on the second page of their manuscript. Conclusions The main conclusions of the study may be presented in a short Math formulae Conclusions section, which may stand alone or form a subsection of a Present simple formulae in the line of normal text where possible Discussion or Results and Discussion section. and use the solidus (/) instead of a horizontal line for small fractional terms, e.g., X/Y. In principle, variables are to be presented in italics. Appendices Powers of e are often more conveniently denoted by exp. Number If there is more than one appendix, they should be identif ed as A, B, consecutively any equations that have to be displayed separately from etc. Formulae and equations in appendices should be given separate the text (if referred to explicitly in the text). numbering: Eq. (A.1), Eq. (A.2), etc.; in a subsequent appendix, Eq. (B.1) and so on. Similarly for tables and f gures: Table A.1; Fig. A.1, etc. Footnotes Essential title page information Footnotes should be used sparingly. Number them consecutively throughout the article, using superscript Arabic numbers. Many Title. • Concise and informative. Titles are often used in information- wordprocessors build footnotes into the text, and this feature may be retrieval systems. Avoid abbreviations and formulae where possible. used. Should this not be the case, indicate the position of footnotes in Author names and afﬁ liations. • Where the family name may the text and present the footnotes themselves separately at the end of be ambiguous (e.g., a double name), please indicate this clearly. the article. Do not include footnotes in the Reference list. Present the authors’ aff liation addresses (where the actual work was done) below the names. Indicate all aff liations with Table footnotes a lower-case superscript letter immediately after the author’s name Indicate each footnote in a table with a superscript lowercase letter. Artwork communications are not recommended in the reference list, but Electronic artwork may be mentioned in the text. If these references are included in the General points reference list they should follow the standard reference style of the • Make sure you use uniform lettering and sizing of your original journal and should include a substitution of the publication date with artwork. either “Unpublished results” or “Personal communication” Citation • Save text in illustrations as “graphics” or enclose the font. of a reference as “in press” implies that the item has been accepted • Only use the following fonts in your illustrations: Arial, Courier, for publication. Times, Symbol. • Number the illustrations according to their sequence in the text. Web references • Use a logical naming convention for your artwork f les. As a minimum, the full URL should be given and the date when the • Provide captions to illustrations separately. reference was last accessed. Any further information, if known (DOI, • Produce images near to the desired size of the printed version. author names, dates, reference to a source publication, etc.), should • Submit each f gure as a separate f le. also be given. Web references can be listed separately (e.g., after the reference list) under a different heading if desired, or can be included A detailed guide on electronic artwork is available on our website: in the reference list. http://www.elsevier.com/artworkinstructions You are urged to visit this site; some excerpts from the detailed References in a special issue information are given here. Please ensure that the words ‘this issue’ are added to any references in the list (and any citations in the text) to other articles in the same Formats Special Issue. Regardless of the application used, when your electronic artwork is f nalised, please “save as” or convert the images to one of the following Reference management software formats (note the resolution requirements for line drawings, halftones, This journal has standard templates available in key reference and line/halftone combinations given below): management packages EndNote (http://www.endnote.com/ EPS: Vector drawings. Embed the font or save the text as “graphics”. support/enstyles.asp) and Reference Manager (http://refman.com/ TIFF: color or grayscale photographs (halftones): always use support/rmstyles.asp). Using plug-ins to wordprocessing packages, a minimum of 300 dpi. authors only need to select the appropriate journal template when TIFF: Bitmapped line drawings: use a minimum of 1000 dpi. preparing their article and the list of references and citations to these TIFF: Combinations bitmapped line/half-tone (color or grayscale): will be formatted according to the journal style which is a minimum of 500 dpi is required. described below.

If your electronic artwork is created in a Microsoft Off ce application Reference style (Word, PowerPoint, Excel) then please supply “as is”. The International Soil and Water Conservation Research (ISWCR) Please do not: follows the Publication Manual of the American Psychological

• Supply f les that are optimised for screen use (like GIF, BMP, PICT, Association (APA) 6th for reference lists and text citation. The author WPG); the resolution is too low; are referred to the Publication Manual of the American Psychological • Supply f les that are too low in resolution; Association, Fifth Edition, ISBN 1-55798-790-4, copies of which may or • Submit graphics that are disproportionately large for the content. be ordered from http://www.apa.org/books/4200061.html APA Order Dept., P.O.B. 2710, Hyattsville, MD 20784, USA. or APA, Color artwork 3 Henrietta Street, London, WC3E 8LU, UK. Details concerning Please make sure that artwork f les are in an acceptable format this referencing style can also be found at http://humanities.byu.edu/ (TIFF, EPS or MS Off ce f les) and with the correct resolution. linguistics/Henrichsen/APA/APA01.html. If, together with your accepted article, you submit usable color f gures then Elsevier will ensure, at no additional charge, that these f gures Examples: will appear in color on the Web (e.g., ScienceDirect and other sites) Text: All citations in the text should refer to: regardless of whether or not these illustrations are reproduced in color One author in the printed version. –Smith (2002) found… Figure captions –(Smith, 2002). Ensure that each illustration has a caption. Supply captions separately, Two Authors: not attached to the f gure. A caption should comprise a brief title (not –Smith and Jones (2003) found… on the f gure itself) and a description of the illustration. Keep text in –(Smith & Jones, 2003). the illustrations themselves to a minimum but explain all symbols and abbreviations used. Three or More Authors –Smith et al. (2001) found… Tables - (Phelps et al., 2004) Number tables consecutively in accordance with their appearance in –Smith et al. (2002) found… the text. Place footnotes to tables below the table body and indicate Groups as Authors: them with superscript lowercase letters. Avoid vertical rules. Be –1st Citation: sparing in the use of tables and ensure that the data presented in tables (American Psychological Association [APA], 2000). do not duplicate results described elsewhere in the article. –Subsequent Citations: References (APA, 2000). Citation in text Anonymous or No Author Please ensure that every reference cited in the text is also present –Use ﬁ rst few words of reference list entry (usually title): in the reference list (and vice versa). Any references cited in the (—Study Finds, 1995) abstract must be given in full. Unpublished results and personal (TEA, 2007) Authors with Same Surname Online resource from group/government –Include initials: S. T. Smith (2000) and J. D. Smith (1999) U.S. Department of Health and Human Services. (2003). Managing Two of more works within the same parentheses asthma: A guide for schools. Retrieved from http://www.nhibi.nih.gov/ –In order alphabetically, as they would appear in references, health/prof/lung/asthma/asth_sch.pdf separated by semi-colons Reference in other Language (Jones, 2003; Thomas, 2010) –If by same author, then by date Hughes, J. C., Brestan, E. V., & Valle, L. A. (2004). Problem-solving (Jones, 2003, 2007) interactions between mothers and children. Child and Family Behavior Therapy, 26(1), 1-16. (In Chinese) References should be arranged f rst alphabetically and then further sorted chronologically if necessary. More than one reference from the Journal abbreviations source same author(s) in the same year must be identif ed by the letters “a”, Journal names should be abbreviated according to “b”, “c”, etc., placed after the year of publication Index Medicus journal abbreviations: http://www.nlm.nih.gov/tsd/ serials/lji.html; Reference to a journal publication List of title word abbreviations: http://www.issn.org/2-22661-LTWA- Carlson, L. A. (2003). Existential theory: Helping school counselors online.php; attend to youth at risk for violence. Professional School Counseling, CAS (Chemical Abstracts Service): http://www.cas.org/sent.html. 6(5), 10-15. Sagarin, B. J., & Lawler-Sagarin, K. A. (2005). Critically evaluating Submission checklist competing theories: An exercise based on the Kitty Genovese murder. The following list will be useful during the f nal checking of an article Teaching of Psychology, 32(3), 167–169. prior to sending it to the journal for review. Please consult this Guide for Authors for further details of any item. Hughes, J. C., Brestan, E. V., & Valle, L. A. (2004). Problem-solving interactions between mothers and children. Child and Family Behavior Ensure that the following items are present: Therapy, 26(1), 1-16. One Author designated as corresponding Author: • E-mail address Journal with more than seven authors • Full postal address Gilbert, D. G., McCleron, J. F., Rabinovich, N. E., Sugai, C., Plath, • Telephone and fax numbers L. C., Asgaard, G., Botros, N. (2004). Effects of quitting smoking on EEG activation and attendtionlast for more than 31 days and are more All necessary f les have been uploaded severe with stress. Nicotine and Tobacco Research, 6, 249-267. • Keywords • All f gure captions Herbst-Damm, K.L., & Kulik, J.A. (2005). Volunteer support, • All tables (including title, description, footnotes) marital status, and the survival times of terminally ill patients. Health Psychology, 24, 225-229. doi: 10.1037/0278-6133.24.2.225 Further considerations • Manuscript has been “spellchecked” and “grammar-checked” Silick, T.J., & Schutte, N.S. (2006). Emotional intelligence and • References are in the correct format for this journal self-esteem mediate between perceived early parental love and adult • All references mentioned in the Reference list are cited in the text, happiness. E-Journal of Applied Psychology, 2(2), 38-48. Retrieved and vice versa from http://ojs.lib.swin.edu.au/index.php/ejap. • Permission has been obtained for use of copyrighted material from Reference to a book other sources (including the Web) Beck, C. A. J., & Sales, B. D. (2001). Family mediation: Facts, myths, • Color f gures are clearly marked as being intended for color and future prospects. Washington, DC: American Psychological reproduction on the Web (free of charge) and in print or to be Association. reproduced in color on the Web (free of charge) and in black-and- white in print Johnson, R. A. (1989). Retrieval inhibition as an adaptive mechanism • If only color on the Web is required, black and white versions of the in human memory. In H. L. RoedigerIII & F. I. M. Craik(Eds.), f gures are also supplied for printing purposes Varieties of memory & consciousness (pp. 309-330). Hillsdale, NJ: Erlbaum. For any further information please visit our customer support site at http://support.elsevier.com. English translation of a book: Lang, P. S. (1951). A philosophical essay on probabilities AFTER ACCEPTANCE (F. W. Truscott & F. L. Emory, Trans.). New York, NY: Dover. (Original work published 1814) Use of the Digital Object Identiﬁ er The Digital Object Identif er (DOI) may be used to cite and link to *In text, cite original date and translation date: (Lang, 1814/1951). electronic documents. The DOI consists of a unique alpha-numeric Dissertations and Theses character string which is assigned to a document by the publisher upon the initial electronic publication. The assigned DOI never Caprette, C. L. (2005). Conquering the cold shudder: The origin and changes. Therefore, it is an ideal medium for citing a document, evolution of snake eyes (Doctoral dissertation). Ohio State University, particularly ‘Articles in press’ because they have not yet received Columbus, OH. their full bibliographic information. The correct format for citing Pecore, J. T. (2004). Sounding the spirit of Cambodia: The living a DOI is shown as follows (example taken from a document in the tradition of Khmer music and dance-drama in a Washington, DC journal Physics Letters B): doi:10.1016/j.physletb.2010.09.059 community (Doctoral dissertation). Retrieved from Dissertations and When you use the DOI to create URL hyperlinks to documents on the Theses database. (UMI No. 3114720) web, they are guaranteed never to change. Caprette, C. L. (2005). Conquering the cold shudder: The origin Proofs and evolution of snake eyes (Doctoral dissertation). Retrieved from One set of page proofs (as PDF f les) will be sent by e-mail to the http://www.ohiolink.edu/etd/send-pdf.cgi?acc_num=osu1111184984 corresponding author (if we do not have an e-mail address then paper proofs will be sent by post) or, a link will be provided in the e-mail so Offprints that authors can download the f les themselves. Elsevier now provides The corresponding author, at no cost, will be provided with a PDF authors with PDF proofs which can be annotated; for this you will f le of the article via e-mail. For an extra charge, paper offprints can need to download Adobe Reader version 7 (or higher) available free be ordered via the offprint order form which is sent once the article from http://get.adobe.com/reader. Instructions on how to annotate PDF is accepted for publication. The PDF f le is a watermarked version of f les will accompany the proofs (also given online). The exact system the published article and includes a cover sheet with the journal cover requirements are given at the Adobe site: http://www.adobe.com/ image and a disclaimer outlining the terms and conditions of use. products/reader/systemreqs. More information about article offprint is available here: http:// If you do not wish to use the PDF annotations function, you may webshop.elsevier.com/ list the corrections (including replies to the Query Form) and return them to Elsevier in an e-mail. Please list your corrections quoting line AUTHOR INQUIRIES number. If, for any reason, this is not possible, then mark the corrections and any other comments (including replies to the Query Form) on a For inquiries relating to the submission of articles (including electronic printout of your proof and return by fax, or scan the pages and e-mail, submission) please visit this journal’s homepage. Contact details for or by post. Please use this proof only for checking the typesetting, questions arising after acceptance of an article, especially those relating editing, completeness and correctness of the text, tables and f gures. to proofs, will be provided by the publisher. You can track accepted Signif cant changes to the article as accepted for publication will articles at http://www.elsevier.com/trackarticle. You can also check only be considered at this stage with permission from the Editor. We our Author FAQs (http://www.elsevier.com/authorFAQ) and/or contact will do everything possible to get your article published quickly and Customer Support via http://support.elsevier.com. accurately – please let us have all your corrections within 72 hours. It is important to ensure that all corrections are sent back to us in one communication: please check carefully before replying, as inclusion of any subsequent corrections cannot be guaranteed. Proofreading is solely your responsibility. Note that Elsevier may proceed with the publication of your article if no response is received.