Parlar Scientific Publications, Angerstr

FEB – Fresenius Environmental Bulletin founded jointly by F. Korte and F. Coulston

Production by PSP – Parlar Scientific Publications, Angerstr. 12, 85354 Freising, Germany in cooperation with Lehrstuhl für Chemisch-Technische Analyse und Lebensmitteltechnologie, Technische Universität München, 85350 Freising - Weihenstephan, Germany Copyright © by PSP – Parlar Scientific Publications, Angerstr. 12, 85354 Freising, Germany. All rights are reserved, especially the right to translate into foreign language. No part of the journal may be reproduced in any form- through photocopying, microfilming or other processes- or converted to a machine language, especially for data processing equipment- without the written permission of the publisher. The rights of reproduction by lecture, radio and television transmission, magnetic sound recording or similar means are also reserved.

Printed in GERMANY – ISSN 1018-4619 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

FEB - EDITORIAL BOARD Environmental Toxicology: Prof. Dr. H. Greim Chief Editor: Senatskomm. d. DFG z. Prüfung gesundheitsschädl. Arbeitsstoffe TU München, 85350 Freising-Weihenstephan, Germany Prof. Dr. H. Parlar Institut für Lebensmitteltechnologie und Analytische Chemie Prof. Dr. A. Kettrup TU München - 85350 Freising-Weihenstephan, Germany Institut für Lebensmitteltechnologie und Analytische Chemie e-mail: [email protected] TU München - 85350 Freising-Weihenstephan, Germany

FEB - ADVISORY BOARD

Co-Editors: Environmental Analytical Chemistry:

Environmental Analytical Chemistry: K. Ballschmitter, D - K. Bester, D - K. Fischer, D - R. Kallenborn, N D.C.G. Muir, CAN - R. Niessner, D - W. Vetter, D – R. Spaccini, I Dr. D. Kotzias Via Germania 29 21027 Barza (Va) Environmental Proteomic and Biology: ITALY D. Adelung, D - G.I. Kvesitadze, GEOR A. Reichlmayr-Lais, D - C. Steinberg, D

Environmental Proteomic and Biology: Environmental Chemistry: Prof. Dr. A. Görg J.P. Lay, D - J. Burhenne, D - S. Nitz, D - R. Kreuzig, D Fachgebiet Proteomik D. L. Swackhammer, U.S.A. - R. Zepp, U.S.A. – T. Alpay, TR TU München - 85350 Freising-Weihenstephan, Germany V. Librando; I

Prof. Dr. A. Piccolo Università di Napoli “Frederico II”, Environmental Management: Dipto. Di Scienze Chimico-Agrarie Via Università 100, 80055 Portici (Napoli), Italy L.O. Ruzo, U.S.A - U. Schlottmann, D

Prof. Dr. G. Schüürmann UFZ-Umweltforschungszentrum, Environmental Toxicology: Sektion Chemische Ökotoxikologie Leipzig-Halle GmbH, K.-W. Schramm, D - H. Frank, D - D. Schulz-Jander, U.S.A. - Permoserstr.15, 04318 Leipzig, Germany H.U. Wolf, D – M. McLachlan, S

Environmental Chemistry: Prof. Dr. M. Bahadir Managing Editor: Institut für Ökologische Chemie und Abfallanalytik TU Braunschweig Dr. G. Leupold Hagenring 30, 38106 Braunschweig, Germany Editorial Chief-Officer:

Prof. Dr. M. Spiteller Selma Parlar Institut für Umweltforschung Universität Dortmund PSP- Parlar Scientific Publications Otto-Hahn-Str. 6, 44221 Dortmund, Germany Angerstr.12, 85354 Freising, Germany

e-mail: [email protected] - www.psp-parlar.de Prof. Dr. Ivan Holoubek RECETOX_TOCOEN Marketing Chief Manager: Kamenice 126/3, 62500 Brno, Czech Republic Max-Josef Kirchmaier MASELL-Agency for Marketing & Communication, Public- Relations Environmental Management: Angerstr.12, 85354 Freising, Germany Dr. H. Schlesing e-mail: [email protected] - www.masell.com

Secretary General, EARTO, Rue de Luxembourg,3, 1000 Brussels, Belgium

Prof. Dr. F. Vosniakos T.E.I. of Thessaloniki, Applied Physics Lab. Abstracted/ indexed in: Biology & Environmental Sciences, P.O. Box 14561, 54101 Thessaloniki, Greece BIOSIS, C.A.B. International, Cambridge Scientific Abstracts, Chemical Abstracts, Current Awareness, Current Contents/ Agricul- Dr. K.I. Nikolaou ture, CSA Civil Engineering Abstracts, CSA Mechanical & Trans- portation Engineering, IBIDS database, Information Ventures, NISC, Organization of the Master Plan & Research Alert, Science Citation Index Expanded (SCI Expanded), Environmental Protection of Thessaloniki (OMPEPT) SciSearch, Selected Water Resources Abstracts 54636 Thessaloniki, Greece

1 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

CONTENTS

PREFACE

SESSION: Spatial and urban planning

REVIEW ARTICLE

EXISTING AND FUTURE TECHNOLOGIES IN NOISE REDUCING PAVEMENTS 2803 Fotini Kehagia and Evangelos Manthos

TOWARD PRECISE SHORELINE DETECTION AND EXTRACTION FROM REMOTELY 2809 SENSED IMAGES WITH THE USE OF WET AND DRY SAND SPECTRAL SIGNATURES Panos Drakopoulos, George Ghionis, Kostas Lazogiannis and Serafim Poulos

EUROPEAN LAND COVER MAPPING SYSTEMS 2814 FOR THE NATURAL AND URBAN ENVIRONMENT Athanasios Moysiadis and Konstantinos Perakis

ORIGINAL PAPER

ASSESSING THE WALKING CONDITIONS IN PEDESTRIAN NETWORKS: 2819 THE CASE OF THE CITY CENTER IN LARISSA, GREECE Anastasios Tsakalidis, Alexandros Sdoukopoulos and Nikolaos Gavanas

THE ENVIRONMENTAL IMPACT OF TRAFFIC CALMING MEASURES 2826 TOWARDS THE REGENERATION OF AN URBAN AREA: THE CASE OF THE INTERVENTIONS ALONG THE AXIS OF AGIA SOFIA IN THESSALONIKI, GREECE Nikolaos Gavanas, Alexandros Sdoukopoulos and Magda Pitsiava-Latinopoulou

INVESTIGATION AND PREDICTION OF TRAFFIC TRAVEL TIME FOR THE 2832 ROAD NETWORK OF THESSALONIKI THROUGH SPATIAL TEMPORAL MODEL Konstantinos Lakakis, Kalliopi Kyriakou and Paraskevas Savvaidis

REDUCTION OF THE UNCERTAINTY IN THE FREQUENCY 2840 CALCULATION IN PORT ACCIDENTS THROUGH FUZZY LOGIC José R. González Dan and Rosa Mari Darbra

SESSION: waste monitoring and management

REVIEW ARTICLE

RATIONAL SOLITARY WELL SPACING IN SOIL REMEDIATION PROCESSES 2847 Marios S. Valavanides and Eugene D. Skouras

ORIGINAL PAPER

EFFECTS OF LEAKAGE OF COMPOUNDS FROM 2852 RADIOACTIVE OILY WASTE ON SOIL MICROBIAL COMMUNITY Svetlana Selivanovskaya, Raushaniya Gumerova and Polina Galitskaya

2800 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

ANTIMICROBIAL PERSONAL CARE PRODUCTS 2859 EFFECTS IN THE TERRESTRIAL ENVIRONMENT Jūratė Žaltauskaitė and Diana Miškelytė

NATURAL AND FE(II)-INDUCED REDUCTION OF HEXAVALENT CHROMIUM IN SOIL 2865 Alja Margon, Massimiliano Valentini, Claudio Mondini, Tania Sinicco, Paolo Sequi and Liviana Leita

GREEN MARKET: A COMPARATIVE STUDY IN ROMANIA 2870 Mirela Panainte, Iuliana Caraman, Valentin Nedeff, Vassilis Inglezakis, Christos Venetis, Philipos Coutsikos, Gabriel Lazar and Narcis Barsan

WASTE PREVENTION CAMPAIGN REGARDING THE WASTE FRAMEWORK DIRECTIVE 2876 AntonisA. Zorpas, Katia Lasaridi, Costas Abeliotis, Irene Voukkali, Pantelitsa Loizia, Anastasios Georgiou, Christina Chroni, Korina Phanou and Nantia Bikaki

ASSESSING RECYCLING POTENTIAL IN LOCAL LEVEL: 2884 THE CASE OF NEAPOLI-SYKIES MUNICIPALITY, GREECE Christos Karkanias, George Perkoulidis, Nikolaos Grigoriadis, Stavros Stafylas, Emmanouil Dagdilelis, Eleni Feleki and Nicolas Moussiopoulos

AN ANALYSIS OF AN INNOVATIVE CONCEPT REGARDING 2890 EXCHANGING RECYCLABLE WITH AGRICULTURAL PRODUCTS Konstantinos Aravossis and Christina Fountzoula

SESSION: Water pollution

REVIEW ARTICLE

ADAPTATION OF PORT WASTE RECEPTION FACILITIES TO 2895 BALLAST WATER TREATMENT SYSTEM: TURKISH PORT PERSPECTIVE Tanzer Satır and Neslihan Doğan-Sağlamtimur

GREY WATER FOOTPRINT OF CROPS AND CROP-DERIVED PRODUCTS: 2899 ANALYSIS OF CALCULATION METHOD Chrysi S. Laspidou

ORIGINAL PAPER

EVALUATION OF WATER QUALITY OF A MOUNTAINOUS STREAM 2904 (PELION, CENTRAL GREECE) USING BENTHIC MACROINVERTEBRATES Christos I. Rumbos and Athanassios Kungolos

ΟIL SPILL DETECTION FROM RADARSAT-1 SYNTHETIC APERTURE 2909 RADAR IMAGERY AT NORTHERN ENTRY OF BOSPORUS STRAIT, TURKEY Burcu Ozsoy-Cicek

CORRELATION BETWEEN CHLOROPHYLL A CONCENTRATION AND SEA SURFACE 2919 TEMPERATURE IN THE EASTERN MEDITERRANEAN SEA USING GIS AND SATELLITE DATA Dimitra Kitsiou and Konstantinos Topouzelis

TOXICOLOGICAL ASSESSMENT OF THE EFFECTS OF 2926 CLOSED LANDFILL ON NEIGHBOURING HYDROECOSYSTEM Gintaras Svecevičius, Nijolė Kazlauskienė, Asta Slučkaitė and Tomas Makaras

2801 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

ASSESSING TOOLS AND METHODS FOR WATER QUALITY MONITORING 2933 IN MEDIUM SIZED RIVERS. THE CASE OF ARCADIKOS RIVER (MESSINIA, GREECE) Konstantinos C. Gritzalis, Evangelia Th. Anastasopoulou, Vasiliki Markogianni, and Nikolaos A. Georgiopoulos

INDEX 2939

2802 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

PREFACE

The present special issues of Fresenius Environmental Bulletin contain selected and revised papers that have been presented at the 4th International Conference on Environmental Management, Engineering, Planning and Economics (CEMEPE) which was held together with the conference of the Society of Ecotoxicology and Environmental Safety (SECOTOX) and took place on the Greek island of Mykonos from June 24 to 28, 2013. The conference was organized by the Department of Planning and Regional Development (University of Thessaly), the Sector of Industrial Manage- ment and Operations Research (School of Mechanical Engineering, National Technical University of Athens), the De- partment of Mechanical Engineering (Aristotle University of Thessaloniki), the Food Technology Department (Techno- logical Educational Institute of Thessaloniki) and the Department of Environmental Sciences (German Research Center for Environmental Health). An international board of valued scientists comprised its scientific committee. Regarding the venue of the conference, Mykonos is one of the best-known Greek islands, acclaimed both for its cosmopolitan environment and its exquisite natural beauty.

CEMEPE congresses have currently grown to be keystone congregations for environmental scientists; they are held on a biennial basis in selected places around Greece characterized by both natural splendor and cultural significance. The aim of these symposia is the presentation and sharing of research related to environmental sciences, covering varied topics of high environmental relevance. 207 oral and 89 poster presentations were included in the five days of the congress in three parallel sessions.

Fifty five (55) papers were initially submitted to be considered for publication in this special issue. The finally selected papers that were included in the two special issues represent some of the most important aspects discussed in the congress and may approximately fall in the categories of “Coastal planning and policy”, “Energy and environment”, “Environmental economics”, “Natural resource management”, “Spatial and urban planning”, “Waste monitoring and management” and “Water pollution”. We would like to thank all the authors for their worthy contributions to this special issue as well as the Editor of the Journal Professor Harun Parlar for his undeterred help throughout the handling of manuscripts. We also thank the reviewers for the time and effort they devoted in order to ensure high standards for the submitted manuscripts.

The next conference, CEMEPE5/SECOTOX, will be held on the island of Skiathos in June 2015 and will be organized by Aristotle University of Thessaloniki, University of Thessaly and SECOTOX. We are looking forward to the participation of all previous CEMEPE participants as well as to new participations from environmental scientists. Until the next meeting, we send you all our warmest regards.

Dr. Christina Emmanouil School of Rural and Surveying Engineering, Aristotle University of Thessaloniki, 54124 Thessaloniki, Greece

Assistant Professor Chrysi Laspidou Department of Civil Engineering, University of Thessaly, Pedion Areos, 38334 Volos, Greece

Professor Athanassios Kungolos Department of Planning and Regional Development, University of Thessaly, Pedion Areos, 38334 Volos, Greece

2803 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

EXISTING AND FUTURE TECHNOLOGIES IN NOISE REDUCING PAVEMENTS

Fotini Kehagia* and Evangelos Manthos

Department of Civil Engineering, Aristotle University of Thessaloniki, 54124, Thessaloniki, Greece

Presented at the 4th International Conference on Environmental Management, Engineering, Planning and Economics (CEMEPE), June 24 to 28, 2013, Mykonos, Greece

ABSTRACT The entry into force of the European Directive on En- vironmental Noise 2002/49/EC known as the Environ- Noise can produce negative effects on people's health mental Noise Directive requires Member States to produce since it interferes with basic activities such as sleeping, strategic noise maps followed by action plans. At national resting, studying and communicating. Road traffic noise is level, approaches differ slightly from country to country [1, the most significant source of environmental noise pollution 2]. By the increasing concern about the excessive cost, in cities. The use of noise reducing asphalts is one of the engineers have turned their attention to deal with the noise means of noise abatement measures applied along exist- problem at its source. Source control strategies include ing national roads as well as in new road construction quieter vehicles and tires, speed control, additional building projects. The main theme of this paper is the presentation insulation, more aggressive building codes for new con- of different types of noise reducing asphalts, specifically, struction, and noise-reducing pavements. porous asphalt, asphalt concrete for very thin layers and Stone Mastic Asphalt (SMA) implemented in Europe and This paper deals with the presentation of different types USA. Additionally, new noise-reducing pavement tech- of noise reducing asphalts. Porous asphalt, asphalt concrete nologies of the future are presented. for very thin layers and Stone Mastic Asphalt (SMA), implemented in Europe and USA, are presented. Their noise

performance in comparison to other asphalts is outlined.

KEYWORDS: Additionally new noise reduction pavement technologies noise, pavement, road construction are described.

2. NOISE ABATEMENT MEASURES 1. INTRODUCTION In Europe, the main noise problems occur along the Complaints of excessive noise are becoming one of the existing road network. The magnitude of the problems in- most predominant nuisances in communities near major creases with traffic volume. Therefore, noise abatement roadways. Noise can produce negative effects on people's along these roads is crucial in order to launch a process health since it interferes with basic activities such as whereby noise exposure is reduced in the long term. As sleeping, resting, studying and communicating. Road traffic early as the 1970s many European countries had begun noise is the most significant source of environmental noise their research toward lowering road traffic noise. Devel- pollution in cities. The EU Commission estimates that up opment and implementation of noise barriers, building to 80 million people (20%) in the EU suffer from unaccepta- insulation, building codes, noise limitation on new roads, ble noise levels. Additionally, 170 million citizens are traffic management and different types of surface asphalt living in so-called “grey areas” where the noise levels are courses that reduced noise were underway in many coun- such as to cause serious annoyance during the daytime. The tries based on resident complaints of excess noise levels. economic damage of environmental noise in the EU could The use of different types of noise abatement measures in reach as much as 38 billion €. Germany estimates that for Europe along new and existing national roads are present- each 1dB(A) increase in noise levels, average property ed in Figure 1: as it can be seen, noise barriers are the values fall by 0.5% [1]. dominant type of mitigation measure, while pavements (noise reduction surface courses) are the second in use in * Corresponding author new road projects and the third in existing roads.

2803

FIGURE 1 - The percentage of European countries that use different types of noise abatement measures along new and existing national roads [adapted from Ref.2].

During the last 15-20 years the asphalt industry has asphalt had limited use in the beginning, but nowadays been faced with growing demands for specialized prod- more than 195 km2 of main road surface is covered with ucts with the additional feature of noise reduction. This porous asphalt, in many European countries, such as fact has led to the development of a range of new surface Netherlands, Denmark, France, Italy, UK [3]. Porous asphalt types and the improvement of existing ones. The asphalt is a gap graded mixture. It’s fundamental charac- noise reducing properties of asphalt depend principally on teristic, in relation to conventional dense asphalt concrete, its porosity, namely the surface macrotexture and air void is the high air void content (> 18%). This high air void content. Therefore, the design of noise reduction surface content allows water and air to penetrate through the asphalts consists of optimizing these factors without com- mixture and as a result, improves driving conditions dur- promising other asphalt characteristics as skid resistance, ing rain by reducing splash and spray and provides better durability, evenness etc. skid resistance in wet conditions. Figure 2 shows a porous Generally, the comparison of noise reducing pave- asphalt surface in comparison to a dense asphalt concrete ments performance is related to a reference pavement. surface. The most commonly used reference pavement is a tradi- As far as noise reduction is concerned porous asphalt tional dense asphalt concrete, namely a continuously diminishes traffic noise in two ways: By reduction of graded mixture where the aggregate structure is filled rolling noise and by noise absorption [3]. In the first case with a mortar of very fine aggregate and binder, which the air is pressed into the pores of the mixture and rolling has a widespread use in most European countries. How- noise is reduced. In the second case the noise generated ever, in the UK the principal surface course has been by vehicles is absorbed by the road structure. Various chipped hot rolled asphalt, which is about 2 dB(A) noisier researchers have showed the beneficiary effect of noise than the conventional asphalt concrete. reduction of porous asphalt. Tesoriere et al [4] showed that the reduction can be up to 6 dB (A) compared to concrete layers, while Nicholls [5] showed a reduction of 3. EXISTING NOISE REDUCING PAVEMENTS 2 to 6 dB (A) compared to hot rolled asphalt (HRA) surfaces. According to the Swiss standards [6], under dry 3.1 Porous Asphalt conditions in a 70 dB (A) area, by using porous asphalt a The first porous asphalt was developed by TRL in the noise reduction of 5 dB (A) can be achieved. 1950s for use on airport runways. The use of this type of

(a) (b) FIGURE 2 - The surface of porous asphalt (a) compared to dense asphalt concrete (b)

2804 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

(a) (b)

FIGURE 3 - Double layered porous asphalt (a) Double layered is applied on top of an existing main road in Copenhagen, Denmark (b) [1, 3].

Additionally studies in the Netherlands showed that of macrotexture and good noise reduction properties. In porous asphalt give, a noise reduction of 4 dB (A) on comparison to standard dense or semi-dense asphalt con- average, compared to a conventional dense asphalt concrete, the ACVTL present a noise reduction in the range crete [3]. It is noted that in the Netherlands, the use of of 3 to 5 dB(A) [3]. Also it has been recorded a noise re- porous asphalt on highways is a national policy guideline. duction of approximately 3dB(A) when compared to HRA In 2008, around 80% of the Netherlands network featured [11, 12]. porous asphalt. In USA porous asphalts are defined as open graded friction courses (OGFC). Studies have shown 3.4 Stone Mastic Asphalt (SMA) that OGFC gives a noise reduction of 3-5 dB (A) in com- Stone mastic asphalt is a gap graded mix, with high parison to dense asphalt [7]. A specialized OGFC is Rub- binder content and a high proportion of coarse aggregate berized OGFC, which apart from the noise reduction prop- providing an interlocking stone-on-stone aggregate skele- erties, similar to those of OGFC, has also enhanced durabil- ton. Common layer thickness varies between 15 (SMA ity [7, 8]. The asphalt rubber technique has also gained 0/6 mm) and 45 mm (SMA 0/16 mm). An SMA 0/16mm interest in European countries. Portugal has extensive expe- surface is shown in Figure 4. SMA surface courses with a rience on this technology. In Sweden, at the end of 2009, maximum aggregate size of 11 mm (0/11 mm) or less about 15 test sections had been constructed using approx- (0/6 mm) have given up to 2 - 3 dB (A) less noise com- imately 57.000 tonnes of rubberized asphalt covering pared with dense asphalt concrete [3, 13]. Research made about 100 lane-km [13, 14]. in Germany concluded in almost the same result, namely reduction of 2-2.5 dB(A) compared to dense asphalt [7]. 3.2 Doubled layered porous asphalt Double layered porous asphalt was developed in the Netherlands in the early 1990s as an evolution of single layer porous asphalt in order to offer improved noise performance compared to the single layer. Double layered consists of two layers of porous asphalt: a coarse, single sized bottom layer (11/16 mm) which acts as a drainage layer and a thin textured top layer (4/8 mm) where the sand fraction is left out which attribute to advantageous acoustic properties Figure 3 shows top and bottom layer of double layered asphalt and application of the mix in Denmark. The thickness of the bottom and upper layer is normally 45 mm and 25 mm, respectively. Noise measurements show considerable noise reduction, about 5-6 dB(A), at high speeds FIGURE 4 - Stone mastic asphalt surface (highways) and 5 dB(A), at low speeds (urban roads) in comparison to dense asphalt [3]. 4. ROAD MAINTENANCE AND 3.3 Asphalt concrete for very thin layers NOISE REDUCING PAVEMENTS Asphalt concretes for very thin layers (ACVTL) are laid to a thickness of 20-30 mm. They were developed, Generally, the choice of the final surface is always a initially, in France in the 1980s, in order to replace the balance in satisfying many characteristics as skid re- more traditional surface courses of 30-40 mm thickness. sistance, evenness, bearing capacity, traffic safety, They are gap-graded mixtures which provide a high level noise. Usually, there could be a conflict in the selection

2805 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

of the surface material when optimizing the characteris- ported to give noise reductions of circa 6 dBA compared tics individually. The design of low noise road pavements to conventional dense asphalt layers [14, 15] (Figure 5). depends on the surface macrotexture and aggregate grad- ing and consists of optimizing these factors without com- 5.2 Modieslab promising other functional pavement characteristics. The Modieslab is a prefabricated road slab manufactured type of pavement has an impact on the tyre/road noise with cement concrete elements, constructed from two dif- level on a given road, therefore, in principle, noise abate- ferent open layers of concrete on a further supporting layer ment can be integrated into road maintenance procedures. of concrete, designed to be anchored to piled foundations In a comprehensive planning approach, the different pa- (Figure 5). The slabs can be constructed with channels for rameters that characterize the surface have to be evaluated utilities connections, and the open concrete surface has and balanced against each other. been designed to have low noise characteristics and low Noise, in Europe, is not generally one of the criteria rolling resistance. Water run-off channels are designed to used when determining which roads need maintenance or allow rainwater to be quickly discharged, and it is claimed, new pavements. Nevertheless, in 65% of the member to have road surface self cleaning properties. The concept states, noise is considered when deciding on the type of was originally introduced to the innovation programme pavement to be used when a road needs repairs or a new ‘Roads of the Future’ established by the Netherlands pavement [1]. However, the use of noise-reducing pave- Public Works and Water Management department. Since ments can be promoted by introducing guidelines, legisla- this time, three trials have taken place, most recently a tion, or recommendations on how and when to use such 100 meters test lane on the A12 motorway near Utrecht, pavement types in the construction of new roads or the involving a full-scale motorway set up under intensive maintenance of existing roads [2]. traffic conditions. Test results have demonstrated shape retention, noise reduction and fast run-off of rainwater [14, 15]. 5. NOISE REDUCING PAVEMENTS OF FUTURE 5.3 Ecotechnic Road Systems Ecotechnic Road Systems is a modular concept of There have been significant initiatives in several coun- mitigating solutions concerning an appropriate combina- tries to reduce the environmental impact of noise both in tion of low noise pavement and anti-noise barrier subsys- terms of actions to reduce at source and also to reduce the tems. “Euphonic pavement” composite multilayer consists associated costs. This has led to the development of a of double porous layer, a micro porous asphalt wearing large supply of new products for pavement surfacing, course (20mm, 0-5mm basaltic) and porous asphalt base especially for heavy traffic roads. course (40mm, 0-25mm limestone), connected by neck to a layer with a Helmholtz resonators system in either the 5.1 Road on a roll - Rollpave third or fourth layer. The Helmholtz resonators are de- The Rollpave concept has been developed from the signed to absorb noise over the range 100 to 250 Hz. Com- Dutch ‘Road to the Future’ Innovation programme, and is pared to the reference pavement, these pavements typically a prefabricated asphalt mat, approximately 30 mm thick, are 2 to 4 dB better from 80 to 250 Hz, 8 to 14 dB better that is laid on top of an existing pavement with sufficient from 315 to 800 Hz, and 2 to 6 dB better from 800 to structural strength to carry the design traffic. Its design 5,000 Hz. The main remedial measures of the acoustic allows for repairs to be carried out 50% faster than con- control/abatement solution designed and implemented for ventional methods, whilst providing a low noise surface the city of Genoa, in the framework of the NR2C project course, comparable to a 2-layer porous asphalt; this is re- [14, 16].

FIGURE 5 - Rollpave Test Section (left). Modieslab section being put into place (right) [14, 15]

2806 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

FIGURE 6 - “Euphonic pavement” [16]

6. CONCLUSIONS REFERENCES

Various noise abatement measures are applied along [1] CEDR (2010) Noise management and abatement. existing national roads as well as in new road construction [2] Danish Road Institute (2008) Use of noise reducing pave- projects. Pavements are a major component of road con- ments-European experience, Technical note 69. struction and the noise generated by or from them, during [3] EAPA (European Asphalt Pavement Association) (2007) their service life, is mainly related to the type of surface Abatement of traffic noise, the arguments for asphalt. asphalt course used. The selection of this asphalt course depends on design, economical and various other reasons [4] Tesoriere G, Canale S and Ventura F. (1989) Analysis of Draining Pavements from a Point of View of Phono- generated by the project’s needs and management. Factors Absorption. Proc. 4th Europian Symp., Madrid, pp.878-881. that are considered include roadside development, road layout, traffic composition and state of road maintenance [5] Nicholls J.C. (1996) Review of UK porous asphalt trials. which potentially impact on the performance of low noise TRL Report 264, Transport Research Laboratory, Crowthorne, UK. surfaces. It is well documented though, that certain types of surface asphalts (porous asphalts, double layered po- [6] SN 640 433b: Drainasphaltschichten, Konzeption Anforder- rous asphalts, ACVTS, SMA) have much better noise ung, Ausführung (2001) Swiss Standards for porous asphalt Original in German and French. reduction performance than others (dense and semi dense asphalt concrete or hot rolled asphalt) and their use has [7] Arizona Department Transportation (2008) Survey of traffic been beneficiary in terms of pavement noise reduction. noise reduction products, materials and technologies, Final Report 584 Besides this beneficiary effect, new technologies such as Rollpave, Modieslab and Ecotechnic road systems are [8] Kehagia F, Mavridou S., Oikonomou N. (2013) Noise meas- under development and are starting to be implemented in urements on pavement made from rubberized asphalt mixture in the city of Lamia Greece, Protection and Restoration of roadworks worldwide. By using the existing and new the Environment XI, 3-6 July, Thessaloniki, Greece technologies noise reducing pavements are becoming one of the major noise abatement measures in urban and rural [9] SILVIA Project Deliverable, Traffic management and Noise Reducing pavements- Recommendations on additional noise highway projects. reducing measures

The authors have declared no conflict of interest. [10] SILENCE (2008) Low-noise surfaces for urban roads and streets, Deliverable F.D3 [11] Cattley D, Towards sustainability Good Practice Case Sheets, The Use of thin surfacing systems – A quiet revolution. The Institution of Civil Engineers, UK. [12] Greer G. (2006) Stone Mastic Asphalt-A review of its noise reducing and early life skid resistance properties, Proceed- ings of ACOUSTICS 2006, Christchurch, New Zealand. [13] Abbot P. G., P. A. Morgan and B. McKell, (2010) A review of current research on road surface noise reduction techniques. Published project report PPR443. [14] FHWA (2005) Quiet pavement systems in Europe, Interna- tional Technology Exchange Program. [15] Martin van de Ven (2010) Innovative pre-fabricated pavement systems. Presentation. Delft University of Technology, Netherlands.

2807 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

[16] NR2C (2008) New Road Construction Concepts, Towards reliable, green, safe & smart and human infrastructure in Eu- rope.

Received: March 19, 2014 Accepted: April 02, 2014

CORRESPONDING AUTHOR

Fotini Kehagia Department of Civil Engineering Aristotle University of Thessaloniki 54124 Thessaloniki GREECE

Phone: +30 2310994380 Fax: +302310995764 E-mail: [email protected]

FEB/ Vol 23/ No 11a/ 2014 – pages 2803 - 2808

2808 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

TOWARD PRECISE SHORELINE DETECTION AND EXTRACTION FROM REMOTELY SENSED IMAGES WITH THE USE OF WET AND DRY SAND SPECTRAL SIGNATURES

Panos Drakopoulos1,*, George Ghionis2, Kostas Lazogiannis2, and Serafim Poulos2

1Department of Optics and Optometry, Technological Educational Institute of Athens, Ag Spyridonos 12210, Athens, Greece 2 Department of Geology and Geoenvironment, University of Athens, Zografou, 15784, Athens, Greece

Presented at the 4th International Conference on Environmental Management, Engineering, Planning and Economics (CEMEPE), June 24 to 28, 2013, Mykonos, Greece

ABSTRACT Within this concept, high resolution remote sensing information from already operational satellite sensors (e.g. Shoreline displacements are due to physical, anthro- SPOT-5, IKONOS, Quickbird, Formosat, WorldView-2, pogenic and environmental factors such climate change. Geoeye, TerraSAR-X, TanDEM-X), recently launched or Remote sensing has become an essential tool for monitor- future missions (e.g. SPOT-6/7, Pleiades, TerraSAR-X2, ing shoreline changes and understanding the physical pro- with accuracy down to 0.5 m) could be used in principle cesses in the coastal environment. However, ground truth- to extract information on the morphological characteris- ing is necessary in order to detect the limits of sea-level tics of the beach (e.g. shoreline length/position, beach inundation and /or wave run up. Within the framework of width and area) and to contribute in a better understand- the THALES-DAPHNE project, we have initiated an ing of the medium-term beach zone dynamics. The proce- investigation regarding the reflectance properties and dure includes the collection of in-situ reflectance meas- spectral signatures of various samples of dry and wet urements, the creation of a local database of spectral sig- beach sands. To this end, spectral reflectance characteris- natures for the differentiation between the wet part of the tics for a collection of sand samples from various Hellenic lower beach face (swash zone) and the dry upper part, and beaches have been measured in the laboratory under dry finally the correlation with corresponding satellite data. and wet conditions. In addition ongoing field measure- This will be a vital step that will lead to an improved de- ments are in process along the beach face of the delta of termination of the shoreline position through the assess- the River Pinios,that debouches at the NW coast of the ment of the satellite information accuracy and a better microtidal Aegean Sea. image classification of the beach sediment nature.

The selection of the deltaic coast of Pinios river (Figure

KEYWORDS: Environmental monitoring; Remote Sensing, shore- 1) as a study area, is based on the following facts: (i) Pinios line position is the largest Greek river having its drainage basin (some 2 10,000 km ) within Greek geographical boundaries; (ii) it is the only river in Greece with very limited flow controls (in 1. INTRODUCTION less than 10% of its catchment; (iii) the Pinios delta is recog- nized as an environmentally sensitive coastal zone and as Deltaic coasts are amongst the most vulnerable coastal such, it is protected by the NATURA 2000 network. environments either because of human impact (e.g. the construction of dams) and/or because of climate change The Pinios River delta has a cuspate shape and is ex- (e.g. sea level rise, changes in storminess [1]). The study of posed to moderate wave activity (mean wave heights ~1 m); the seasonal and decadal changes of shoreline position is the highest (although rare) offshore waves approach from the one of the key processes in understanding the morphody- SE and because of the long fetches, can locally exceed 4 m namic evolution of such sedimentary environments. In in height [2]. The deltaic coast is characterized by sandy particular, for the Pinios delta, recent observations (Thalis- beaches with low sand dunes, consisting (mostly) of me- DAPHNE research project) along the deltaic coast revealed dium – sized sand abundant in quartz and feldspars. seasonal changes in shoreline position in the order of few The scope of this contribution is to develop a tool that meters, with the exception of its mouth, where its south- will be used to identify the shoreline position with high ern part retreated by more than 10 m after the SE storm of accuracy. The results will be verified using satellite imag- January 2012. es calibrated through ground truthing on both the wet lower shoreface and the adjacent dry (beyond the wave * Corresponding author reach) part of the upper beach face. This work also con-

2809 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

FIGURE 1 - Map of the study area and sampling site (image source: Google Earth).

tributes to an effective coastal zone management facilitat- data were collected throughout the 350 -850 nm range ing a rapid and cost-effective assessment of shoreline with 1 nm resolution. retreat induced by changes in mean sea level and storm The wavelength dependent reflectance measured is surges, in order to identify hot-spots of beach erosion. defined as:

LLsand()  dark () RR()  g (1) 2. MATERIALS AND METHODS LLspec()  dark () Where L (λ) is the radiance from the sand sample, For the development of the procedure, reflectance sand Lspec(λ) the radiance from the spectralon™ plate and Rg its spectra were initially acquired in the laboratory. For this known reflectance. reason the sand samples were placed in ~7 cm diameter, 1.5 cm deep tin plates. The measurements were performed For classification purposes, the following indices by means of a reflectance probe placed at a 45o monitor- were estimated [3,4]: 1/2 ing and illumination angle thus avoiding possible specular R222GB(2 RGB  ) components. The sensing output of the probe was fed to BI, HI (2) an HR4000 miniature spectrometer while the illumination 3()GB input was coupled to a tungsten-halogen light source where BI, the brightness index, is indicative of the operating at 3100 K (all components supplied by Ocean average reflectance magnitude and HI the hue index Optics). Reference reflectance was provided by a 99% related to the dominant colour of the sample. We have diffuse reflectance standard spectralon™ plate. For the also introduced the simple slope index: field measurements, the various radiances were probed (3) under daylight conditions normal to the surface through a SI () R B 6o FOV Gershun tube attached via an optical fiber to a This index is proportional to the finite element first portable Jaz (Ocean Optics) field spectrometer. Spectral derivative (slope from red to blue) and is closely tied to

2810 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

the water content of the sample. The parameters R, G, and cending slope of the spectrum towards longer wave- B denote the reflectances at red, green and blue wave- lengths. Colored sand samples generally have spectral lengths respectively. The indices were evaluated at the curvatures associated to the dominant color (Figure 2). center of the corresponding band of the IKONOS satellite Grain size seems to contribute to the overall albedo which (R=665 nm, G=551 nm, B=480 nm) using the appropriate increases with decreasing grain size. boxcar bandwidth. Experiments including dry and wet samples showed In addition we evaluated the Y – luminance compo- that increased wetness lowers, proportionally, the albedo of nent of the sampled spectra which corresponds to the the samples across the whole spectrum (Figure 3a). Field pixel value of the panchromatic images: measurements of the spectral reflectance of the wet and dry 780 sediments of the foreshore zone along the deltaic coast of YRyd  ()()  (4) the River Pinios produced similar results (Figure 4a). 380 Two spectral indices were found to be highly efficient in differentiating between wet and dry spectra: dry sam- The laboratory measurements were carried on samples have high Brightness Index (average reflectance ples from 11 beaches all around Greece. The samples magnitude) and Slope Index; increased water content in were initially dry and gradually the moisture content was the samples results in proportionally lower Brightness increased while the reflectance was monitored. Pinios Index and Slope Index (3b, 4b). Moreover the presence of beach in situ reflectance spectra were also collected at water shifts significantly the dominant colour (3c, 4c). station 66, located 5 km north of river’s outfall. In addi- Therefore, depending on the water content, two separate tion, beach sediment samples were analysed granulomet- classes of signatures exist in the three dimensional space rically by dry-sieving (for material coarser than 0.0625 of BI, SI and HI and can be used to classify beach sand mm) and characterised according to Folk’s (1980) no- samples as dry or wet (Figure 4d). Evaluation of a similar menclature [4]. set of classification indices proposed by Ouillion et al. [6]

but for dry samples only, did not exhibit as high sensitivi-

ty for different moisture contents. 3. RESULTS AND DISCUSSION In Table 1 the granulometric results for the location The measured reflectance spectra showed that differ- where in-situ measurements were performed are shown. ent kinds of sand have dissimilar optical properties and The mean grain size is similar significantly throughout the spectral signatures, as expected. Whitish sands have high sampling region and therefore no variation in the albedo reflectance (~70% in the visible part of the spectrum), is expected due to this factor. while dark grey sands low (~20%), with an overall as-

FIGURE 2 - Reflectance spectra of different sand types as measured in the laboratory. To avoid congestion only four representative spectra are displayed.

2811 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

DRY AND WET SAND REFLECTANCES 0.14 3.1 VERY WET 0.4 3.05 WET 0.13 DRY 0.35 3

0.12 0.3 2.95

2.9 0.25 0.11

2.85 0.2

0.1 Index Hue Reflectance Index Slope 2.8 0.15 0.09 2.75 0.1 2.7 VERY WET 0.08 VERY WET 0.05 WET WET 2.65 DRY DRY 0 2.6 500 600 700 800 0.1 0.15 0.2 0.25 0.3 0.35 578 579 580 581 582 wavelength (nm) Brightness Index λd

FIGURE 3 - (a) Sand reflectance spectra in the visible for variable moisture content. (b) Variation of slope index and brightness index with moisture content. (c) Variation of hue index with water content – here plotted as a function of the dominant colour. The sample is from the Glyfada beach (Near Athens).

0.08 1.675

0.25 1.67 0.075

1.665 0.2 0.07 1.66

0.065 0.15 1.655

Hue Index 1.65

Reflectance Index Slope 0.06

0.1 1.645 0.055

1.64 0.05 DRY 0.05 DRY 1.635 DRY WET WET WET

0 0.045 1.63 400 500 600 700 800 0.1 0.15 0.2 0.25 572.2 572.3 572.4 572.5 Wavelength(nm) Brightness Index λd

1.675

1.67

1.665

1.66

1.655

HI DRY 1.65 WET 1.645

1.64

1.635

1.63 0.08 0.075 0.07 0.22 0.2 0.065 0.18 0.06 0.16 0.055 0.14 0.05 0.12 0.045 0.1 SI BI FIGURE 4 - (a-c) Similar to Figure 3 but for field measured spectral reflectance under daylight conditions (Delta of Pinios). (d) Wet and dry properties in 3D index space.

2812 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

TABLE 1 - Granulometric results for the beach sediments at Station be the association and comparison of the above indices 8 (Pinios river delta). with indices estimated from satellite images acquired Mean grain size Characterization concurrently with the in-situ measurements. Sample (Mz) (Folk Class) 66S (Swash zone) 1.45 slightly gravelly sand 66B (Beach face) 1.55 slightly gravelly sand ACKNOWLEDGEMENTS

This work is supported by the research program THALES - DAPHNE (MIS: 375908) that is funded by the 0.6 data Operational Programme “Education and lifelong learning, linear fit 2007-2013” of the Ministry of Education and Religious Affairs, Culture and Sports

0.5 The authors have declared no conflict of interest.

0.4 REFERENCES

[1] Vörösmarty C.V., Meybeck M., Fekete B., Sharma K., Green P. and Syvitski J. (2003). Anthropogenic sediment retention: major

Brightness Index Brightness global impact from registered river impoundments, Global and 0.3 Planetary Change, (39), 169–190. [2] Soukissian, Τ., Hatzinaki, Μ., Korres, G., Papadopoulos, Α., Kal- los, G. and Anadranistakis, Ε. (2007). Wind and Wave Atlas of the Hellenic Seas, Hellenic Centre for Marine Research Publ., 0.2 300 pp., ISBN-987 960 86651 9-4. [3] Mathieu R., Pouget M., Cervelle B. and Escadafal R. (1998). Re- lationships between Satellite-Based Radiometric Indices Simulat- ed Using Laboratory Reflectance Data and Typic Soil Color of an Arid Environment, Remote Sens. Environ., (66), 17-28. 0.1 100 150 200 250 300 350 400 450 [4] Ray S.S., J.P. Singh, G. Das and S. Panigrahy, (2004). Use of Y (luminance) High Resolution Remote Sensing Data for Generating Site- FIGURE 5 - Brightness index of the eleven sand samples plotted specific Soil Mangement Plan, in ISPRS Archives Vol XXXV against their panchromatic luminance. (B7), 132-137. [5] Folk, R.L. (1974). The petrology of sedimentary rocks: Austin, An interesting correlation is that arising between CIE Tex., Hemphill Publishing Co. panchromatic luminance (Y) and brightness index –BI [6] Ouillion S., Lucs Y. and Gaggelli J. (2002) Hyperspectral detec- (Figure 5). As is indicated, higher spatial resolution pan- tion of sand, Proceedings, 7th International Conference on Re- mote Sensing for Marine and Coastal Environments, Miami Flor- chromatic data can be used to support lower resolution ida sand brightness index data and thus increase the accuracy [7] Kaya S., Seker D., Yasa N., Kabdasli S. and Tanik A. (2007). in the identification of the shoreline position. Utility of remote sensing technology for determining characteristics of sandy beaches, Clean 2007, 35(6), 654 – 659

4. CONCLUSIONS Received: March 19, 2014 This paper presents some of the preliminary results of Accepted: April 02, 2014 an investigation on the reflectance properties and spectral signatures of various samples of dry and wet beach sand. Future work may include monitoring of the in-situ sedi- CORRESPONDING AUTHOR ment moisture [7] as well as an identification of the min- eralogical composition of each sample. Additional indices Panos Drakopoulos comprising the water- sensitive near infrared band will be Department of Optics and Optometry investigated. Nevertheless, initial laboratory and field Technological Educational Institute of Athens results seem encouraging enough that a database with Ag Spyridonos distinct spectral signatures for various sand types and 12210 Athens conditions of sand can be built. Having established typical GREECE ranges of spectral indices for dry -wet state of a particular beach from in-situ measurements, remotely sensed images Phone: +30 2105385747 can be potentially used for the automated identification E-mail: [email protected] and extraction of the shoreline position by engaging neural network models. The next major step to undertake will FEB/ Vol 23/ No 11a/ 2014 – pages 2809 – 2813

2813 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

EUROPEAN LAND COVER MAPPING SYSTEMS FOR THE NATURAL AND URBAN ENVIRONMENT

Athanasios Moysiadis* and Konstantinos Perakis

Department of Planning and Regional Development, School of Engineering, University of Thessaly, 38334 Volos, Greece

Presented at the 4th International Conference on Environmental Management, Engineering, Planning and Economics (CEMEPE), June 24 to 28, 2013, Mykonos, Greece

ABSTRACT database in a way to protect the environment. The first map of its kind was created to depict the entire planet with its Since 1985, European land cover mapping projects complete land cover using 1km spatial resolution satellite have been established to provide maps, a detailed and up-to- images obtained through the use of the VEGETATION date portrayal of land cover from small to large scales. These instrument, on-board the SPOT 4 satellite [1]. Five years programs offer technological and scientific support both later, the aim of the GlobCover, an ESA initiative, was to globally and regionally, from major ecological systems to develop a service capable of delivering global composites artificial features. The European Commission’s Global and land cover maps using input observations from the Land Cover 2000 and the GlobCover (2004–2009), an ESA 300m MERIS sensor on-board the Envisat satellite mis- initiative, provide a global coverage of land cover maps, sion for the periods between 2004 and 2009 [2]. The whereas the CORINE programme, initiated 1985 till 2006 CORINE programme, initiated by the European Union in focuses on land cover maps at a scale of 1:100000. Initial 1985, was a prototype project focused on many different operations of the Global Monitoring for Environment and environmental issues, such as an inventory of land cover in Security (GMES) land monitoring service, the “GIO land”, 44 classes presented as a cartographic product, at a scale of include a land cover change product between 2006 and 1:100000, for the years 1990, 2000 and 2006 [3]. The 2012 and a land cover product for 2012. All these systems objective of the Global Monitoring for Environment and use original geospatial data such as ground truth data, aerial Security (GMES) land monitoring service is to provide land and satellite imagery and a key feature of these maps, is cover information to users in the field of environmental and that they are available at no cost to the end users, through other terrestrial applications. Initial operations of the ser- specially designed geo-databases, to be easily implemented vice (GIO land) include a land cover change product be- in various earth-related applications. tween 2006 and 2012, a land cover product for 2012, pan- European high-resolution layers on land cover characteristics, artificial surfaces, forest areas, agricultural areas

KEYWORDS: Mapping systems; remote sensing; GIS; geo- (grasslands), wetlands and small water bodies [4]. database, natural environment.

2. EUROPEAN LAND COVER MAPPING SYSTEM 1. INTRODUCTION This chapter presents the four most known European It is nearly the last 30 years that European land cover land cover mapping systems (Global Land Cover 2000, mapping projects have been established to provide maps, GlobCover, CORINE, GMES - GIO land) which provide an accurate and detailed and up-to-date portrayal of land a global and a regional cover of the natural and urban envi- cover in small and large scales. The aim of these programs ronment, outlining their importance, their key points and is to offer technological and scientific support both globally complementation and their evolution until today. and regionally, from major ecological systems such as forests, grasslands, and cultivated areas to artificial features Global Land Cover 2000 such as urban coverage, road networks and other man- The objective of the Global Land Cover 2000 was to made structures. provide for the year 2000 a harmonized land cover data- European Commission’s Global Land Cover 2000 base over the whole globe. It was initiated by the Joint was initiated to provide a global harmonized land cover Research Centre (JRC) of the European Commission, in the vision to propose ways to protect the environment * Corresponding author through technological and scientific support [1].

2814 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

The year 2000 was considered a reference year for the This map illustrates the distribution of surface mate- assessment of the environment, as the Global Land Cover rials or “land cover” over the entire globe, showing major (GLC) 2000 database was designed to give a detailed and ecological systems and their boundaries, such as forests, up-to-date global portrayal of the land cover. Global Land grasslands, and cultivated areas [5]. The released datasets Cover 2000 Map (Figure 1), the output of this database, for all regional windows of the world, as well as the glob- was the first one to depict the entire planet with its com- al land cover classification are now available to download plete land cover using 1km resolution satellite images ob- in various formats, through a geo-server, both at full reso- tained through the use of an instrument known as VEGE- lution [6]. TATION, which was on-board the SPOT 4 satellite [1].

FIGURE 1 - The digital Global Land Cover 2000 Map and the respective legend (TerraViva GeoServer).

FIGURE 2 - The Globcover product, the first 300m global land cover map for the period December 2004-June 2006 (MEDIAS-France, 2008).

2815 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

2.2 GlobCover  The regional Globcover Land Cover maps for the GlobCover is an ESA initiative, began in 2005. The period (2005-2006), produced by the same automatic aim of the project was to develop a service capable of and regionally-tuned classification of a MERIS FR delivering global composites and land cover maps, using time series. an automated processing chain, using input observations from the 300m MERIS sensor on board the ENVISAT 2.3 CORINE Land Cover satellite mission. ESA makes available the land cover In 1985, European Union initiated the CORINE Land maps, which cover the periods (December 2004 - June Cover (CLC) programme, a prototype project focusing on 2006) (Figure 2) and (January - December 2009) [2, 7]. different environmental issues, such as an inventory of The methodological chain involves two main steps: land cover in 44 classes presented as a cartographic prod- the “pre-processing” leading to global mosaics of land uct, at a scale of 1:100000, for the years 1990, 2000 and surface reflectance at 300m resolution in 13 spectral 2006 for most areas of Europe [3] (Figure 3). The Euro- bands and the “classification” achieving the final land pean Environment Agency (EEA) has contracted the cover map at 300m resolution [8]. CORINE databases and several of its programmes. The main characteristics of CORINE include the use of the The GlobCover Portal [2] provides free access to the MSS sensor of the Landsat satellites mission, a working products, of the GlobCover project, for any non- scale of 1:10000 and a photo interpretation of false-colour commercial use. These are the following [8]: composites of satellite imagery. For the effectiveness of  The GlobCover (2009) land cover product, the second this prototype project it was necessary [9]: 300-m global land cover map produced from an auto-  to clearly characterise the unit area for land cover mated classification of MERIS FR time series. mapping,  The Globcover Land Cover map (2005-2006), pro-  to formulate a Community land cover nomenclature duced by an automatic and regionally-tuned classifica- hierarchically in three levels based on the unit area tion of a MERIS FR time series. Its 22 land cover definition, classes are defined with the UN Land Cover Classifi- cation System (LCCS).  to refine some of the methods tried out in the feasibility study.

FIGURE 3 - The CORINE Land Cover Map.

2816 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

Another objective of the CORINE programme was to providing very high resolution information on specific bring together all the attempts which have been made over areas of interest and access to a reference data building on the years to obtain the land cover changes on the environ- INSPIRE2 architecture and useful for several GMES ment at various levels, from international to national and services. regional scale. In this scope, CORINE has provided devis- The “Pan-European component” is coordinated by the ing procedures for collating, standardising and exchanging European Environment Agency (EEA) and will produce a data on the environment in the EC Member States. It also European mosaic of high resolution satellite imagery, an has created a geographical information system to provide update of the CORINE Land Cover dataset (CLC) to the the information on the environment which is essential when reference year 2012, and five high resolution layers preparing and implementing Community policies. (HRL) describing characteristics of the main land cover The CORINE land cover maps for the years 1990, types. 2000 and 2006 as well as the land cover changes between The “global component” is coordinated by the EC’s 1990, 2000 and 2006 are available at no cost via the Data Joint Research Centre (JRC). It will produce data across a & maps section of the European Environment Agency wide range of biophysical variables at a global scale, website [10]. which will describe the state of vegetation, the energy budget and the water cycle. Products are generated at a 2.4 GMES land monitoring service - GIO land 1km resolution within 3 days after the end of each 10-day The objective of the GMES (now Copernicus) land period of acquisitions. monitoring service is to provide land cover information to The objective of the “local component” will be to users in the field of environmental and other terrestrial provide more detailed information complementary to the applications. GMES initial operations (2011-2013) of the pan-European component on specific areas of interest land monitoring service or “GIO land” focus on the pri- (e.g. urban areas, biodiversity). It will primarily be based ority for multi-purpose information common to a large on very high resolution images collected between 2011/13 community of users (i.e. land cover/land cover change at in combination with other available datasets (high resolu- various scales and periodicity; biophysical variables for tion and medium resolution images) over the pan- dynamic land monitoring; and improved access to refer- European area. A first local component will focus on ence data) [4]. biodiversity and ecosystem policy. A second local com- In GMES, four components have been identified [11]: ponent will build on the initial GMES local component, a pan-European land cover, land cover change and land such as the European Urban Atlas, to capture urban land cover characteristics, a “global” component producing cover and land use, including low density urban fabric, biophysical variables at global scale, a “local” component and a resolution that is 100 times higher than CORINE

FIGURE 4 - The Urban Atlas product for Brussels, (Mapping Guide for a European Urban Atlas, European Union, 2011).

2817 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

land cover (Figure 4). The first version of the Urban Atlas [7] Defourny P., Vancutsem C., Pekel J.F., Bicheron P., Brock- was made on the basis of 2006 satellite imagery [12]. The mann C., Niño F., Schouten L. and Leroy M. (2006). To- wards a 300 m global land cover product – the Globcover ini- results of the first Urban Atlas are available via the Data tiative. In: Braun, M. (ed.) Second Workshop of the EARSeL & maps section of the European Environment Agency Special Interest Group on Land Use and Land Cover: Appli- website [13]. cation & Development. Universtätsclub Bonn, Bonn (Germa- ny), September 2006. All of the GMES services need access to in-situ data [8] UCLouvain and ESA (2011) GLOBCOVER 2009, Products in order to ensure an efficient and effective use of the Description and Validation Report. GMES space data. The EEA has been tasked with the coordination of the GMES in-situ component through the [9] CORINE land cover report, Methodology & Nomenclature, European Environment Agency/EEA. GISC project. In support of the GMES Bureau, the project explores how a future management (governance, architec- [10] Datasets - European Environment Agen- cy/EEA,http://www.eea.europa.eu/data-and- ture) of the in-situ component during an operational phase maps/data#c11=&c17=CLC1990&c5=all&c0=5&b_start=0 might best be achieved. (accessed March 31, 2013).

[11] GIO Land EEA Briefing, 2012.

3. CONCLUSIONS [12] Urban Atlas - European Environment Agency/EEA, http://www.eea.europa.eu/data-and-maps/figures/urban-atlas (accessed April 5, 2013). Land cover mapping systems have been implemented since 1990 and until today, to provide an accurate and [13] Urban Atlas, GIS data - European Environment Agen- cy/EEA, http://www.eea.europa.eu/data-and- detailed portrayal of the natural and urban environment maps/data/urban-atlas (accessed April 5, 2013). from a global to a regional level. This work outlines the main features of those mapping systems that meet the [14] European Union (2011). Mapping Guide for a European Ur- ban Atlas. above need; the Global Land Cover 2000, the GlobCover, the CORINE programme, and the initial operations of the GIO land, part of the Global Monitoring for Environment and Security (GMES) land monitoring service. Even though each product of these systems has each own specifications in terms of accuracy and scale, it is however common, that the final product comes from geospatially processed data (ground truth data, aerial and satellite imagery). Most importantly, a key feature of these maps is that they are available at no cost to the end users for non-commercial reasons, through specially designed geo-databases, to be easily implemented in earth-related applications.

The authors have declared no conflict of interest.

REFERENCES

[1] Global Land Cover 2000, European Commission-Joint Re- Received: March 18, 2014 search Centre, Accepted: April 02, 2014 http://bioval.jrc.ec.europa.eu/products/glc2000/glc2000.php (accessed March 26, 2013).

[2] GlobCover - European Space Agency/ESA, http://due.esrin.esa.int/globcover (accessed March 26, 2013). CORRESPONDING AUTHOR [3] CORINE Land Cover - European Environment Agency/EEA, Athanasios Moysiadis http://www.eea.europa.eu/publications/COR0-landcover (accessed March 26, 2013). Department of Planning and Regional Development School of Engineering [4] GIO Land - European Environment Agency/EEA, http://www.eea.europa.eu/about-us/what/seis-initiatives/gio- University of Thessaly land (accessed March 27, 2013). 38334 Volos GREECE [5] Fritz S., E. Bartholomé A. Belward A. Hartley H-J Stibig H. Eva, (2003) Harmonization, mosaicking, and production of the Global Land Cover 2000 database, Ispra, Italy, Joint Re- Phone: +30 697 398 6456 search Center. Fax: +30 24210 74371 [6] Global Land Cover 2000 (IES) - GeoServer E-mail: [email protected] http://geoserver.isciences.com:8080/geonetwork/srv/en/meta data.show?id=55 (accessed March 30, 2013). FEB/ Vol 23/ No 11a/ 2014 – pages 2814 – 2818

2818 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

ASSESSING THE WALKING CONDITIONS IN PEDESTRIAN NETWORKS: THE CASE OF THE CITY CENTER IN LARISSA, GREECE

Anastasios Tsakalidis1, Alexandros Sdoukopoulos1, and Nikolaos Gavanas2,*

1Transport Engineering Laboratory, Faculty of Engineering, Department of Civil Engineering, Aristotle University of Thessaloniki, 54124 Thessaloniki, Greece 2Department of Infrastructure Engineering, School of Technological Applications, Technological Educational Institution of Larissa, 41110 Larissa, Greece

Presented at the 4th International Conference on Environmental Management, Engineering, Planning and Economics (CEMEPE), June 24 to 28, 2013, Mykonos, Greece

ABSTRACT varies from 21% to 25% in the European countries with tradition in integrated transport planning, such as Austria, Reconsidering the conventional approach of transport Finland, France, Germany, the Netherlands, Norway, Swe- planning, transport stake-holders have recently begun to den and the United Kingdom [3]. shift their focus from the private car towards sustainable Walking is the primary mode of human transportation transport systems, among which walking is included. Plan- ning and design of walking infrastructure is a complex pro- and the way that each trip begins and ends [4, 5]. As op- cess as pedestrian movements are affected by a variety of posed to motorized modes, it does not produce any kind pollution or require fuel consumption, while certified skills, parameters. The present research aims at the investigation of the pedestrians’ perception regarding the parameters e.g. driving license, are not necessary. Moreover, walking which are considered significant for an effective pedestrian is clearly the most economical way to travel, since there are no running and parking costs, taxation or any other form network. Towards this purpose, a comprehensive analysis of the walking conditions in a Greek medium-sized city was of payment. It is a common physical and recreational activity with documented benefits for personal health and social conducted through an extensive questionnaire survey (including a sample of over 1000 valid questionnaires) on interaction, enhancing the socialization of individuals and selected sidewalks and walkways in the city center of helping to restore the relationship between the citizen and the public space through the stimuli received from the Larissa, Greece. adjacent sidewalk environment. Furthermore, the pedestrian infrastructure occupies only a small percentage of

KEYWORDS: Questionnaire survey, Larissa city, pedestrian the urban space with minimum construction and mainte- infrastructure, walking conditions nance costs in comparison to motorized infrastructure. Thus, walking can be considered as the most sustainable mode of

urban transportation in terms of economic, social and 1. INTRODUCTION environmental impacts [4, 6-8]. Despite the significant advantages of walking, in most The conventional approach of transport planning fo- countries only a small share of trips is conducted on foot. cuses on the provision of additional road infrastructure in Depending on the case, the main reasons for the degrada- order to deal with the increasing mobility demand. Never- tion of walking in contemporary cities are the lack of theless, this approach proved to be inefficient, since high continuity and accessibility for all to the pedestrian net- levels of congestion continue to be observed while induced work and the low operating characteristics such as inade- traffic enhances the degradation of the built and natural quate width and manufacturing defects. In addition, the environment and the quality of life [1]. On the other hand, existence of obstacles within the pedestrian corridors due to the contemporary planning approach shifts the focus on the elements of urban furniture hampers their movement [9]. promotion of multimodality, in the context of which The motorized traffic features, such as the speed of the walking plays a key role [2]. According to a recent study, vehicle in the lane next to the sidewalk, the traffic noise the share of walking in the distribution of trips per mode and pollution, relate to the feeling of comfort and security that pedestrians feel while moving. The lack of trees that * Corresponding author provide shade and protection from the weather and also

2819 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

adorn the urban landscape, combined with indifferent ar- In this context, the current research focuses on the identi- chitectural building facades, reduce the interest and desire fication of the pedestrian perception on these features and of people for walking. Furthermore, the absence of pedes- the analysis of the walking conditions in a medium-sized trian crossings at regular intervals increases walking dis- city of Greece, through an extensive questionnaire survey tances and often causes confusion about choosing the on representative sidewalks and walkways in the city appropriate route. Inadequate lighting, resulting in the center of Larissa. existence of dark spots, significantly reduces the feeling of pedestrian safety and is one of the major causes that prevent walking during night hours [9]. 2. METHODOLOGICAL APPROACH On the other hand, the limited time available for travel prevents people from moving on foot, as the result of In order to investigate the parameters that the pedes- the well established perception that motorized transport is trians consider significant for the effective operation of less time consuming. Another predominating perception walking networks, a methodological approach was im- in some developing countries is that pedestrians hold a plemented including three stages: a) Identification of the lower social position than car drivers while car occupancy main pedestrian segments, b) Collection of data and c) is linked to the concept of economic welfare. These phe- Processing of data. nomena coupled with the large scale development of road The identification of the segments that are representa- infrastructure and land use zoning observed in many cit- tive of the main sidewalks and walkways of the Larissa ies, such as in North America, contribute to the rapid city centre was conducted using the road map of Larissa increase in distances traveled, favoring private car use. and a series of on-site observations. Thus, the study area According to the above, it is common for stakeholders covers the main pedestrian infrastructure of the city centre to focus on the evaluation and enhancement of motorized (Figure 1). An extensive questionnaire survey was im- urban networks rather than pedestrian ones. Thus, there is a plemented during the second stage for the collection of relatively small portion of the international research expe- primary data. The questionnaire comprised four parts: a) rience directed towards the examination of the factors that “Socio-demographic characteristics” so as to provide affect the quality of the walking environment (walkabil- anonymous information on the respondent, i.e. sex, work- ity), in comparison to the corresponding research referring ing condition, age, driver’s license, car occupancy, b) to motorized transportation [10]. “Trip characteristics” in order to describe the daily travel Considering the numerous factors that affect the habits of the respondent, i.e. frequency of use of public walking conditions, the assessment of the level of service and active transport modes in daily transport, average of sidewalks and walkways is a complex process. Since duration of walking at a daily basis, objectives of walking the pedestrian experience entails more than just a com- trips and main reasons for walking, c) “Pedestrian condi- muting function, it is significant for transport planners to tions” by using a stated preference survey for ranking the be able to identify the features that define walkability [7]. most crucial features that affect the walking conditions

FIGURE 1 - Presentation of the study area (Source: [13], own elaboration)

2820 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

and the attractiveness of the examined pedestrian infra- the data processing stage, the valid questionnaires were structure based on the advantages of walking, the design indexed, forming a common database. The statistical analy- principles of pedestrian infrastructure and the dominant sis of the sample was followed by the management and perceptions regarding active and motorized transportation analysis of the results. and d) “Suggested improvements”, where the respondent is asked to propose up to six from a predefined set of improvements, which are based on the aforementioned 3. MAIN RESULTS OF THE QUESTIONNAIRE design principles, i.e. increase of pavement width, better SURVEY ON THE PEDESTRIAN NETWORK maintenance, removal of obstacles, restoration of continu- OF THE LARISSA CITY CENTRE ity, proximity to public bus stops, easier access for vulnerable users, effective separation from roadways, control This section refers to the discussion of the results of the of illegal parking on pavements and crossings, better questionnaire survey, which are organized into four parts. cleaning of pedestrian ways, more trees and vegetation and better lighting infrastructure. 3.1 Socio-demographic characteristics The survey was conducted on site during two periods The population of the survey comprises 50.25% women (in autumn 2011 and spring 2012) according to the Trav- and 49.75% men, a ratio which is similar to the correspond- eler Intercept Survey method. The specific method was ing national ratio [14]. Approximately 60.00% of the re- adopted aiming to capture the traveler’s response while spondents are up to the age of 30 years old while the making a real trip. In order to overcome the drawback of 33.00% are adults between the ages of 31 and 60. More this method, which refers to biased results due to the fact than half of the survey’s sample (52.70%) work either in a that people do not like to be interrupted during their trip full-time or part-time occupation. In spite of the fact that [11], a relatively large sample of questionnaires was col- the 65.58% of the respondents have a driving license, the lected, i.e. a sample of 1600 questionnaires out of which percentage of car ownership does not exceed the 46.02%. 1017 were valid. The format of the questionnaire was based on the corresponding questionnaire used for the case study 3.2 Trip characteristics of Thessaloniki on the development of a model for the A significant share (65.78%) of the survey respond- estimation of pedestrian level of service [12]. The ques- ents state that they use walking as a means of daily tionnaire survey was conducted by the students of the De- transport while almost half of them (48.97%) use the public partment of Infrastructure Engineering of the Technologi- bus daily or weekly. On the opposite, 34.51% of the partic- cal Educational Institution of Larissa as part of a teamwork ipants in the survey never use cycling for their trips as a project in the context of the course: “Traffic Engineer- result of the inadequate bicycle network and the lack of ing/Railways” in the academic period 2011-2012. During “cycling culture”. In Figure 2 there is a presentation of the

FIGURE 2 - Frequency of use per transport mode for private car owners and non-owners

2821 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

100 100%

90 82.44%

80 82.24% 70.09% 76.65% 76.46% 70 65.85% 76.21% 64.11% Combined trip 60 60.77% Recreation 55.84% Shopping 50 Commuting Physical exercise 40 31.78% 30.73% 30 30.24% 26.28% 20

17.77% 10

0 0 < 30 min 30 min -60 min 1 h - 2 h > 2 h FIGURE 3 - Cumulative frequency chart of daily walking trip duration in relation to trip purpose

frequency of use of each transport mode in relation to car cumulative frequency chart concerning the total walking ownership. According to the Figure, a little more than one duration per day according to the pedestrian’s trip pur- quarter of private car owners use the bus and just above pose. More than a quarter of daily trips for commuting 21% the bicycle for daily or frequent trips, while a little and approximately one third of daily trips for recreation, less than 85% choose walking, either exclusively or com- shopping or combined transport are concluded in less than bined with other modes. In the case of the respondents 30 min. In the same time period, less than 18% of daily that do not own a car, walking represents more than 95% walking trips for physical exercise are accomplished. of their choice for daily or frequent trips. Furthermore, it takes from one to two h to complete more According to the questionnaire survey, the 40.71% of than 82% of the total daily pedestrian trips for combined pedestrians conduct commuting trips. Trips for shopping, transport or recreation and over 76% of the daily trips recreation and physical exercise present similar shares (in with the purpose of shopping, commuting or exercise. the range of 15.00 – 19.00%). Walking with the purpose There are two main reasons for the respondents’ of accessing a public transport stop or a private car corre- choice of walking as a means of transport: a) Physical sponds to the 8.16%. A significant share of daily walking exercise (31.17%) and b) money saving (27.53%). As it is trips (42.87% of the sample) has a duration that ranges described in Figure 4, money saving comprises a strong from 30 – 60 min. In Figure 3 there is a presentation of a motivation for younger and older pedestrians. However,

FIGURE 4 - Main reasons for selecting walking as a transport mode according to age class

2822 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

the main reason for walking in the case of people over 61 the respondents believe that the presence of vegetation is years old is physical exercise. It should be also highlight- the least important feature (Figure 6). ed that a significant share of respondents in the age class of 31 – 60 years old declare that the main reason for choosing walking is the lack of parking spaces, which is a common fact in the central areas of cities in Greece.

3.3 Pedestrians’ perception on factors affecting walking trips In this part of the survey, the participants were asked to rank in a scale from 1 (most significant) to 3 (less significant) the following factors in terms of their importance in a walking trip (Figure 5): a) Quality of infrastructure, FIGURE 5 - Ranking of main factors that affect walking trips b) Mobility and access to land uses and c) Proximity to processed with Mode function road traffic. The survey also indicated that narrow and poor main- The respondents also ranked the mobility features that tained infrastructure is considered to have the most nega- may discourage them from using the sidewalk or walkway tive effect on the comfort level of a pedestrian trip. On the (Figure 7). It is worth mentioning that approximately 25% contrary, it should be pointed out that more than 40% of of them state that the most significant negative factor in

Absense of trees and 6.40% 9.64% 10.22% 16.92% 16.71% 40.11% plants

Absense of curb 10.13% 13.17% 20.84% 21.83% 24.39% 9.64% ramps

Poor maintaned pedestrian 28.91% 21.23% 19.08% 16.62% 8.85% 5.31% infra structure

Absense of guide 9.44% 18.68% 18.78% 17.79% 21.24% 14.06% for blinds

Low free height due 16.13% 19.77% 19.18% 13.17% 16.41% 15.34% to obstacles

Narrow pedestrian 29.01% 17.50% 12.19% 13.67% 12.09% 15.54% infra structure

1 (More important) 2 3 4 5 6 (Less important) FIGURE 6 - Infrastructure features with negative impact on pedestrian comfort

Absense of public transport / bicycle 8.85% 6.88% 12.29% 12.19% 13.86% 45.92% infrastructure Formation of a queue in bus stops / 11.11% 14.36% 16.13% 18.10% 29.40% 10.91% intersectios Often maneuvres to avoid vertical 14.16% 21.63% 18.39% 22.12% 15.14% 8.56% movements

Often maneuvres to 21.53% 22.52% 18.00% 18.97% 11.02% 7.97% avoid obstacles

Sense of insecurity 19.27% 16.52% 19.96% 15.25% 17.99% 11.01% due to low lighting

Heavy pedestrian 25.07% 18.09% 15.24% 13.38% 12.58% 15.63% volumes

1 (More important) 2 3 4 5 6 (Less important) FIGURE 7 - Mobility features that discourage the use of sidewalks and walkways

2823 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

their decision making to use a specific infrastructure is the inadequacy and low quality of pedestrian infrastructure, existence of high pedestrian flows and saturation. On the which rarely comprises the core of urban transport plan- other hand, the lack of immediate access to the bus and ning. In this context, the first step towards the promotion bicycle network comprises the least important feature for of walking is the identification and analysis of the precise over 45% of the sample. elements that attract or discourage people from using the pedestrian network. Then, the findings can be exploited 3.4 Suggested improvements effectively in the context of integrated transport planning. In the final part of the survey, the participants were The current methodological approach for assessing walk- asked to briefly describe up to six suggestions which, ability, although focused on the city Larissa, could be according to their opinion, will improve the overall quali- easily adapted to the specific characteristics of other me- ty of service in the city’s pedestrian network. Among a dium-sized Greek cities while it is compatible to the main variety of suggestions, the most common ones involve the principles for contemporary pedestrian networks. repair and regular maintenance of the network and the removal of all (permanent or not) obstacles that obstruct The authors have declared no conflict of interest. pedestrian movements, which accounted for the 55.65% and 47.59% of the total suggestions respectively. Other suggestions refer to the implementation of a strict policy REFERENCES against illegal parking (40.70%), the widening of narrow sidewalks (38.84%), the restoration of infrastructure for [1] Byrd J., Sisiopiku V. (2006) Comparison of Level of Service methodologies for pedestrian sidewalks. Proceedings of the disabled pedestrians (37.17%) and the adequate illumina- 85th Annual Meeting of Transportation Research Board. Pa- tion during the night (33.62%). per No. 06-0898. Washington D.C. [2] Dixon L. B. (1996) Bicycle and pedestrian level-of-service performance measures and standards for congestion manage- 4. DISCUSSION ment systems. Transportation Research Record: Journal of the Transportation Research Board, 1538(1), 1-9. Worldwide, the relevant state-of-the-art suggests as [3] Buehler R., Pucher J. (2012) Walking and Cycling in Western most crucial features of walkability the adequate width of Europe and the United States: Trends, Policies and Lessons. the sidewalk (minimum 1 m), the existence of obstacles, TR News, No. 280, 2012, pp 34 – 42. such as illegally parked vehicles, along the pedestrian [4] LSA Associates Inc. (2003) Kansas City Walkability Plan. City routes, the efficient separation from the roadway, the speed of Kansas City, Missouri. City Planning & Development De- partment. Online Document. Accessed October 10, 2012. of nearby motorized vehicles and the use of pedestrian http://www.kcmo.org/idc/groups/cityplanningdevelopmentdi infrastructure by cyclists [15-18]. The results of the present v/documents/cityplanninganddevelopment/walkability.pdf. research partially coincide with the aforementioned fea- [5] Krambeck H. (2006) The global walkability index. PhD thesis, tures. More specifically, the research indicates the first Massachusetts Institute of Technology, Massachusetts. two features as highly important while the poor mainte- [6] Jaskiewicz, F. (2003) Pedestrian Level of Service Based on nance of the infrastructure, the lighting conditions and the Trip Quality. Urban Street Symposium. Transportation Re- congested pedestrian segments were also highlighted. The search Board. Online Document. Accessed April 2, 2013. research also highlights that walking constitutes a critical http://www.urbanstreet.info/2nd_sym_proceedings/Volume% element of Larissa’s transport system regarding the daily 201/Ec019_g1.pdf. trips to work as well as to other activities such as shop- [7] Saelens B. E., Handy S. L. (2008) Built environment correlates ping and recreation. According to the respondents’ perspec- of walking: A review. Medicine and Science in Sports and Exercise, 40(7 Suppl.), 550–566. tive, the main reason for selecting the specific mode is the low cost followed by the need for physical exercise espe- [8] Walsh R. (2012) NCHRP Synthesis 436: Local Policies and Practices That Support Safe Pedestrian Environments: A cially for older age classes. These major socio-economic Synthesis of Highway Practice. National Cooperative High- benefits which derive from walking should be set as priori- way Research Program (NCHRP) Project 20-05. ties by the local stakeholders in order to cope with the [9] N.Z. Transport Agency. (2009) Pedestrian planning and de- aforementioned prohibitive features. sign guide, N.Z. Transport Agency, Wellington.

[10] Landis B. W., Vattikuti V. R., Ottenberg R. M, McLeod D. S., Gutterplan M. (2001) Modeling the roadside environment: 5. CONCLUSIONS A pedestrian level of service. Transportation Research Rec- ord: Journal of the Transportation Research Board, 1773(1), Walking is progressively gaining significance in sus- 82-88. tainable urban mobility, either as a mode of small and [11] Dowling, R., Reinke, D., Flannery, A., Ryus, P., Vandehey, medium distance door-to-door trips or as a major part of M., Petritsch, T., Landis, B., Rouphail, N., Bonneson, J. (2008) Multimodal Level of Service Analysis for Urban the multimodal transport system. The main obstacles in Streets. Transportation Research Board. National Coopera- this process derive from the society’s perceptions regard- tive Highway Research Program Report 616. Transportation ing the efficiency and prestige of private motorized Research Board. Online Document. Accessed April 1, 2013. transport as opposed to public and active transport and the http://onlinepubs.trb.org/onlinepubs/nchrp/nchrp_rpt_616.pdf

2824 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

[12] Christopoulou P. (2012) Development of a methodology for the assessment of the pedestrian level of service in the framework of multimodal transport. PhD thesis, Faculty of

Civil Engineering, School of Engineering, Aristotle Universi- ty, Thessaloniki. [13] Bing Maps. Accessed March 19, 2013. http://www.bing.com/maps/. [14] Hellenic Statistical Authority (2011) Announcement of inter- im results of the 2011 Population Census. (In Greek). Main Page EL.STAT. 22 July 2011 Press Release. Hellenic Statisti- cal Authority. Online Document. Accessed March 19, 2013. http://www.statistics.gr/portal/page/portal/ESYE/BUCKET/ General/A1602_SAM01_DT_DC_00_2011_01_F_GR.pdf. [15] Gallin N. (2001) Quantifying pedestrian friendliness – Guidelines for assessing pedestrian Level of Service, Austral- ia: Walking the 21st Century Conference, Perth, Western Australia. [16] Mozer D. (1997) Calculating Multi-Mode Levels-of-Service. International Bicycle Fund. [17] Muraleetharan T., Adachi T., Hagiwara T., Kagaya S., Ka- wamura S. (2004) Method to determine overall Level-of- Service of pedestrians on sidewalks and crosswalks based on total utility value. Proceedings of the 83rd Annual Meeting of Transportation Research Board, Paper No. 04-4422, Wash- ington D.C. [18] Transportation Research Board (2000) Highway Capacity Manual 2000. National Research Council, Washington D.C.

Received: November 11, 2013 Revised: May 15, 2014; May 30, 2014 Accepted: June 18, 2014

CORRESPONDING AUTHOR

N. Gavanas Department of Infrastructure Engineering School of Technological Applications Technological Educational Institution of Larissa 41110 Larissa GREECE

Phone: +302310994160 E-mail: [email protected] [email protected]

FEB/ Vol 23/ No 11a/ 2014 – pages 2819 - 2825

2825 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

THE ENVIRONMENTAL IMPACT OF TRAFFIC CALMING MEASURES TOWARDS THE REGENERATION OF AN URBAN AREA: THE CASE OF THE INTERVENTIONS ALONG THE AXIS OF AGIA SOFIA IN THESSALONIKI, GREECE

Nikolaos Gavanas*, Alexandros Sdoukopoulos and Magda Pitsiava-Latinopoulou

Transport Engineering Laboratory, Faculty of Engineering, Department of Civil Engineering, Aristotle University of Thessaloniki, 54124 Thessaloniki, Greece

Presented at the 4th International Conference on Environmental Management, Engineering, Planning and Economics (CEMEPE), June 24 to 28, 2013, Mykonos, Greece

ABSTRACT non-motorized users of the road [1]. Isolated examples of traffic calming interventions can be found in the cities of The minimization of negative externalities and the central Europe since the mid 20th century [2]. Nowadays, achievement of the three main goals of sustainable devel- organized traffic calming measures are often used in the opment, i.e. social, economic and environmental sustaina- context of urban regeneration schemes for the restoration bility, are essential elements for the successful regenera- of the environment, the stimulation of the local socio- tion of the urban environment. The regeneration of the economic activity, and the improvement of safety condi- area surrounding the Agia Sofia axis, which is located in tions [3]. Thus, these measures are compatible with the the historical center of the city of Thessaloniki, was the main sustainability goals of the European policy for urban objective of a recent national architectural competition. mobility [4, 5]. More specifically, in Thessaloniki, the The present paper aims to assess the environmental im- transport infrastructure Master Plan 2020 is based on the pacts from the application of a proposal awarded with the aforementioned sustainability goals aiming specifically at competition’s 2nd prize through a ‘before and after’ study. the alleviation of through-traffic and the promotion of envi- In specific air, emissions were assessed based on the ronmentally-friendly transport in the city centre [6]. In this area’s traffic features using the COPERT 4 model for the framework, the current paper focuses on the assessment periods “before and after” the proposed regeneration scheme. of the change in pollutant emissions from a proposed This application led to the evaluation of the changes in the network of low-speed and pedestrian ways along the road environmental conditions due to the traffic rearrangement axis of Agia Sofia in the historical centre of Thessaloniki. of the study area’s road network on the wider catchment The calculations refer to the workday’s peak-hour aver- area. age. The proposed network is part of the comprehensive proposal that was awarded with the 2nd prize in the Na- tional Architectural Competition for the regeneration of

the Agia Sofia axis [7]. KEYWORDS: Urban regeneration, traffic calming, environmental impacts, Agia Sofia axis

2. DESCRIPTION OF THE STUDY AREA

Thessaloniki accommodates the second largest population among Greek cities, while it comprises the most 1. INTRODUCTION significant urban centre of Northern Greece and a major commercial gateway for the Balkans and the wider East- Traffic calming measures are moderate interventions ern Mediterranean region. The concentration of popula- which combine mainly physical measures with the aim to tion and activity in the wider urban area imposes signifi- reduce the negative effects of vehicle use and alter the cant pressures on the city’s transport system in combina- driver’s behavior in order to improve the conditions for tion with the following factors: a) Road transport (private car and bus) is the dominant mode while approximately * Corresponding author one out of two residents owns a car [8]; b) There is insuf-

2826 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

ficient integration between transport planning and urban the Agia Sofia axis and expands to a surface of approxi- development strategies. The above incompatibility is criti- mately 0.285 km2. The current traffic situation of the axis cal due to the landscape features of the city, as the coast- as well as the proposed traffic calming measures expected line at the south and the mountainous terrain at the north to be implemented after the metro system’s operation are forces urban sprawl towards the eastern and western represented in Fig. 1 (1a and 1b, respectively). gions. In specific, the proposed interventions refer to the According to the latest General Transport Study Agia Sofia axis from Filippou St. in the north to Nikis Av. (completed in 2000) and the updated estimations, the city in the south. They were designed in a way that local traf- centre of Thessaloniki is the origin or/and destination of fic would be serviced, while motorized traffic through the 6 the 25% of approximately 2.3 10 daily trips. The main centre from north to south and opposite would be discour- arteries that are linked to the city centre accommodate a aged. This would divert this kind of low volume trips to significant share of through-traffic. The congestion in the the neighbouring road segments, resulting thus to a rela- city centre and along main axes results to great pressures tively small increase of their traffic flows. Furthermore, it on the area’s air quality, and thus, on the ability to follow should be highlighted that the role of the main arteries the corresponding EU directives [9]. that serve through-traffic from/to east to/from west The study area, located in the core of the city centre, (Egnatia St., Tsimiski St., Mitropoleos St. and Nikis Av.) is defined as the area affected by traffic interventions along is maintained according to the scope of the specific study.

Source: own elaboration

FIGURE 1 - Current traffic organization (a) and proposed traffic calming interventions (b).

2827 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

3. METHODOLOGICAL APPROACH ventions along the road segments, which will be operating under the new network classification, was based on the The methodology is a ‘before and after study’ con- application of the growth factor method. In order to apply sisting of the following steps: a) Analysis of the study the method, apart from the segments illustrated in Fig. 1, area’s traffic features in the two time periods, b) Estima- the segments along Iasonidou and Halkeon St. were in- tion of pollutant emissions for the current situation and cluded in the analysis. The resulted traffic volumes are future scenario, and c) Assessment of the changes in the presented in Table 1. area’s environmental features. Furthermore, in Table 2, there is a presentation of the modal distribution along the examined segments in the 3.1. “Before and after” traffic features in the study area current situation. The estimation of future traffic flows is The current traffic features in the study area were based on the hypothesis that the modal distribution for the measured during a workday’s peak-hour by on-site observ- through-traffic from north to south remains unchanged ers. The observation showed that the actual peak-hour of since the new metro system will serve trips from east to traffic in the city centre is from 11:00 a.m. to 12:00 a.m., west of the city centre. Moreover, the measurements of and does not correspond to the typical morning and after- traffic speed with specialized equipment along representa- noon peak-periods. The estimation of traffic volumes for tive segments during the peak period showed an average the period after the implementation of the proposed inter- value of 22 km/h with small variations (in the range of 5-

TABLE 1 - Traffic volumes before interventions and their change ratio after traffic calming measures.

Road Segment (Flow direction) Agia Sofia Agia Sofia Agia Sofia Agia Sofia Agia Sofia Agia Sofia (from Iasonidou Iasonidou (from Egnatia (from Ermou (from Egnatia (from Tsimiski (from Mi- Koromila to (from Egnatia (from Filip-

to Filippou) to Egnatia) to Ermou) to Mitropoleos) tropoleos to Nikis) to Filippou) pou to Koromila) Egnatia) Traffic volume (veh/h) Before traffic 238 248 179 187 178 153 568 511 calming Change ratio (%) -100.0% -100.0% 0.0% 0.0% -100.0% -73.2% 8.7% 4.6% After traffic calming Road Segment (Flow direction) Halkeon Platonos* P. Ioakim K. Ntil (from K. Ntil (from P.P. Germanou P. Mela (from M. Iosif (from Egnatia (from Filip- (from Ermou Ermou to Tsimiski to (from Tsimiski to P.P. Ger- (from Koro-

to Filippou) pou to to Egnatia) Egnatia) Ermou) Egnatia) manou to mila to Egnatia) Tsimiski) Nikis) Traffic volume (veh/h) Before traffic 1200 143 113 100 372 229 612 380 calming Change ratio (%) 5.5% 0.0% 20.9% 22.6% 22.6% 11.3% 9.4% 29.4% After traffic calming * After interventions, the specific segment of Platonos St. will be converted from one-way to two-way road. The estimated traffic volume of Platonos (from Egnatia to Filippou) is 238 veh/h.

TABLE 2 - Distribution of traffic per vehicle category.

Modal distribution (% veh.) Road Segment (Flow direction) Private cars & taxis Motorcycles Busses Trucks Agia Sofia (from Egnatia to Filippou) 63% 34% 0% 3% Agia Sofia (from Ermou to Egnatia) 64% 31% 2% 3% Agia Sofia (from Egnatia to Ermou) 72% 25% 0% 3% Agia Sofia (from Tsimiski to Mitropoleos) 60% 36% 1% 3% Agia Sofia (from Mitropoleos to Koromila) 69% 27% 1% 3% Agia Sofia (from Koromila to Nikis) 75% 20% 0% 5% Iasonidou (from Egnatia to Filippou) 79% 19% 0% 2% Iasonidou (from Filippou to Egnatia) 82% 17% 0% 1% Halkeon (from Egnatia to Filippou) 73% 25% 0% 2% Platonos (from Filippou to Egnatia) 69% 22% 0% 9% P. Ioakim (from Ermou to Egnatia) 65% 35% 0% 0% K. Ntil (from Ermou to Egnatia) 60% 36% 0% 4% K. Ntil (from Tsimiski to Ermou) 58% 41% 0% 1% P.P. Germanou (from Tsimiski to Egnatia) 76% 24% 0% 0% P. Mela (from P.P. Germanou to Tsimiski) 67% 29% 1% 3% M. Iosif (from Koromila to Nikis) 69% 27% 0% 4%

2828 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

TABLE 3 - Road traffic volumes before and after the operation of the metro along main arteries.

Road Segment (Flow direction) Egnatia (from K. Egnatia (from Tsimiski (from Mitropoleos Nikis (from Ntil to Georgiou Georgiou Kupriou M. Iosif to K. (from K. Ntil to K. Ntil to M. Kupriou Ch.) Ch. to K. Ntil) Ntil) M. Iosif) Iosif) Traffic volume (veh/h) 838 1012 3209 483 1890 Before the operation of metro system Traffic volume (veh/h) 696 840 2663 401 1569 After the operation of metro system

10%) among different segments. The implementation of sis of the wider transport network exceeds the scope of interventions is expected to decrease the mean speed along the specific research. the traffic calmed segments of Agia Sofia to 20 km/h, according to the international specifications [1]. 3.2. Estimation of pollutant emissions Concerning the traffic changes along the main arter- The calculation of emissions is conducted using the ies which cross the study area, gross estimations result to COPERT (COmputer Programme to calculate Emissions the following: the operation of the metro system, which from Road Transport) 4 model which obtains the emission coincides with the implementation of the traffic calming factors as a function of the average cycle speed [11]. The measures is estimated to generate an overall reduction of basis for the model’s calculations is an original and up- 17.5% in traffic flows (Table 3) [10]. Nonetheless, the dated database for Greece. The original database was ad- metro is expected to change the traffic composition on the justed according to the traffic data of Tables 1 and 2, i.e. examined segments due to the shift from car to metro, the traffic volume per vehicle category (veh/h), mean traffic reorganization of the public bus routes etc. It should be speed (km/h), and length of examined segments (km). pointed out herein that a more comprehensive traffic analy-

TABLE 4 - Pollutant emissions and fuel consumption before and after traffic calming interventions.

Road Segment (Flow direction) CH4 (g) PM10 (g) NOX (g) CO2 (g) FC (g) Before traffic calming interventions Agia Sofia (from Egnatia to Filippou) 2.4 1.5 13.0 7777.2 2460.4 Agia Sofia (from Ermou to Egnatia) 2.3 1.6 15.6 8175.9 2586.1 Agia Sofia (from Egnatia to Ermou) 1.5 1.0 8.9 5770.9 1825.6 Agia Sofia (from Tsimiski to Mitropoleos) 0.8 0.5 4.6 2542.3 804.2 Agia Sofia (from Mitropoleos to Koromila) 0.5 0.3 3.2 1891.7 598.4 Agia Sofia (from Koromila to Nikis) 0.4 0.3 2.6 1619.8 512.4 Iasonidou (from Egnatia to Filippou) 15.6 10.4 101.9 68527.6 21677.6 Iasonidou (from Filippou to Egnatia) 13.9 9.4 98.3 63029.0 19938.0 Halkeon (from Egnatia to Filippou) 11.4 7.5 68.3 43918.7 13893.4 Platonos (from Filippou to Egnatia) 1.3 0.9 10.1 5266.9 1666.1 P. Ioakim (from Ermou to Egnatia) 1.1 0.6 4.4 3267.4 1033.8 K. Ntil (from Ermou to Egnatia) 1.0 0.6 5.2 3003.9 950.4 K. Ntil (from Tsimiski to Ermou) 4.3 2.6 17.8 12046.7 3811.7 P.P. Germanou (from Tsimiski to Egnatia) 5.3 3.3 27.3 20466.0 6474.7 P. Mela (from P.P. Germanou to Tsimiski) 9.8 6.7 66.1 36583.8 11572.0 M. Iosif (from Koromila to Nikis) 3.9 2.6 23.9 14616.3 4623.8 After traffic calming interventions Agia Sofia (from Egnatia to Filippou) 0.0 0.0 0.0 0.0 0.0 Agia Sofia (from Ermou to Egnatia) 0.0 0.0 0.0 0.0 0.0 Agia Sofia (from Egnatia to Ermou) 1.5 1.0 9.0 6008.7 1900.8 Agia Sofia (from Tsimiski to Mitropoleos) 0.8 0.5 4.7 2646.7 837.2 Agia Sofia (from Mitropoleos to Koromila) 0.0 0.0 0.0 0.0 0.0 Agia Sofia (from Koromila to Nikis) 0.1 0.1 0.7 452.9 143.3 Iasonidou (from Egnatia to Filippou) 17.0 11.3 110.0 74379.0 23528.7 Iasonidou (from Filippou to Egnatia) 14.6 9.9 102.6 65972.7 20869.3 Halkeon (from Egnatia to Filippou) 12.0 7.9 72.6 46368.1 14668.2 Platonos (from Filippou to Egnatia) 1.3 0.9 10.2 5482.3 1734.2 P. Ioakim (from Ermou to Egnatia) 1.3 0.8 5.3 4095.5 1295.8 K. Ntil (from Ermou to Egnatia) 1.2 0.8 6.3 3845.2 1216.5 K. Ntil (from Tsimiski to Ermou) 5.2 3.2 21.9 15355.5 4858.6 P.P. Germanou (from Tsimiski to Egnatia) 5.9 3.7 30.3 22798.9 7212.8 P. Mela (from P.P. Germanou to Tsimiski) 10.7 7.4 72.7 40180.1 12709.5 M. Iosif (from Koromila to Nikis) 5.0 3.4 31.2 19622.3 6207.4 Platonos (from Egnatia to Filippou) 2.4 1.5 13.2 8093.5 2560.4

2829 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

TABLE 5 - Main pollutant emissions and fuel consumption along the Agia Sofia axis and in the study area before and after traffic calming interventions.

Before traffic calming interventions After traffic calming interventions Area CH4 (g) PM10 (g) NOX (g) CO2 (g) FC (g) CH4 (g) PM10 (g) NOX (g) CO2 (g) FC (g) Agia Sofia 7.8 5.2 47.9 27777.6 8787.1 2.4 1.6 14.4 9108.2 2881.3 axis Study area (excl. main 75.2 50 471.2 298504 94428.7 78.9 52.3 490.7 315301 99742.8 arteries) Study area 207.5 159.2 1906.9 869928 275106 188.7 142.6 1669.5 778586 246230

More specifically, the main adjustment concerns the after’ study concerning the impacts of the proposed traffic composition of the vehicle fleet. The original features de- calming measures on pollutant emissions and fuel con- scribing the fleet composition of Greece are divided by sumption showed that these measures do not result to the vehicle category and type of engine, and they include area’s better environmental features by their own, while specific vehicle categories that are not used in the city their combination with the metro’s operation could coun- centre’s road network (i.e. heavy duty articulated trucks terbalance or even reverse the situation. The aforemen- as well as urban CNG and biodiesel busses). These cate- tioned case also indicates the complex effects of traffic gories were extracted from the original sample and both calming measures on sustainable mobility. In order to cope the original fleet volume and composition were re-cal- with this complexity, it is highlighted the significance of a culated. Moreover, an adjustment was also necessary for holistic transport planning approach based on interdisci- the calibration of the fleet composition concerning the diesel plinary cooperation. fueled taxis, as on-site traffic measurements gave an average of 20% of taxis in the total passenger car volume, while the corresponding share of diesel fueled passenger cars in the original database was approximately 1.5%. In ACKNOWLEDGEMENT order to cope with this incompatibility, the on-site composition of passenger cars was used in the model’s calcu- We would like to thank Chariton Kouridis (EMISIA lations. The results of the model’s implementation per S.A.) for his valuable contribution to our research and, segment “before and after” the measures are presented in more specifically, for his guidance and the provision of Table 4. access to the most recent available data referring to the calculation of emission factors in Greece. 3.3. Assessment of the change in pollutant emissions The comparative study for the assessment of changes The authors have declared no conflict of interest. in pollutant emissions after the implementation of the interventions refers to three cases: i.e. i) exclusively along the Agia Sofia axis, ii) in the study area without considering the main arteries, and iii) including the main arteries REFERENCES that are affected by the metro’s operation The cumulative [1] ITE and FHWA. (1999) Traffic calming. State of the prac- emissions generated during a workday’s peak-hour are tice. Institute of Transportation Engineers, Washington. presented in Table 5. It can be seen that the traffic calming measures will reduce emissions along the Agia Sofia [2] Kjemtrup, K., Herrstedt, L. (1992) Speed management and traffic calming in urban areas in Europe: a historical view. axis up to 70%, whereas an increase of approximately 5% Accident: Analysis and Prevention, 24(1), 57-65. is estimated for the study area (case ii) due to the speed decrease along segments that will take on diverting vol- [3] Department for Transport. (2007) Traffic calming. Depart- ment for Transport - GOV.UK. Local transport note 1/07. umes from the traffic calmed axis. Finally, the combina- March 2007. The Stationery Office (TSO). Online document. tion of the proposed measures with the operation of the Accessed on March 5, 2013. metro system (case iii) will lead to an overall decrease of http://assets.dft.gov.uk/publications/local-transport-notes/ltn- the pollutant emissions and fuel consumption as the result 1-07.pdf. of the traffic reduction along the main arteries. [4] European Commission. (2009) Action Plan on urban mobility (COM(2009) 490 final). Brussels. [5] European Commission. (2007) Green Paper towards a new 4. CONCLUSIVE REMARKS culture for urban mobility (COM(2007) 551 final). Brussels. [6] Ministry of Infrastructure, Transport and Networks. (2010) The proposed interventions aim at the study area’s Master Plan of transport infrastructure for Thessaloniki’s regeneration through traffic calming measures safeguard- wider area (in Greek). YME.gr. Online document. Accessed ing meanwhile the service of local traffic. The ‘before and on April 3, 2013. http://www.yme.gr/getfile.php?id=3155.

2830 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

[7] Ministry of Environment, Energy and Climate Change. (2012). Diavgeia.gov.gr. Decision ΑΔΑ: Β4ΛΗ0-ΚΔ4 (in Greek). Online document. Accessed on March 2, 2013. http://static.diavgeia.gov.gr/doc/%CE%924%CE%9B%CE% 970-%CE%9A%CE%944. [8] Pitsiava, M. (2008) The transport system of an emerging me- tropolis: Development trends and access inequalities. Chap- ter in Thessaloniki on the edge: The city from the perspective of change (eds. G. Kafkalas, L. Labrianidis and N. Papami- hos), Kritiki Publishing, Athens, 489-523. [9] Moussiopoulos, N., Vlachokostas, Ch., Tsilingiridis, G., Douros, I., Hourdakis, E., Naneris, C., Sidiropoulos, C. (2009) Air quality status in greater Thessaloniki area and the emission reductions needed for attaining the EU air quality legislation. Science of The Total Environment, 407(4), 1268– 1285. [10] TREDIT. (2005) Study of economic feasibility of Thessalo- niki’s metro (in Greek). Attiko Metro S.A., Athens [11] Gkatzoflias, D., Kouridis, C., Ntziachristos, L. and Samaras, Z. (2007) COPERT 4: Computer programme to calculate emissions from road transport. User manual (Version 5.0). Emisia SA. Laboratory of Applied Thermodynamics, Aristo- tle University Thessaloniki and European Environment Agency. Online document. Accessed on April 9, 2013. http://www.emisia.com/docs/tech01.pdf.

Received: November 05, 2013 Revised: March 09, 2013 Accepted: April 02, 2014

CORRESPONDING AUTHOR

Nikolaos Gavanas Transport Engineering Laboratory Faculty of Engineering Department of Civil Engineering Aristotle University of Thessaloniki 54124 Thessaloniki GREECE

Phone: +302310994160 E-mail: [email protected] [email protected]

FEB/ Vol 23/ No 11a/ 2014 – pages 2826 - 2831

2831 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

INVESTIGATION AND PREDICTION OF TRAFFIC TRAVEL TIME FOR THE ROAD NETWORK OF THESSALONIKI THROUGH SPATIAL TEMPORAL MODEL

Konstantinos Lakakis*, Kalliopi Kyriakou and Paraskevas Savvaidis

Department of Civil Engineering, Aristotle University of Thessaloniki, 54006, Thessaloniki, Greece

Presented at the 4th International Conference on Environmental Management, Engineering, Planning and Economics (CEMEPE), June 24 to 28, 2013, Mykonos, Greece

ABSTRACT A traffic prediction system would remarkably contribute to this solution, predicting the segments of road Traffic congestion is a vulnerable point for every ur- network that are crowded and proposing the optimum ban area creating insupportable situations for citizens. A route by estimating the travel time. The purpose of current traffic prediction system could ensure better quality trans- study is to create a traffic prediction system using a space- portation. The current study aims at predicting travel time model and investigate the traffic flow. The Kalman times for every single day for one of most central roads of Filter method has been chosen among the many available Thessaloniki using Kalman Filter and historic data which methods. Another aim of the project is to use the minimum are provided from taxi’s GPS. Homogenous time zones necessary equipment and automating the whole process. are created for each day with regard to land uses that The numerous required traffic data for the current predominate at the current time period. Aiming at a traffic study has been provided by taxi-vehicles which use GPS prediction system each phase of study has been carried receivers to navigate in the city. out with ArcGIS and Matlab. The levels of traffic congestion are represented on maps through color gradient where the traffic congestion points are easily spotted. The con- 2. MATERIALS AND METHODS tribution of this study can be described in terms of a smart urban navigation system, but also as a tool which learns The methodology presented in this paper is specifi- every day and supports the decision makers on emergency cally developed and implemented in order to investigate and predict the traffic conditions in the study area. The responses or unexpected traffic events. process consists of the four phases presented below:

2.1 Selection of Data KEYWORDS: Thessaloniki’s urban zone; Traffic prediction per time period; Space-time model The road which is studied is Tsimiski road which is one of the most central road in the road network of Thes- saloniki city (Figure 1, created by authors). The total length is approximately 1755m. Aiming at more accurate results, 1. INTRODUCTION the road has been divided into eleven segments starting and ending at nodes with traffic lights (Figure 2, created The rapid growth of demand for transportation, high by authors). Thus, it was possible to calculate the travel levels of car dependency caused by the urban sprawl have times for each segment excluding traffic lights times. The exceeded the slow increments in transportation infrastruc- lengths of segments are shown at Table 1. ture supply in many areas causing severe traffic congestion. Negative effects of traffic congestion, such as fuel The traffic data has been provided by taxi-vehicles wasted in idling vehicles lead to increasing air pollution which use GPS receivers to navigate in the city. While the and carbon dioxide emissions; time and associated cost. A taxi-vehicles travel along the road network of Thessaloni- more viable approach to cope with the congestion prob- ki (Figure 1) following the general traffic flow they pro- lem is to monitor traffic congestion and traffic manage- vide a large amount of GPS points giving vehicle’s ID, ment systems to avoid the traffic congestion [1]. coordinates, velocity, orientation, time and date. It must be noted that all the taxi vehicles used travel every single day without any restriction for the needs of current study. * Corresponding author The coordinates of GPS points were transformed from

2832 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

WGS84 that GPS uses to Greek Grid (EGSA87) by 2.2. The process of Map-Matching means of the software ArcGIS. Thus, higher accuracy of The data has to be referenced spatially, so a map- the results is achieved. In addition, the points that were matching procession was necessary. This procedure is a assigned to Tsimiski road should be selected from the vast crucial issue especially in urban areas as Thessaloniki data set. Thus, there were selected Tsimiski’s points giv- with narrow roads, tall buildings and other signal degrading a margin of 80 m right and left of the road through ing environments. There are various procedures proposed ArcGIS. Another excluding criterion was the vehicle’s for map matching. Many of them have been used in urban orientation. Since Tsimiski has only one direction, the and suburban areas and the percentage of correct links points should have similar orientation. So a corridor of ranges between 86% and 93% and the 2-D horizontal values was defined for the orientation and points that did accuracy range between 15m and 10.6m [2]. In this study, not follow this criterion were excluded. one map-matching method has been selected which has

TABLE 1 - The 11 segments of Tsimiski road

Segment’s ID 1 2 3 4 5 6 7 8 9 10 11 Length (m) 270 173 98 226 143 126 143 138 102 91 78

FIGURE 1 - The city of Thessaloniki in northern Greece

FIGURE 2 - Study-area: Tsimiski road and its 11 segments, downtown Thessaloniki

2833 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

been accepted by the international community and it is  Select points with same ID at the same period of day based on the idea of using as spatial street network only Calculation of travel time from point p to p+1 the road segments which are measured by the system  itself and not by external sources (i.e., electronic city  Exclude travel times with more than some minutes maps produced by other surveying methods). The mathe- depending on each segment’s length. matical algorithms which have been used were the linear  Calculation of travelled distance and a further calcula- regression model and its prediction zone. The method has tion of time travel in relation to the segment’s length. been tested at different confidence levels (95%, 97.5% and 99%) and in the case of stand-alone GPS observations 2.4. Definition of Time Periods gave map-matching accuracy range between 15m and 21m. In the case of differential GPS observations the At this phase, the working days and weekends were range was 1.5m to 2.0m [3], [4]. The general mathematic separated into 4 homogenous time zones as shown in model of linear regression is the following: Table 2. The criteria were the hours when land uses such as shopping, entertainment and use of offices and services Υ =  + X +  (1) predominate (Tsimiski road is found in the central busi- And its confidence interval or prediction zone for ness area of Thessaloniki). The land uses for each time every road segment is given by the following relationship: zone are represented. The time zones are coloured in a grey scale. Homogenous time zones are coloured with ˆ ˆ same gradation of grey in order to be easily observed. Y0 -tα/2 s ˆ  Y0  Y0 tα/2 s ˆ (2) Y0 Y0 Furthermore, at each homogenous time zone the land uses Normally, the prediction zone is defined by two hy- that are firstly mentioned are those that predominate in perbolic curves while the closest point is defined imple- this time zone. For more accurate results the working menting the equation (2) for x mean. hours should have also been divided into various time zones however not enough points were available because 2.3. Calculation of Travel Times of many exclusion criteria.

At this stage, it was feasible to calculate travel times 2.5. Traffic Investigation per road segment during all the day for every single day. The implementation was performed through Matlab soft- Having the travel times for each homogenous time ware because the numerous data constituted a critical zones it is possibility to investigate the traffic at Tsimiski problem. The process of calculation was performed as Road. The investigation is performed through maps that following: have been created on ArcGIS. The levels of traffic con gestion are represented on maps by color gradient where  Identify TAXI’s ID that appeared among the road the traffic congestion points are easily spotted. segment at least twice in order to be possible to calculate its travel time

TABLE 2 - Definition of Time Periods: Homogenous time zones based on land uses

Monday Tuesday Wednesday Thursday Friday Saturday Sunday Houses Houses Houses Houses Houses Houses Houses Cafes Cafes Restaurants Restaurants 0:00– 7:59 7:59 0:00– Shops Shops Shops Shops Shops Restaurants Restaurants

Offices Offices Offices Offices Offices Cafes Cafes Road Land uses in Tsimiski Services Services Services Services Services Houses houses Restaurants Restaurants Restaurants Restaurants Restaurants Cafes Cafes Cafes Cafes Cafes 8:00 – 15:30 – 15:30 8:00 Houses Houses Houses Houses Houses Offices Shops Offices Shops Shops Restaurants Restaurants Restaurants Offices Restaurants Offices Offices Cafes Cafes Cafes Restaurants Cafes Restaurants Restaurants Houses Houses Houses Cafes Houses Cafes Cafes Houses Houses Houses 15:31– 21:30 21:30 15:31–

Houses Houses Houses Restaurants Cafes Cafes Cafes Restaurants Restaurants Restaurants Cafes Restaurants Restaurants Restaurants Cafes Cafes Cafes Houses Houses Houses Houses

21:31 - 23:59 23:59 - 21:31

2834 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

Time Update ‐ Predictor

e(p+1)=var3d*cf(p+1) (3) P(p+1)=rtpv(p)*cf(p+1)+cca(p+1)*rtday1(p+1) (4)

Measurement Update ‐ Correction

cf(p+1)=ep+varP/[var3d+varP+e(p)] (5)

cca(p+1)=1-cf(p+1) (6)

2.6. Travel Time Prediction initial value. Therefore the Kalman Filter is called a re- After the calculation of travel times, a travel time cursive filter [9]. prediction is feasible. The method of Kalman Filter was performed through Matlab. Kalman Filter is a set of mathe- matical equations that implement a predictor-corrector type 3. RESULTS estimator [6]. This filter has two steps: the prediction step, where the next state of the system is predicted given the 3.1 Map-Matching previous measurements. At the second step, the update As mentioned previously, the prediction zone which is step, the current state of the system is estimated given the generated from the linear regression is defined by two hy- measurement at that time step [7]. In this case, the filter perbolic curves. However, at the current study, the difference takes into consideration historical data of travel times for between maximum and minimum points is negligible giving the last three days before the day of prediction. These data the possibility to assume the hyperbolic curve as a straight are related to a specific road segment for a place from p to line [10]. After implementing this method a set of spatial p+1. In addition with the current observation of place p, it models are revealed, each one for a segment and its predic- is possible to predict the travel time for the place p+1. tion zones. This means that any point of a vehicle’s route The basic typology follows: If p denotes the place where located within the zone, will be matched with the regression the vehicle is situated then we get1: line. The points that did not belong to each prediction zone It is commonly assumed the different between the pre- were excluded as it is shown in Figures 3 and 4. diction and the observation as zero, this this difference is referred in the bibliography as innovation or residual vec- 3.2 Traffic Investigation tor. So it is resulting that the measurement is equal to ob- As it is shown in Figure 5 for a weekday from 8:00 to servation.[6] Taking into consideration this admission it is 15:30 traffic congestion problems are intensive as it was ensuing that var3d=varP the equation (5) is transformed expected. During this time period many land uses pre- as [8]: dominate which attract numerous vehicles. On the contra- cf(p+1)= [e(p)+var3d)]/ (2*var3d+e(p)) (7) ry, the same day from 21:30 to 23:59 no traffic problem could be observed since the land uses that predominate at The whole process is briefly described from the two this time period are fewer (Figure 6). following steps “Measurement Update – Correction” and “Time Update – Prediction”. The procedure is repeated It is also interesting to investigate the taxi points with for each time step using the state of previous time as an regard to velocity. Defining 35 km/h as a free-flow (un- congested), speed velocity lower than 35 km/h is assumed

as delay. For Monday on 17th of September traffic con- 1  cf: contribution from filtering, tells how much the gestion problems are revealed for the road segments 1 to 4 predictions should be corrected on the next loop from the above map. It should be noted that every point  cca: contribution from the cycle action that is figured has been assigned to a road segment ac-  e: error of filtering cording to the map-matching process and equation (1).  P: prediction More precisely, 84% of taxi vehicles have speed lower  rtpvp: Real Time of Previous vehicle than 35 km/h (Figure 7). Almost, the same situation is  rtpdp1: Real Time of Previous Day noted for Tuesday 18 September for the same road seg-  var3d: Variance of the 3 last Days ments (Figure 8). This could be explained either by a  varP: Prediction’s variance before seeing the measure- radical change of the road geometry or by a vulnerable ment segment of this road network of Thessaloniki. Knowing  rtday1p1, rtday2p1, rtday3p1 real time of that nothing changed at that period, it is concluded that it vehicle the previous day, 2 days ago and 3 days ago re- is a traffic congestion point. spectively

2835 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

FIGURE 3 - Taxi points at Tsimiski road before the process of map-matching

FIGURE 4 - Taxi points at Tsimiski road having implemented the process of map-matching

FIGURE 5 - Traffic map for Thursday 27/09/2007 from 08:00 to 15:30 FIGURE 6 - Traffic map for Thursday 27/09/2007 from 21:30 to 23:59

2836 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

FIGURE 7 - Traffic map for Monday 17/09/2007 from 15:30 to 21:30 FIGURE 8 - Traffic map for Tuesday 18/09/2007 from 15:30 to 21:30

3.2 Travel time prediction for each homogenous time zone Implementing the whole process, through Matlab, some instructive results are revealed. For each day a comparative diagram between prediction and tracking travel time has been created. It should be noted that the process is smart as it is being ameliorated and after some loops the predictions tend significantly to tracking time. As shown in Figures 9 and 10, the difference between tracking and prediction is insignificant. The effectiveness of the method is being ameliorated when the points are numerous. When the road segment is short, the effectiveness is being deteriorated because the taxi-points are restricted as it is shown in the following figure.

For the 1st time period (00:00 - 7:59), the predictions were not satisfying because of the limited number of FIGURE 10 - Prediction for and tracking travel time for a random points. In this period the traffic flow is negligible, so this Thursday, for road segment 5, for time period 8:00-15:30 result was expected. In contrast to the 2nd time period (8:00-15:30), the peak period where there are abundant

FIGURE 9 - Prediction and tracking travel time for an incidentally FIGURE 11 - Prediction and tracking travel time for a random Thursday, for road segment 4, for time period 15:31-21:30 Friday, for road segment 7, for time period 08:00-15:30

2837 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

the current study was performed through programming in Matlab. Without programming, the process would be quite difficult to be implemented because of the heavy data load. So the whole process contributes to an integrated and smart traffic prediction system.

Within future studies it would be also interesting to incorporate the field of vehicles’ velocity in the process of Kalman Filter. Finally, it would be also interesting to investigate the efficiency of a traffic prediction system using different method such as fuzzy logic, exponential smoothing, and artificial neural networks etc.

The authors have declared no conflict of interest.

FIGURE 12 - Prediction and tracking travel time for a random Friday, for road segment 2, for time period 08:00-15:30 REFERENCES points. For the 4th time period (21:31-23:59) in some cases it was not feasible to implement the whole process [1] Gidófalvi G. and Morán C. (2010) Estimating Traffic Per- formance in Road Networks from Anonymized GPS vehicle because the number of points was insufficient. Probes", Proceedings of Workshop on Movement Research: Are you in the flow?. 13th AGILE International Conference on Geographic Information Science 4. DISCUSSION AND CONCLUSIONS [2] Wuk, K., Gyu-In J. and JangGyu L. (2000) Efficient use of digital road map in various positioning for ITS. Proceedings The benefits of an efficient traffic prediction system of IEEE Symposium on Position Location and Navigation, are very significant. This system can affect positively and San Diego, CA., pp.170-176 has been applied in many sectors such as social, environ- [3] Lakakis, K., Savvaidis, P., Ifadis, I. and Doukas, I. (2004) mental and financial [11]. One of most significant appli- Quality of map-matching procedures based on DGPS and cation is to control the traffic especially at rush-hour stand-alone GPS positioning in an urban area. Proceedings of when the risk of the occurrence of traffic jam is high International conference, FIG Working Week 2004 Athens, Greece. enough [12]. So, the local Stakeholders based on the information provided from the traffic prediction system can [4] Lakakis, K. and Savvaidis, P. (2008) Map-matching proce- initiate certain traffic control methods to avoid the traffic dures at urban grade intersections. Proceedings of 10th Inter- national Conference on Applications of Advanced Technolo- jams. Furthermore, it is a useful tool which indicates the gies in Transportation, Athens, Greece. weaknesses of the road network and provides knowledge of where it is crucial to improve the infrastructures. This [5] Daskalakis N. (2008) Telematics systems using GPS for mass transit vehicles & pattern predictions in real time. Di- system could also be used as a tool for urban planning in ploma thesis, NTUA, Greece. the city defining the land uses according to traffic congestion spots. This traffic prediction system could also help [6] Georgiadou E. (2009). Control of visual servoing robot ma- to increase the safety on the roads by diminishing the nipulators with implementation of Kalman Filter. Diploma thesis, NTUA, Greece number of accidents involving injury [13]. Several approaches have been proposed for travel [7] Hartikainen J., Solin A. and Särkkä S. (2011) Optimal filtering with Kalman filters and smoothers. Dept. of Biomedica time estimation and prediction. The combination of the Engineering and Computation Science, Aalto University methods that are used in the current study can be assumed School of Science, Finland as an innovative process to predict travel times in a dense [8] Yunlong Z. and Zhirui Y. (2008) Short-Term Traffic Flow urban area such as Thessaloniki city while most of previ- Forecasting Using Fuzzy Logic System Methods. Journal of ous studies are based on the applications of regression, Intelligent Transportation Systems, Vol. 12 (3), pp.102-112 bayesian and time series models [13]. However, a signifi- [9] Kleinbauer R. (2004) Kalman Filtering Implementation with cant condition for an effective performance of the whole Matlab Study Report in the Field of Study Geodesy and method is the existence of over 20 GPS points per road Geoinformatics at Universtat Stuttgart, Germany.. segment per day even if the difference between tracking [10] Lakakis K. (2000) Land Vehicle Navigation in an Urban Ar- and prediction is minor. The basic requirements for the ea by Using GPS and GIS Technologies. PhD Dissertation, study are the computational and processing strength and a Aristotle University of Thessaloniki, Thessaloniki, Greece large amount of GPS points. So, it constitutes a cost-effective endeavour to mitigate a crucial problem that degrades qual- [11] Lin H-E, Zito R. and Aptaylor M. (2005) A review of travel- time prediction in transport and logistics. Proceedings of the ity of living and deteriorates day by day. The endeavour to Eastern Asia Society for Transportation Studies 5, pp.1433- automate the process is also successful since each stage of 1448

2838 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

[12] Kriegel H-P, Renz M., Schubert M. and Zuefle A. (2008) Statistical density prediction in traffic networks. In Proceed- ings of the 8th SIAM International Conference on Data Min- ing (SDM), Atlanta, pp.692-703 [13] Ta-Yin, H. and Wei-Ming, H. (2010) Travel Time Prediction for Urban Networks: the comparisons of Simulations-based and Time-Series Models. Proceedings of the 17th ITS World Congress, Busan, pp.1-11

Received: June 04, 2014 Accepted: June 18, 2014

CORRESPONDING AUTHOR

K. Lakakis Department of Civil Engineering Aristotle University of Thessaloniki 54006, Thessaloniki GREECE

Phone: +30 2310995720 Fax: +302310996159 E-mail: [email protected]

FEB/ Vol 23/ No 11a/ 2014 – pages 2832 - 2839

2839 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

REDUCTION OF THE UNCERTAINTY IN THE FREQUENCY CALCULATION IN PORT ACCIDENTS THROUGH FUZZY LOGIC

José R. González Dan1 and Rosa Mari Darbra1,*

1Centre d’Estudis del Risc Tecnològic (CERTEC). Chemical Engineering Department. Universitat Politècnica de Catalunya. Diagonal 647. 08028-Barcelona. Catalonia. Spain.

Presented at the 4th International Conference on Environmental Management, Engineering, Planning and Economics (CEMEPE), June 24 to 28, 2013, Mykonos, Greece

ABSTRACT property, vessels included. To ensure safety in this area, it is necessary to conduct a risk assessment [2]. This as- Seaports play a very important role in the internation- sessment relies on a method of hazard evaluation based al commerce. They have influence on very large geograph- on numerical estimation of incident frequency and conse- ical areas and carry out very complex operations with large quences [3]. In recent decades, the way that consequences volumes of goods. Maintaining high efficiency in the trans- are treated in risk assessment has changed. New formulas portation of goods as well as a low level of risk within the for more accurate data results have been created, but the port area is very important. In order to reduce this risk, it is estimation of frequencies still appears to be largely based necessary to analyze the different components of the risk on past data; it is common to use values of frequencies assessment: frequency and consequences. These two from decades ago to conduct the risk calculations. This is parameters are influenced by numerous and imprecise the reason why the present work focuses on the improve- factors. In the present work, an analysis of the principal ment of these frequencies through the use of a modifier. In uncertainty sources present in the calculation of the fre- order to obtain the frequency of an accident for any dan- quency rate for a maritime accident is presented. Taking gerous scenario different techniques are available; different into consideration the variables that can influence the parameters are considered in each one of them to calculate frequency and a standard value of this parameter, a modi- the frequency. However, the no inclusion of other relevant fication is made in order to get better assessment of the parameters for the frequency calculation can lead to the frequency of accidents in ports. The developed method is increase of the uncertainty when assessing the frequency based on a fuzzy modifier which reduces the uncertainty of an accident. The elements of the risk assessment have in the seaports accident frequency, improving as a conse- associated uncertainty most of the time. This parameter will quence the final risk assessment. never be eliminated; it will always be present in the frequency calculation, however it can be reduced. One of the best ways to deal with all the types of uncertainty including

KEYWORDS: lack of knowledge and vagueness is through fuzzy logic port; uncertainty; fuzzy logic; frequency; accidents. [4]. This theory allows including new variables not considered by the traditional methods of frequency estimation and deal with the uncertainty involved. It allows representing the uncertainty of a measure in a human language. 1. INTRODUCTION This technique is very useful for concepts such as risk, hazard and quality. Recent works link this methodology Port terminals are critical infrastructures that play a with the risk assessment field. For example, Markowski et key role in the transportation of goods and people [1]. al. [5] used an application of the fuzzy logic to an explo- Most of the goods transported through seaports are part of sion risk assessment. Zhou [6] also used the fuzzy method the daily life. However, seaports do not have the same for a risk assessment for major hazard installations storing recognition with the general public as airports. Under fixed flammable gas. Gürcanli and Müngen [7] made and occu- shipping lines, the growing number of vessels, may lead to pational safety risk analysis at a construction sites using the increase of maritime accidents. Most frequently marine fuzzy sets. casualties occur near ports, where most of the vessels ma- noeuvres occur. A port is considered a place of risk where This paper introduces a fuzzy modifier developed to the danger can be directed to persons, environment and/or reduce the uncertainty associated to the accident frequency estimation. An initial frequency is required; this tradi- * Corresponding author tional frequency is modified according to the influence of

2840 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

a set of variables that were not considered initially but 3. FUZZY LOGIC that are relevant for the scenario selected. Fuzzy logic is employed to incorporate the expert knowledge in the Fuzzy logic is oriented to the rationality of uncertainty frequency calculation in order to reduce the associated due to imprecision or vagueness and resembles human uncertainty. In order to test this modifier a specific sce- reasoning in its use of approximate information and uncer- nario was selected: the accidents in ports. Two different tainty to generate decisions. In Figure 1, the different case studies were analyzed. They were based on a paper steps of the fuzzy methodology are presented. First of all, in which 828 accidents were analyzed through a historical it is necessary to identify the variables that are relevant to analysis in order to predict the frequency of accidents in the system (initial inputs). After this, it is necessary to port areas [8]. Through the fuzzy modifier method, new establish fuzzy sets to represent their numerical value; this parameters are included in the frequency calculation tak- process is called fuzzification [10]. Next, this methodolo- ing in to account the uncertainty generated and as a con- gy and the following steps are explained. sequence improving the accidents frequency assessment. 3.1 Fuzzy sets and membership functions

2. FREQUENCY CALCULATION A fuzzy set is a class of objects with a continuum of grades of membership. These fuzzy sets transform the Risk is calculated by multiplying the frequency in numerical values to linguistic parameters such as low, which an event occurs (or will occur) by the magnitude of medium or high. Such a set is characterized by a member- its probable consequences [3]. Since the frequency of an ship function, in which a grade of membership is assigned event is a very important part of any kind of risk assess- to each object ranging between 0 and 1. Furthermore, a ment and also the starting point for this work, a specific fuzzy set is an extension of a crisp set used in classical subsection is devoted to its calculation. The frequency logic. A fuzzy set in the universe represented by “U” can calculation depends on the quality of the failure rate data be defined in the following equation: used and it is notoriously difficult to collect. Therefore, in many cases no sufficient information is available. The un- Eq (1) certainty present can be related to the lack of real time data for equipment failure rates, human errors, and the Where is the membership function, “U” is the wrong selection of the variables to analyze. There is a universe and “x” is a finite set of points contained in the common opinion that the frequency calculation, especially universe. in maritime fields depends also on other variables that are A fuzzy membership function is an expression defin- no taken into account in the accident databases. It is clear ing the grade of membership of an element “x” in a set that the calculation of the frequency of maritime accidents “A”. For each type of membership function, some param- in ports should not only be based on the parameters eters and equations have to be defined to represent the shown in an accident database. There exist different vari- function. For example, in the case of the triangular mem- ables that may affect the calculation of this frequency and bership function the main parameters are: they need to be studied in order to be included later on the final calculation. According to Romer et al. [9] the marine accident frequencies depend also on visibility, brightness, geographical environment, age of vessel and size of the vessel. They found that the collision and grounding frequencies were found to increase with decreasing visibility, brightness and more restricted waters. Fuzzy logic is a methodology which allows including these new variables and working with uncertainty, taking into account the Eq. (2) imprecisions and the real-world vague data.

Initial Inputs Final Output

Fuzzy Sets Fuzzification Defuzzification Membership functions

Fuzzy Rules

FIGURE 1 - Steps of the fuzzy logic methodology

2841 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

3.2 Fuzzy rules 4. CASE STUDIES Once defined the sets and membership functions for each variable, the inputs and outputs need to be connected The scenario selected is based on 828 accidents that by rules which are normally based on suppositions like “if were analyzed through a historical analysis of accidents … then… else” [11]. The main advantage of the applica- occurred in port areas [8]. From these data, the sequence tion of the fuzzy “if – then” rules is its capability to per- followed by the accidents was identified and allowed to form inferences under partial matching. This matching define a probability event tree with different frequency of degree is combined with the consequence of the rule to occurrence. For this study, the port operation selected was form a conclusion inferred by the fuzzy rule. In that way, the approach followed by manoeuvre. This kind of opera- a fuzzy rule associates a given condition (antecedent) to a tion occurs when a vessel enters or exits port waters; it conclusion (consequent). Therefore, in a rule-base scenar- involves crossing several ships on its way when proceeds io a rule R can be denoted by: towards or departs from a berth or jetty of a port. Then,

“IF u1 is A1 AND u2 is A2 AND u3 is A3, using frequency data of port accidents taken from litera- THEN v is B” Eq (3) ture it was possible to obtain two different scenarios: 1. Initial frequency for a single hulled ship – ship collision 3.2 Inference process (Figure 2) and, 2. Initial frequency for the collision of two In order to get an output when applying the rules, it is fuel oil tankers (Figure 3). Multiplying the initial frequen- necessary the inference process. The Mamdani model [12] cy with the probability of each branch obtained from the is the most common inference fuzzy model which involves event trees (both data from literature), frequencies are the implication and aggregation of the rules outputs [13]. obtained (in black colour). Those frequencies have uncer- The outputs of the model are fuzzy estimates, that is, de- tainty associated because those are standard values based grees of certainty of different possible outcomes. on a collection of past accidents from different sources obtained by statistical methods. The objective of this 3.2 Defuzzification method paper is to reduce this uncertainty through the modifica- Hence, in order to transform the fuzzy results into a tion of these frequencies through a modifier which will precise output it is necessary the defuzzification process include the information on the valuable variables into the which transforms an aggregate output fuzzy set into a final frequency value (red colour). single number. The method generally used for this process is the centroid method [14]. With this crisp value a final assessment of the output variable can be provided.

Freq. of initi- Release No further consequence -06 -06 ating event 8.59x10 7.64x10 9.47x10-06 -05 -05 1.00x10 1.06x10 -05 1.24x10 Fire No further consequence -07 -07 -07 6.03x10 5.60x10 4.52x10 5.60x10-07

Gas cloud

-08 Explosion 1.00x10 1.24x10-08 -07 1.51x10 1.87x10-07

Explosion -07 2.51x10 3.11x10-07

Fire No further consequence -07 -07 8.54x10 -07 -06 6.03x10 7.48x10 1.05x10

Explosion -07 2.51x10 3.11x10-07

Explosion -07 5.53x10

6.85x10-07

FIGURE 2 - Corrected frequencies for a single hulled ship-ship collision based on [8]

2842 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

Freq. of initi- Release No further consequence -07 -07 ating event 1.27x10 1.13x10 1.40x10-07 -07 -07 1.48x10 1.57x10 -07 1.83x10 Fire No further consequence -09 -08 -09 8.88x10 1.10x10 -09 6.66x10 8.25x10

Gas cloud -09 Explosion -09 -09 1.48x10 1.83x10 2.22x10 2.75x10-09

Explosion -09 3.70x10 4.58x10-09

Fire No further consequence -08 -09 1.25x10 -08 -08 8.88x10 1.10x10 1.55x10

Explosion -09 3.7x10 4.58x10-09 Explosion -09 8.28x10 -08 1.027x10

FIGURE 3 - Corrected frequencies for a collision of two fuel oil tankers based on [8].

5. METHODOLOGY logic process, the first step is to determine the variables that affect the frequency according to the selected scenar- In order to reduce the uncertainty associated to the io (approach followed by manouvre operation). The vari- frequency, it is necessary to take in to account new varia- ables that are taken into account are: pilot experience, bles that can affect the frequency of an accident in port weather condition, traffic and safety measures (see Figure areas, specifically in the approach followed by manouvre 4). They need to be transformed from numerical values to scenario. This can be done through the use of a fuzzy linguistic variables through a fuzzification process, where frequency modifier. This concept was already used to fuzzy sets and membership functions are defined. All this estimate frequency of hazardous materials transport acci- has been done using the “fuzzy toolbox” in the mathemat- dent through fuzzy sets [15]. From an initial frequency ical program Matlab [16]. Each one of these variables and value and the fuzzy modifier, a new frequency value is its fuzzification process is explained next. obtained which includes the influence of the relevant variables for the selected scenario. Following the fuzzy

Fuzzy Frequency Modifier

Safety Traffic Weather Pilot Measures Conditions Experience

Harbour Marine Management Services

FIGURE 4 - General scheme for the fuzzy frequency modifier

2843 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

5.1 Pilot experience variable modifier and its fuzzy sets are presented in table 1: Calm A pilot is the person responsible to enter a ship safely seas (good weather), storm (heavy weather, bad weather, into a port. According to the main EU regulations on mari- heavy seas, rough seas, squall), typhoon (hurricane, cy- time safety of ports, pilotage is a very important activity in clone, tornado, freak weather conditions, freak seas [21]. ports in order to ensure safety. Therefore, the pilot’s role is The weather conditions are measured in a scale from 1 to essential to reduce the frequency of accidents in a port area. 10; if the condition of the weather is optimal to navigation However, as stated in a survey conducted to 77 maritime (sunny day) the value measured is going to be 1 and 10 pilots around Australia and New Zealand [17] pilots will be referred to extreme weather conditions (i.e. hurri- are exposed to commercial pressure and long hours work- cane, heavy rain). ing, which could lead then to a potential risk of accident. Therefore, the experience of the pilots is considered as an 5.4 Safety measures variable important variable in order to reduce the uncertainty of It is logical to think that if ports have poor safety the frequency calculation. The fuzzification of this varia- measures there will be an increase of the frequency of the ble is done through three fuzzy sets which represent the accidents and also of the magnitude of the consequences. different expertise of the pilots: novice, medium and experienced (Table 1). The existing international regulations and also the national regulations established the minimum safety measures TABLE 1 - Fuzzy sets for the input variables that a port needs to have in order to keep the risk at minimum levels. Of all the requirements of the different regula- Variable Fuzzy sets tions that a port must have, the safety measures can be Pilot experience Novice Medium Experience divided in [22]: “Harbor management measures” (activities Range: 0 - 25 required to maintain the port in a safe condition for naviga- 0-15 10-18 15-25 years tion) and “Marine services measures” (provision of com- Traffic Low Medium High mercial services within the port.).

Range: 0 - 45 000 0- 20 000 15 000- 30 000 25 000- 45 000 vessels/year Fuzzy sets have been defined for each variable, now, membership functions need to be defined for them. In this Weather Calm seas Storm Typhoon work the selected type of membership functions has been Range: 1 - 10 1 - 5 3 - 7 6 - 10 the “Gaussian function”. After this, the variables are connected through rules and the inference process is applied. Finally, the centroid method is used to defuzzify and get 5.2 Traffic variable the final output, which provides the value of the modifier. Traffic in a port can be defined as the number of ves- These last two steps can be seen in Figure 5 in which the sels going in and out. It was established that when ships areas of the activated rules are joined and a final numeri- are navigating into the port area the vessel traffic increas- cal value is obtained. es whereas the space reduces, leading to more often marine casualties [18]. An investigation of the risk caused by traffic port through a study of accidents in Hong Kong 6. RESULTS AND DISCUSSION [19] established that the traffic level is one of the main indicators for improvements in the efficiency and service After the combination of the 4 variables into the in ports. This study pointed out that the most busy sea- fuzzy modifier (see Figure 4), the frequency modifier ports, face risk issues (i.e. increase of the number of acci- gives a final value of 1.24 (Figure 5). In the risk assess- dents) due to the increase of traffic level, since trade con- ment methodologies, it is a normal practice to correct the tinues to grow. Therefore this variable is also taken into value of the frequency by multiplying by several factors account in the fuzzy modifier and its sets are presented in (potential domino effect, the working hours of the equip- Table 1. ment) and then, continue with the study to obtain an overall value of risk. Under this concept, this modifier, repre- 5.3 Weather condition variable sented as mi in the equation 4, modifies the literature fre- It is widely known that transport systems on the whole quency (finitial) by multiplying it. From here a final frequen- perform worse under adverse and extreme weather condi- cy is obtained (ffinal). tions. Weather continues to have a major effect on the safety of ships in ports. Improvements in weather forecast- ffinal = finitial x mi Eq (4) ing can have significant effects on reducing weather- related port accidents [20]. Although there exist high level From the study of Ronza et al. [8] an event tree of of safety measures against bad weather in ports, the un- probabilities for each case study was obtained. For each predictability and force that the weather may have, make of them, a new initial frequency was calculated using the difficult to control the damages that the weather may frequency modifier (1.24): 1.24x10-05 harbour movement-1 cause. Therefore this variable is taken into account by the for the singe hulled ship-ship collision and 1.83x10-07 har-

2844 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

FIGURE 5 - Obtention of the numerical value of the modifier

bour movement-1 for the collision of two fuel tankers. If a more specific approach wants to be followed, it is These new frequencies come from the multiplication of even possible to modify the frequency of each single the initial frequency (1.00x10-05 harbour movement-1 and accidental scenario (e.g. release – gas cloud), including 1.48x10-7 harbour movement-1) by the fuzzy modifier the effect of more specific variables into their particular value (1.24). With these new initial frequencies it is pos- frequency calculation. For example, if the aim is to reduce sible to obtain new final frequencies for each accidental the frequency of occurrence of a release – gas cloud acci- scenario (i.e. fire, gas cloud, explosion), this can be seen dent, variables related to the dispersion of toxic gases in figure 2 and 3 in which the modified frequencies (in must be included in the calculation of the fuzzy modifier. red colour) are next to the initial frequencies. The new The results obtained through the use of the modifier rep- final frequencies are slightly higher than the previous resent more realistic values of frequencies since they ones, with this increase, there has been a reduction of the include the influence of other variables into their fre- uncertainty since new variables have been taken into quency calculation. The relatively higher result of the account. In most of the cases the variation is not greater frequencies implies a more conservative approach lead- than one order of magnitude. This is normal, since the ing to the increase of safety measures. objective was to improve the frequency not to change it drastically. However, changing a little bit the frequency can improve the whole risk assessment in an important 7. CONCLUSIONS way and reduce the associated uncertainty. In figure 2 it can be also observed that the most common accidental It is undeniable that uncertainty is present when a fre- scenario that can occur is the release without no further quency calculation is carried out. The methodology pro- consequences occurring about once each 9.47x10-06 har- posed in this paper allows reducing the frequency uncer- bour movements and the scenario with the lowest value of tainty through the use of a fuzzy modifier which includes frequency is the release followed by gas cloud occurring the information of specific variables. Two case studies once every 1.24x10-08 harbour movements. In the second related with port accidents have been selected, particular- case study (figure 3) the most common accidental scenar- ly focusing in the accidents related to the approach and io is also the release with no further consequences occur- manoeuvre of a vessel. The initial frequency of occurrence ring one each 1.40x10-7 harbour movements and the sce- has been modified introducing new variables that were not nario with the lowest value of frequency is the release+ considered. This has allowed a reduction of the uncertainty gas cloud occurring once every 1.83x10-9 harbour move- associated. The new frequency does not represent a major ments. change in the value, because although the variables influ-

2845 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

ence in the results, the aim is to obtain a more accurate [14] Yen, J., Langari, R. (1999). Fuzzy Logic: Intelligence, Con- value more alike to the real-world, and not a major change trol, and Information. Pearson Education, USA. in the frequency value. However, these results allow reduc- [15] Qiao, Y., Keren, N., Mannan, M.S. (2009). Utilization of ac- ing the uncertainty associated to the frequency estimation cident databases and fuzzy sets to estimate frequency of and provide a more conservative approach which will lead HazMat transport accidents. J. Hazard. Mater., 167, 374-382. to the increase of safety measures, since the frequency is [16] Mathworks (2011). Fuzzy Logic Toolbox. Available from: higher than the previous one. The fuzzy sets theory has http://www.mathworks.com/help/toolbox/fuzzy/fuzzy_produ ct_page-html(accessed 08.07.2011) been growing in numerous applications, in this case in the risk assessment field and in particular in the frequency [17] Darbra, R.M., Crawford, J.F.E., Haley, C.W., Morrison R.J., estimation. However, future research needs to be made in (2007). Safety culture and hazard risk perception of Australi- an and New Zealand maritime pilots. Marine Policy, 31, the consequence part of risk assessment as well. This 736–745. approach can be also applied to different situations in [18] Wen-Kai K.H. (2011). Ports’ service attributes for ship navi- which uncertainty is present. gation safety. Safety Science, 50, 244–252. [19] Yip, T.L. (2008). Port traffic risks-A study of accidents in The authors have declared no conflict of interest. Hong Kong waters. Transportation Research Part E: Logis- tics and Transportation Review, 44, 921-931. [20] Dutton, K. (1998). Forecasting Improved Safety and Cost- Savings, Dock and Harbour Authority. 79, 88. REFERENCES [21] http://promo.lloydslistintelligence.com/lloyds-list- intelligence-casualty-reporting-service/ (accesed July 05, [1] Mokhtari K., Ren J., Roberts C., Wang J. (2011). Application 2011). of a generic bow-tie based risk analysis framework on risk management of sea ports and offshore terminals. J. Hazard. [22] New Zealand Maritime Safety Authority (2004). Guidelines Mater., 92, 465-475. for Port & Harbourd Risk Assessment and Safety Manage- ment Systems in New Zealand. [2] Chlomoudis, C.I., Kostagiolas, P.A., Lampridis, C.D. (2011). Quality and safety systems for the port industry: empirical evidence for the main Greek ports. European Transport Re- search Review, 1-9. [3] Casal, J. (2007). Evaluation of the Effects and Consequences of Major Accidents in Industrial Plants. Elsevier Science. [4] Markowski, A.S., Mannan, M.S. (2009). Fuzzy logic for pip- ing risk assessment (pfLOPA). J Loss Prev Process Ind., 22, 921-927. [5] Markowski, A.S., Mannan, M.S., Kotynia, A., Pawlak. H. (2011). Aplication of fuzzy logic to explosion risk assessment. J Loss Prev Process Ind 24, 780-790. [6] Zhou, J. (2010). SPA-fuzzy method based real-time risk assessment for major hazard installations storing flammable gas. Safety Science. 48, 819-822.

[7] Gürcanli, G. E., Müngen, U. (2009). An occupational safety risk analysis method at construction sites using fuzzy sets. Int Received: June 05, 2014 J Ind Ergonom. 39, 371-387 Accepted: June 18, 2014 [8] Ronza, A., Félez, S., Darbra, R., Carol, S., Vílchez, J., Casal, J. (2003). Predicting the frequency of accidents in port areas by developing event trees from historical analysis. J Loss CORRESPONDING AUTHOR Prev Process Ind., 16, 551-560. [9] Rømer, H., Petersen, H., Haastrup, P. (1995). Marine acci- Rosa Mari Darbra dent frequencies–review and recent empirical results. J. Nav- Centre d’Estudis del Risc Tecnològic (CERTEC) ig., 48, 410-424. Chemical Engineering Department [10] Zadeh, L.A. (1965). Fuzzy sets. Information and control, 8, Universitat Politècnica de Catalunya 338-353. Diagonal 647 [11] Darbra, R. M., Mure, S. (2008). Preliminary risk assessment 08028 Barcelona of ecotoxic substances accidental releases in major risk in- SPAIN stallations through fuzzy logic. Process Safety and Environ- mental Protection., 86, 103-111. Phone: +34 4934010811 [12] Jang, J. (1997). Fuzzy Inference Systems. Uper Saddle River, Fax: +34934011932 NJ: Prentice-Hall. E-mail: [email protected] [13] Dubois, D., Prade, H. (1980). Fuzzy sets and Systems – The- ory and applications. Academic Press. FEB/ Vol 23/ No 11a/ 2014 – pages 2840 – 2846

2846 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

RATIONAL SOLITARY WELL SPACING IN SOIL REMEDIATION PROCESSES

Marios S. Valavanides1,* and Eugene D. Skouras2

1 Dept. of Civil, Survey & Geoinformatics Engineering, TEI Athens, 12210 Athens, Greece 2 FORTH/ICE-HT, Stadiou Str., Platani, P.O.Box 1414, GR-26504 Patras, Greece

ABSTRACT remediation technologies have been developed to treat soil, leachate, wastewater, and groundwater contaminated The present work addresses design aspects of soil re- by various pollutants, including in-situ and ex-situ meth- mediation processes and suggests a rational well-spacing ods. An extended review of biological, physical & chemi- scheme to optimize the operational efficiency of systems cal site-restoration technologies when soil and groundwa- and processes, considered in terms of pollute extracted per ter are contaminated by petroleum and related products, kW of mechanical power dissipated in pumps. The optimi- furnishing the respective processes, their applicability, zation scheme is proposed on the provisions of a universal advantages, limitations and concerns, the site-specific pa- map for steady-state two-phase flow in porous media, rameters as well as the related costs, can be found in [1]. demarcating the process operational efficiency over the According to the type of contaminated site, there are domain of the process operational parameters, i.e. the pol- two major groups of applicable technologies: (A) Soil lute/water flow-rate ratio and the capillary number. Process remediation technologies: these comprise soil washing, efficiency is considered over a formation volume contain- soil vapour extraction (SVE), land-farming, soil flushing, 2 ing a single well. Pertinent DeProF model scaling law solidification/stabilization (waste fixation), biopiles, phy- predictions for the reduced mechanical power dissipation are toremediation, bioslurry systems, bioventing, encapsula- integrated across the formation control volume. Then, tion, aeration, and (B) Groundwater treatment technol- global values of energy utilization are estimated in terms of ogies: these comprise air sparging, groundwater pump- the process design parameters, i.e. the capillary number at and-treat technology, passive/ reactive treatment walls, the vicinity of the well-bore (Ca0), the duty oil/water flow- bioslurping, ultraviolet-oxidation treatment, biosparg- rate ratio (r), and the radius of influence ( c ). Results ing, groundwater circulation wells, in-well air stripping, indicate that, given the maximum permissible value of Ca0, in-situ air and vapour stripping and vacuum vapour ex- the overall efficiency of the process is increased with traction, natural attenuation. decreasing c . Many of these technologies implement injection/ extraction wells [1]. In soil flushing, contaminated soils are

flushed in-situ by passing an extraction fluid through soil KEYWORDS: two-phase flow in porous media, operational effi- using an injection or infiltration process. Contaminated ciency, rational well spacing, soil remediation, near well flow groundwater and extraction fluids are captured and pumped to the surface using standard groundwater extraction wells. In groundwater pump-and-treat, extraction wells are introduced at various locations in a contaminated aquifer, 1. INTRODUCTION and the contaminants are removed with the pumped water. Groundwater circulation wells comprise a developing Pollutants, mainly in the form of complex chemical technology, used for removing contaminants from ground- mixtures – such as total petroleum hydrocarbons (TPH), water and saturated soils, implementing a process that polychloro biphenyls (PCBs), polycyclic aromatic hydro- continuously removes VOCs from the groundwater with- carbons (PAHs), heavy metals, and pesticides - enter the out bringing it to the surface. This circulation pattern is environment directly as a result of accidents, spills during created in an aquifer by drawing water into- and pumping transportation, leakage from waste disposal or storage it through- a well, and then reintroducing it without reach- sites, or from industrial facilities, and pose potential dan- ing the surface. gers to both the human health and the environment. Many The successful and cost-efficient treatment of a contaminated site depends on selection, design, and adjust- * Corresponding author ment of the appropriate remediation technology, consider- 2 DeProF is the acronym for Decomposition in Prototype Flows, see ing the properties of the contaminants, the soil and the section 1.1

2847 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

~ potential performance of the associated process which, in capillary pressure, defined as Ca  μ~ Uw /~γ (μ~ is the turn, depend on the proper system design as well as on the w ow w ~ w selection of conditions providing maximum operational viscosity of water, U is the superficial velocity of wa- ~ efficiency (optimum operating conditions). ter, and ow is the interfacial tension), the oil/water flow- ~ ~ rate ratio, r  qo / qw , the oil/water viscosity ratio, The core physical process in the aforementioned ~ ~ technologies is immiscible two-phase flow in porous   o / w , the advancing and receding contact angles 0 0 media whereby one fluid (“water”) is wetting the solid in  and  , and a parameter vector, xpm, composed of presence of the other, non-wetting fluid (pollutant or “oil”). A R all the dimensionless geometrical and topological parame- Possible flow modes are: simultaneous, concurrent, forced ters of the porous medium affecting the flow (e.g. porosi- flow of wetting and non-wetting fluids at preselected flow- ty, genus, coordination number, normalized chamber and rates, imbibition, i.e., displacement of a non-wetting by a throat size distributions, chamber-to-throat size correla- wetting fluid and drainage, i.e., displacement of a wetting tion factors, etc), including the absolute permeability of by a non-wetting phase or air. These types of flow have ~ two key characteristics: the disconnection of the non- the porous medium, k . (Note: a tilde ~ indicates a di- wetting phase (pollutant /oil or air) into small fluidic mensional variable.) entities -called ganglia and droplets (or bubbles)- and the Extended simulations of steady-state two-phase flow formation of the respective wetting/non-wetting interfac- in a three-dimensional (3D) pore network of the chamber- es. Ganglia are fluidic entities (non-wetting blobs) that and-throat type, using the DeProF model, showed that, in occupy the void volume of a few (1-10) pores. Droplets general, the reduced mechanical power dissipation may be (or bubbles) are smaller fluidic entities, and any pore may analytically expressed as a scaling law, contain many of these. It has been experimentally ob- A r served [2, 3] that during the simultaneous flow of two   ~ ~~  6 Br immiscible phases, the disconnected phase (non-wetting) Wkw 10 Ca r  rlim Ca WCa,r    (1) contributes significantly (and in certain cases of practical ~ 2 2 ow Ca  interest, even exclusively) to the total flow. Furthermore,  n / a r  rlim Ca the flow-rate vs pressure gradient relation is found to be ~ strongly non-linear, and to be strongly affected by the where, W is the specific - per unit porous medium vol- physical parameters that pertain to the fluid-fluid interfac- ume - mechanical power dissipation, and es. 3 i Ai logr i0 The scope of the present work is to furnish a rational Ar 10 , Br  B0  B1 logr and methodology for designing more efficient soil/ groundwa- Dk  r Ca  10C Ca (2) ter remediation processes, specifically for those remedia- lim   tion schemes where injected /extracted fluids are adminis- are functions of r, and Α0, Α1, Α2, Α3, Β0 and Β1 are ap- tered through systems of rectilinear wells. The methodol- propriate scaling coefficients that take values according to ogy implements the provisions of the DeProF theory for the system parameters; rlim(Ca) is the maximum value of two-phase flow in porous media. the flow-rate ratio for which steady-state two-phase flow can be maintained. Typical values of the scaling coeffi- 1.1 The DeProF theory for steady-state two-phase flow in cients as well as W(Ca, r) diagrams pertaining to Eqs. (1) porous media and (2) are provided in [4]. The mechanistic model DeProF for immiscible steady- state two-phase flow in pore networks, developed by 1.2 Optimum operating conditions (OOC) for steady-state Valavanides & Payatakes [4], predicts the relative perme- two-phase flow in porous media abilities of the two-phases using the concept of decompo- The efficiency of the process, expressed as “the non- sition in prototype flows. Combining effective medium wetting phase transport per kW of mechanical power theory with appropriate expressions for pore-to-macro scale supplied to the system”, may be assessed by the values of consistency for oil and water mass transport, the model the energy utilization coefficient, f , a macroscopic quanti- takes into account the pore-scale mechanisms and the EU ty defined by Valavanides & Payatakes [4] as follows: network-wide cooperative effects as well as the sources of non-linearity (caused by the motion of interfaces) and fEU = r/W (3) other complex effects. Simulations implementing the DeProF model suggest Using the DeProF model, one can obtain the solution that conditions of optimum operation (read: maximum to the problem of steady-state two-phase flow in porous efficiency) exist for processes of steady-state two-phase media in terms of the following set of independent varia- flow in pore networks. The term ‘optimum operating conditions’ (OOC) interprets those values of Ca and r (the bles, Ca, r; , 0 , 0 , x i.e., the capillary number, Ca, A R pm operating parameters), for which the process efficiency which is a measure of the ratio of viscous stresses over takes one (or many) locally maximum values. As is de-

2848 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

~ ~ ~ picted by the hump-shaped surface of Fig. 6 in [4], for where,    0 is the reduced radius ( 0 is the well every fixed value in Ca, there exists a unique value in r, radius), r*(Ca), for which fEU attains a locally maximum value. These provisions are described by a universal map for Ca0 = Ca(1) is the Ca value at the well-bore walls, (5) steady-state two-phase flow in porous media, demarcating and r0 is the oil-water flow-rate imposed by the process the process operational efficiency [5]. setup.

Detecting and setting such conditions in a real process could eventually increase its efficiency. Now, when ~q one-dimensional flow conditions are considered, the en- o gineer may decide to operate the process at a certain value ~q of the capillary number, Ca, dictated either by the power ~ w capacity of the available equipment or the maximum al- 0 lowable stress of the underground formation (porous medium); then, he may tune the process to optimum operating ~ ~ conditions in terms of the flow-rate ratio, r*(Ca). Neverthe-  Uo ~ less, there are applications whereby isolated wells are c used and one-dimensional flow conditions can only be ~ ~ U attained far away from the well bore. The flow in the w vicinity of the well bore is axisymmetric. Therefore, the Darcy velocity, hence the capillary number, change as 1/  , to the radial distance,  , away from the well bore. In such cases, OOC can only be attained at a certain distance FIGURE 1 - The representative cylindrical volume where the analy- from the well bore. The engineer must then decide on the ~ sis is carried out. The direction of macroscopic flow of oil, Uo , and particular operating conditions for which the overall effi- ~ water, , is radial; oil and water are extracted with flow-rates ciency of the process is maximized. Uw ~ ~ q o and qw through the rectilinear well (perpendicular to figure) acting as a sink. 2. PROBLEM DESCRIPTION AND ANALYSIS Given the physicochemical characteristics of the sys- Consider a horizontal, contaminated underground tem (pollutant, water and porous medium characteristics), formation, say an aquifer. Consider the case whereby the an expression for the process efficiency in terms of the ~ formation is homogeneous and isotropic, and is confined radius of influence, c , can be derived. The efficiency of between two horizontal impermeable planes. A typical the process depends: (a) on the values of the design and soil flushing process is accomplished by passing an ex- operating parameters that can be tuned over a wide range traction fluid through the soil using an injection or infil- and may be considered as “soft”, and (b) on the values of tration process in order to mobilize the “oily” pollutants the oil-water-p.m. system parameters, i.e. the physico- (say, TPH, PCBs, PAHs) that are trapped in the smaller chemical properties of the two phases and the p.m. char- pores of the formation, drive it to extraction wells, and acteristics, that can be tuned only over a narrow range and then, to the surface. Injection and production wells may may be considered as “rigid”. For example, the engineer be arranged in various patterns depending on the mor- has to decide on the extent of the recovery (formation) phology, extent and structure of the contaminated for- volume per well, dictated by c , a process set-up parame- mation. ter. Similarly, the values of the imposed operating param- In the present work, we consider the simple case of eters, Ca0 and r0, can be adjusted so that the process is one or many parallel isolated extraction wells (acting as appropriately tuned. Note that the value of the admissible rectilinear sinks), surrounded by injection wells distribut- (or attainable) Ca0 is limited by the strength of the for- ed over the contaminated reservoir field. The wellbore mation material near the wellbore or the power capacity radius of any extraction well is ~ . of the equipment (pumps etc). 0 Consider also a normal cylinder, of radius ~ , coaxial The process operating cost is directly related to the re- c duced (over the unit p.m. volume) rate of irreversible trans- to the extraction well, such that the macroscopic flow ~ ~ formation of mechanical energy into heat (the reduced velocities Ui , or flow-rate of each phase, qi , i=o,w with- ~ ~ ~ mechanical power dissipation), Wc ,, z , needed to main- in this cylinder may be considered – within acceptable tain two-phase flow over the recovery volume (the right error – to be radial (Fig. 1). Mass balance for each phase circular cylinder, coaxial to the production well, extending imposes that from the wellbore radius,   1 , to the influence radius c .

r  r0 and Ca  Ca 0  ,   1 (4) The mechanical power dissipation is then given by

2849 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

and design,  , parameters on the process efficiency. To ~ ~ ~ c Wc Ca 0 ,r0 ;c  WdV this end, a typical system with oil/water viscosity ratio, ~ V (6)  1,45 (with scaling coefficient values given in [4]) was ~ 2 c ~ 2 ~ ow 2 1 considered. Results are provided in Figure 2.  20 z ~ Ca 0 W d ~   k w 1 Examination of the diagram of Figure 2 reveals that: The reduced, with respect to the reduced rate for one- (A) Under any fixed values of Ca0 and  there cor- phase flow (1 ) power losses, are estimated as c responds a specific value of the oil/water flow-rate ratio, ~ c Wc 1 1 r*(Ca0), for which the process attains maximum efficien- Wc Ca 0 ,r0 ;c  ~  W d (7) W1 ln   cy. This is more noticeable for low Ca0 values and for c c 1 recovery cylinders of relatively small diameter. Note that

The idea is to set the values of the process design pa- there are cases (for relatively high values of Ca0 and c ) rameters, i.e. the values of Ca0 and r0 on the bore wall, where such a maximum is not attainable: the cylindrical and the radius, c , of the extraction volume (right cylin- surface outlined perpendicular at the far right end of the -9 der), such that the process operates at its overall maxi- fc=10 plane represents the (Ca0, log r0) domain where mum efficiency. two-phase flow is physically not sustainable.

A measure, fc, of the efficiency of the particular pro- (B) The global efficiency of the process, fc, decreases cess, may be defined by writing the ratio of the recov- significantly with increasing radius, c , of the influence ered cylinder. This is a direct consequence of the fact that as oil flow-rate, ~q , over the total mechanical power dissi- o Ca drops sharply with increasing well distance,  , the ~ pation, Wc c , within the recovery volume of radius c , effect of interfaces become pronounced (hindering the and then, reduce it by the respective efficiency ratio for migration of non-wetting blobs -ganglia and droplets), equivalent one-phase flow conditions (in terms of total and the respective local efficiency of the process is drasti- flow-rate), to produce the global energy utilization coeffi- cally reduced. cient for a recovery cylinder of radius c , -1 ~ 10 f ~q o W  c f Ca ,r ;  c c  c 0 0 c ~ w ~ 1 q W  10 c c -3 3 1 (8) 10 10 c  1 r0  r ln  W d  0 c   W  6  1  c c -5 10 10 The benefit of using Eq. (8) is that all system characteristics (pertaining to water-pollutant-porous medium) -7 are normalized-out, and the effects of the geometrical 10 aspects of the process are revealed. Incorporating the scaling law of the DeProF predictions for W from Eqs. (1) and (2), into Eq. (7) for the 2 reduced specific mechanical power consumption over a C 4 recovery cylinder of radius  , yields a 6 c 0 2 x 8 1 Br0  1 0 0 1 Ar0  Ca 0 Br - 0 6 -1 Wc Ca 0 ,r0 ;c  c 1 (9) 10 6Br0 logr ln c Br0 10 0 and the energy utilization factor, Eq. (8), may now be FIGURE 2 - Effect of the values of the operational parameters, i.e. expressed as capillary number at well, Ca0, and flow-rate ratio, r0, and for different radii, c1 , c2 , c3 , of the recovery cylinder, on the global ln c Br0 6 Br0 f Ca ,r ;  r 10 Ca (10) energy utilization coefficient, fc, (flow-rate of extracted pollutant per c 0 0 c 0 Br0 0 c 1 Ar0 kW of mechanical power consumed).

3. APPLICATION AND RESULTS 4. CONCLUSIONS Τhe problem of designing more efficient soil remedi- With the analytical expressions for W (Ca , r ;  ) and c 0 0 c ation processes implementing rectilinear sink (s) was fc(Ca0, r0; c ) available through Eqs. (9) and (10), it is considered under the provisions of the DeProF theory possible to investigate the effect of the operational, Ca0, r0 predictions.

2850 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

The energy utilization coefficient, i.e. the flow-rate of REFERENCES recovered pollutant per unit mechanical power spent - originally determined for one-dimensional flow - was [1] Khan, F.I., Husain, T., Hejazi, R. (2004) “An overview and extended for radial flow conditions to account efficiency analysis of site remediation technologies” Journal of Envi- ronmental Management 71, pp. 95–122. aspects of near-wellbore flow. [2] Avraam D.G., Payatakes A.C. (1995) “Flow Regimes and When a solitary sink is present, the energy efficiency Relative Permeabilities during Steady-State Two-Phase Flow of the process - for fixed values of the system parameters in Porous Media” J. Fluid Mech. 293, pp. 207-236. (physicochemical characteristics of the two fluids and [3] Tallakstad, K.T., Knudsen, H.A., Ramstad, T., Løvoll, G., porous medium) - depends not only on the values of the Maløy, K.J., Toussaint, R., Flekkøy, E.G. (2009) “Steady- operational parameters Ca and r, but also on the size of State Two-Phase Flow in Porous Media: Statistics and the influence /remediation volume over which process Transport Properties” Physical Review Letters 102, 074502 efficiency is evaluated. The effective energy utilization pp. 1-4. coefficient, fc, is drastically reduced with the influence [4] Valavanides, M.S. (2012) “Steady-State Two-Phase Flow in radius,  , of the flushed p.m. volume. Therefore, when a Porous Media: Review of Progress in the Development of the c DeProF Theory Bridging Pore- to Statistical Thermodynam- distribution of wells is considered, fc increases with the ics- Scales”, Oil & Gas Science and Technology, 67(5), number of wells per unit area, or, well density. The pre- pp.787-804 [http://dx.doi.org/10.2516/ogst/2012056] sented analysis shows that the efficiency of the process is [5] Valavanides, M.S. and Kamvyssas, G. (2013) “Operational more sensitive in the adjustment of the flow-rate ratio, r0, Efficiency Map of Steady-State Two-Phase Flow in Porous than in the adjustment of the reference capillary number Media Processes” InterPore2013, 5th International Confer- near the well, Ca . The proposed design methodology has ence on Porous Media Prague, 22-24 May 0 [http://users.teiath.gr/marval/publ/Valavanides_Kamvyssas_I a greater added value when processes implementing rela- nterPore2013.pdf] -6 tively low values for Ca0 (<610 ) are considered. When the formation material is weak and cannot sustain high pressure gradients across the well-bore, the engineer, having an upper limit for th Ca0 value to cope with, may tune the flow-rate ratio, r0, to a value that would optimize the overall operational efficiency of the process.

So long as only unit operational costs are examined, a “dense” well field should be sought for; nevertheless, the associated investment and managerial cost would increase with the number of wells (equipment, commission- ing/decomissioning costs, management costs etc.). There- fore, in order to perform a global technoeconomic evaluation of any proposed field design, one should carry out a break-even analysis considering the fixed costs associated with the number of wells and the accompanying equipment against the operational efficiency of each well.

ACKNOWLEDGEMENTS

This research work has been co-funded by the Euro- Received: November 05, 2013 pean Union (European Social Fund) and Greek national Revised: February 24, 2014 Accepted: April 02, 2014 resources under the framework of the “Archimedes III: Funding of Research Groups in TEI of Athens” project of the “Education & Lifelong Learning” Operational Pro- CORRESPONDING AUTHOR gram. The assistance of Dr. Alex Kalarakis with the compu- Marios S. Valavanides tations is acknowledged. Dept. of Civil, Survey & Geoinformatics Engineering TEI Athens The authors have declared no conflict of interest. 12210 Athens GREECE

Phone: +30 2105387205 E-mail: [email protected]

FEB/ Vol 23/ No 11a/ 2014 – pages 2847 – 2851

2851 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

EFFECTS OF LEAKAGE OF COMPOUNDS FROM RADIOACTIVE OILY WASTE ON SOIL MICROBIAL COMMUNITY

Svetlana Selivanovskaya*, Raushaniya Gumerova and Polina Galitskaya

Kazan Federal University, Kremleyvskaya street 18, Kazan, 420008, Russia

Presented at the 4th International Conference on Environmental Management, Engineering, Planning and Economics (CEMEPE), June 24 to 28, 2013, Mykonos, Greece

ABSTRACT 1. INTRODUCTION

The study deals with the effects of disposal of oily Oil production, transportation, processing and refin- waste containing natural radionuclides on chemical and ing leads to the generation of oily wastes, which consist biological properties of different soil layers. Two wastes of oily compounds, water and solids [1-4]. Oily com- were used in a laboratory experiment: raw oily waste H pounds of wastes pose a hazard to humans, plants and (TPH - 575±121 g kg-1, 226Ra - 4403±312, 232Th - 2848±211, animals [5,6]. Furthermore, solid phase of wastes can 40K - 1276±133 Bq kg-1) from a production yard, and contain TENORM (technologically enhanced naturally treated waste R obtained by eluting oily components from occurring radioactive materials); and this raises their level waste H. The wastes were disposed on soil columns (H- of environmental hazard [7-10]. and R-columns), at the amount equalized by the concentra- Despite a wide spectrum of methods for treatment of tion of radionuclides. C-columns without waste disposed oily wastes, these are often disposed on land, especially in were used as a control. After 30 days of irrigation, soil prop- developing countries [1,2,6]. Such disposal can lead to the erties of layers 0-20, 20-40 and 40-60cm were estimated. leaching of pollutants from waste into soil and groundwa- TPH content in all the layers of H-columns was signifi- ter. Column leaching methods have been reported to be cantly higher than in C- and R-columns. Activity concentra- realistic in simulating the leaching processes occurring in 226 232 the field. They allow estimating the impact of pollutants tions of Ra and Th in R0-20 samples were 3.5 times higher than in control samples. Soil microbial biomass de- disposed on the surface, on different soil layers. The use creased from the upper to the lower layer in all the col- of columns, and the leaching method to investigate the umns. Significant reduction of microbial biomass in the migration of metals and hydrocarbons into soil was previ- upper layer of column H was observed in comparison with ously described [11,12]. No data concerning the leaching -1 of toxic elements from oily waste containing TENORM columns C and R (12, 26 and 22 mg Cmic kg correspond- are available. ingly). Respiration activity in samples H0-20 was 2.4 and 1.5 times higher in comparison to C0-20 and R0-20 samples. Hy- Migration of toxic components into soil leads to the drocarbon oxidizing bacterial counts was significantly higher increase of their concentration and therefore to the change in all the samples of H-column, whereas there were no of soil properties. Microorganisms are one of the most im- differences in total bacteria counts. On the basis of cluster portant components of soil communities, because of their analysis of the data obtained, it was concluded that namely importance in cycling organogenic elements and regulating the oily compounds cause the alteration in microbial com- active nutrient pools in soils [13]. Changes in microbial munities, especially in the upper layer of soil. The effect of community cause loss of soil productivity. Among the bio- radionuclides on soil microbiota was not observed. logical methods reported to be sensitive to soil pollution, soil respiration and microbial biomass have to be mentioned [14,15]. To estimate the properties of soil contaminated by oil, total and hydrocarbon oxidizing bacterial KEYWORDS: oily waste, technologically enhanced naturally counts are often used [16]. Thus, these methods were used occurring radioactive materials (TENORM), microbial biomass, microbial respiration, soil column. to estimate the quality of soils polluted by hydrocarbons or to analyze the biodegradation of hydrocarbons [5, 13, 17, 18]. The aim of the study was to estimate the migration of TPH and radionuclides into soil due to rain precipitation * Corresponding author and to analyze their effects on soil microbial properties.

2852 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

TABLE 1 - Chemical properties of the waste samples

Activity concentration, Bq kg-1 Moisture Sample TPH, g kg-1 pH 226Ra 232Th 40K content, % H 575.2±121.0 4402.6±312.1 2848.1±211.3 1276.7±133.4 12.0±1.1 7.2±0.2 R 1.6±0.2 7861.7±516.2 2085.9±198.3 2277.9±205.4 9.5  1.3 7.1±0.2

2. MATERIALS AND METHODS 2.3 Microbiological analysis Soil microbial biomass (Cmic) was determined after 2.1 Experimental design the fumigation of the samples with C2H5OH-free CHCl3 Raw waste sample (H) was collected at Tikchonovskii and extraction with 0.5 M K2SO4. The soluble C content oil production yard (Tatarstan, Russia) from tanks during was determined by bichromate oxidation in accordance cleaning and maintenance routines. Treated waste sample with ISO 14240-2 [19]. (R) was obtained by multiple washing of sample H with Basal respiration rates (CO production with no sub- benzine and carbon tetrachloride to decrease the oil con- 2 strate addition) were determined with soil samples of 10 g tent. Characteristics of these two wastes are presented in adjusted to 60% water holding capacity and incubated at Table 1. 20°C. CO2 was trapped in 1M NaOH and measured by The soil (Haplic greyzem) columns were collected at titration with 0.1 M HCl after the addition of BaCl2 and the forest nursery “Matyushenski”, Tatarstan, Russia indicator solution [20]. The number of colony forming (latitude - 55°48'07"С, longitude - 49°16'13"E). heterothrophic bacteria (CFU) was determined by the serial dilution method as described in [21]. To enumerate hydro- Waste samples R and H were disposed of on the top carbon-oxidizing bacteria, the most probable number tech- of the columns (R- and H-columns, correspondingly). The nique was used [22]. Crude oil at 0.5% (v/v) was used as amounts of waste samples were equalized according to carbon source. radioactive elements content. Columns without any waste on the top were used as control (C-column). Each experi- 2.4 Statistical analysis ment was conducted in two replicates. During 30 days, Sampling and chemical analyses were carried out in rainfall was imitated. The amount of water was calculated triplicate, biological analyses in five replicates. All results according to the average atmospheric precipitation for the were expressed on an air-dry soil basis. The confidence of European part of Russia. After a month of incubation at data generated in the present investigations has been ana- 25oC, the soil of each column was divided into three parts lyzed by standard statistical methods to determine the mean (layers 0-20, 20-40 and 40-60 cm) and analyzed (samples values and standard errors (S.E.). The hierarchical cluster H , H , H , R etc.). 0-20 20-40 40-60 0-20 analysis method based on a Euclidean Distance Matrix

2.2 Chemical parameters between normalized vectors of values of parameters, was used. The Ward-method was implemented for clustering The total hydrocarbon content (TPH) of the waste and [23]. soil samples was determined by IR-spectrometry with an analyzer AN-2 (LLC “NEFTEHIMAVTOMATIKA-SPb”, Saint-Petersburg, Russia). The samples were analyzed for 3. RESULTS AND DISCUSSION natural radioactivity (226Ra, 232Th, 2 and 40K) by γ- spectroscopy. Each sample was dried for 24h in an oven at Characteristics of the samples from different levels of 110°C, homogenized and finally sieved through 0.8 mm the soil column taken after the 30 days of overhead irriga- meshes. The sieved samples were weighed and packed in tion are presented in Table 2. TPH content in the upper Marinelli-type beakers (1000 mL capacity) to be analyzed layer of C-columns was 0.40 g kg-1 and slightly decreased using a gamma spectrometer “Progress” (SPC “Doza”, -1 with the depth (0.36 and 0.24 g kg in C20-40 C40-60, corre- Zelenograd Moscow, Russia). Samples were carefully spondingly) (Table 2). No significant difference was shown sealed and stored for 4 weeks to reach secular equilibrium between TPH content in soil samples of R- and C-columns. 226 between Ra and its progenies. The gamma spectrometry The higher content of hydrocarbons was observed in H- measurements were carried out using the gamma-spectro- columns. Thus, TPH content in H0-20 samples was 3.5 times meter “Progress” based on an NaI(Tl) detector with the higher than in C0-20 samples. It decreased with the depth up resolution of 30 keV at 662 keV Cs-137 gamma line. to 0.53 g kg-1, but remained higher than in control sam- Soil pH-values and electrical conductivity were meas- ples. ured in an aqueous extract 1:5 (w/v). To determine the Activity concentration of 226Ra, 232Th and 40K in con- moisture content, 100 g of sample was dried at 600C oven trol column was measured to be 10.5, 21.5 and 294.5 Bq for overnight. Afterwards, the weight loss was measured kg-1, correspondingly. These values do not exceed their and calculated as the moisture content. natural levels in soil or sedimentary rocks [24-26]. In the

2853 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

TABLE 2 - Chemical properties of the soil samples

TPH, Activity concentration, Bq kg-1 Sample EC, µSm cm-1 pH g kg-1 226Ra 232Th 40K

С0-20 0.40±0.07 10.5±2.6 22.1±1.2 294.5±9.6 146.8±0.4 7.5±0.1

С20-40 0.36±0.02 17.4±2.9 21.0±0.8 335.1±6.3 63.2±8.0 7.3±0.3

С40-60 0.24±0.02 14.8±0.2 22.8±0.3 316.8±8.9 61.6±5.2 7.9±0.5

Н0-20 1.39±0.02 15.1±3.0 17.6±1.2 275.5±30.0 527.4±23.0 7.0±0.2

Н20-40 1.00±0.01 12.7±4.4 23.1±2.5 314.1±1.1 159.6±10.4 7.6±0.6

Н40-60 0.53±0.02 12.1±1.8 22.6±1.2 326.1±11.0 119.4±5.1 7.6±0.3

R0-20 0.39±0.03 64.7±4.2 39.1±2.5 269.0±26.1 221.0±15.8 7.4±0.0

R20-40 0.26±0.01 17.2±1.7 24.1±1.9 314.7±2.9 130.8±9.2 7.4±0.2

R40-60 0.19±0.01 14.3±0.3 21.7±2.2 302.7±3.6 112.2±1.8 7.1±0.1

upper layer of R-column, the concentration of radioactive from 146 µSm cm-1 in the upper level to 62 µSm cm-1in isotopes was higher than in control: 226Ra – 3.5 times the lower level of C-columns. The highest level of electri- 232 higher and Th – 1.6 times higher. In other layers no cal conductivity was revealed in sample H0-20. This is significant difference in comparison to control was ob- probably a result of leaching of chlorides and sulphates, served. The concentration of 40K did not exceed control contained in waste, into the upper soil layers. The pres- levels in none of the samples. Despite the fact that the ence of chlorides and sulphates in oily wastes was amount of waste samples disposed on soil columns was demonstrated by Mishra et al. [33]. Besides, radioactive equalized according to radioactive elements content, no elements can be presented in waste in the forms of sul- increase of concentration of radioactive isotopes in H- phates, carbonates and silicates [27]. columns was recorded. This is probably due to the bind- Biological characteristics of soils are presented in ing of these isotopes with oily components of the waste Figs. 2-5. The effects of migrated compounds on soil which are hardly water soluble and have less ability to microbial communities were monitored by soil respiration migration [27]. tests, microbial biomass and bacteria enumeration. As it is Disposal of waste samples did not change pH level, shown in Fig. 2, the level of microbial biomass in sample -1 which was measured to be 7.1-7.9. pH levels of soils C0-20 was 25.6 mg Cmic kg and it decreased in layers C20- polluted by oily wastes was reported to range between 7 40 and C40-60 by 1.5 and 4.6 times, correspondingly. Such and 8 [28-31], which is optimal for heterotrophic and effect connected with the reduction of organic matter hydrocarbon-degrading microorganisms [2,32]. Electrical content was reported by other authors [14, 34, 35]. conductivity reflecting the level of soil salinity varied

FIGURE 2 - Microbial biomass in the soil samples

2854 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

The same trend was revealed in soil samples of H- case of the increase in the volume of waste, microbial column. However, in sample Н0-20 the absolute value of community could be affected by radionuclides leaking microbial biomass was significantly lower than in control, from wastes. -1 it was estimated to be 12.1 ±2.0 mg Сmic kg . The respiration activity of samples from C-columns There were no significant differences in microbial bi- did not depend on depth and ranged between 1.0-1.2 mg -1 -1 omass between all the samples of column R and samples CO2-C kg h (Fig. 3). Respiration activity of soil micro- of control column. Based on the results it was assumed flora in H0-20 samples was 2.4 times higher than in C0-20 that the radioactive elements did not negatively effect on samples. Most often the increase of respiration activity microbial biomass despite of its higher content in column level in oily polluted soils was explained by microbial R in comparison with column C. This result is in accord- consumption of hydrocarbons as C-source [5,13,29,41- ance with the results presented in studies conducted by 44]. However, taking into account the reduced level of [36-38], which demonstrated that a negative effect on microbial biomass in sample H0-20, the increase of respira- microorganisms was revealed at doses 40-50 Gy. In our tion activity is more probably attributed to an overcoming experiment the weighted dose of 226Ra for bacteria in soil of stress by microbial community which leads to intensi- sample calculated by means of Spreadsheet-based model fication of metabolic processes. Such reaction of microor- R&D 128 (terrestrial model) from the England and Wales ganisms on stress was reported previously [42,45]. For Environment Agency [39,40] was equal to 0.002 Gy. H20-40 and H40-60 and all R-samples, no significant differ- However in case of waste storage for a longer period or in ence with control was found.

FIGURE 3 - Respiration activity of the soil samples

FIGURE 4 - Total bacterial counts in the soil samples

2855 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

FIGURE 5 - Hydrocarbon oxidizing bacterial counts

Height

0246810

(1)H20-40 C0-20 (2)C0-20 R0-20 (2) H0-20 (1)H0-20 (2)H0-20 H40-60 (1)H40-60 R20-40 (1)R20-40 C0-20 (1)C0-20 R20-40 (2)R20-40 C40-60 (2)C40-60 (2)R40-60

R0-20 (1) C20-40 (1)C20-40 H20-40 (2)H20-40 (2)H40-60 (2)C20-40 C40-60 (1)C40-60 (1)R40-60 FIGURE 6 - Dendrogram of hierarchical clustering

Total bacterial count in the upper layer of C-columns tween the same layers of different columns. To analyze these was estimated to be 4,5 logCFU g-1, in the middle and the results and to reveal the effect of the wastes disposed, the lower layers it decreased insignificantly. Migration of waste cluster analysis was implemented. Each soil sample was components into the soil did not effect the counts of het- considered as an object for the cluster analysis. The num- erotrophic bacteria, except for H0-20 and H20-40 samples ber of the objects was eighteen – three types of columns where the increase of the counts was revealed. (H, R and C), three layers, and two replications in each. The analysis of hydrocarbon oxidizing bacterial revealed Each object was characterized by a set of normalized a significant difference between its counts for control sam- values of biological parameters. ples and the samples of column H. The maximum increase As is presented in Fig. 6, the dendrogram of hierar- in the number of these bacteria was shown for samples chical clustering contains three main clusters. The first H0-20: its ratio was about 20% of total bacterial counts. cluster includes three objects – all from column H: two of This effect which is connected with the input of hydro- them are replications from the upper layer (H0-20(1) и H0- carbons was mentioned in studies conducted by [44,46]. 20(2)) and one from the middle layer (H20-40(1)). The second The decrease of hydrocarbon oxidizing bacterial count cluster includes three objects as well, two of them are correlated with the decrease of TPH (R=0.92). replications С0-20(1) and С0-20(2) and one C20-40(1). All other Results obtained show that the values of chemical and objects are consolidated in the third cluster which, in its biological parameters of soils vary both between the sam- turn, is subdivided into two groups, the first one includes ples taken from different layers of one column, and be- samples from the lower layers of H-column, the second

2856 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

one consolidates all the R-samples and C-samples from REFERENCES the middle and low layers. The result of clustering shows that untreated waste H influences significantly on biologi- [1] Zhang J, Dai JL, Chen HR, Du XM, Wang WX, et al. (2012) Pe- troleum contamination in groundwater/air and its effects on farm- cal parameters of the soil. This effect is more obvious in land soil in the outskirt of an industrial city in China. Journal of the upper layer, in middle and low layers it decreases and Geochemical Exploration 118: 19-29. becomes insignificant. The treated R waste without oily [2] Wang X, Wang QH, Wang SJ, Li FS, Guo GL (2012) Effect of components does not effect significantly on soil biologi- biostimulation on community level physiological profiles of mi- cal activity. croorganisms in field-scale biopiles composed of aged oil sludge. Bioresource Technology 111: 308-315. Yet, it should be stressed, that despite the fact that the [3] Galitskaya PY, Gumerova RK, Selivanovskaya SY (2014) Bio- negative effect of natural radionuclides on the soil was remediation of oil waste under field experiment. World Applied not revealed in this one-month experiment, it is not possi- Science Journal 30: 1689-1693. ble to exclude such effect in a long-term outlook. Larger [4] Gumerova RK, Galitskaya PY, Badrutdinov OR, Selivanovskaya amounts of wastes can cause negative effects in a shorter SY (2013) Changes of hydrocarbon and oil fractions contents in period of time. oily waste treated by different methods of bioremediation. Neftyanoe Khozyaistvo - Oil industry 9: 118-120. [5] Caravaca A (2003) Assessing changes in physical and biological 4. CONCLUSIONS properties in a soil contaminated by oil sludges under semiarid Mediterranean conditions. Geoderma 177: 53-61. In this study, the process of leakage of hazardous [6] Liu WX, Luo YM, Teng Y, Li ZG, Christie P (2009) Prepared bed bioremediation of oily sludge in an oilfield in northern China. compounds from oily waste was examined. Raw waste Journal of Hazardous Materials 161: 479-484. containing oily compounds and TENORM and waste, from which oily compounds were washed out, were dis- [7] Bakr WF (2010) Assessment of the radiological impact of oil re- fining industry. J Environ Radioact 101: 237-243. posed on the surface of soil columns. Precipitation was imitated during 30 days. It was shown that the disposal of [8] Abo-Elmagd M, Soliman HA, Salman Kh A, El-Masry NM (2010) Radiological hazards of TENORM in the wasted petrole- raw waste caused the increase of TPH content in the up- um pipes. J Environ Radioact 101: 51-54. per layer of soil by 3,5 times. Radioactive elements [9] Selivanovskaya SY, Gumerova RK, Galitskaya PY (2013) As- leaked from both types of wastes but their migration was sessing the efficiency of methods for the bioremediation of oil pro- higher from the treated waste with low hydrocarbon con- duction wastes. Contemporary Problems of Ecology 6: 542-548. tent. Disposal of both waste types increased the electrical [10] Selivanovskaya SY, Kuritsin IN, Akhmetzyanova LG, Galitskaya conductivity of the upper layer of soil. This effect was PY, Solovjev DA (2012) Use of biological actibity index for de- related to chlorides and sulfates presented in wastes and it termination of the oil polluted area meant for remediation. was more distinct in the H-columns. Neftyanoe Khozyaistvo - Oil industry 6: 102-103. According to the data obtained by means of cluster [11] Enell A, Reichenberg F, Warfvinge P, Ewald G (2004) A column method for determination of leaching of polycyclic aromatic hydro- analysis, it can be concluded that namely the oily com- carbons from aged contaminated soil. Chemosphere 54: 707-715. pounds cause an alteration in microbial communities. The [12] Li W, Yu LJ, He QF, Wu Y, Yuan DX, et al. (2005) Effects of effect consisted in the decrease of microbial biomass, the microbes and their carbonic anhydrase on Ca2+ and Mg2+ mi- increase of respiration activity and counts of hydrocarbon gration in column-built leached soil-limestone karst systems. Ap- degrading bacteria. This effect was more pronounced in plied Soil Ecology 29: 274-281. the upper layer of soil. During the experiment, the effect [13] Labud V, Garcia C, Hernandez T (2007) Effect of hydrocarbon of radionuclides on soil microbiota was not observed. pollution on the microbial properties of a sandy and a clay soil. However, according to our calculations, in case of waste Chemosphere 66: 1863-1871. storage for a longer period of time or in case of the waste [14] Babujia LC, Hungria M, Franchini JC, Brookes PC (2010) Mi- volume increases, microbial community may be affected crobial biomass and activity at various soil depths in a Brazilian oxisol after two decades of no-tillage and conventional tillage. by radionuclides leaking from wastes. Soil Biology & Biochemistry 42: 2174-2181.

[15] Bastida F, Moreno JL, Hernandez T, Garcia C (2006) Microbio- logical degradation index of soils in a semiarid climate. Soil Bi- ACKNOWLEDGEMENTS ology & Biochemistry 38: 3463-3473.

[16] Petrović O, Knežević P, Marković J, Rončević S (2004) Screen- The work is performed according to the Russian ing method for detection of hydrocarbon-oxidizing bacteria in Government Program of Competitive Growth of Kazan oil-contaminated water and soil specimens. Journal of Microbio- Federal University. logical Methods 74: 110-113. [17] Li JH, Zhang JT, Lu Y, Chen YQ, Dong SS, et al. (2012) Deter- The authors have declared no conflict of interest. mination of total petroleum hydrocarbons (TPH) in agricultural soils near a petrochemical complex in Guangzhou, China. Envi- ronmental Monitoring and Assessment 184: 281-287.

[18] Li H, Zhang Y, Irina K, Xu H, Mang CG (2007) Dynamic changes in microbial activity and community structure during biodegradation of petroleum compounds: A laboratory experiment. Journal of Environmental Sciences-China 19: 1003-1013.

2857 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

[19] ISO (2009) 14240-2: Soil quality - Determination of soil micro- [38] Buchan D, Moeskops B, Ameloot N, De Neve S, Sleutel S (2012) bial biomass - part 2: Fumigation-extraction method. Geneva, Selective sterilisation of undisturbed soil cores by gamma irradia- Switzerland: Intentational Organization for Standardization. tion: Effects on free-living nematodes, microbial community and nitrogen dynamics. Soil Biology & Biochemistry 47: 10-13. [20] Schinner F, Ohlinger R, Kandeler E, Margesin R, editors (1995) Methods in Soil Biology. Berlin: Heidelberg: Springer-Verlag. [39] Coppelstone DB, S.; Jones, S.R.; Patton, D.; Daniel, P.; Gize, I. (2001) Impact assessment of ionising radiation on wildlife. Bris- [21] Peng SW, Zhou QX, Cai Z, Zhang ZN (2009) Phytoremediation tol: R & D Publication. of petroleum contaminated soils by Mirabilis Jalapa L. in a greenhouse plot experiment. Journal of Hazardous Materials 168: [40] Jones S, Coppelstone D, Zinger-Gize I (2003) A methode of im- 1490-1496. pact assessment for ionising radiation on wildlife. Protection of the environment from ionizing radiation: the development and [22] Peressutti SR, Alvarez HM, Pucci OH (2003) Dynamics of hy- application of a system of radiation protection for the environ- drocarbon-degrading bacteriocenosis of an experimental oil pol- ment: Darwin (Australia): IAEA (Vienna), Austria. pp. 248-260. lution in Patagonian soil. International Biodeterioration & Bio- degradation 52: 21-30. [41] Balba MT, Al-Awadhi N, Al-Daher R (1998) Bioremediation of oil-contaminated soil: microbiological methods for feasibility as- [23] Ward JH (1963) Hierarchical grouping to optimize an objective sessment and field evaluation. Journal of Microbiological Meth- function. Journal of the American Statistical Association 58: 236- ods 32: 155-164. 244. [42] Franco I, Contin M, Bragato G, De Nobili M (2004) Microbio- [24] Starkov VD, Migunov VI (2003) Radiation ecology (in Russian). logical resilience of soils contaminated with crude oil. Geoderma Tumen (Russia): FGU IPP "Tumen". 121: 17-30.

[25] Vera Tomé F, Blanco Rodrı́guez P, Lozano JC (2002) Distribu- [43] Bundy JG, Paton GI, Campbell CD (2004) Combined microbial tion and mobilization of U, Th and 226Ra in the plant–soil com- community level and single species biosensor responses to moni- partments of a mineralized uranium area in south-west Spain. tor recovery of oil polluted soil. Soil Biology & Biochemistry 36: Journal of Environmental Radioactivity 59: 41-60. 1149-1159. [26] Shawky S, Amer H, Nada AA, Abd el-Maksoud TM, Ibrahiem [44] Lee SH, Oh BI, Kim JG (2008) Effect of various amendments on NM (2001) Characteristics of NORM in the oil industry from heavy mineral oil bioremediation and soil microbial activity. Bio- Eastern and Western deserts of Egypt. Applied Radiation and Iso- resource Technology 99: 2578-2587. topes 55: 135-139. [45] Chander KB, P.C. (1991) Microbial biomass dynamics during the [27] Gazineu MHP, de Araujo AA, Brandao YB, Hazin CA, Godoy decomposition of glucose and maize in metal-contaminated and JMD (2005) Radioactivity concentration in liquid and solid phas- non-contaminated soils. Soil Biology and Biochemistry 23: 917- es of scale and sludge generated in the petroleum industry. Jour- 925. nal of Environmental Radioactivity 81: 47-54. [46] Tahhan RA, Abu-Ateih RY (2009) Biodegradation of petroleum [28] Ouyang W, Liu H, Murygina V, Yu YY, Xiu ZD, et al. (2005) industry oily-sludge using Jordanian oil refinery contaminated Comparison of bio-augmentation and composting for remediation soil. International Biodeterioration & Biodegradation 63: 1054- of oily sludge: A field-scale study in China. Process Biochemis- 1060. try 40: 3763-3768. [29] Marin JA, Hernandez T, Garcia C (2005) Bioremediation of oil refinery sludge by landfarming in semiarid conditions: Influence on soil microbial activity. Environmental Research 98: 185-195. [30] Tang JC, Lu XQ, Sun Q, Zhu WY (2012) Aging effect of petroleum hydrocarbons in soil under different attenuation conditions. Agriculture Ecosystems & Environment 149: 109-117. [31] Margesin R, Zimmerbauer A, Schinner F (2000) Monitoring of bioremediation by soil biological activities. Chemosphere 40: 339-346.

[32] Ayotamuno MJ, Okparanma RN, Nweneka EK, Ogaji SOT, Probert SD (2007) Bio-remediation of a sludge containing hydro-

carbons. Applied Energy 84: 936-943. [33] Mishra S, Jyot J, Kuhad RC, Lal B (2001) Evaluation of inocu- lum addition to stimulate in situ bioremediation of oily-sludge- Received: June 05, 2014 contaminated soil. Applied and Environmental Microbiology 67: Accepted: June 18, 2014 1675-1681. [34] Ros MP, J.A.; Garcia, C.; Hernandez, M.T.; Insam, H. (2006) Hydrolase activities, microbial biomass and bacterial community CORRESPONDING AUTHOR in a soil after long-term amendment with different composts. Soil Biology and Biochemistry 38: 3443-3452. Svetlana Selivanovskaya [35] Fall D, Diouf D, Zoubeirou AM, Bakhoum N, Faye A, et al. (2012) Effect of distance and depth on microbial biomass and Kazan Federal University mineral nitrogen content under Acacia senegal (L.) Willd. trees. Kremleyvskaya street 18 Journal of Environmental Management 95: S260-S264. Kazan, 420008 [36] Jones HE, West HM, Chamberlain PM, Parekh NR, Beresford RUSSIA NA, et al. (2004) Effects of gamma irradiation on Holcus lanatus (Yorkshire fog grass) and associated soil microorganisms. J Envi- ron Radioact 74: 57-71. Phone: +7 843 2337515 Fax: +7 843 2315417 [37] Yardin MR, Kennedy IR, Thies JE (2000) Development of high quality carrier materials for field delivery of key microorganisms E-mail: [email protected] used as bio-fertilisers and bio-pesticides. Radiation Physics and Chemistry 57: 565-568. FEB/ Vol 23/ No 11a/ 2014 – pages 2852 – 2858

2858 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

ANTIMICROBIAL PERSONAL CARE PRODUCTS EFFECTS IN THE TERRESTRIAL ENVIRONMENT

Jūratė Žaltauskaitė* and Diana Miškelytė

Department of Environmental Sciences, Vytautas Magnus University, Vileikos 8-223, LT-44404 Kaunas, Lithuania

Presented at the 4th International Conference on Environmental Management, Engineering, Planning and Economics (CEMEPE), June 24 to 28, 2013, Mykonos, Greece

ABSTRACT wastewater and sewage sludge application. The efficiency of TCS removal from the wastewater can be highly vari- Triclosan (TCS) is a broad-spectrum antimicrobial and able depending on the wastewater treatment system and antifungal agent and it is a common constituent of house- commonly the removal efficiency is higher than 90 % [2, hold and personal care products. TCS widespread use has 3]. Low levels of TCS are discharged with effluents to resulted in its introduction into environment and has raised water bodies and may pose a risk to aquatic biota. Be- concerns regarding possible adverse effects on terrestrial cause of its hydrophobic nature (log Kow = 4.8), signifi- and aquatic organisms. Earthworm Eisenia fetida was cho- cant amounts (30-50 %) of TCS during wastewater treat- sen as representative organism to evaluate the impact of ment sorb to sludge and biosolids [3, 4]. The range of TCS on soil organisms. Adult E. fetida were exposed to 10- TCS concentration in sewage sludge is wide ranging from 750 mg/kg of triclosan for 8 weeks. The impact on surviv- 20 to 55 mg kg-1 [5]. Sewage sludge and biosolids land al, growth and reproduction was evaluated. TCS had low application transfers TCS to soils and elevated levels of acute toxicity to adults, though subchronic exposure se- TCS in amended soils are observed (up to >800 µg kg-1) verely affected the survival of earthworms (8-week LC50 [6, 7]. Moreover, biodegradation of TCS is slower in 252.88±86.84 mg kg-1). None of the tested TCS concen- biosolids amended soils relative to non-amended soils [7]. trations evoked earthworm weight loss during the experi- Numerous studies have investigated TCS toxicity to ment, but the earthworms treated with TCS grew slower aquatic organisms. Orvos et al. [8] investigated TCS toxici- and this resulted in lower weight at the end of the experi- ty to several species of algae, duckweed (Lemna spp), ment. Reproduction was shown to be the most susceptible daphnids (Daphnia magna, Ceriodaphnia dubia) and sev- endpoint to TCS exposure. Cocoon production in the soils eral species of fish. Algae were shown to be the most sus- spiked with 500-750 mg kg-1 was reduced by 90% in ceptible organisms to TCS impact. Several studies have comparison with control group. also shown TSC to be toxic to terrestrial organisms: plants, invertebrates and vertebrates. It was reported that

TCS inhibits plant growth, soil respiration and it poses low toxicity to birds and mammals [9, 10]. KEYWORDS: earthworm, growth, reproduction; mortality, toxicity, triclosan. Since the usage of wastewater for irrigation and sewage sludge or biosolids application is increasing, the potential ecological risk to soil dwelling organisms is increasing due to TCS contamination. Earthworms are soil keystone species constituting the dominant biomass of the 1. INTRODUCTION soil fauna. Earthworms are important in terrestrial food chain, energy and nutrient cycling and play a significant Triclosan (TCS, 5-chloro-2-(2,4-dichlorophenoxy)- role in soil community structure and ecosystem functions. phenol) is a broad spectrum antimicrobial agent. It is com- Earthworm acute and chronic toxicity tests are used to monly used in medical, household and personal care prod- assess the potential hazards of various environmental ucts such as toothpastes, detergents, soaps, cosmetics, deo- pollutants to soil ecosystems and they are used in risk dorants, etc. Due to its widespread use, triclosan is found assessment procedures. Although researches about TCS worldwide in the environment: air, soil, surface and even toxicity to soil invertebrates are limited, there is no doubt ground waters, sediments, living organisms and humans that earthworms may be adversely affected by TCS [11]. as well [1]. The main route TCS enters the environment is The aim of the study was to determine the subchronic effects of the different total triclosan concentrations in the * Corresponding author soil on earthworm Eisenia fetida growth and survival.

2859 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

2. MATERIALS AND METHODS [13]. All the statistical analysis was carried out using Statis- tica and R version 2.15.2 (R Development Core Team, The toxicity test was carried out according to the modi- 2004) software. fied OECD guidelines for the testing of chemicals [12] Earthworms growth (measure as fresh weight) rate with earthworm Eisenia fetida. The effects on survival, during the study period was determined by linear regres- growth and reproduction were determined. sion and the slope of the curve (b) was used as a prediction The artificial soil was prepared with the following of growth rate (g week-1). Significance of difference be- composition (by dry weight): 70 % quartz sand (more than tween the linear regression slopes for different TCS con- 50 % of the particles between 0.05-0.2 mm), 20 % clay and centrations was assessed using Z-test [14]: 10 % Sphagnum peat. The constituents of artificial soil b  b were air dried, mixed thoroughly and weighted (500 g) Z  1 2 , (1) into plastic boxes. The soil pH was adjusted to 6.0 ± 2 2 KCl SEb1  SEb2 0.5 with powdered calcium carbonate (CaCO3). The dry soil was moistened with distilled water to obtain approxi- where b1 and b2 are the slopes of the linear regression mately half of the final required water content. TCS was for the different metal concentrations, SEb – standard purchased from Alfa Aesar GmbH & Co KG. Solutions of error of b coefficient. The differences between the slopes TCS (TCS was solved in acetone (0.01 ml L-1)) were were considered to be statistically significant (p<0.05) mixed with soil to obtain the final required water content when Z-test exceeded the critical value of 1.96. (45-50 % of the maximum water holding capacity) and TCS concentrations in soil. The same volume of distilled water was added to control. The concentrations of TCS in 3. RESULTS AND DISCUSSION soil were (in mg kg-1 of soil): 10, 100, 250, 500 and 750. Three replicates were used for each of the test concentra- No significant effect (ANOVA, p>0.05) of TCS on tion and control. Two controls were executed in parallel: the survival of earthworms was found during the first two only water added and a solvent (acetone) control. No statis- weeks of exposure. It indicates that TCS does not evoke tical significant differences were observed between posi- acute toxicity to the survival of the earthworms. Longer tive and negative controls and in further statistical analy- exposure to TCS (from the 3rd week) had significant ef- sis only acetone control was used. fect (F>3.02, p<0.05) on the survival of earthworms (Fig- ure 1A). In the 10-100 mg TCS kg-1 treatments, the mor- The test earthworms were taken from a breeding cul- tality of the earthworms was relatively low (17-23 %) and ture kept in the laboratory of the Vytautas Magnus Uni- did not differ significantly from the control group versity. All selected worms were adult and fully clitellate. (p>0.05). The mortality of earthworms treated with 250- The selected earthworms were acclimatized for 7 days. 750 mg kg-1 differed significantly from the control group Ten washed and weighed earthworms were added to each (Dunnett’s test, p<0.05) and was higher than 50%. container and after 3 hours they were checked to ensure that all the worms had burrowed into the soil. All contain- The mortality of earthworms significantly increased ers were covered to prevent water loss. with time at all treatment concentrations (r>0.86, p<0.05). The experiment was run for 8 weeks at 20 ºC under Especially sharp increase in mortality was observed after continuous illumination (600 lux). Mortality and growth 7 weeks of exposure. An LC50 value after 8 weeks of were measured on a weekly basis (every 7 days) by count- treatment was 251.88 ± 86.84 mg kg-1. Regression analy- ing and weighing the surviving earthworms in each con- sis revealed positive significant relationship between TCS tainer. The worms were considered alive if they were able concentration in the soil and mortality of earthworms after to respond to mechanical stimulus. Any cocoons and juve- 8 weeks of treatment (R2=0.45, p=0.002) (Figure 1B). niles produced during the experiment were collected by The results indicate that TCS do not evoke acute tox- sorting twice through the soil at the end of the exposure. To icity, but longer exposure to this compound leads to lethal ensure earthworms growth, the earthworms were weekly effects. Positive significant relationship between E. fetida supplied with approximately 5 g of oatmeal per container. mortality exposed to 250-750 mg kg-1 of TCS and time of The unconsumed food was removed prior to resupplying a exposure indicate that subchronic or chronic exposure new portion. may result in reduced survival of E. fetida. No lethal acute A one-way analysis of variance (ANOVA) was used toxicity of TCS (0-1026 mg kg-1) on Eisenia fetida during to assess the concentration effect on estimated endpoints 14-days of exposure was reported by Reiss et al. [10]. (survival, growth, reproduction). Significant differences Data of chronic or subchronic TCS toxicity to earthworms between control and contaminated samples were deter- are limited. Acute toxicity of TCS was only tested to mined by the Dunnett test and p<0.05 were considered to Eisenia andrei and 14-days LC50 was determined to be be significant. Significant differences between treatments 866 mg kg-1 and inhibition of reproduction was observed were determined by Student’s test and p<0.05 were con- during 56 days of exposure [15]. 28 days exposure to -1 sidered to be significant. Lethal concentration (LC50) values biosolids-borne TCS additions (0.05-1 mg kg ) did not were calculated using the logistic dose-response model affect earthworm survival [16]. Because of relative low

2860 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

120 A 100

% 60

Survival, 0 10 20 100 250 0 500 750 -20 12345678 Week

100 B

80 %

, 60 y t i l a t r o 40 M

0 01234567 TCS log concentration

FIGURE 1 - (A) The survival of earthworms (Eisenia fetida) exposed to triclosan (0-750 mg kg-1 of soil) in OECD artificial soil for 8 weeks and (B) the relationship between the mortality of earthworms (Eisenia fetida) and TCS concentrations (in logarithmic (ln) scale) in OECD artificial soil to which earthworms were exposed for 8 weeks

concentrations of TCS in biosolids or sewage sludge and evoke weight loss during the experiment, but the earth- biodegradation of TCS in the soil, we may presume that worms treated with TCS were significantly smaller than adverse effects of TCS mediated toxicity to soil-dwelling those in the control group. Analysis of variance (one-way organisms could be only due to long-term exposure. ANOVA) showed that both TCS concentrations and time Moreover, as TCS degradation products such as methyl- had significant effects on weight changes of E. fetida triclosan or dioxin by-products [17, 18] are more lipo- (p<0.05). The pattern of earthworms growth exposed to philic and persistent in the environment, the increase in low TCS concentrations was very similar to that in the toxicity with time may be related with these TCS degra- control group, however the earthworms exposed to TCS dation products. grew slower than those of control and resulted in the final E. fetida weight changes during the test period are lower fresh weight than that of control earthworms (Dun- shown in Figure 2. None of the tested TCS concentrations nett, p<0.01).

2861 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

320 0 10 300 100 t

h 250 g

i 280 500 e

w 750

l 260 a i t i n

i 240

f o 220 %

e 200 g n

a 180 h c

t 160 h g i

e 140 W 120

100

012345678 Week

FIGURE 2 - Changes in fresh weight of earthworms Eisenia fetida exposed to triclosan (TCS) (mg kg-1) in OECD artificial soil for 8 weeks

0.6

0.5

0.4

0.3

0.2 * Cocoons per earthworm Cocoons

0.1 * *

0.0 Control 10 100 250 500 750 TCS concentrati on, mg kg-1

FIGURE 3 - Cocoons production by earthworms Eisenia fetida during 8 weeks incubation in OECD artificial soil contaminated with TCS (10 – 750 mg kg-1). Asterisk * indicates significant differences from the control (Dunnett, p<0.05)

Earthworm’s weight growth rate (g week-1) was de- TABLE 1 - Growth rate of Eisenia fetida exposed to TCS in OECD termined by linear regression b coefficient value. Earth- artificial soil for 8 weeks worm’s growth rate decreased with TCS concentration in TCS concentration, mg kg-1 Growth rate, g week-1 the soil (Table 1). Earthworm’s growth rate was shown to 0 0.1478 be sensitive parameter of soil pollution by heavy metals 10 0.0923* and various organic compounds [19, 20]. 100 0.0992* 250 0.0833* The highest growth rate was characteristic for control 500 0.0682* -1 earthworms (0.1478 g week ). Comparison of growth 750 0.0983* rates (with Z-test) indicated significant differences be- Asterisk * indicates growth rate significantly different from the control tween the treatments with all TCS exposure concentra- (determined by Z-test)

2862 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

tions and control. The growth rate of earthworms exposed [4] Waria, M., O’Connor, G.A. and Toor, G.S. (2011) Biodegra- to 10-250 mg kg-1 and 750 mg kg-1 was very similar and dation of triclosan in biosolids-amended soils. Environmental Toxicology and Chemistry, 30, 2488-2496. no significant differences between the treatments were observed. The lowest growth rate was observed at the [5] Heidler, J. and Halden, R.U. (2007) Mass balance assessment -1 of triclosan removal during conventional sewage treatment. treatments of 500 mg kg . The results indicate that sub- Chemosphere, 66, 362-369. chronic exposure to TCS in soil may result in slower growth of earthworms and we may presume that chronic [6] Kinney, C.A., Furlong, E.T., Kolpin, D.W., Burkhardt, M.R., Zaugg, S.D., Werner, S.L., Bossio, J.P. and Benotti, M.J. earthworms exposure to TCS may evoke even more pro- (2008) Bioaccumulation of pharmaceuticals and other an- nounced effects. thropogenic waste indicators in earthworms from agricultural soil amended with biosolid or swine manure. Environmental The reproduction was estimated as the number of the Science and Technology, 42, 1863-1870. cocoons produced by one earthworm and was estimated [7] Cha, J. and Cupples, A.M. (2009) Detection of the antimi- only at the end of the study (Figure 3). crobials triclocarban and triclosan in agricultural soils following land application of municipal biosolids. Water Research, Triclosan had a significant effect on the cocoons pro- 43, 2522-2530. duction (F=7.74, p=0.004) and at the highest TCS concentrations (500-750 mg kg-1) the cocoon production was [8] Orvos, D.R., Versteeg, D.J., Inauen, J., Capdevielle, M., Rothenstein, A. and Cunningham, V. (2002) Aquatic toxicity reduced by 90%. A decrease in the cocoon production rate of triclosan. Environmental Toxicology and Chemistry, 21, along with TCS concentration in the soil was observed 1338-1349. (R2=0.59, p<0.01). In the control and in the treatment -1 [9] Liu, F., Ying, G.G., Yang, L.H. and Zhou, Q.X. (2009) Terres- with 10 mg kg , emerged juveniles were found at the end trial ecotoxicological effects of the antimicrobial agent triclo- of the exposure, respectively 0.1 juveniles per earthworm san. Ecotoxicology and Environmental Safety, 72, 86-92. and 0.076 juveniles per earthworm. The study shows that [10] Reiss, R., Mackay, N., Habig, C. and Grifﬁn, J. (2002) An reproduction was the most sensitive endpoint and it sug- ecological risk assessment for triclosan in lotic systems fol- gests that even TCS residues in the soil may adversely lowing discharge from wastewater treatment plants in the affect earthworm reproduction. United States. Environmental Toxicology and Chemistry, 21, 2483-2492. [11] Higgins, C.P., Paesani, Z.J., Abbott Chalew, T.E., Halden, 4. CONCLUSIONS R.U. and Hundal, L.S. (2011) Persistence of triclocarban and triclosan in soils after land application of biosolids and bioaccumulation in Eisenia foetida. Environmental Toxicology Results indicate that 10-750 mg kg-1 of TCS does not and Chemistry, 30, 556-563. pose acute toxicity to soil key species earthworm Eisenia [12] OECD guideline for the testing of chemicals 207. Earth- fetida. Mortality was shown to increase in time of expo- worm, acute toxicity tests. sure. Though impact on the growth was rather significant [13] Ritz, C. and Streibig, J.C. (2005) Bioassay analysis using R. and subchronic E. fetida exposure to TCS in soil signifi- Journal of Statistical Software, 12, 1-22. cantly reduced the growth rate of earthworms. It suggests [14] Clogg, C.C., Petkova, E. and Haritou, A. (1995) Statistical that earthworms under such TCS pollution level are at risk methods for comparing regression coefficients between mod- enough to cause impairment of growth. Reproduction was els. Americal Journal of Sociology, 100, 1261-1293. shown to be the most susceptible to TCS exposure. Cocoon -1 [15] Amorim, M.J.B., Oliveira, E., Soares, A.M.V.M. and Scott- production in the soils spiked with 500-750 mg kg was Fordsmand, J.J. (2010) Predicted no effect concentration reduced by 90%. It may be concluded that the main threat (PNEC) for triclosan to terrestrial species (invertebrates and to earthworms from TCS may arise only due to long-term plants). Environment International, 36, 338-343. exposure to TCS in the soil. [16] Pannu, M.W., O’Connor, G.A. and Toor, G.S. (2012) Toxici- ty and bioaccumulation of biosolids-borne triclosan in terres- The authors have declared no conflict of interest. trial organisms. Environmental Toxicology and Chemistry, 31, 646-653. [17] Lindström, A., Buerge, I.J., Poiger, T., Berqvist, P.-A., Mül- ler, M.D. and Buser, H.-R. (2002) Occurrence and environ- REFERENCES mental behavior of the bactericide triclosan and its methyl derivative in surface waters and in wastewater. Environmen- [1] Allmyr, M., Adolfson-Erici, M., McLachlan, M.S. and Sand- tal Science and Technology, 36, 2322-2329. borgh-Englund, G. (2006) Triclosan in plasma and milk from [18] Latch, D.E., Packer, J.L., Arnold, W.A. and McNeill, K. Swedish nursing mothers and their exposure via personal (2003) Photochemical conversion of triclosan to 2,8- care products. Science of the Total Environment, 372, 87-93. dichlorodibenzo-p-dioxin in aqueous solution. Journal of [2] McAvoy, D.C., Schatowitz, B., Jacob, M., Hauk, A. and Photochemistry and Photobiology A: Chemistry, 158, 63-66. Eckhoff, W.S. (2002) Measurement of Triclosan in [19] Van Gestel, C.A.M., Dirven-Van Breemen, E.M., Wastewater Treatment Systems. Environmental Toxicology Baerselman, R., Emans, H.J.B., Janssen, J.A.M., Postuma R. and Chemistry, 21, 1323-1329. and Van Vliet P.J.M. (1992) Comparison of sublethal and le- [3] Bester, K. (2003) Triclosan in a sewage treatment process – thal criteria for nine different chemicals in standardized tox- balances and monitoring data. Water Research, 37(16), 3891- icity tests using the earthworm Eisenia andrei. Ecotoxicology 3896. and Environmental Safety, 23, 206-220.

2863 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

[20] Spurgeon, D.J., Svendsen, C., Kille, P., Morgan, A.J. and Weeks, J.M. (2004) Responses of earthworms (Lumbricus rubellus) to copper and cadmium as determined by measurement of juvenile traits in a specifically designed test system. Ecotoxicology and Environmental Safety, 57, 54-64.

Received: June 05, 2014 Accepted: June 18, 2014

CORRESPONDING AUTHOR

Jurate Žaltauskaitė Department of Environmental Sciences Vytautas Magnus University Vileikos 8-223 44404 Kaunas LITHUANIA

Phone: +370 37327904 Fax: +370 37327916 E-mail: [email protected],

FEB/ Vol 23/ No 11a/ 2014 – pages 2859 - 2864

2864 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

NATURAL AND Fe(II)-INDUCED REDUCTION OF HEXAVALENT CHROMIUM IN SOIL

Alja Margon*, Massimiliano Valentini, Claudio Mondini, Tania Sinicco, Paolo Sequi and Liviana Leita

CRA-RPS Consiglio per la Ricerca e Sperimentazione in Agricoltura, Centro per lo Studio delle Relazioni tra Pianta e Suolo, Via Trieste 23, 34170 Gorizia, Italy

Presented at the 4th International Conference on Environmental Management, Engineering, Planning and Economics (CEMEPE), June 24 to 28, 2013, Mykonos, Greece

ABSTRACT Heavy metals can adversely affect soil quality and endanger the surrounding ecosystems. Several studies have In this study the influence of iron sulphate (FeSO4) on shown that short or long-term exposure to heavy metals the dynamics of Cr(VI) – Cr(III) reduction in Cr(VI) results in reduction of microbial diversity and activity, spiked soil was evaluated, together with the concomitant which play a fundamental role in soil quality [3]. Chromium assessment of the effect of the different treatments on soil is a heavy metal largely used in industrial processes, ulti- microbial biomass size and activity. The results indicate mately making its way into soil through industrial spills or that during the incubation period, the concentration of inadequate disposal of industrial wastes [4]. The trivalent Cr(VI) in soil decreased over time, thus evidencing a Cr(III) and the hexavalent Cr(VI) chemical forms found in natural attenuation capacity of soil towards chromate soil have contrasting toxicity, mobility, and availability in reduction. However, soil microbial biomass and respira- the soil-plant-food chain system. The highly toxic Cr(VI), tion rate were both negatively affected by the presence of even though being a strong oxidant, can remain in soil Cr(VI), and did not recover during the experiment. Soil solution or get weakly adsorbed onto solid surfaces, thus amendment with FeSO4 increased the Cr(VI) reduction displaying remarkable mobility and posing a considerable rate and mitigated its negative effect towards soil micro- environmental and health hazard, while the much less toxic bial biomass and activity. Cr(III) is mainly insoluble and thermodynamically stable

[5]. Regulatory authorities have therefore placed significant emphasis on remediation and rehabilitation of Cr(VI)- KEYWORDS: Hexavalent chromium; chromate reduction; iron contaminated ecosystems through reduction of soluble sulphate; soil microbial biomass. Cr(VI) to insoluble Cr(III). Besides the known catalytic influence of natural compounds involved in the attenuation of Cr(VI) in contaminated soil, there is also evidence 1. INTRODUCTION of the effectiveness of technologies that make use of added reductants. Zero-valent iron [6] and Fe(II) [7] have Soil plays a central role in the food chain, water pro- been shown to be successful reducing agents of Cr(VI) in tection and the exchange of gases, not to mention its im- mineral matrices. portance as a biological habitat, gene pool, and element of In the present experiment, we evaluated the use of the landscape and cultural heritage. Any loss in soil quali- FeSO4 as a Cr(VI) reductant in Cr(VI) spiked soil. FeSO4 ty (e.g. natural fertility) therefore has proportionate ef- was obtained from a spent pickling solution provided by fects on social and natural systems [1]. the local steel industry. Soil is a complex heterogeneous system characterised by interrelated physical, chemical, and biological processes which influence soil functions, such as filtering, buffer- 2. MATERIALS AND METHODS ing and transformation capacities towards contaminants. Soil natural attenuation concerns the reduction of the A composite sample of a meadow soil from Northern amount/ bioavailability of pollutants at contaminated sites Italy was collected at 5-30 cm depth. After removal of [2]. However, natural processes are usually time- living plant residue, the soil was sieved to < 3 mm and demanding, especially when soil quality has already been stored moist at 4 °C. The sampled soil was a sandy loam affected. Mollic Hapludalf (USDA) with 67% sand, 21% silt, 12% clay, pH (H2O) 6.85, 2.2% organic C, 0.21% total N and * Corresponding author 10.5 organic C / total N ratio.

2865 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

2.1 Soil treatments attenuate Cr(VI). This is consistent with previous findings Pre-conditioned soil was treated with a Cr(VI) (as reporting that reduction of Cr(VI) to Cr(III) in subalkaline K2CrO4) solution in order to achieve a concentration of soils occurs naturally and in a relatively short period of -1 10 µg Cr(VI) g soil and then split in two portions; the time [5]. However, addition of FeSO4 to the Cr(VI) first was left untreated, while the second was amended spiked soil significantly increased the initial Cr(VI) re- with iron sulphate (Fe(II) 95%; 5% Mn and Si impurities) duction rate (50% reduction after 4 h) and resulted in from spent pickling solution provided by the local steel almost complete Cr(VI) reduction in 7 days, thus signifi- industry. Since the stoichiometry of the reduction reaction cantly shortening the time necessary for reduction. of Cr(VI) with Fe(II) requests a 1:3 ratio in order to achieve successful Cr(VI) reduction [7], we decided to apply at 1:5 [Cr(VI)] to [Fe(II)] ratio to ensure full conversion.

2.2 Extraction chromium and determination of total Cr, Cr(VI) and Cr(III) Cr was extracted from soil following the method described by James and Bartlett [8] and analysed for total and Cr(VI) content. The concentration of total Cr in clear soil solutions was measured with a Perkin Elmer ICP-OES Optima 3000, while the Cr(VI) content was estimated by the diphenylcarbazide colorimetric assay [9]. After colour development, absorbance was measured at 540 nm using a Varian Cary 1E UV-Visible spectrophotometer.

2.3 Soil microbial biomass C Soil microbial biomass C (MBC) was determined according to the fumigation-extraction method [10]. The FIGURE 1 - Cr(VI) reduction dynamics in soil over time: only Cr(VI) ( ), Cr(VI) and FeSO4 ( ). K2SO4 extractable organic carbon (OC) was determined with an OC analyser (Shimadzu TOC-VCSN). MBC was 3.2 Microbial biomass C calculated from the equation MBC = K x EC, where EC = The toxic effect of Cr(VI) towards soil microorgan- [(OC extracted by K SO from fumigated samples) - (OC 2 4 isms has been widely reported in the literature [16-19]. As extracted by K SO from non-fumigated samples)] and K 2 4 a susceptible soil component, microbial biomass has been (conversion factor) = 2.22 (see also [11]). therefore widely used as an indicator of soil quality and

2.4 Soil respiration the degree of restoration in degraded and/or contaminated soils [20-22]. Aliquots of control and treated soils were incubated at 20 °C in plastic jars continuously aerated at a constant -1 flow rate (20 ml min ). CO2 evolution was measured every 4 h for 2 weeks by means of an automated system for gas sampling and measurement [12].

2.5 Total enzymatic activity in soil We assayed the activity of two different enzymes: arylsulfatase (E.C. 3.1.6.1) in acetate buffer 0.5 M pH 5.2 using p-nitro-phenyl sulphate (PNPS) as substrate [13] and β-glucosidase (E.C. 3.2.1.21) in modified universal buffer (MUB) at pH 6.0 buffer using p-nitrophenyl β-D- glucoside (PNG) as substrate [14]. The enzymatic activity was determined according to the rapid fumigation method according to Fornasier [15].

3. RESULTS AND DISCUSSION FIGURE 2 - Dynamics of soil microbial biomass size in Cr(VI) spiked soil samples: control sample ( ), Cr(VI) ( ), Cr(VI) and 3.1 Reduction of hexavalent chromium by iron sulphate FeSO4 ( ). As shown in Figure 1, reduction of Cr(VI) in otherwise untreated soil reached about 20% after 4 h and was Since our results showed FeSO4 to increase Cr(VI) almost complete at the end of the incubation period. This reduction rate, we foresaw that FeSO4 would eventually indicates that the used soil had the capacity to naturally also mitigate the overall negative effect of Cr(VI) towards

2866 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

soil biota and ameliorate environmental conditions in the after the first 4 h of incubation and reaching 72% of the contaminated soils. After one day of incubation, MBC in initial value at the end of the incubation period. These the Cr(VI) spiked soil decreased by about 20% with re- findings are in agreement with Kandeler et al. [25] who spect to the untreated soil and remained constant until the observed that even low concentrations of heavy metals end of the experiment (Figure 2). Addition of FeSO4 allevi- significantly affected arylsulfatase activity in soil. ated the negative effect, as MBC decreased only by 10%. However, even when FeSO4 was added, MBC did not recover during the duration of the experiment. These results are in agreement with the findings of Renella et al. [23] who observed that freshly added metals have an effect of direct toxicity on microbial biomass.

3.3 Soil respiration As shown in Figure 3, cumulative respiration at the end of the incubation period in the Cr(VI) treated soil and the soil treated with both Cr(VI) and FeSO4 was 37 and 21% lower than that in the control soil, respectively. Ad- dition of FeSO4 to Cr(VI) spiked soil therefore resulted in a decrease in the toxic effect of Cr(VI). These observations are in agreement with Landi et al. [24], who asserted that heavy metals may reduce soil respiration by forming complexes with the substrates or by killing the microorganisms, as the observed dynamics of soil CO2 evolution in the control and treated soils clearly indicated the toxic FIGURE 4 - Effect of Cr(VI) on soil arylsulphatase activity: control sample ( ), Cr(VI) spiked soil ( ), soil amended with both effect of Cr(VI) towards soil biota. Cr(VI) and FeSO4 ( ).

On the other hand, Cr(VI) had only a transitory adverse effect on total β-glucosidase activity (Figure 5) that reached its lowest point after 1 day of incubation, with a decrease of 41%. However, the enzyme’s activity soon began to recover and reached a level not significantly different to that observed in uncontaminated soil within 5 days. In the soil treated with both FeSO4 and Cr(VI) the overall negative impact of Cr(VI) was even lower with an initial decrease of 25%. These findings are in agreement with Renella et al. [26] who observed that presence of various trace elements, also in high concentration, did not appear to have any significant effect on soil β-glucosidase activity.

FIGURE 3 - Cumulative CO2 evolution in Cr(VI) spiked soil samples: control sample ( ), Cr(VI) ( ), Cr(VI) and FeSO4 ( ).

Just as a MBC, soil respiration therefore appears to be a highly susceptible to soil contamination and may be therefore also useful as an early indicator of soil pollution.

3.4 Soil enzymatic activity Figure 4 shows that Cr(VI) had a significant negative effect on total arylsulfatase activity. In the Cr(VI) treated soil the enzymatic activity decreased to only 16% of the original value already after 4 hours of incubation. Even though it soon began to recover, it comprised only 55% of that observed in the control soil at the end of the incubation period. In the soil treated with both FeSO and Cr(VI) 4 FIGURE 5 - Effect of Cr(VI) on soil β-glucosidase activity: control the overall negative impact of Cr(VI) was lower with the sample ( ), Cr(VI) spiked soil ( ), soil amended with both enzymatic activity decreasing to 30% of the initial value Cr(VI) and FeSO4 ( ).

2867 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

Arylsulfatase therefore appears to be a better indica- [10] Vance, E.D., Brookes, P.C. and Jenkinson, D.S. (1987) An tor of soil pollution than β-glucosidase. extraction method for measuring microbial biomass C. Soil Biology and Biochemistry 19, 703-707.

[11] Cayuela, M.L., Sinicco, T., Fornasier, F., Sanchez-Monedero, 4. CONCLUSIONS M.A. and Mondini, C. (2008) Carbon mineralization dynamics in soils amended with meat meals under laboratory conditions. Waste Management 28(4), 707-715. Our study provides relevant information for the implementation of in situ treatment of Cr(VI)-contaminated [12] Mondini, C., Sinicco, T., Cayuela, M.L. and Sanchez- soil. Results indicate that albeit Cr(VI) reduction occurs Monedero, M.A. (2010) A simple automated system for naturally, the process can be significantly enhanced with measuring soil respiration by gas chromatography. Talanta 81, 849-855. additional FeSO4 supplementation. Microbial biomass and respiration rate were both negatively affected by the pres- [13] Tabatabai, M.A. (1994) Soil enzymes. In: R.W. Weaver et al. ence of Cr(VI) and did not recover during the experiment. (Eds.), Methods of soil analysis. Part 2. Microbiological and However, these two parameters clearly indicate a positive biochemical properties, Soil Science Society of America, Madison, 775-833. effect of FeSO4 addition in ameliorating environmental conditions. On the other hand, the influence of Cr(VI) on [14] Eivazi, F. and Tabatabai, M.A. (1988) Glucosidases and ga- soil enzymatic activity varied from enzyme to enzyme. Our lactosidases in soils. Soil Biology and Biochemistry 20(5), 601-606. findings therefore suggest that addition of FeSO4 significantly increases Cr(VI) reduction and mitigates the nega- [15] Fornasier, F. (2002) Indirect approaches for assessing intra- tive effects on the microbial biomass size and activity. cellular arylsulfatase activity in soil. In: A. Violante, P.M. Huang, J.-M. Bollag, and L. Gianfreda (Eds.), Developments The authors have declared no conflict of interest. in soil science 28B, Elsevier Science BV, Amsterdam, 345- 351.

[16] Ross, D.S., Sjogren, R.E. and Bartlett, R.J. (1981) Behavior REFERENCES of chromium in soils: IV. Toxicity to microorganisms. Jour- nal of Environmental Quality 10(2), 145-148. [1] Blum, W.E.H., Barcelo, D., Büsing, J., Ertel, T., Imeson, A. [17] Wong, P.T. and Trevors, J.T. (1988) Chromium toxicity to and Vegter, J. (2004) Scientific basis for the management of algae and bacteria. In: Nriagu, J.O. and Nieboer, E. European soil resources – Research agenda. Verlag Guth- (Eds.),Chromium in the natural and human environments, mann-Peterson, Wien. Wiley, New York., 305-315. [2] USEPA (1999) Use of monitored natural attenuation at super- [18] Cervantes, C., Campos-Garca J., Devars S, Gutierrez-Corona, fund, RCRA corrective action and underground storage tank F., Loza-Tavera, H., Torres-Guzman, J.C. and Moreno- sites. OSWER Directive N. 92004-17P, Washington (DC) Sanchez, R. (2001) Interactions of chromium with microorgan- [3] McGrath, S.P., Zhao, F.J. and Lombi, E. (2001) Plant and isms and plants. FEMS Microbiology Reviews 25, 335-347. rhizosphere processes involved in phytoremediation of metal- contaminated soils. Plant and Soil 232, 207-214. [19] Leita, L., Margon, A., Sinicco, T. and Mondini, C. (2011) Glucose promotes the reduction of hexavalent chromium in [4] Fendorf, S.E. and Li, G. (1996) Kinetics of chromate reduc- soil. Geoderma 164, 122-127. tion by ferrous iron. Environmental Science and Technology 30(5), 1614-1617. [20] Brookes, P.C. (1995) The use of microbial parameters in monitoring soil pollution by heavy metals. Biology and Fer- [5] Palmer, C.D. and Puls, R.W. (1994) Natural attenuation of tility of Soils 19, 269-279. hexavalent chromium in ground water and soils (U.S. EPA/540/S-94/505). Environmental Protection Agency, Cin- [21] Giller, K.E., Witter, E. and McGrath, S.P. (1998) Toxicity of cinnati. heavy metals to microorganisms and microbial processes in agricultural soils: a review. Soil Biology and Biochemistry [6] Puls, R.W., Paul, C.J. and Powell, R.M. (1999) The applica- 30, 1389-1414. tion of in situ permeable reactive (zero-valent iron) barrier technology for the remediation of chromate-contaminated [22] Clemente, R., de la Fuente, C., Moral, R. and Bernal, M.P. groundwater: a field test. Applied Geochemistry 14, 989- (2007) Changes in microbial biomass parameters of a heavy 1000. metal-contaminated calcareous soil during a field remediation experiment. Journal of Environmental Quality 36, 1137- [7] Pettine, M., D’Ottone, L., Campanella, L., Millero, F.J. and 1144. Passino, R. (1998) The reduction of chromium (VI) by iron (II) in aqueous solutions. Geochimica et Cosmochimica Acta [23] Renella, G., Chaudri, A.M. and Brookes, P.C. (2002) Fresh 62(9), 1509-1519. additions of heavy metals do not model long-term effects on microbial biomass and activity. Soil Biology and Biochemis- [8] James, B.R. and Bartlett, R.J. (1983) Behavior of chromium try 34, 121-124. in soils. VII. Adsorption and reduction of hexavalent forms. Journal of Environmental Quality 12, 177-181. [24] Landi, L., Renella, G., Moreno, J.L., Falchini, L. and Nan- nipieri, P. (2000) Influence of cadmium on the metabolic [9] Bartlett, R.J. and James, B.R. (1996) Chromium. In: D.L. quotient, L-:D-glutamic acid respiration ratio and enzyme ac- Sparks (Ed.), Methods in soil analysis. Part 3. Chemical meth- tivity: microbial biomass ratio under laboratory conditions. ods, Soil Science Society of America, Madison, 682-701. Biology and Fertility of Soils 32, 8-16.

2868 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

[25] Kandeler, E., Kampichler, C. and Horak, O. (1996) Influence of heavy metals on the functional diversity of soil microbial communities. Biology and Fertility of Soils 23, 299-306.

[26] Renella, G., Mench, M., van der Leile, D., Pietramellara, G., Ascher, J., Ceccherini, M.T., Landi, L. and Nannipieri, P. (2004) Hydrolase activity, microbial biomass and community structure in long-term Cd-contaminated soils. Soil Biology and Biochemistry 3, 443-451.

Received: March 19, 2014 Accepted: April 02, 2014

CORRESPONDING AUTHOR

Alja Margon CRA-RPS Consiglio per la Ricerca e Sperimenta- zione in Agricoltura Centro per lo Studio delle Relazioni tra Pianta e Suo- lo, Via Trieste 23 34170 Gorizia ITALY

Phone: +39 0481522041 Fax: +39 0481520208 E-mail: [email protected]

FEB/ Vol 23/ No 11a/ 2014 – pages 2865 - 2869

2869 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

GREEN MARKET: A COMPARATIVE STUDY IN ROMANIA

Mirela Panainte1,*, Iuliana Caraman1, Valentin Nedeff1, Vassilis Inglezakis2,3, Christos Venetis3, Philipos Coutsikos3, Gabriel Lazar1 and Narcis Barsan1

1 "Vasile Alecsandri" University of Bacau, Romania, 600115, Bacau, Romania 2 Nazarbayev University, School of Engineering, Chemical Engineering Department, 53 Kabanbay Batyr Ave., 010000, Astana, Kazakhstan 3 SC European Focus Consulting SRL, Banatului 16, Bacau, Romania

Presented at the 4th International Conference on Environmental Management, Engineering, Planning and Economics (CEMEPE), June 24 to 28, 2013, Mykonos, Greece

ABSTRACT studies, which were run in the period of 2010-2012. Ac- cording to the methodology developed under the project, The green products market emerged as a response to 20 product categories were monitored. The study was per- the negative impact of production and consumption of formed in two stages, at the beginning of the project in conventional products on the environment. The need for April-May 2010 and the middle of the project during greener production and consumption eventually lead to April-May 2012. the need of the development of a green market, especially in Europe and US. Green products are made by using environmental-friendly processes and materials taking also KEYWORDS: into consideration energy efficiency. Ideally, a green prod- green products, eco label, market research, Romania uct follows environmental-friendly path throughout its life, from raw materials to disposal. Some examples are the use of recycled paper for packaging, phosphate-free detergents, containers with less plastic content, and utilization of the 1. INTRODUCTION equipment more efficient in terms of energy consumption. Naturally, the studies on identification and comparison of Natural resource conservation and environmental pro- green products in relation to conventional ones gained tection is a primary objective of humanity. The climate more interest in the last decade, and are important as they change impact and the implications of the widespread envi- offer environmental information to consumers as well as ronmental pollution drew the attention of the private and to producers on the benefits of green market. Further- public sector producers and of various global organizations more, the relevant legislative initiatives as the application to support environmental sustainability [1]. Eco-labeling of the eco labeling in EU is increasing the environmental programs attempt to influence consumers' decisions by es- awareness, establishing high standards for producers and tablishing sustainable consumption patterns to change the assisting the consumers to identify and use green products. production patterns [2-4]. The environmental labeling serves The green products market in Romania is at the beginning as a market tool designed to inform consumers on the envi- as the eco labeling was introduced only in 2000 and in ronmental benefits of the products. Environmental labeling 2011 the European eco-label catalog included only five has also become a useful tool for world governments to Romanian green products. To support the consumers on encourage good environmental practices and to determine identifying green products and boost the green marketing producers in identifying and establishing green products in Romania, a LIFE+ project was funded by the European market. By offering environmentally friendly products, Commission, LIFE 08/INF/RO/000507 "Promoting Green the companies can create a positive image towards the Products" (2010-2013). The main objective of the project consumers with positive impact on sales [2, 5]. is to implement an integrated awareness raising campaign for consumers and producers in order to promote the con- Eco-labels along with other types of labelling, such as sumption and production of green non-food products in ISO standards, statements, environmental information etc., Romania. One of the project activities aims to conduct allows consumers to select products and services which market analysis in order to identify green products availa- have low or an insignificant impact on the environment. ble in Romania and to analyze the several labels they Moreover, these label schemes can be used as instruments carry in conformity to international and EU standards. of information, depending on the recipient, as follows [6]: The present paper presents the results of these  For governments: providing tools that help manufacturers and consumers to encourage behavior change * Corresponding author towards sustainable development.

2870 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

TABLE 1 - Comparison of the main key performance indicators (KPIs) from 2001 to 2012 according to Eurobarometer [7].

KPIs 2001 2005 2011 2012 Number of Companies 83 250 887 ~1000 Number of Licenses 95 279 1357 1671 Number of Products No statistics No statistics 17935 17176 Number of People who have seen/heard No statistics 11% 37% No statis- or bought Eco-labelled products (in 2006) (in 2009) tics available

 For companies: providing the opportunity to promote vast list of product categories available on the Romanian and sell the products with high environmental perfor- market, 20 categories of products were chosen for moni- mance. toring. The monitored categories of products are: television  For consumers: influencing the decision of purchasing sets, washing machines, refrigerating machines, electric eco-labelled products, providing data about infor- lamps, clothes, caper, paints and varnishes, dish washing mation behind the product. detergents, soaps and shampoos, light bulbs, IT equipment, bed linen, tourist accommodation, footwear, universal EU Strategy for the period 2011 – 2015 regarding detergents and WC detergents, lubricants, vacuum ma- eco-labels aims to strengthen its position, market visibil- chines, soil drainage products, rigid floor materials and ity of Eco-label, confirming the label as a reference stand- heat pumps. ard in labelling of organic products [7]. The parameters for measuring the visibility of Eco-label, according to The methodology of choosing product categories for Eurobarometer surveys, are divided into three groups [7]: the project was based on Eco-label scheme requirements and the respective national legislation as well as on ISO  Companies: the number of companies that have prod- 14020, 14021, 14024 and 14025 standards. According to ucts awarded with the Eco-label or number of licenses the standards listed above, the labels can be classified into these companies hold. the following categories [7-13]:  Products: the number of products for which these companies have the Eco-label.  Type I: Eco-labels. The Eco-labels take into account  Awareness: number of people who have seen/heard or all aspects of the environmental impact of product bought eco-labelled products (data from Eurobarome- throughout its lifecycle, like water and energy con- ter surveys). sumption, emissions, recovery etc. These labels are awarded by national authorities following specified The paper presents the results of green products mar- strict criteria. The EU Eco-label (green flower), the ket survey during 2010 to 2012. The study was conducted German Blue Angel and Nordic Environmental Label by GREEN PRO project team led by "Vasile Alecsandri" are some characteristic examples. University of Bacau in representative supermarket and specialty stores of Romania. It should be noted that the  Type I-like: Labels similar to those of Type I label. GREEN PRO market research is not restricted to eco-label These special groups of labels are referring to a single but includes all relevant product labels. issue instead of the whole life cycle. Single-issue labels focus specifically on one negative environmental impact, therefore, cannot be considered as eco-labels 2. MATERIALS AND METHODS [14]. A single-issue label can be based on a pass/fail criterion, for example setting a maximum level of en- In order to create a real image on the dynamics of vis- ergy consumption for electric appliances (like the En- ibility of green products in Romania, the project planned ergy Star label), or guaranteeing a responsible man- three market surveys over consecutive years 2010–2013. agement of the world forests (like the Forest Steward- The comparative data for the first two market studies ship Council). Other single-issue labels assess the per- carried out in two stages, July-August 2010 (at the begin- formance of the product on a range, for example grad- ning of the project), and the project halfway (April-May ing its energy or water efficiency. 2012), are presented in the paper. The surveys were con-  Type II: Environmental claims. These labels are ducted in seven supermarkets belonging to the most rep- developed within companies and can take the form of resentative/widespread chains in Romania and four spe- statements, logos and advertising, referring to one of cialty stores. The survey objectives were defined taking the company's products. The labels belonging to this into account the following criteria: the size and position group do not share the usual characteristics of envi- of stores in the market and impact on buyers, the diversity ronmental labels, the main difference being that they of products the store offers and stores offering a product are not awarded by an independent authority. category under this study.  Recycling labels. These are very common labels and The categories of monitored products were estab- are essentially producers’ self-declarations. They refer lished after well-documented methodology [7]. From the to the recyclability of the product, or its package, or in

2871 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

the portion of the product, or packaging that comes 3. RESULTS AND DISCUSSION from recycled material. According to the data from the second survey, in  Type III: Environmental information labels. This 2012, 1460 products were analyzed in comparison to 662 category of labels refers to qualified information based products in 2010. The comparative results for green prod- on impact of product lifecycle. The environmental pa- ucts identified on the Romanian market in 2010 and 2012 rameters are defined by a qualified independent organ- are presented in Table 3. ization. The companies gather information about the

environment in form of a report, and the data from re- TABLE 3 - Green products identified in Romania in 2010 and 2012. ports are independently verified. Label type 2010 2012  Green-wash. This is an environmental claim which is Type I: Eco-label 17 54 unsubstantiated or irrelevant. There are products on Type I-like labels 235 303 the market, which are promoted as green. However, Type II: Environmental Claims 84 468 they are not green and pose as green, in order to attract Type III : Environmental Information 6 20 consumers, with low awareness and/or knowledge on Recycling label (exclusively for pack) 99 280 environmental issues. Environmental Management Systems 10 27 (ISO 14001 and EMAS)  Environmental Management Systems. Eco-Manage- No-label 195 291 ment and Audit Scheme (EMAS) refers to the manage- Green-wash 21 17 ment of an organization's environmental programs in a comprehensive, systematic, planned and documented way. EMAS and ISO 14001 are most relevant labels for this project, and they are both volunteer schemes.

The identification of products from selected product categories was performed by the GREEN PRO project team led by “Vasile Alecsandri” University of Bacau through visits to stores. The observation sheet was filled- in for each product found on the shelves. The survey includes information on Product (model, company) and Label type: Type I: Eco-label; A label similar to those of Type I: Environmental Labels; Type II: Environmental Requirements; Recycling label (exclusively for pack); Environmental Management Systems (ISO 14001 and EMAS); No-label and Green-wash.

The evaluation of each category of products was based on its environmental performance as indicated by the product’s label(s). The assessment score for each label found on the product has been assigned according to Ta- ble 2. Finally, an average score for each product category was calculated applying the following equation [8, 14]:

Total score Average score = (1) Number of products in the category

TABLE 2 - Scoring system for labels [8].

Label type Score Type I: Eco-label 20.00 Type I-like labels 7.50 Type II: Environmental Claims 5.00 (a) (b) Type III : Environmental Information 2.50

Recycling label (exclusively for packaging) 1.00 FIGURE 1 - Distribution of products identified by label type: (a) – Environmental Management Systems 1.00 2010, and (b) – 2012. (ISO 14001 and EMAS) No-label 0.00 As it can be seen from Fig. 1, there is an increase of Green-wash -5.00 eco-labelled products (Type I) from 3% to 4%, and a

2872 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

decrease of green-wash cases from 3% to 1%. Also, unla-  IT equipment. The results show a decrease of Type II beled products decreased from 29% to 20%. These data and Type I-like labels and an increase of no-label show a positive trend towards a greener market. Another products. This result seems to notify an overall de- conclusion is that there is a shift between Type I-like and crease of green products in this category. However, it Type II, while the sum of Type I-like and Type II prod- is very promising percentage of almost 11% of eco- ucts is almost constant (48%-51%). In 2010, eco-labeled labelled products. The average category score is de- (Type I) products from 7 product categories, out of 20 creased to 6.20 from 6.33 in 2010. analyzed, were identified, and in 2012, this class increased  Refrigerators. The result in this category is un- to 10 categories. changed as these products are entirely labeled with The highest scores are for light bulbs, floor rigid ma- Type I-like products. The average category score is terials and IT equipment (Fig. 2). stable at 7.5 (the same as in 2010).  Light bulbs. The results show that these products are It is evident that for almost all the products there is an entirely dominated by Type-I like labels in 2012 while increase of mean score (Fig. 2). Total average score for all in 2010 several other labels were found. Naturally, the the products is higher in 2012 compared with 2010, and it average category score is increased to 7.50 from 5.32 increased from 2.92 to 3.86, i.e. about 32%. The results in 2010. per product category are summarized as follows:  Electric lamps. The results show a considerable de-  TV sets. The results show a slight decrease of eco- crease of green-wash and no-label cases but an increase labelled products and, at the same time, a huge in- of recycling labels. The large increase is noticed for crease of Type I-like products. No label products per- Type II and Type I-like labels. Furthermore, Type III centage is greatly decreased, no green-wash products and, more significantly, Type I (eco-labeled) products are found on the market and Type II percentage is in- appeared on the market. Naturally, the average category creased. Naturally, the average category score is in- score is increased to 2.45 from 0.36 in 2010. creased to 7.44 from 3.16 in 2010.  Clothing. The results show a slight decrease of no-label  Washing machines. The results show that this catego- cases and Type III cases. At the same time, appearance ry is almost entirely dominated by Type-I like labels, of Type II labels has a very high percentage. It is clear which is practically unchanged in comparison to 2010 that this category is dominated by no-label cases. Natu- research. The average category score is increased to rally, the average category score is increased to 1.12 7.5 from 7.23 in 2010. from 0.078 in 2010.

FIGURE 2 - Distribution of average scores for each product category.

2873 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

 Paper. The results show a considerable decrease of Type I-like are decreased. The average category score green-wash and no-label cases, which seems to be is increased from 3.91 to 2.83 in 2010. substituted by the big increase of recycling labels and  Lubricants. These products appeared in much bigger the appearance of EMS labels. The Type II and Type I numbers in 2012 in comparison to 2010. Furthermore, like changes are slightly similar to that of the Type I it is clear that Type II is dominant in 2012, while in labels. The average category score decrease to 4.33 2010, the only products found were with no-label. The from 5.00 in 2010. average category score is increased to 3.90 from 0.00  Paints and varnishes. This is a category dominated in 2010. by Type II labels. The results show green-wash cases  Vacuum cleaners. This category is dominated by to be stable and no-label cases to be considerably de- recycling labels, which are almost stable in percentage creased. A considerable percentage is covered by in both 2012 and 2010 research rounds. A slight in- EMS cases, stable within the years investigated. Re- crease is noticed in Type II labels, and the appearance cycling labels appeared in a high percentage, Type III of Type III labels is promising. The average category labels decreased and Type II increased. Type I-like score is increased to 1.72 from 1.61 in 2010. and Type I labels slightly decreased. The average score category is almost stable at 3.89, in comparison  Soil fertilizers. In 2010 the dominating label is that of to 3.49 in 2010. recycling which, in 2012, is substituted by Type II labels. At the same time, green-wash is greatly de-  Dishwashing detergents. While this category in 2010 creased but no-label products in relatively high per- was dominated by no-label cases in 2012, recycling centage appeared on the market. The average category and Type II labels are dominating. Green-wash and score is increased to 3.15 from 2.83 in 2010. no-label cases are considerably lower and recycling labels much more while Type II labels appear on the  Rigid materials for floors. While this category was market with a very high percentage. This very positive dominated by no-label cases in 2010, the situation is picture is challenged by the apparent decrease of Type changed in 2012 with the considerable increase of I-like and Type I labels. The average category score Type II labels and the appearance of Type I-like and in 2012 is 2.70 compared to 1.90 in 2010. Type I labels. The average category score is increased from 2.04 to 1.04 in 2010.  Soaps, shampoos and conditioners. While this category was dominated by no-label cases in 2010, now is In the framework of the market research, a financial dominated by Type II labels. At the same time, green- and environmental analysis was conducted for 375 elec- wash and no-label cases considerably decreased as trical products [15]. The analysis demonstrated that: well as recycling labels. Type II labels exhibit much  The purchase price is higher for green (or greener) higher percentage in 2012 in comparison to 2010, but versions for six products, following the order: TVs eco-labelled products percentage is almost half. The (44%), washing machines (43%), dishwashers (26%), average category score increased to 2.75 from 1.10 in light bulbs (21%), refrigerators (8%) and laptop com- 2010. puters (0.2%).  Bed linen. While no-label cases are increased and  The purchase price is lower for green (or greener) recycling labels are decreased, there is an equally con- versions for two products, following the order: desk- siderable increase in Type II labels. Probably, recy- top computers (-22%), computer displays (-4%). cling labels are substituted or joined by Type II labels. The average category score is increased to 2.11 from  The purchase and use of green (or greener) versions 1.56 in 2010. results in financial benefit (from higher to lower benefit) for: light bulbs (-72%), desktop computers (-24%),  Tourist accommodation services. This category is refrigerators (-12%) and computer displays (-9%). dominated by no-label cases and only Type I label ap-  The purchase and use of green (or greener) versions pears. The average category score is decreased to 2.40 results in financial damage (from lower to higher from 3.15 in 2010. damage) for: laptop computers (0.5%), dishwashers  Footwear. The recycling label was dominating in (8%), washing machines (10%) and TV sets (15%).

2010 as well as Type II in 2012, while no-label cases  In terms of absolute CO2 emissions per year, the envi- are slightly increased. The average category score is ronmental damage order is (from higher to lower dam- increased to 2.66 from 1.00 in 2010. age): refrigerators, washing machines, dishwashers,  Universal detergents and cleaners for sanitary use. light bulbs, TV sets, desktop computers, computer dis- This category is dominated by Type II labels for both plays and laptop computers. For light bulbs, the as- research periods. Green-wash and no-label cases are sumption is that lighting for three rooms is applied. considerably decreased but recycling label and Type II  In comparison to the non-green counterparts, the abso- labels increased. While Type I labels are increased, lute environmental benefit is clear for all products,

2874 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

with the exception of laptop computers (31%), follow- [3] Lavallee, S. and Plouffe, S. (2004) The eco label and sustainable development. International Journal of life cycle assessment, 9, ing the order (from higher to lower benefit): light bulbs 349-354. (-79%), desktop computers (-57%), TVs (-45%), com- [4] Rex, E. and Baumann, H. (2007) Beyond eco labels: what green puter displays (-43%), refrigerators (-28%), washing marketing can learn from conventional marketing. Journal of machines (-15%) and dishwashers (-9%). Cleaner Production, 15, 567-576. [5] Hazel, A.N. (2009) The European Commission's sustainable con- According to the opinion of research studies conduct- sumption and production and sustainable industrial policy action ed in Romania, in the same period (2010-2012), 44% of plan. Journal of Cleaner Production, 17, 496-498. respondents answered that the eco-label has an important [6] Global Ecolabelling Network (2004) Information paper: Intro- duction to Ecolabelling. role in buying choices and 54% have identified green [7] European Commission (2013) EU Ecolabel Work Plan for 2011 - products on the Romanian market [16]. However, only 2015, Version 1.7. 42% of the citizens trust the producers’ statements about [8] Inglezakis, V.J., Zorpas, A., Nedeff, V., Lazar, G. and Panainte M. the environmental performance of their products and 51% (2010) Green products and eco-labelling: an EU-project in Roma- of the respondents think that better information can im- nia. Second International Symposium on Green Chemistry for En- prove identification and access to the green products. vironment and Health, September 26-29, Mykonos, Greece. These results, in combination to the market research re- [9] Lazar, G., Nedeff ,V., Panainte, M., Caraman, I., Mosnegutu, E., Muraru, C., Inglezakis, V.J., Zorpas, A., Venetis, C. and Cout- sults presented above, show that the evolution of the sikos, P. (2011) A market research for eco-labeled products in green market in Romania exhibits positive trends but Romania. Proceedings of the 12th International Conference On more effort is needed in order to be further developed and Environmental Science & Technology, 8 – 10 September, Rho- to reach the level of other countries with well-developed dos, Greece, 1046 – 1053. green market. [10] ISO 14020:2000 on Environmental labels and declarations - Gen- eral principles, International Standardization Organization.

[11] ISO 14021:1999 on Environmental labels and declarations - Self- declared environmental claims (Type II environmental labelling), 4. CONCLUSIONS International Standardization Organization. [12] ISO 14024:1999 on Environmental labels and declarations - Type The analysis of data from two market surveys reveals I environmental labeling - Principles and procedures, Internation- that between 2010 and 2012 an increase of 33% of eco- al Standardization Organization. labeled products occurred while the amount of green- [13] ISO 14025:2006 on Environmental labels and declarations - Type III environmental declarations - Principles and procedures, Inter- wash products significantly decreased to 67%. New eco- national Standardization Organization. labelled products appeared on the market (2010, products [14] Guide to environmental labels for procurement practitioners of from 7 categories were identified as eco-labeled and in the United Nations system, UNOPS, 2009. 2012 this number increased to 10). Furthermore, it is [15] Inglezakis, V.J., Venetis, V.C., Coutsikos, P., Lazar, G., Nedeff, encouraging that a significant increase occurs in the num- V., Cojoc, C., Paliuc B. and Mocanasu, M. (2012) Comparison of ber of products labeled with ISO 14000 or EMAS. The green and non-green products: Romanian market case study. GreenChem 2012 Symposium, 3-5 October, Skiathos, Greece. appearance of products labeled with ISO 14000 could be [16] GREEN PRO (2012) Action B1: Market Research, Part D: Con- explained by the manufacturer's environmental concerns. sumers survey study, “Promoting Green Products” project In overall, it can be concluded that the Romanian market (LIFE+08/INF/RO/000507), Bacau, Romania. becomes greener although, in some cases, the trends are not yet stabilized. Received: January 14, 2014 Revised: March 06, 2014 ACKNOWLEDGEMENTS Accepted: April 02, 2014

The “Promoting Green Products” project (LIFE+08/INF/RO/000507) co-financed by the European CORRESPONDING AUTHOR Commission by 49.83% through the LIFE+ program. Mirela Panainte The authors have declared no conflict of interest. "Vasile Alecsandri" University of Bacau Faculty of Engineering Department Environmental Engineering and Mechan- ical Engineering REFERENCES 600115 Bacau ROMANIA [1] Dendler, L. (2014) Sustainability Meta Labelling: an effective measure to facilitate more sustainable consumption and production?. Journal of Cleaner Production, 63, 74-83. Phone: +40234580170 Fax: +40234580170 [2] Utgikar, V.P., Chaudhary, N., Koeniger, A., Tabak, H.H., Haines, J.R. and Govind R. (2004) Toxicity of metals and metal mixtures: E-mail: [email protected] analysis of concentration and time dependence for zinc and copper. Water Research, 38, 3651-3658. FEB/ Vol 23/ No 11a/ 2014 – pages 2870 - 2875

2875 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

WASTE PREVENTION CAMPAIGN REGARDING THE WASTE FRAMEWORK DIRECTIVE

Antonis A. Zorpas1,*, Katia Lasaridi2, Costas Abeliotis2, Irene Voukkali3, Pantelitsa Loizia3, Anastasios Georgiou4, Christina Chroni2, Korina Phanou4, and Nantia Bikaki3

1 Cyprus Open University, Faculty of Pure and Applied Science, Environmental Conservation and Management, P.O.Box 12794, 2252, Latsia, Nicosia, Cyprus 2 Harokopio University, Department of Geography, Kallithea, Athens Greece 3 EnviTech Ltd, (Institute of Environmental Technology and Sustainable Development, Department of Research - Development) Paralimni, Cyprus 4 Municipality of Paralimni, Cyprus.

Presented at the 4th International Conference on Environmental Management, Engineering, Planning and Economics (CEMEPE), June 24 to 28, 2013, Mykonos, Greece

ABSTRACT distance to travel in achieving its full potential. Tradition- ally, stakeholders have trusted and relied on technological Waste prevention will be defined as the prevention of means to alleviate waste production. Although technologi- waste at source through avoidance, reduction and reuse, cal advancement has been useful in the efforts to conserve but excluding off site recycling. However, the Waste natural resources and to reduce environmental impacts, Framework Directive (WFD) and especially in Article 3, waste generation is still increasing and one of the reasons is clause 12-13 states that prevention means measures taken the weak and inadequate individuals’ engagement in waste before a substance, material or product has become waste. management policies [1]. Through a Large LIFE+ Project (Project LIFE10 ENV/GR/ The amounts of municipal solid waste have been in- 000622, that is co funded from EU) which focuses on creasing for years in many countries. In the European Un- Development and Demonstration of a Waste Prevention ion (EU27), a 15% increase between 1995 and 2008 was Support Tool for Local Authorities WASP Tool, Munici- observed with an annual increase of 1% in recent years [2, pality of Paralimni has developed a waste prevention aware- 3]. In the Organisation for Economic Co-operation and De- ness campaign and several waste minimization practices. velopment (OECD) countries, an increase of 18% has been Those activities have to do with the (a) home composting reported between 1995 and 2007 [4]. Many governmental (b) Promoting waste prevention practices like seminars, agencies and international organisations have established activities with students from primary school. Also, through waste related policies to reduce the environmental impacts the project, the Municipality has prepared and submitted of waste management, including reducing the amounts of the first study in order to receive an official Licence for the waste. Studies of the relationship between quantity of waste Management of Waste from the Environmental Authorities. generated and economic growth indicators have suggested These targets, promote the waste management strategic that a decoupling is needed in order to reduce the increas- plan that several insular communities have the ability to ing burden from waste management [5, 6]. Waste preven- implement. tion has the highest priority in waste policy in several

regions, including the European Union, as stated in Di- rective 2008/98/EC [7], but so far, absolute decoupling KEYWORDS: waste prevention, waste campaign, home com- between waste growth and economical growth has not posting, materials exchange been demonstrated in Europe, in absolute terms, as indi-

cated by Mazzanti and Zoboli [8,9]. In Europe, waste prevention has been a key part of 1. INTRODUCTION waste management policy. In 2008, it was integrated in the legislation waste framework directive 2008/98/EC [7] There is no doubt that waste prevention has major po- and stated that prevention is the first priority of waste tential benefits to an effective management of waste by management, being at the top of the waste hierarchy, with addressing the problem of landfill shortage and resource a requirement from European member states to produce savings. While this is true, the practice has a considerable waste prevention plans. In the UK, the Government has funded a large research program in waste prevention. It * Corresponding author includes a review of evidence analysing the behavioural

2876 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

opportunities and barriers in household waste prevention, behaviour of households and businesses and on new para- associated with the effectiveness of various policy digms in industrial processes and product design [19]. measures [10]. The impact of waste prevention campaigns [11], and methods to monitor and evaluate waste prevention through mass reduction and behavioural studies [12] 2. MATERIALS AND METHODS were also developed. Current decoupling indicators of waste prevention were reviewed [13]. This unprecedented gov- 2.1. Area Description ernment led research initiative has demonstrated the key The Municipality of Paralimni is based in the Eastern issues of waste prevention from a behavioural perspec- Region of Cyprus and according to the last inventory tive. It is, however, quite critical also to assess the envi- study which was carried out from the Cyprus Statistical ronmental significance of waste prevention as this can Services on November 2011 the permanent population is reinforce evidence for policy development. In the United 18601 people [20, 21]. However, as the Municipality States, the US EPA has undertaken a program to support consists of the main economical lung of the island due to local authorities and waste management organisations to the fact that in this area there are the largest hotel resorts quantitatively assess the consequences of waste preven- the permanent population increases during the tourist tion on global warming potential (GWP) reduction with period (April – October) to 50000/d. There is not any the Waste Reduction Model (WARM) but does not in- major water consuming industry in the project area, and clude other environmental impact categories [14]. Some according to the available development plants the situa- countries, such as Taiwan [15], South Australia [16] and tion will remain the same in the future. The main activities New Zealand [17] have adopted the target of ‘‘zero waste’’ of the municipality [20-23] are: 115 Hotels and apartments as a form of strategic waste prevention. Regrettably, zero are presented according to the Cyprus Tourist Organization waste is usually interpreted as zero waste to uncontrolled (2010) [20, 21], almost 6000 houses (of which 70% belong disposal or landfill, mostly includes recycling and generally to the more Urban Area (inland), 5 petrol stations and more excludes environmental assessment. than 12 cars cleaning services, approximately 20 machinist's Waste is a key environmental, social and economic is- craftsmanship, small industries like bakeries, confec- sue as well as a growing problem. Approximately, 3 billion tioneries, car wash, food suppliers, supermarkets, tonnes of waste - 100 million tonnes of it hazardous - are schools (5 primary, 2 secondary have the ability to guest thrown away in the European Union annually. This amounts almost 2000 students), clinical laboratories, 2 private to about 6 tonnes of solid waste for every European. The link clinics and 1 public hospital, football fields and athletic between economic growth and waste generation is begin- activities, chicken farms (approximately 30000 chicken/y), ning to weaken in the EU. The generation of municipal 2 big laundries, 1 concrete plant and some small industrial waste per capita has slightly decreased in the EU-27 since activities which do not produce liquid waste. 2000 (-4%), while the GDP increased by 33% between 2000 and 2013. Exceptionally, waste generation in new 2.2 Municipality Waste Data Member States has remained relatively stable by weight According to Zorpas et al. [24] the current waste since the 1990s. This may be due to a reduced incidence management plan consist of the collection of the waste of heavy mining and construction waste and increased (door by door) twice a week and their transmission in the lighter paper and packaging waste. Decoupling economic Koshis Municipal Waste Treatment Plant (KMWTP) growth from the environmental impacts associated with which is about 65 km from the Municipality of Paralimni waste generation is a key objective of the EU’s revised traveling to the Capital Town. The total amount of waste Waste Framework Directive [7, 18]. Stabilising waste according to 2011 data was 15093.08 t while at the end of generation is no longer enough, waste growth in the EU 2012 it was up to 16250.11 t. The recycle materials are must now reverse [18]. Waste prevention, the most effi- collected from Green Dot Cyprus (GDC) twice a month cient option in approaching this goal, can contribute to: (i) and the total amount for the year 2011 were 962615 kg reduction of environmental impacts induced by waste while for the year 2012 were 1032256. The total cost management, e.g. greenhouse gas emissions reductions, (gate fee) for the year 2011 was up to 1.47m € while for methane released from landfill sites in particular, as well the year 2012 was up 1.59m €. KMWTP charge 54.8 €/t as carbon dioxide from incineration; (ii) improved resource for the mix waste, 46.8 €/t for the green waste, 80.80 €/t efficiency through energy savings and reduced material for the recyclable waste and 100.80 €/t for the residual use, as well as the “hidden” environmental impacts of re- waste [24]. source extraction, manufacturing and distribution; (iii) reductions in the production of hazardous waste and therefore 2.3 Waste Compositional Analysis improved conditions for public health. Until the end of 2011 there was not any composition- A broad objective, waste prevention affects and deal analysis of the municipality. The waste characterization pends on a very wide range of stakeholders. It benefits study was carried out for a period of summer (2 weeks from national targets and local authority engagement, but during August, 2012), autumn (2 weeks October and No- it depends fundamentally on changes in the attitudes and vember of the year 2012), winter (2 weeks during Decem-

2877 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

ber 2012 and January 2013) and spring 2013 (2 weeks 3. WASTE PREVENTION POLICY during March). A total amount of 80-100 plastic packs were collected sapling every day, door-by-door covering Waste prevention encompasses a range of policy op- all the main urban area of the Municipality, but not the tions and has a broad range of benefits. Targeting at- tourist area. It is very important to note, that, the tourist source waste production, it reduces the amount and toxici- area from the urban area has up to 10 Km distance. ty of waste before recycling, composting, energy recovery For the composition analysis, the household waste and landfilling become options. Waste prevention also was separated into 13 several main categories as presented includes measures to reduce the adverse impacts of the in Table 1. Those categories include Plastics, Metals and generated waste on the environment and human health. Drinking Cartons (PMD), Plastic film, plastics non- Waste prevention can be achieved by reducing the quanti- recyclable, aluminium packages, Papers, class, toilet ty of material used in the creation of products and increas- papers, Food Waste (category A), Food Waste (catego- ing the efficiency with which products, once created, are ry B), products that can be composted (mainly green used. Preventing waste by limiting unnecessary consump- and yard waste), stationery and others. The compositional tion and by designing and consuming products that gener- analysis was undertaken according to the standard EN ate less waste are forms of strict avoidance of waste. 14899:2005 [25]. Waste prevention also encompasses actions that can be

TABLE 1 - Waste compositional analysis and proposed targets to minimize each waste stream

Main Categories Sub categories Compositional Proposed Implemented plan to reach analysis % target % by the target the end of 2015

Plastics, Metals Plastic bottles/pots, metal packages, tetra pack 10 50 Public awareness campaign re- and Drinking (like milk, juices) garding the recycling program Cartons (PMD)

Plastic film (nylon) 5 -

Plastics Straws, yoghurts plastics, butter pots 2 - non-recyclable

Aluminium Aluminium papers, tins/cans 1 100 Public awareness campaign regarding the recycling program

Paper Package, Newspapers, magazines, offices, adver- 11 20 Public awareness campaign tised

Class Bottles, others 5 50 Public awareness campaign

Toilet-kitchen 12 - papers

Food Waste (A) Bakery’s, confectionery, dairy-farming, meat, 15 10 Development of Cooked book fish, cocked handbook using as raw materials left over

Food Waste Yogurt, wine, cooking oil, olives, eggs, banana, 5 25 Public awareness campaign to (B Whole, apples, pears, peaches, pomegranates, grapes, reuse those foods ready to eat) watermelons, oranges, passions fruits, mandarins,

potatoes, girasol, tomatoes, lemons, cucumber, carrots, onions, breads, pasta

Compost Vegetables, skin fruits, green waste, dust, soil 26 50 Public awareness campaign (products that can Home composting be composted)

Others Toys, textile, shoes, Pens, pencils medicines, 8 25 Exchanged materials syringe, spays, CDs, kitchen brush, lamps, poly- Public awareness campaign styrene, batteries, chandlery, stones, metals (spoons, knifes, pans, screws)

2878 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

undertaken once a product reaches its end-of-life: rather 4.1 How to promote waste prevention in households than discarding the product, the final user should consider Informational strategies are likely to form the basis of re-use, repair or refurbishment as options. Extending a waste prevention policy towards households. Products in product’s lifetime or considering options like reuse are household waste streams that can be effectively targeted forms of prevention though diversion of waste flows [18, include food, garden clippings, junk mail, nappies, batter- 19, 26]. ies, plastic bags, waste electrical and electronic equipment (WEEE) and textiles. These are constructively addressed Waste prevention plays a key role in sustainable through awareness campaigns accompanied by easy to waste management and is seen as a beneficial waste man- follow instructions for reducing and/or reusing each waste agement option which should be integrated to any waste product. Taking into consideration the reduction of haz- strategy. The Waste Strategy for England suggests that ardous materials that are transferred to the landfills, aware- local authorities and policy makers should now put more ness campaigns will be critical to implement prevention emphasis on waste prevention and re-use in order to activities focusing on households that encourage the sepa- achieve a reduction of disposal costs, CO2 emissions and a ration, sorting and applicable disposal practices of hazard- decrease in the demand for natural resources [18]. ous materials. Cleaning products, batteries, paints, pesticides, compact fluorescent light bulbs and WEEE, should There are environmental as well as socio-economic, fi- be an important part of household prevention programmes nancial and legal reasons for being concerned about waste as well as PMD (Plastics, Metals and Drinking Cartons) prevention [27] which are: (i) Environmental reasons: Mu- and green waste. Furthermore, the promotion of the sub- nicipal waste is merely the final evidence of consumption stitution of less hazardous products should be included which entails environmental footprints which are rarely here. The EU Ecolabel, for example, makes it easy to iden- sustainable; (ii) Socio-economic reasons: Production of tify products with reduce hazardous properties in a range of waste appears to be the result of wasting natural re- categories, including paints and cleaning products. sources and these resources are both limited and une- qually distributed; (iii) Financial reasons: Reducing the The waste minimization prevention campaign as pre- quantities of waste produced means it should be possible sented and proposed in Figure 1 includes among several to reduce the budget required for the collection, transpor- activities (a) home composting, Backyard and On-Site tation and treatment of waste products; (iv) Legal reasons: Composting. The composting and management of organic The European framework directive on waste requires na- materials where they are produced is known as “On-Site tional waste prevention programmes to be drawn up. Composting”. For homeowners, this is typically a small- scale backyard operation that may include leaves, grass, pruning, garden materials, and household fruit and vege- 4. PROPOSED WASTE table materials, (b) development of traditional recipe book PREVENTION CAMPAIGN using as raw materials, leftovers (uneaten edible remains of a meal after the meal is over, and everyone has finished A waste prevention programme should not be a mere eating); thus will reduce the food wastes that are gathers list of top-down-measures prescribed by the national gov- to landfills, (c) materials exchange. The goal of a material ernment but a plan which integrates the interests and con- exchange program is to connect potential users of dis- cerns of interested and affected parties (stakeholders) based carded materials with the businesses and individuals that on a commitment to carry on the programme. There are produce them, (d) Promoting waste prevention practices hundreds of potential or existing waste prevention like rent and repair, materials exchange, and the use of measures, instruments and initiatives which could be thrift shops and flea markets is part of a waste prevention integrated into the programme. The main task of the pro- education strategy. gramme therefore is to provide a vision and a framework that builds up on existing initiatives and to add the most effi- Table 1 presents the target regarding the waste mini- cient and effective complementary measures. mization prevention campaign. The targets were set out after the compositional analysis took place the last year in Drawing public awareness to waste prevention is a the study area (Paralimni Municipality). The composi- fundamental first step in stimulating behavioural change. tional analysis covers all the seasons through the analysis Effective awareness campaigns often focus on a specific of the waste during August (summer period), October- waste stream and offer practical, easy to follow guidance November (autumn period), December (winter period), on waste preventing actions. Waste preventing actions are and March (spring period) 2013. In reverse series the in fact much more environmentally beneficial, but often analysis consists of 26% products that can be composted, not as obvious. Using materials more efficiently and buy- 20% Food waste (both categories A and B), 12% Toilets ing less go relatively unseen. Campaigns may choose to and Kitchen Papers including sanitary, 11% papers, 10% focus on visible actions, such as the use of canvas bags PMDs, 8% others, 5% plastic films and glasses, 2% plas- instead of plastic bags, the use of ‘no junk mail’ stickers tics non-recyclable and 1% aluminium. Very, worrying is or the implementation of home composting [28]. the fact that 5% of the total compositional analysis con-

2879 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

FIGURE 1 - Implemented waste minimization prevention campaign

2880 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

TABLE 2 - Proposed actions to reach waste minimization

Area Action Waste Prevention The Local Authority i s gravely concerned about the year-on-year growth in waste and would need greater action from the Government to minimise waste.

The Local Authority will actively support Business Networks encouraging demonstrably effective waste prevention and minimisation amongst local business. Waste Avoidance The Local Authority will seek external funding or regional support to develop a packaging waste prevention campaign with local manufacturing companies.

The Local Authority will seek external funding to run waste prevention public awareness campaigns across the Municipality throughout the period of this strategy . The Local Authority will share good practice on waste prevention activities and will have regard to the effects on Waste Reduction waste arising when introducing new waste services. The Local Authority will consider the opportunities presented by offering incentives and rebates to residents for reducing waste and will review the need for direct and differential charging for waste during the implementation of this strategy.

The Local Authority will continue to support actively the development of best practice in waste reuse (and will encourage the development of community sector and other partnerships to deliver effective reuse services. Waste Reuse The Local Authority will continue to support bids for external funding of reuse services and will seek to develop a means of rewarding effective reuse services directly through a reuse credit, to reflect the avoided or deferred cost of disposal.

The Local Authority will provide a concerted and on-going promotional campaign to encourage home composting throughout the period of this strategy, offering residents purpose-built bins at subsidised rates and providing Home Composting support to residents wishing to compost at home. The Local Authority will aim to ensure that 15% of all the permanent residents with gardens compost at home by 2018 to divert approximately 4000 tonnes from the waste stream

sists of foods (of category B) that are safe to be used like one of the largest components of the waste stream from whole and high quality fruits, vacuumed and packed Paralimni Municipality households. To reduce the overall foods which have not been expired (like pasta, cans, rise, generation of household food waste and green waste, the frozen, cakes etc.). Using those foods as raw materials Waste Prevention Plan is focusing on promotion of food high quality recipes could be made. For example, several waste reduction messages and home composting of green fruits can be used to make a homemade marmalade. On waste. Implementation of «Best before of Good before», the other hand, several left-overs from the food waste of home composting, community presentations, competitions category A can be also used to generate safe food for the and numerous local events across the Authority area. The next day. For example, (i) with delicatessen, tomatoes, Authority would be helping residents identify how to save and cheese we can prepare a homemade pizza (ii) with the money as well as how to reduce food waste (and the asso- rest of the cooked chicken we can use it to prepare salad ciated environmental impacts); this would be a key mes- or sandwich, (iii) with the vegetables we can make soups sage to encourage take-up of the lower waste producing or fresh broth for other foods, etc. behaviour patterns.

Several actions must be taken into account in order to Furniture reuse is proposed as a second focal area due reach the proposed targets. Those actions (Table 2) are to its high tonnage diversion and increase in visibility and related to waste avoidance, waste reduction, waste reuse, consequently participation to complement the develop- and home composting which form the top half of the ment of the Reuse Network, a not-for-profit social enter- waste hierarchy. On local level several actions were orga- prise. The Plan will continue to be complemented by nized in order to reach the proposed targets set on Table 1 routinely undertaken, smaller scale initiatives which will which totally aim to reduce the waste sent to landfill by promote reduction of other waste streams such as unwant- 2018 up to 35% v/v. ed mail and textile reuse. It is proposed that the Waste Prevention Plan includes more action-oriented activities 4.2 Setting Priorities focusing on informational, educational and promotional When looking into the best options for the Municipal- initiatives that can be applied on a local level on the basis ity of Paralimni it is important to keep in mind that the that regulatory and legislative strategies are occurring on Municipality is a single purpose authority focused on a national level. It is anticipated that drawing attention to waste reduction and any decisions made by the Council waste prevention will result in a significantly raised pro- Members of the Authority must demonstrably focus on file for the Authority and its key messages, and thereby achieving that purpose. Food waste and green waste is still contribute to a reduction in total waste arising. All actions

2881 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

proposed will be supported by a cost benefit analysis and REFERENCES evaluation methodology. One of the most commonly used evaluation methods is by estimating the quantity of waste [1] Bortoleto, P.A., Kurisu, H.K., Hanaki, K. (2012) Model de- reduced at source. velopment for household waste prevention behaviour, Waste Management, 32, 2195–2207.

Estimating the quantity assumes that there is a con- [2] Eurostat, (2010) Municipal waste generated, 1000 tonnes, stant relation between private consumption and waste 1995-2008 (update 11/03/ 2010). Environmental Data Centre on Waste [WWW Document] URL generation although the complexity of factors determining http://epp.eurostat.ec.europa.eu/portal/page/portal/waste/docu waste generation is high. Progress in waste prevention can ments/Municipal_%20waste_%20generated_ be effectively tracked through “rates”. For the purpose of 1000t_%20update_%20website180112.mht. this plan, diversion rates were used where possible to [3] Gentil, C.E., Gallo, D., Christensen, H.T. (2011) Environ- indicate the portion of a waste stream which is managed mental evaluation of municipal waste prevention, Waste through waste prevention as opposed to disposal. Cost Management 31, 2371–2379. benefit analysis was also carried out in order to measure [4] OECD (2010) Total Amount Generated of Municipal Waste. further the benefits arising from the implementation of the In. OECD Factbook 2010: Economic, Environmental and So- waste prevention plan. The Cooked Green Kitchen hand- cial Statistics. Environment – Air and Land Municipal waste. OECD Publishing, Paris, France. book was used to assess the campaign benefits. However, it should be noted that even though some impacts can be [5] OECD (2002) OECD workshop on waste prevention: toward performance indicators (No. ENV/EPOC/WGWPR/SE measured with a satisfactory degree and accuracy, some (2002)1/ FINAL). Paris, France. others can only be estimated. There are a number of actions for which the impact in terms of tonnage diversion [6] Jacobsen, H., Kristoffersen, M. (2002) Case studies on waste minimisation practices in Europe. European Topic Centre on and/or cost benefit analysis could not be carried out be- Waste, Copenhagen, Denmark. cause they are either policy measures without direct tonnage implications or for some actions baseline data could [7] The European Parliament and the Council of the European Union, 2008. Directive 2008/98/EC of the European Parlia- not be established. However, these actions may result in ment and of the Council of 19 November 2008 on waste and changes in policies and practices, increase in knowledge repealing certain Directives. Official Journal 312, 3–30. of waste prevention or attitudinal change, all of which [8] Mazzanti, M., Zoboli, R. (2006) Waste prevention, waste will benefit Municipality residents in the long term. management and landfill policies effectiveness. European Topic Centre on Resource and Waste Management, Copen- hagen, Denmark. 5. CONCLUSION [9] Mazzanti, M., Zoboli, R. (2008) Waste generation, waste disposal and policy ffectiveness: evidence on decoupling Waste is a key environmental, social and economic from the European Union. Resources, Conservation and Re- issue and a growing problem. Waste prevention encom- cycling 52, 1221–1234. passes a range of policy options and has a broad range of [10] Cox, J., Giorgi, S., Sharp, V., Strange, K., Wilson, D.C., Bla- benefits. Targeting at-source waste production, it reduces key, N. (2010) Household waste prevention – A review of the amount and toxicity of waste before recycling, com- evidence. Waste Management Research 28, 193–219. posting, energy recovery and landfilling become options. [11] Sharp, V., Giorgi, S., Wilson, D.C. (2010) Delivery and im- Waste prevention also includes measures to reduce the pact of household waste prevention intervention campaigns (at the local level). Waste Management Research 28, 256– adverse impacts of the generated waste on the environ- 268. ment and human health. The waste minimization prevention campaign is in priority in order to achieve all the [12] Sharp, V., Giorgi, S., Wilson, D.C. (2010) Methods to monitor and evaluate household waste prevention. Waste Man- targets set out from the Municipality in order to reduce agement Research 28, 269–280. the waste production. Municipality’s policy is to reach the proposed targets by the end of 2015 and then to re- [13] Fell, D., Cox, J., Wilson, D.C. (2010) Future waste growth, modelling and decoupling. Waste Management & Research establish a new prevention target as well as to set out a 28, 281. complete zero waste approach. In general, the results could be helpful for improving policy actions to the bene- [14] USEPA (2006) Solid waste management and greenhouse gases: a lifecycle assessment of emissions and sinks. Wash- fits about recycling, prevention and reuse for other munic- ington DC, USA. ipalities also. [15] Young, C.Y., Ni, S.P., Fan, K.S. (2010). Working towards a zero waste environment in Taiwan. Waste Management & The authors have declared no conflict of interest. Research 28, 236.

[16] Zero Waste South Australia (2007) South Australia’s Waste Strategy 2005–2010. Benefit Cost Assessment. Volume 1: Summary Report. Melbourne, Australia [17] Ministry for the environment, 2007. Targets in the New Zea- land Waste Strategy 2006: Review of Progress. New Zealand Ministry for the Environment, Wellington, New Zealand.

2882 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

[18] Rappou D. (2012). North London Waste Prevention Plan, North London Waste Authority

[19] Zorpas, A., Lasaridi, K. (2013), Measuring Waste Preven- tion. Waste Management, 33, pp 1047-1056 [20] Zorpas, A.A., Coumi, C., Drtil, M., Voukalli, I. (2011). Mu- nicipal sewage sludge characteristics and waste water treatment plant effectiveness under warm climate conditions. De- salination and Water Treatment, 36, 1-15.

[21] Zorpas, A.A., Lasaridi, K., Voukkali, I., Loizia, P., Ingleza- kis, V.J. (2012). Solid waste from the hospitality industry in Cyprus. WIT Transactions on Ecology and the Environment, 166, 41-49.

[22] Zorpas, A.A, Voukkali, I., Loizia, P. (2013). Proposed treatment applicable scenario for the treatment of domestic sew-

age sludge which is produced from a sewage treatment plant

under warm climates conditions. Desalination and Water Treatment, 51 (13-15), 3081-3089.

[23] Zorpas, A.A. (2014). Chemical oxidation and MBR reactor

for the treatment of household heating wastewater. Desalina- tion and Water Treatment, In Press [24] Zorpas, A.A. Lasaridi, K., Voukkali, I., Chroni, C., Loizia,

P., Foitiri, L. (2012). Waste minimization and prevention activities in insular communities. The case studies of a Munici- . rd pality in Cyprus. 3 International Conference on Industrial and Hazardous Waste Management "CRETE 2012" Septem- ber 12th - 14th, 2012, Chania, Oral Presentation.

[25] EN 14899:2009 Waste Compositional Analysis Standard

[26] EU, (2012), Preparing a waste reduction programme; Guid-

ance Document, Directoral General Research [27] Defra (2007). Waste Strategy for England

[28] Guide Lines on Waste Prevention program, http://ec.europa.eu/environment/waste/prevention/pdf/Waste %20Prevention_Handbook.pdf, Access on March 2013)

Received: March 19, 2014 Accepted: April 02, 2014

CORRESPONDING AUTHOR

Antonis A. Zorpas Cyprus Open University Faculty of Pure and Applied Science Environmental Conservation and Management P.O.Box 12794 2252 Latsia, Nicosia CYPRUS

Phone: +357-22411935 E-mail: [email protected] [email protected]

FEB/ Vol 23/ No 11a/ 2014 – pages 2876 – 2883

2883 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

ASSESSING RECYCLING POTENTIAL IN LOCAL LEVEL: THE CASE OF NEAPOLI-SYKIES MUNICIPALITY, GREECE

Christos Karkanias*, George Perkoulidis, Nikolaos Grigoriadis, Stavros Stafylas, Emmanouil Dagdilelis, Eleni Feleki and Nicolas Moussiopoulos

Department of Mechanical Engineering, Laboratory of Heat Transfer and Environmental Engineering, Aristotle University of Thessaloniki, Box 483, 54124, Thessaloniki, Greece

Presented at the 4th International Conference on Environmental Management, Engineering, Planning and Economics (CEMEPE), June 24 to 28, 2013, Mykonos, Greece

ABSTRACT landfilling is the most used waste management method in Greece, thus methane production has reached high level. A The administrative division of Greece has been re- proposed way to tackle the aforementioned problem as formed through mergers with the Law 3852/2010, which several studies have indicated is to increase the recycling came into existence on 1st January 2011. The enforcement rate. More specifically, Knut [2] indicates that in the case of additional charges regarding the gate fees for waste dis- that the European Union achieves 50% of recycling the posal to sanitary landfills has led to a rethinking of adopting avoided emissions will meet the 247 million tones of CO2. rational waste management systems in order for the 325 The administrative system of Greece was overhauled new municipalities to avoid the rapidly increasing cost. in accordance with the Kallikratis plan from 1 January However, only a few Greek municipalities have fore- 2011. The former system of 1.033 municipalities and com- seen the need to change their waste management policy munities was replaced by a new one consisting of 325 mu- for achieving both cost saving and natural resources con- nicipalities. The municipality of Neapoli–Sykies, in North- servation. This study is focused on the municipality of ern Greece was formed by the merger of the former munic- Neapoli-Sykies, in the Greek region of central Macedonia, ipalities of Agios Pavlos, Neapoli, Pefka and Sykies. which is one of the first that has adopted rational ways to manage their waste. The municipality of Neapoli-Sykies has started adopting rational Solid Waste Management (SWM) policies in For the evaluation of the current situation regarding order not only to contribute in reducing methane produc- waste management in municipality, a Geographic Infor- tion but also conforming to the EU and national framework mation System Database was developed. First, the exact [3, 4]. In this notion, local authorities adopted a rational number of waste bins was identified with sticker numbers. policy to increase the recycling rate in its area. The daily Next, a data collection form was utilized, where the loca- production of Municipal Solid Waste (MSW) in the munic- tion, type, collection frequency, filling rate and other crite- ipality of Neapoli-Sykies is estimated at 85 kt. Packaging ria of the bins for mixed and recyclable waste was record- recycling is achieved in big (1,100 L) blue recycling bins of ed. The developed database could be utilized by decision Hellenic Recovery Recycling Corporation (HERRCO), makers as a simulation tool for decreasing the municipal who provide bins and vehicles for MSW collection. waste landfilling through the improvement of the current recycling system. The methodology for increasing the recycling rate is based on the implementation of a series of actions including the recording of waste bins located in municipality

KEYWORDS: area, identification of each waste bin (i.e. by using sticky GIS; recycling; waste bins; collection; decision making. numbers), data analysis regarding the waste bins charac-

teristics (location, type, filling rate) as well as optimiza-

1. INTRODUCTION tion of the waste bins (mixed and recyclable) location.

1.1 Review of the literature According to Air Pollution Information System, 32% of the total methane production in the European Union (EU) The enforcement of the new national waste frame- derived from sanitary landfills, while this rate reached 7.5% work based on the law 4042/2012 led the municipal au- in worldwide level [1]. It is widely known that sanitary thorities in Greece to rethink their strategies concerning the waste management issue. The additional charges on the gate fees for waste landfilling has made the imple- * Corresponding author mentation of rational waste management schemes the best

2884 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

solution to avoid the highly increased cost. According to small business, as well as with the new location and cal- Karkanias et al. [5], a large number of municipalities in culation of the number of containers needed in the streets Greece have already started adopting such rational systems for both commercial and non-commercial use due to the in order to reduce waste generation and thus the waste large amount on cardboard waste deposit in them in five management cost. shopping areas of the city of Leganes (Community of Furthermore, numerous selected districts participated Madrid, Spain). in a similar programme organised and implemented by the Furthermore, a case study of battery recycling in cen- government in England. In this case, multiple cities, towns tral Taiwan was presented by Lin et al. (2010) [15] to and rural communities developed innovative practices to demonstrate the effectiveness of a developed integrated confront waste management issue targeting to zero waste [6]. system with GIS. Finally, a similar case study was pre- On the other hand, several studies mentioned the fact sented by Perakis et al. (2004) [16] concerning ways to that the adoption of new policies and technologies by the confront the waste management issue. local authorities must be based on the local characteristics as well as on the current environmental and socio-economic aspects to meet the targets of the programme and promote 2. MATERIALS AND METHODS the notion of urban sustainability [7, 8]. The methodology that was utilized for the evaluation The indicators of urban sustainability which devel- of the current situation concerning waste management in oped in the frame of CAT-MED project include commu- Neapoli – Sykies municipality was based on the following nity sectors such as territorial management and urban aspects: design, mobility and transport, natural resources management and social and economic cohesion. The adoption Specification of form for waste bins/inventory. The of rational waste management schemes is a step for the fields of waste bins characteristics included identification transition from current industrialized cities to sustainable number, type, construction material, colour, capacity (L), and viable ones. existence of wheels, age, construction year, manufacturer, fullness (full, empty), status and odors emission. This paper used the Geographic Information System (GIS) to evaluate the current situation regarding waste Inventory of waste bins and waste collection routes. management in the municipality of Neapoli – Sykies. A The characteristics of each waste bin located in the mu- number of studies have also used this system for such pur- nicipality area were identified and recorded in printed pose. GIS was used for the reduction of operation costs and forms during October - December 2012. At the same pollutant emissions by definition of new collection sched- time, waste collection routes were given in printed form uling and optimization of MSW collection routes in Bar- by the municipality’s waste management and environment reiro, Portugal [9]. Moreover, Rada et al. (2013) [10] ana- department. lyzed a Web-GIS based system where 80% of waste was Database creation in GIS environment. Inventory da- source separated for recycling purposes. Using GIS, Clarke ta (of waste bins and collection routes) was set down in a and Maantay (2006) [11] developed an index for improv- developed GIS database. Attribute table of waste bins ing recycling education, awareness and participation in (points) was compromised by date fields such as the in- New York City. ventory days from October – November 2012, text fields In addition, Anghinolfi et al. [12] indicated that a GIS such as the number of sticker, address and name of former model was adopted to simulate the replenishment of the municipality where each waste bin was located, con- bins over the whole optimization horizon. Another exam- structed material of waste bin, year of construction, manu- ple comes from Tavares et al. [13], mentioning that MSW facturer name as well as colour numerical fields such as collection routes were optimised for minimum fuel con- the capacity of waste bin (L). sumption using 3D GIS modelling. Lopez et al. (2008) [14] Filling rate and waste collection frequency. Waste presented a methodology for the design of routes for the collection routes were first traced in Google Earth ® and “bin to bin” collection of paper and cardboard waste from then each route was traced in Arc GIS ®. Attribute table

TABLE 1 - 4 out of 16 rows of 1st waste collection route database, where the name and length of 4 parts and the number and identification of 1,100 L waste bins per each part are given.

ID* Name of part of route Length (m) Number of 1,100 L bins Sticker numbers of 1,100 L bins 1 Venizelou El. 1,210 66 4, 6, 9, 10, 11, 13, 16, 14, 17, 18, 20, 22, 23, 24, 25, 26, 27, 29, 31, 32, 33, 34, 37, 39, 36, 41, 43, 46, 44, 47, 49, 50, 413, 51, 52, 56, 54, 61, 63, 65, 64, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 91 2 Faistou 76 2 93, 94 3 Kekropos 106 4 95, 96, 97, 98 4 Neochoriou 143 6 99, 100, 101, 102, 103, 104 * ID: Identification number

2885 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

of waste collection routes (lines) was compromised by below show the location of mixed and recycling waste text fields such as the name of each part of route and the bins in the selected area. On the other hand, Figure 3 number of stickers of waste bins located in that part. Fur- shows the areas (A and B) that have very few recycling thermore, numerical fields such as the length (m) of each bins compared to the rest of the municipality’s area and part of route and the number and type of waste bins per need to be supplied with more in order to achieve higher part were included in each route’s attribute table. rates of recycling. Optimization of waste bins location. The use of GIS to locate the position of waste bins was based on conven- ient distance from households and maximum service coverage. The index KJ for calculating the number of bins every 50 m was utilized (j=1, …, 11 - total number of collection routes). Perspectives of recycling. Estimation of the total amount of the produced packaging waste and evaluation of the location of blue recycling bins located in the municipality area.

3. RESULTS AND DISCUSSION

3.1 Record of waste bins

The record of waste bins concerned the characteristics FIGURE 1 - Location of waste bins in the municipality of Neapoli- mentioned above while a unique sticky number corre- Sykies. sponded to each bin (for mixed and recyclable waste) to be easily identified. The recording showed that 2,965 and 910 of mixed and recycling waste bins respectively were located in the municipality of Neapoli-Sykies. The capacity of waste bins was 2,688.58 m3 or 4.12 kg/cap and 995.52 m3 or 0.88 kg/cap for mixed and recyclable materials respectively. Furthermore, the capacity of the majority of mixed and recycling waste bins were of 1,100 L while a small part of them were 660 and 330 L. Regarding the construction material the vast majority of mixed and recycling waste bins are made of plastic while only a small part of them are metallic. Moreover, almost half of the mixed and the total number of recycling waste bins use wheels. Re- garding the age of the bins almost half of the mixed waste bins are constructed between 2004 and 2010, while the FIGURE 2 - Location of recycling bins in the municipality of Nea- majority of blue bins constructed in 2010 and 2011. poli-Sykies. On the other hand, the recording showed that only a few mixed and recycling waste bins out of the total were Furthermore, decision makers can easily identify the in bad status meaning that it needs to be replaced by new destroyed bins and all their characteristics (location, type, ones. Besides, the municipality of Neapoli-Sykies had size, etc.) by using this database as shown in figure 4. purchased waste bins from multiple companies. This GIS database can be used as a simulation tool from decision makers. The areas A in Neapoli and B in Sykies According to the results of the recording, mixed and shown by GIS in Figure 3 were localities in the munici- recycling waste bins had low level of fullness as the vast pality without recycling bins in their area. majority of them seem to be half full. The above is due the frequent (every day) waste collection programme. The The waste collection service in Neapoli – Sykies mu- frequent waste collection also affected the odors that were nicipality consisted of eleven (11) collection routes in emitted from only a very small part of the total number of 2012. Each collection route per day was starting close to waste bins. the vehicle depot and was finishing at the same place with a fully loaded vehicle. Waste collection routes took into 3.2 Waste bins database account widths and congestion caused by parked cars or The data collected in the study area were used for the trucks due to the fact that drivers had their own reasons creation of database regarding waste bins of municipality for preferring the particular routes. The distance travelled of Neapoli-Sykies in GIS environment. Figures 1 and 2 for each waste collection load was recorded in GIS envi-

2886 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

FIGURE 3 - Areas A in Neapoli and B in Sykies without recycling bins in streets Karavaggeli, Mitropoleos, Nikitara, Aishylou, Larisis, Leonidou, Souliou, Korai, Zarifi and Thermopylon.

FIGURE 4 - Identification of destroyed waste bins in GIS environment in the area of Neapoli-Sykies municipality.

FIGURE 5 - Part of a waste collection route in GIS where identification number (ID), name, length (m) and number of recycling bins (RB) are given.

2887 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

FIGURE 6 - Calculated index K1 for waste collection route 1 in Neapoli-Sykies municipality.

ronment, along with specific data as the route’s identifica- In addition, local authorities will use this valuable in- tion number, the name of the route’s part, the length (m) formation to identify the need for more collection routes of this part, the quantum as well as the stick number of when possible in specific streets or areas or even re- blue and green waste bins (Figure 5). organize the collection system to both adequately cover In addition, convenience index KJ was utilized for cal- the whole municipal area and reduce the waste collection culating the number of the waste bins (1,100 L) per 50 m, cost. In this case, the main target of Neapoli-Sykies mu- where J=1-11 (11 waste collection routes). Waste collec- nicipality was the expansion of recycling through the tionroute 1 was characterized by 13 stops (1-4, 7-9, 11- placement of new bins for recyclable waste as well as the 16). Moreover, the total length of waste collection route 1 optimization of vehicle routes for the recyclable material was 7,13 km whereas the number of waste bins was 214. collection. Index K1 was greater than 1, while in 6 stops (5, 6, 10, 17-19) K1 was equal to zero. Combining the GIS database and values of the calculated K1, the number of green 4. CONCLUSIONS waste bins (1,100 L) could be changed and/or replaced by blue bins (1,100 L) (c.f. Table 1, Figure 6). The aim of this paper was to present a methodology for encouraging recycling in Neapoli-Sykies municipality. 3.3 Discussion There were four areas with 16 roads where the number of The database developed using the results from waste blue bins should be increased. The label of each bin al- bins survey can be used as a support tool by decision lowed the utilization of GIS for the digitization of four makers. In the case of Neapoli-Sykies municipality, local municipal units of Agios Pavlos, Neapoli, Pefka and Sykies. authorities obtained valuable data regarding the waste Decision makers from municipality can change the size of bins – for mixed and recyclable material – located in their waste bins or replace them with recycling bins, while area. This information accurately describe the current must provide recycling bins where needed. waste management situation in the municipality of Nea- The calculation of index KJ for the 11 collection poli-Sykies, while at the same time the vast majority of routes indicated that 11 parts of route 5, 9 parts in route 6, Greek municipal authorities have no record of their waste 7 parts in route 7 and 17 parts in route 8 could reduce the generation and collection system. number of green bins and to increase the number of blue In this sense, the referred municipality acquires all (recycling) bins in the former municipality of Sykies. In the needed data to design a new waste management sys- Pefka there were only 2 parts of routes 9 and 10 where tem that will be implemented based on the waste charac- green (mostly organic waste) bins could be reduced by teristics of each separate area in the municipality. An recycling bins. Finally, in Agios Pavlos, there were only example would be the removal of waste bins from areas 2 parts of route 11, where existed waste bins could be re- that data show low amounts of collected waste and trans- placed by recycling bins. The created database will be fer to areas where waste bins overflow problem is com- used as a tool for decision makers in order to decrease monly faced. municipal waste landfilling and achieve cost saving for the municipality of Neapoli-Sykies.

2888 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

ACKNOWLEDGEMENTS [12] Anghinolfi, D., Paolucci, M., Robba, M. and Taramasso, A. C. (2013) A dynamic optimization model for solid waste recycling, Waste Management (33), 287-296. This study was implemented in the frame of Urban Empa- thy project (MED Operational Programme 2007-2013), [13] Tavares, G., Zsigraiova, Z., Semiao, V., Garvalho, M.G. (2009) Optimisation of MSW collection routes for minimum co-financed by the European Regional Development Fund fuel consumption using 3D GIS modelling, Waste Manage- and coordinated by Malaga City Council. ment (29), 1176-1185. [14] Lopez, J.V., Aguilar, M., Fernandez-Carrion, S. and Jimenez del Valle, A. (2008) Optimising the collection of used paper from small business through GIS techniques: The Leganes ABBREVIATIONS case (Madrid, Spain), Waste Management (28), 282-293.

EU European Union [15] Lin, H.-Y., Chen, G.-H., Lee P.-H. and Lin, C.-H. (2010) An interactive optimization system fort he location of supple- GIS Geographic Information System mentary recycling depots, Resources, Conservation and Re- HERRCO Hellenic Recovery Recycling Corporation cycling (54), 615-622. MSW Municipal Solid Waste [16] Perakis, K., Malliaros, D., Soulakelis, N., Silleos, N. and SWM Solid Waste management Kungolos, A. (2004) Remote Sensing and Statistics for the investigation of uncontrolled landfill sites: a case study in The authors have declared no conflict of interest. Lesbos island, Greece, Fresenius Environmental Bulletin 5(13), 378-384.

REFERENCES [1] Air Pollution Information System, UK (2009) http://www.apis.ac.uk/overview/pollutants/overview_CH4.htm [2] Knut, S. (2008) Climate Protection Potentials of EU Recy- cling Targets, Ökopol GmbH, Germany. [3] Law 2939/01 on ‘Packaging and the Alternative Management of Packaging and Other Products’, 2001.

[4] European Parliament and Council Directive 94/62/EC on packaging and packaging waste, 1994. [5] Karkanias, C., Karagiannidis, A., Antonopoulos, I.S. and Samaras, P. (2012) Adopting rational waste management schemes: The case of Preveza municipality, Economics and Policy of Energy and the Environment (3), 65-79. [6] Phillip, P.S., Tudor, T., Bird, H. and Bates, M. (2011) A critical review of a key Waste Strategy Initiative in England: Ze- ro Waste Places Projects2008-2009, Resources, Conservation and Recycling (55), 335-343.

[7] Jha, A. K., Singh, S. K., Singh, G. P., Gupta, P. K. (2011) Sus- tainable municipal solid waste management in low income group of cities: a review, Tropical Ecology 1 (52), 123-131.

[8] Weigand, H., Fripan, J., Przybilla, I., Marb, C. (2003) Com- Received: June 05, 2014 position and contaminant loads of household waste in Bavria, Accepted: June 18, 2014 Germany: Investigation effects of settlement structure and waste management practice, Proc. of the 9th International Waste Management and Landfill Symposium, Cagliari. CORRESPONDING AUTHOR [9] Zsigraiova, Z., Semiao, V., Beijoco, F. (2013) Operation costs and pollutant emissions reduction by definition of new collection scheduling and optimization of MSW collection Christos Karkanias routes using GIS. The case study of Barreiro, Portugal, Waste Aristotle University of Thessaloniki Management (33), 793-806. Box 483 [10] Rada, E.C., Ragazzi, M., Fedrizzi, P. (2013) Web-GIS ori- 54124 Thessaloniki ented systems viability for municipal solid waste collection GREECE optimization in developed and transient economies, Waste Management (33), 785-792. Phone: +30 2310 994167 [11] Clarke M. J. and Maantay J. A., 2006, Optimizing recycling Fax: +30 2310 996012 in all of New York City’s neighborhoods: Using GIS to develop the REAP index for improved recycling education, E-mail: [email protected] awareness, and participation, Resources, Conservation and Recycling 46, 128-148. FEB/ Vol 23/ No 11a/ 2014 – pages 2884 - 2889

2889 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

AN ANALYSIS OF AN INNOVATIVE CONCEPT REGARDING EXCHANGING RECYCLABLE WITH AGRICULTURAL PRODUCTS

Konstantinos Aravossis and Christina Fountzoula*

Sector of Industrial Management and Operations Research, School of Mechanical Engineering, National Technical University of Athens, Athens, Greece

Presented at the 4th International Conference on Environmental Management, Engineering, Planning and Economics (CEMEPE), June 24 to 28, 2013, Mykonos, Greece

ABSTRACT specially created so waste can be put into the ground with little or no harm to the natural environment through pollu- Over the last few years various practices are promot- tion [2]. Another way for the waste disposal is by using ed by municipalities in order to reduce and combat the waste combustors. This involves the burning of municipal huge problem regarding the management of municipal waste at extremely high temperatures to reduce waste vol- waste, which very often ends-up in uncontrolled land- ume, control bacteria, and sometimes generate electricity. filling. Therefore, this article suggests the exchange of However, the solution to the problem does not lie so recyclable municipal waste with coupons correspond- much in the final stage of the disposal, but rather in reducing to vegetables, a method which firstly was successfully ing waste at the source, by means of prevention and recy- applied in Mexico and explores the possibility of applying cling. The recyclable municipal wastes are separated and this method in Greece. The possible applications of this recycled mainly by use of blue bins. This, as well as the project in the Greek municipalities, could bring a broader method of exchanging recyclable material for coupons economic development and an incentive to promote the corresponding to money, are the two most widely used production and marketing of agricultural products. Final- methods in Greek municipalities. The exchange of recy- ly, a comparative study is conducted, for the widespread clables with coupons corresponding to money, as well as recycling methods, as well as a cost-benefit analysis of recycling through recycle bins has played an important the specific method. part in this direction.

In this article it is suggested, in addition to the above KEYWORDS: exchange; recyclable products; agricultural prod- actions, the exchange of recyclables with vegetables through ucts; comparative study; cost-benefit analysis special coupons as a recycling method. This recycling method first made its appearance in Mexico City. The

main goal is to study this innovative method and its potential application in Greece, as well to compare it with 1. INTRODUCTION the methods mentioned above.

Municipal waste, commonly known as trash or garbage, is a combination of all of a city's solid and semisol- 2. ANALYSIS OF THE METHOD OF EXCHANG- id waste [1]. It includes mainly household or domestic ING RECYCLABLES WITH VEGETABLES waste, but it can also contain commercial and industrial waste with the exception of industrial hazardous waste As stated above, there are two basic methods of recy- (waste from industrial practices that causes a threat to cling that are being applied in a municipal level. Recy- human or environmental health) [2]. cling through recycle bins is the most known one, at least The last few years there is a remarkable effort being in Greece, but there are municipalities that have not applied made at municipal level to increase public awareness that yet. However, citizens are not aware even of this towards recycling. However, there are many different method and they are not properly applying it, as tightly tied ways in which cities dispose of their waste. Dumps and bags, dirty packages, wrong materials that cannot be recy- landfills are the most well-known. These are areas that are cled and very big materials end up in the recycle bins. Furthermore, recycling in exchange for coupons cor- * Corresponding author responding to money is a method in which the civilian is

2890 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

trading a quantity of recyclable goods for money that allo- on the market's opening day. The city estimates the mar- cates to the quantity and type of the recyclable. This aims ket brought in nearly 11 tons of recyclables in its first day. to raise public awareness and their active participation in But beyond the waste management, the project was a big recycling, given that they have adequate motive to imple- success for local farmers as well, because of the extra ment it. However, an innovative method, which includes income and benefits for the producers. Farmers that par- the exchanging of recycling goods with vegetables, is ap- ticipate in the recyclables trade market ensure with this plying in Mexico in order to motivate citizens about recy- particular procedure an extra income from the precinct of cling and reducing the volume of the Municipal waste. environmental management of the municipality, which is funded for this purpose from the cooperated recycling carriers [5]. 3. THE EXAMPLE OF MEXICO CITY However, the government sells the raw materials to manufacturers, but doesn't collect quite enough to cover Mexico City is converting rubbish to food. The envi- all the food costs. Therefore, the municipality subsequently ronmental agency recently launched Mercado de Trueque, has expanded the program to include used electronics and a trade market where the recyclable goods are traded for other types of recyclables. In addition to mitigating the fresh food in order to support parcels in the city [3]. city's waste problem, the market provides residents access This innovating program has been designed to show to an array of fresh fruits and vegetables [6]. citizens in an immediate and practical way how rubbish is converted to raw materials. If the solid wastes are separated properly, they can still be valuable. Market accepts 4. APPLICATION IN GREEK REALITY glass, paper and cardboard, soda cans made from aluminum, plastic bottles and returns “green coupons” that are The collection and sorting of recyclable goods is al- exchanged for agricultural products which are being cul- ready an organized activity and obligation of the munici- tivated inside and around Mexico city, including lettuce, palities in Greece, but is not a habit for many households. prickly pears, spinach, tomatoes, plants and flowers [3]. Showing a direct link between sorting and exchanging the Its aim is to encourage and support the existence of wastes and a sustainable food supply, a new approach for local production, in order to increase public awareness for recycling can be reached. The innovative recycling method the local offertory. It is important for local goods to be of Mexico, could help drastically in decreasing the waste consumed in order to avert big shipments of goods, to and increasing the agricultural productivity. This way, lower the carbon footprint, to strengthen the local trade garbage that end up in junk yards are obviously less re- and to maintain the agricultural areas south of the city. ducing environmental pollution, while the household’s weekly budget needed for purchasing fresh vegetables de- The trade mechanism is simple. In a big white tent, creases, which contributes to reducing the level of poverty. the employees of the environmental agency of Mexico City and volunteers in tables accept cardboards, paper, Thus, the proposed program can be implemented di- aluminum, plastic bottles, electronic appliances and other rectly in large or even provincial municipalities as fol- wastes from consumers, they are weighing the goods and lows: they give their customers receipts, which the citizens are The primary objective is to transform the local mar- then handing to special kiosks, from where they receive kets into “green spots”, via the customization of suitable the according coupons that correspond to the credits they places that will serve to the collection of recyclables from collected from the recyclables [4]. Using those coupons citizens. In this effort the primary role will be that of the they can buy tomatoes, potatoes, lettuce, lemons and other municipality’s employees, as the program’s supervisors, products of Mexico City. The quantities of the recyclables after having suitable training on the way of organizing are exchanged according to the credits that were collect- and managing this program. ed. For example, one kilo of plastic bottles equals 24 Afterwards, a specific day of the week will be chosen credits, one kilo of aluminum equals 16 credits and one for the citizens, in order to exchange the collected recy- kilo of cardboard or glass equals 3 credits each. The final clable goods with vegetables. Giving the recyclable goods, procedure is the encashment of the credits with vegeta- they will receive a special coupon that will match the cred- bles. In this way, for example, 20 credits can buy half a its they earned and after that they will be able to purchase kilo of tomatoes or potatoes. products from the local market. Nevertheless, recycling in the country of Mexico is The application of this specific program will be still in primary stage. Over 40 million tons of garbage is achieved through the agreement between the three con- produced each year in the country of 112 million people tracting parties: the United Association of Waste Man- from which only 15% is recycled. In Mexico City respec- agement (HERCO or Collective System of Rewarding Al- tively, 13.000 tons of solid waste is produced each day, ternative Packaging Management), the municipality and the from which only 12% is recycled [3]. traders - producers of the local market. The carrier will The results of this method were surprisingly optimis- fund the municipality according to the quantity of the recy- tic. More than 3,000 families lined up with bags of waste clable goods which was able to gather from the citizens

2891 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

FIGURE 1 - The flow rate of the co-operation between the carriers

FIGURE 2 - The flow of the recyclable goods

and then the municipality will fund the traders according Their essential difference lies in the fact that the proposed to the coupons they gathered from the citizens for the method makes the citizen unconsciously feel the obliga- purchase of vegetables. Below is the flow rate of the co- tion to recycle the same way he feels obliged to purchase operation between the carriers (Figure 1), as also the flow the necessary goods for his nutrition, because the pur- of the recyclable goods (Figure 2). chase of vegetables in the local market presupposes the recycling of the household’s wastes.

The problem with waste management and especially 5. COMPARISON TO THE OTHER with the municipal solid recyclables has many aspects in METHODS APPLICABLE AT THIS TIME: the application of waste management programs as well as ADVANTAGES-DISADVANTAGES-ANALYSIS in the consequences of its application. As far as the application of waste management programs is concerned the The proposed recycling method presents important problem is appearing with its political, social and tech- advantages in relation to the method of recycling through nical aspect [9]: recycle bins. The collection of recyclable goods from recycle bins is not required, as well as separation of materials Environmental aspect: Citizens with their given ac- and their afterward promotion for recycling is avoided [7]. tion can improve their city’s image, as well as cause great The separation is being held from the citizens them- problems to the health and the people’s quality of life. selves, who present the recyclable goods already separat- Single-handed or collective collection of waste is included in paper, glass, plastic. The gathering and transporting ed in these actions. The rational management of waste has cost is obviously less, as the gathering is being held in to be the goal of citizens and collective carriers specific areas – kiosks and it is not scattered in a large number of recycle bins in each municipality. Therefore, Economic aspect: Waste management is connected to this method has a smaller application cost [8]. a significant financial cost. Its objective must be saving Regarding the comparison of the proposed method in money through economizing as well as energy and mate- relation to the procedure of exchanging goods for money rial recovery. Α better quality of recyclable products con- there are some quite similar features and benefits revealed. tributes to the increase of revenue from their sale.

2892 © by PSP Volume 23 – No 11a. 2014 Fresenius Environmental Bulletin

TABLE 1 - Advantage-disadvantage analysis of the proposed method

Advantages Disadvantages 1. Reduction of the cost of managing waste. 1. Considerable time is required for the organization of this project.

2. Saving energy and natural resources, as the citizens carry 2. Increase of the cost of the municipality until the completion of this process. their own materials for recycling.

3. Increase in vegetable production. 3. Increased expenses for updates and promotion.

4. Promotion of trading. 4. Possible increase of operational and administrative costs.

5. Development of solidarity and increase in public awareness. 5. Obstacles in its application near local market areas.

6. Insurance of food to the economically challenged citizens. 6. Uncertain and possibly uncontrolled reaction of the citizens.

7. Increase in the efficiency and effectiveness of the cleaning 7. Political doubts or upheavals. service.

8. Increased collection of recyclables and increase of income through their sale.

9. Decrease in the negative effect the wastes have on the environment.

10. Greater efficiency of the waste management service and ensuing promotion of a more reliable image

Regarding the energy recovery, it is important to re- 6. CONCLUSIONS duce the energy consumed in various recycling processes, for example to reduce the use of garbage trucks. Finally, In summary, recycling of municipal waste is a basic the effective waste management strategies assist in mini- obligation of municipalities and the citizens while a low mizing or avoiding adverse impacts on the environment level of household awareness of waste management issues and human health, while allowing economic development occurs. This low level of awareness translates into a low and improvement in the quality of life level of participation in domestic waste management, such as recycling and disposal of domestic waste. This Social aspect: The environmental problems cause so- low level of participation is related to various factors, cial collisions. The increase in the number of citizens among which socio-economic and institutional factors. participating in this effort is essential. They are chal- An innovative method of exchanging recyclable ma- lenged in a daily basis with the problems caused by the terials with vegetables in Mexico, resulted in public poor waste management such as noise pollution, poor awareness as well as in reducing the volume of municipal aesthetics etc. waste of the country. The implementation of this method A relative comparison and assessment of the existing in other countries, such as Greece, could bring similar methods with the proposed method was made from where effects, given time, where the main goal of all municipali- it can be concluded that the proposed approach can be a ties in Greece is the reduction of their waste management viable alternative to the existing methods. Below is the cost and the strengthening of the local economy. advantage-disadvantage analysis of the proposed method In this article, is proposed the conversion of the local (Table 1.). markets into “green spots” through the exchange of recy- As can be seen from the table above, the proposed clables with vegetables. The implementation of this meth- method presents many advantages compared to disad- od could contribute effectively to the growth of agricul- vantages. Especially in Greece, the advantages presented tural production, the promotion of trade, the reduction of above are much important at the moment, wherein the non-recyclable waste. reduction of the municipal waste management cost, the The benefits of public awareness would be of great promotion of trading and the strengthening of the house- importance, as well as the fact that the application of such holds are great issues. a procedure could be the starting point of promoting other products or even activities through the exchange, contributing in an economic development in general.

The authors have declared no conflict of interest.

REFERENCES

[1] Lavee D. (2007). Is municipal solid waste recycling economically efficient? Environmental Management, 40:926-943 [2] Briney A. (2013). An overview of municipal waste and landfills: how cities deal with garbage, recycling, landfills, and dumps, Available at: http://geography.about.com/od/globalproblemsandissues/a/m unicipalwastelandfills.htm (accessed on:15/3/2014)

[3] Losada, Η, Cortes, Rivera, J. and Vargas, J. (2011) Recycling of solid wastes in Mexico City in livestock and agricultural production systems as a sustainable alternative. The Journal of Field Actions Vol 5 [4] McKenzie-Mohr Associates (2002). Development of a Social Marketing Strategy for Residential Waste Reduction for the City of Waltham. (Report prepared for the City of Waltham, Massachusetts). Waltham, MA [5] Tuckman J. (2013). Mexicans flock to recycle plastic bottles in exchange for food vouchers. The Guardian, Available at: http://www.theguardian.com/world/2013/jul/18/recycling- popular-mexicans-food-vouchers (17/3/2014)

[6] Godoy E. (2012). Swapping trash for fresh produce in Mexi- co City. Inter Press Service. Available at:

http://www.ipsnews.net/2012/08/swapping-trash-for-fresh- produce-in-mexico-city/ (17/3/2014

[7] Oskamp, S., Harrington, M., Edwards, T., Sherwood, D., Okuda, S. and Swanson, D. (1991). Factors influencing household recycling behavior. Environment and Behavior, 23 (4), 494-519 [8] DeYoung, R. (2000). Expanding and evaluating motives for environmentally responsible behavior. Journal of Social Is- sues 56 (3), 509-526 [9] Greene K. and Tonjes D., (2014). Quantitative assessments of municipal waste management systems: Using different indicators to compare and rank programs in New York State, Waste Management 34 (4), 825–836

Received: June 04, 2014 Accepted: June 18, 2014

CORRESPONDING AUTHOR

Christina Fountzoula

Sector of Industrial Management and Operations Re-

School of Mechanical Engineering

National Technical University of Athens

Athens

GREECE

E-mail: [email protected]

FEB/ Vol 23/ No 11a/ 2014 – pages 2890 – 2894

ADAPTATION OF PORT WASTE RECEPTION FACILITIES TO BALLAST WATER TREATMENT SYSTEM: TURKISH PORT PERSPECTIVE

Tanzer Satır1,* and Neslihan Doğan-Sağlamtimur2

1 Department of Maritime Transportation and Management Engineering, Maritime Faculty, Istanbul Technical University (ITU), 34940 Tuzla-Istanbul, Turkey 2 Department of Environmental Engineering, Engineering Faculty, Nigde University, 51245, Nigde, Turkey

Presented at the 4th International Conference on Environmental Management, Engineering, Planning and Economics (CEMEPE), June 24 to 28, 2013, Mykonos, Greece

ABSTRACT 1. INTRODUCTION

The introduction of invasive marine species into a Ballast is any material used to add weight to balance new environment by ballast water attached to ship hulls an object, such as vessel. It is the additional weight neces- has been identified as one of the four greatest threats to sary to bring the vessel to a suitable draft and trim, reduce the world’s oceans. Ballast water is water carried by ships stresses, and improve stability 1. Ships are specifically to ensure stability, trim, and structural integrity. Shipping designed and built to move safely through the water while transports over 80% of the world’s commodities, and each carrying cargo. When a ship is travelling either without year transfers approximately 3 to 5 billion tons of ballast cargo, or only partially laden, it must take additional weight water internationally, which is, absolutely essential to the on board to enable it to operate effectively and safely by safe and efficient operation of modern shipping, provid- keeping the ship deep enough in the water to ensure effi- ing balance and stability to unladen ships. However, it cient propeller and rudder operation. In the past, many may also pose a serious ecological threat to the marine solid materials were used for ballast, including sand, soil, environment. The transferred species including bacteria, and stones. In the mid 19th century, water was used as microbes, small invertebrates, eggs, cysts and larvae of ballast, and since the 1950’s, it has completely replaced various species may survive to establish a reproductive solid ballast in ships used to carry heavy loads. Ballast population in the host environment, becoming invasive, water systems are now an integral part of ship design, and out-competing native species, and multiplying into pest they contribute to stability and balance as well as struc- proportions. tural integrity of the hull. Ballast water is pumped into specially designed tanks distributed throughout the hull as Ballast water treatment is a technology for the treat- ships are offloaded, and pumped out again on arrival to a ment of ship ballast water from aquatic invasive species. port where cargo is to be loaded (Fig. 1) 1, 2. Port waste reception facility should be modified to in- It is estimated that 10 billion tons of ballast water is clude ship ballast water treatment during this transitional transferred globally each year, and that 7,000 invasive spe- phase until 2017. It provides onshore facilities in ports or cies are carried around in ballast water every day 2. There terminals to transfer ballast water for cleaning or storage. are thousands of marine species that may be carried in This paper describes ballast water management, in gen- ship ballast water; basically anything that is small enough eral, and gives perspectives for the Turkish ports, and to pass through a ship’s ballast water intake ports and briefly suggests that the waste reception facilities must be pumps. These include bacteria and other microbes, small modified for ballast water treatment. invertebrates, and the eggs, and cysts and larvae of vari-

ous species. The problem is compounded by the fact that

all marine species virtually have life cycles that include a

KEYWORDS: Port waste reception facility, ballast water, aquatic planktonic stage or stages. invasive species, ship-generated waste

2. AQUATIC INVASIVE SPECIES

Marine plants, animals, and microbes are carried * Corresponding author around the world attached to the hulls of ships and in ship

FIGURE 1 - Ballast Water Cycle between ports

ballast water. When discharged into new environments, 3. INTERNATIONAL APPROACHES FOR they may become invaders and seriously disrupt the na- THE BALLAST WATER MANAGEMENT tive ecology and economy. They displace native aquatic life, alter habitat, ecosystem and food web, and cause se- The International Convention for the Control and Man- vere fouling problems on infrastructure and vessels. Intro- agement of Ships’ Ballast Water and Sediments (BWM) is a duced pathogens may cause diseases and death in humans. new international convention to prevent the potentially dev- International Maritime Organization (IMO) Globallast has astating effects of the spread of harmful aquatic organisms identified ten of the dangerous species. These are cholera carried by ship ballast water, and it has been adopted by the (Vibrio cholerae, some cholera epidemics appear to be IMO, the United Nations agency responsible for the safety directly associated with ballast water), cladoceran water and security of shipping and the prevention of marine flea (Cercopagis pengoi), mitten crab (Eiocheir sinensis), pollution from ships. The instrument was adopted at an toxic algae (red, brown, and green tides, may form harm- international conference held from February 9-13, 2004 at ful algal blooms depending on the species), round goby IMO's London headquarter. The Convention will require (Neogobius melanostomus), North American comb jelly all ships to implement a Ballast Water and Sediments (Mnemiopsis leidyi), North Pacific seastar (Asterias amuren- Management Plan. All ships will have to carry a Ballast sis), zebra mussel (Dreissena polymorpha, fouls all available Water Record Book and will be required to carry out hard surfaces in mass numbers), Asian kelp (Undaria pinnat- ballast water management procedures to a given standard, ifida), and European green crab (Carcinus maenas) 1. following a phase-in period. A number of guidelines has been developed to facilitate the implementation of the The Straits of Istanbul and Çanakkale are among the Convention. The Convention is divided into Articles; and busiest seaways around the globe. Approx. 50.000 ships an Annex, which includes technical standards and require- passed these straits and 22.300 ships visited Turkish ports ments in the Regulations for the control and management during 2011. The majority of the ballast water (BW) vol- of ships' ballast water and sediments 4. The Convention ume discharged to Centroid Moment Tensor (CMT) area will enter into force 12 months after ratification by 30 coun- was from the Mediterranean Sea (69%), which was fol- tries, representing 35 % of world merchant shipping tonnage. lowed by North East Atlantic (11%), North West Atlantic As of 30 September 2013, there were 38 contracting (10%), and Indo-Pacific Ocean (6%). The highest BW states, and 30.38% of world merchant shipping tonnage discharge volume belongs to the Western Mediterranean ratified the convention 4. Sea with ~14.061.894 tones 3. Several strategies have been developed for reducing the risk of spreading Invasive Alien Species (IAS) through ballast water, such as:

a) Minimizing uptake of organisms into ballast water  which have ships generating oily bilge water and other tanks. Avoiding ballast water uptake in shallow and turbid residues that cannot be discharged into the sea, areas, e.g., where propellers can stir up sediment, and  which are loading crude oil, avoiding uptake at night when many organisms migrate vertically to feed, thereby reducing the number of organ-  loading and discharging bulk cargo in respect to oil isms that enter ballast water tanks; residue from combination carriers,  in which 1000 tons/day oil other than crude oil is b) Removing ballast sediment. Routine cleaning of loaded, ballast water tanks and removal of sediment in mid-ocean or at specific facilities provided in ports reducing the  having ship repair yards and providing tank-cleaning number of organisms that are transported; facility 6.

c) Avoiding unnecessary discharge of ballast water. 3.2 Ballast Water Reception Facility Where cargo handling demands uptake and discharge of ballast water within a port, water taken up in another area A ballast water reception facility (BWRF) should be should not be discharged, if avoidable; capable of receiving ballast water from ships so as not to create a risk to the environment, human health, property d) Ballast water exchange. Ballast water can be ex- and resources arising from the release to the environment changed between ports, mid-ocean, and in deep water, in of harmful aquatic organisms and pathogens. A facility order to reduce the risk of organisms carried in the water should provide pipelines, manifolds, reducers, equip- finding a suitable environment on discharge; ment and other resources to enable, as far as practica- e) Treatment of ballast water. Several methodologies ble, all ships wishing to discharge ballast water in a port that seek to remove or render harmless organisms in bal- to use the facility 6. last water while in tanks and on ships are in development or being piloted. This includes mechanical treatment (e.g., 3.3 Sediment Reception Facility filter or cyclonic separation), physical treatment (e.g., The BWM, under Article 5, sediment reception facili- ultraviolet, ultrasound, or heat treatment), chemical treat- ties (SRFs) undertake to ensure that ports and terminals ment (e.g., the use of disinfectants or biocides), biological where the reception facilities have adequate units for treatment, or a combination of these; cleaning or repairing of ballast tanks to receive the sedi- f) Discharge to reception facilities. Discharge of bal- ments 6. last water to reception facilities prevents organisms transported in ballast water from releasing into the wild 2. 3.4 Reception Facilities at the Turkish Ports Turkey has 295 port facilities including tanker termi- 3.1 Reception Facilities nal, bulk terminal, fishing and yacht port. The Turkish The establishment of adequate waste reception facili- National Railway System (TCDD) is managing 7 ports of ties (WRFs) is a necessary step to reduce and eliminate them, 17 ports are being managed by the Turkish Mari- ship-generated pollution 5. WRFs should be designed time Lines (TDI), 2 tanker terminals are managed by the by taking into account the ship types that may be antici- Turkish Petroleum Refineries Corporation (TUPRAS), and pated to use them. Consideration should be given to the 20 ports are being managed by other local administra- requirements for ballast tank cleaning that may take place tions. And also 50 small ports are being managed by local and of repair facilities in the area(s) the reception facility municipalities, 53 ports are being managed by the private serves. The capabilities and any capacity limitations of sectors, 13 yacht harbors are being managed by the Minis- reception process (facilities and equipment’s) should be try of Tourism, and 128 fishing harbours are being man- made available to ships wishing to use the facility. The aged by the local co-operatives and municipalities 7, 8. details made available to ships should include but not be In Turkey, a total of 177 ports have port WRFs. Mo- limited to: 1) maximum capacity (volume or weight) of bile waste reception ships and vessels are working for sediment, 2) maximum volume or weight that can be small ports. WRFs in Turkey are facilities where ship- handled at anyone time, 3) packaging and labeling re- generated wastes of Annex I (oil), Annex IV (sewage), quirements, 4) hours of operation, 5) ports, berths, areas and Annex V (garbage) of MARPOL (73/78) are stored, where access to the facility is available, 6) ship-to-shore treated, and disposed in. Capacity of these facilities and transfer details, 7) if ship or shore crew are required for treatment methods may vary because of regional- the transfer, 8) contact details for the facility, 9) how to economic reasons. WRFs started to be established a few request use of the facility including any notice period and years ago 5. At present, they do not receive and treat what information is required from the ship, 10) all appli- ballast water and sediment from ships. These additional cable fees, and 11) other relevant information 6. treatment systems are expected to demand one more year. Even though all port WRFs must be ready to receive A port WRF needs to be introduced to seaport and ballast water and sediment from ships in 2017, when all terminals entertaining ships with sludge tank, in addition ships must have installed ballast water treatment systems to all ports (BWTSs) at 2017.

4. OUTLOOK

IMO adapted the BWM at 2004 but it is not in force yet. All new and existing ships will install BWTSs until 2017. Some shipping companies installed the BWTSs to their ships. 60 BWTSs were in the maritime sector at 2010 but new ones are expected to come every year. BWTSs can effectively accomplish sterilization for the ballast water flowing into the ballast tank of ships and, thereby, prevent all kinds of pollution to seawater related with ballast water discharge by removing aquatic invasive species from ballast water. BWTS, a system for treatment of ballast water used in ships, is very expensive; therefore, ballast and sediment reception facilities are alternative solutions for ships` owners. If most of the Turkish ports will be effective to use ballast water reception at their WRFs, it will be the main and most practical solution for ballast water treatment until 2017.

ACKNOWLEDGEMENTS

Authors thankfully acknowledge guest editor and ref- erees for comments, Prof. Dr. Temel Oğuz and Veysel Yahşi for improving the text.

The authors have declared no conflict of interest.

REFERENCES

1 GloBallast, The Problem, Available from: http://globallast.imo.org/index.asp?page=problem.htm&men u=true. 2 Tamelander, J., Riddering, L., Haag, F., Matheickal, J. (2010), Guidelines for Development of National Ballast Wa- ter Management Strategies, London: GloBallast Monographs No. 18. 3 Olgun, A., Dönertaş, A.S., Aydöner, C., Gümüşlüoğlu, Y. Received: January 27, 2014 (2011), Assessing of the Ballast Water Risk in Ceyhan Ma- Revised: March 19, 2014 rine Terminal, Proceedings of the Global R&D Forum on Accepted: April 02, 2014 Compliance Monitoring and Enforcement the Next R&D Challenge and Opportunity, October 26-28, Istanbul, Turkey, pp. 167-174. CORRESPONDING AUTHOR 4 International Maritime Organization (IMO), [Internet], Available from: www.imo.org. Tanzer Satır 5 Subaşı, E., Doğan-Sağlamtimur, N. (2013), Performance Department of Maritime Transportation and Man- Evaluation of the Martaş Port Waste Reception Facility Treatment Plant (Tekirdağ, Turkey), Desalination and Water agement Engineering Treatment, 51, 3040-3046. Maritime Faculty Istanbul Technical University (ITU) 6 IMO (2013), Guidelines for Ballast Water Reception Facili- ties (G5), Annex B-4, IMO, London. 34940 Tuzla-Istanbul TURKEY 7 Satir, T., Alkan, G.B., Can, S., Bak, O.A. (2007) Port Recep- tion Facilities: Using Multi Criteria Decision Making, IC- MRT’07, Italy, pp. 25-28. Phone: +90 216 446 8490 Fax: +90 216 4468491 8 Satir T., Demir H., Alkan B. G., Ucan O. N., Bayat C. (2008), Ship Waste Forecasting at the Botas LNG Port Using E-mail: [email protected] Artificial Neural Networks, Fresenius Environmental Bulle- tin, 17, 2064-2070. FEB/ Vol 23/ No 11a/ 2014 – pages 2895 - 2898

GREY WATER FOOTPRINT OF CROPS AND CROP-DERIVED PRODUCTS: ANALYSIS OF CALCULATION METHOD

Chrysi S. Laspidou*

Civil Engineering Department, University of Thessaly, Pedion Areos, 38334 Volos, Greece

Presented at the 4th International Conference on Environmental Management, Engineering, Planning and Economics (CEMEPE), June 24 to 28, 2013, Mykonos, Greece

ABSTRACT creasing demands due to increased global population, as it is intimately connected to food production, economic The explosive increase in world population, along growth [1] and energy production — the so-called water- with the fast socio-economic development, have led to an energy nexus. increased water demand; the impact of global water re- Anthropogenic nitrogen flows are a major component source consumption is mapped with the “water footprint”, of the earth’s biogeochemical cycles, and have become a which is the total volume of freshwater used to produce topic of increasing research attention; nitrogen loading to the goods and services consumed. “Virtual water” is the land surface has doubled since the pre-industrial time [2]. volume of water used to produce a commodity, and is Such changes in nitrogen flows significantly affect the sus- classified as green, blue and grey. “Blue water” refers to tainability of food production and the environment, since the volume of fresh water consumed along the supply excessive nitrogenous fertilizer use causes eutrophication, chain of a product; “green” quantifies the volume of pre- loss of diversity, water and air pollution, and others; there- cipitation stored as soil moisture, and “grey” measures the fore, poor management in nitrogen application can seriously volume of water required to dilute pollutants associated affect environmental health. Crop production is by far the with a product's production chain to meet ambient water single largest cause of human alteration of the global nitro- quality standards. When applied to agricultural crops, the gen cycle, thus raising global sustainability concerns [3-5]. calculation of grey water requires the following quantities: fertilizer application rate, crop yield, fertilizer leach- The area and/or volume of water required for the ing fraction, maximum allowed pollutant concentration in needs and the productive activities of a given population natural waters, and in the water where pollution is reject- during a specific period of time represent the so-called ed. Much of this data is unavailable and researchers seem water footprint of that specific population. Hence, the to proceed with a series of assumptions to calculate grey water footprint corresponds to the total volume of water water figures. In this article, the common assumptions are used to produce the goods and services consumed by reviewed, and prominence is given to weaknesses and the individuals and communities, or produced by firms, in a sensitivity of the calculation method. defined area [6]; for several countries, it plays an important role in the optimization of water resources man-

agement. The water footprint can be divided into three KEYWORDS: Water footprint, grey water footprint, nitrogen components: the blue, green and grey water footprint. pollution, nutrient leaching fraction Each component is related to a specific type of water

resource, and the volumes of that type of water that 1. INTRODUCTION should be consumed for household or productive purposes. The blue component is related to freshwater used from World population increase, followed by an increase in surface water and ground water, the green to rain water, food production and energy consumption, has exerted an and the grey to polluted water, in the sense that this vol- intense stress on energy and water resources. Furthermore, ume of water is required to dilute pollutants in order to over the next three decades, projected population and eco- reach water quality standards [7]. nomic growth levels will exert stress on freshwater and Mekonnen and Hoekstra [8] present a global analysis energy demand, while water pollution and water quality of the water footprint of crop production for a period of degradation have become important issues for many re- ten years. They found that a large part (around 80%) of gions. People continue flocking in urban areas, creating the global water used is green, confirming the importance more pressure on local water resources. Water faces in- of rain for agriculture, while blue and grey water footprint had a smaller share (about 10% each). The quality and * Corresponding author reliability of these results are based on the quality of input

data and underlying assumptions for these calculations. f = 10% Regarding grey water, Mekonnen and Hoekstra [8] stated cN,max = 10 mg/L of nitrate-nitrogen (NO3-N) that their estimate is rather conservative, since it is based c = 0 on the analysis of the required assimilation volume for N,nat leached nitrogen fertilizers only, leaving out relevant Calculations of the same figures of grey water foot- pollutants, such as phosphorus and pesticides, that may be print volume with different f values are presented in Table equally important on the quality degradation of the aquat- 1, in order to test its sensitivity to this factor that is often ic environment. Naturally, the amount of nitrogen that is assumed arbitrarily to be 10%. leached to the aquatic environment is also a quantity that Sensitivity analysis is also performed for the natural is difficult to determine, and the quality of data reported nitrogen concentration cN,nat. Keeping the f value constant may be questionable. A discussion of the problems and at 10% and cN,max at 10 mg/L NO3–N, grey water volumes weaknesses embedded in the calculation of grey water in are re-calculated for the same 14 crops increasing cN,nat to agriculture is the focus of this article. 2 mg/L NO3-N, and the sensitivity of grey water numbers to this parameter is assessed. Liu et al. [3] presented a detailed analysis of which cN,nat concentrations are appro- 2. MATERIALS AND METHODS priate to use, depending on the nutrient of interest (nitrogen or phosphorus, inorganic or organic, particulate or Out of the three types of water footprints (blue, green dissolved, etc.) for each water-body. In this paper, a sim- and grey) and especially in the sector of agriculture, the pler analysis is presented using only a single natural con- grey is the one that has been focused on the least. Mekon- centration of NO3-N. The results of this analysis are pre- nen and Hoekstra [8] presented a list of 17 studies, in sented in Table 2, as previously. which no calculation of the grey water footprint is included. They presented calculations of the global water footprint of crop production for the period 1996-2005, 3. RESULTS AND DISCUSSION and found that the share of grey water footprint is relatively small (around 10% of the total). As shown in Table 1, an important parameter in the According to Mekonnen and Hoekstra [8], the grey calculation of the grey water footprint of crops and crop- water footprint is calculated by quantifying the volume of related products is the fraction of nitrogen that leaches, or water needed to assimilate the nutrients that reach ground- runs off to surface waters. The number of 10% is used or surface water. Nutrients leaching from agricultural throughout the literature for all countries and crops [8, 10]. fields are a main cause of non-point source pollution of Undoubtedly, this is a crude assumption, since leaching surface and subsurface water-bodies. In most studies, the fraction f depends on crop type, crop-specific nitrogen grey water footprint is quantified as related to nitrogen fixation, soil composition and condition, meteorologi- use only — although recently, Liu et al. [9] focus on cal conditions, such as atmospheric precipitation, local phosphorus as well. The grey component of the water climate and area topography and relief, among others. The footprint (GWF) is calculated by multiplying the fraction sensitivity analysis performed herein, which involves the (f) of nitrogen that leaches or runs off by the nitrogen recalculation of grey water footprint for different values application rate (LN), and dividing this by the difference of f, indicates that grey water is quite sensitive to the f between the maximum acceptable concentration of nitro- value. In fact, doubling f doubles the grey water footprint, gen (cN,max) and the natural concentration of nitrogen tripling f triples the grey water footprint, etc. As shown in (cN,nat) in the receiving water-body and by the actual crop Table 1, in the case of vegetables, for example — a crop yield (Y) (Equation 1). Naturally, to obtain an accurate with a high grey water footprint — the part of total foot- figure for the grey water footprint of a crop, one needs to print that is grey can be as high as 41% once the large f have a good estimate of the nitrogen fertilizer application value of 20% is used. This compares to the respective rate for the specific crop, as well as the applied nitrogen percentage of 26% that results from using the widely fertilizer lost through leaching. assumed value of f = 10%. There is evidence in the literature that shows how f (1) may vary for different crops and different locations. Liu et al. [3] defined the nitrogen recovery rate as the ratio of nitrogen uptaken by crops to nitrogen input, and found it A sensitivity analysis on this grey water footprint cal- to be equal to 59% globally. This number, which is direct- culation is performed using the global average water ly related to the leaching fraction f is highly variable and footprint volumes for 14 primary crop categories, as cal- ranges from 20% in Asia to 70% in Europe and North and culated by Mekonnen and Hoekstra [8] for the 1996 to South America. It is related to crop yields, to the extent of 2005 period. The authors present green, blue and grey leguminous crops that incorporate most of the available water footprint volumes, in m3 per ton of crop, while grey nitrogen and to crop type, with rice having the lowest water calculations are performed for nitrogen, using the recovery rate due to difficulties in controlling nitrogen following assumptions: losses. On the other hand, humid and temperate climates

result in cropland soils that remain wet for most of the body, but the concentration that would occur, if there year. Such soils, together with high fertilizer application were no human influences. Given the recent population rates can result in a high level of nitrate leaching. Liu et increase, water scarcity, pollution problems and, thus, the al. [3] showed that the highest nitrogen leaching occurs in amount of stress to our water-bodies has been under, Western Europe, where a large amount of mineral fertiliz- during the last couple of centuries, the concentration that ers is applied in temperate agro-ecological zones. Good soil would occur, if there were no human influences, is obvi- and fertilizer management practices can reduce nitrate ously hard to determine, and may even be irrelevant in leaching to a certain extent; in tile-drained systems, howev- some cases. Although the grey water footprint of a prod- er, nitrate losses are unavoidable, since they remain high, uct refers to the volume of freshwater that is required to even with good management practices. A point to be assimilate the load of pollutants based on existing ambient emphasized herein is that it is essential that more research water quality standards, using as cnat the concentration and data gathering is done to define the leaching fraction f that would occur, should there exist no human influence, better, and to record it for different areas in the world and does not take into account the existing water quality of the various crops and soil types in order to obtain more accu- water-bodies. For the same crop and nutrient load, we rate grey water footprint numbers. would get the same grey water footprint, no matter what the water quality of the water-body that receives the crop Of critical importance to the grey water footprint cal- runoff is. A major weakness of the method is that alt- culation are the values of c and c in the receiving N,nat N,max hough the impact would be different if the runoff would water-bodies. Until recently, most published articles that flow in a high water-quality high-flow river, or in a low present calculations of the grey water footprint would water-quality low-flow river, the grey water footprint consistently report the value of zero for c due to lack N,nat would be the same in both cases. of data [10, 11]. For the maximum concentration, cN,max, Mekonnen and Hoekstra [11] stated that the recommended Table 2 shows that increasing the natural nitrogen maximum value of nitrate in surface and groundwater by concentration from 0 to 2 mg/L NO3-N results in a 25% the World Health Organization and the European Union is increase of the average grey water footprint for all crops 3 50 mg nitrate (NO3) per L, and the maximum value rec- (from 306 to 383 m /ton). The percentages vary depend- ommended by US-EPA is 10 mg/L measured as nitrate- ing on crop and grey water footprint size but, in general, nitrogen (NO3-N); thus, in most studies, the standard of we see that the natural concentration is significant in the 10 mg/L NO3-N is used following Chapagain et al. [10], calculation of grey water footprint. Liu et al. [9] reached which is a reasonable assumption. the same conclusion, although they followed a different calculation procedure. Great care needs to be taken in In a recent publication, Liu et al. [9] offered a differ- defining the natural pollutant concentration figures for ent point of view for c , stating that it is not zero, since N,nat each water-body in order to obtain accurate grey water all rivers naturally transport some nutrients and that it is footprint figures for each crop. not the actual concentration of the river or another water-

TABLE 1 - Global average water footprint of 14 primary crop categories for the 1996 to 2005 period (calculations are based on the following assumptions: cN,max = 10 mg/L of nitrate-nitrogen (NO3-N); cN,nat = 0; LN and Y are the same for each crop category, while f is variable).

Crop category Water footprint (m3/ton)* Water footprint (m3/ton) ** Water footprint (m3/ton) ** f = 10% f = 20% f = 5% Grey Total % Total Grey Total % Total Grey Total % Total in Grey in Grey in Grey Sugar 15 197 7.61% 30 212 14.15% 7.5 189.5 3.96% Fodder 20 253 7.91% 40 274 14.60% 10 244 4.10% Vegetables 85 322 26.40% 170 407 41.77% 42.5 279.5 15.21% Roots & tubers 43 387 11.11% 86 429 20.05% 21.5 364.5 5.90% Fruits 93 967 9.62% 186 1060 17.55% 46.5 920.5 5.05% Cereals 184 1644 11.19% 368 1828 20.13% 92 1552 5.93% Oil crops 121 2364 5.12% 242 2485 9.74% 60.5 2303.5 2.63% Tobacco 700 2925 23.93% 1400 3626 38.61% 350 2576 13.59% Fibers, veg. origin 300 3837 7.82% 600 4138 14.5% 150 3688 4.07% Pulses 734 4055 18.10% 1468 4789 30.65% 367 3688 9.95% Spices 432 7048 6.13% 864 7480 11.55% 216 6832 3.16% Nuts 680 9063 7.50% 1360 9743 13.96% 340 8723 3.90% Rubber, gum, waxes 422 13748 3.07% 844 14169 5.96% 211 13536 1.56% Stimulants 460 14443 3.18% 920 14903 6.17% 230 14213 1.62% AVERAGE 306.4 4375.2 10.62% 612.7 4681.6 18.53% 153.2 4222.1 5.76% * Source: Mekonnen and Hoekstra (2010 [8]; ** Author calculations

TABLE 2 - Global average water footprint of 14 primary crop categories for the 1996 to 2005 period. Calculations are based on the following assumptions: cN,max = 10 mg/L NO3-N, f = 10%; LN and Y are the same for each crop category.

Crop category Water footprint (m3/ton)* Water footprint (m3/ton) ** cN,nat = 0 cN,nat = 2 mg/L NO3 -N Grey Total % Total Grey Total % Total in Grey in Grey Sugar 15 197 7.61% 18.8 200.8 9.3% Fodder 20 253 7.91% 25 259 9.7% Vegetables 85 322 26.40% 106.3 343.3 31% Roots & tubers 43 387 11.11% 53.8 396.8 13.5% Fruits 93 967 9.62% 116.3 990.3 11.7% Cereals 184 1644 11.19% 230 1690 13.6% Oil crops 121 2364 5.12% 151.3 2394.3 6.3% Tobacco 700 2925 23.93% 875 3101 28.2% Fibers, veg. origin 300 3837 7.82% 375 3913 9.6% Pulses 734 4055 18.10% 917.5 4238.5 21.6% Spices 432 7048 6.13% 540 7156 7.5% Nuts 680 9063 7.50% 850 9233 9.2% Rubber, gum, waxes 422 13748 3.07% 527.5 13852.5 3.8% Stimulants 460 14443 3.18% 575 14558 3.9% AVERAGE 306.4 4375.2 10.62% 382.9 4451.9 12.8% * Source: Mekonnen and Hoekstra [8]; ** Author calculations

An important issue for the calculation of the grey wa- density) or modern (higher-density) plantations making ter footprint for crops is that fertilizer application rates per systematic use of fertilizers and pesticides, or in mountain- crop per country are not available for most crops. The ous plantations that are usually exclusively rain-fed with no figures that are available are usually gross estimates based chemical inputs [13]. Obviously, for olive groves in the on crude assumptions, or are based on outdated figures that same country, it may be difficult to generalize on the rate may be irrelevant when used. Terres et al. [12], in a Joint of fertilizer application. Future research should focus on Research Center European Commission report, presented better defining the fertilizer application rate for each crop an analysis on the agricultural nitrogen quantity for EU and geographical area. river basins based on Nomenclature of Territorial Units for Statistics (NUTS) census level data for the years 1990 The choice of pollutant related to grey water footprint to 1997. Thus, the quality of estimates in 2002 is already calculation is also important. Although most studies focus questionable since it was based on data that are over five on nitrogen, other pollutants could be equally important: years old. To deal with the lack of data, Mekonnen and pesticides, herbicides such as atrazine, organics — usual- Hoekstra [8] used rates, which are based on different ly related to industrial crop-processing plants — and sources and a number of assumptions. Thus, the same thermal pollution are some of the polluting factors that are crop in a country is assumed to receive the same fertilizer usually not given enough attention by researchers. Pro- application rate everywhere in the country. cessing of crop-derived products often results in large pollution loads ending up in water-bodies and the simple This assumption becomes problematic for countries emphasis on nitrogen pollution for such products results with large and diverse agricultural areas, since they may in the underestimation of the grey water footprint of these have differences in soil structure and fertilization needs. It crops to the environment. Olive-mill waste is a good exam- also becomes problematic for countries with bold topo- ple, as researchers focus mostly on nitrogen grey water for graphical relief that may grow crops in quite different the olive crop [14], when high-BOD olive processing waste agricultural settings. Irrigated crops generally receive creates big environmental problems in major olive-producing more fertilizer than rain-fed ones. Besides, most small countries that often lack specific regulations regarding the subsistence farmers likely use no or less fertilizer. The discharge of this waste. In Greece for example, in practice, grey water footprint is estimated based on a simplified all the generated olive-mill wastewater results in creeks approach, which gives a rough estimate; it leaves out local (58%), or in sea and rivers (11.5%), or in soil (19.5%) [14]. factors that influence the precise leaching and runoff rates, Grey water footprint analysis of crops and crop-derived such as rainfall intensity, irrigation patterns, soil property, products should, in a systematic way, be expanded to in- slopes, and the amount of already mineralized nitrogen in clude all aspects of water pollution, and not remain limited the upper soil layer. Systematic comparison of the esti- in the environmental impact of nutrient leaching. mate from such simplified approach with other regression models [3] might be required to test the uncertainties and A point to be made regarding current sophisticated limitation of this approach. A good example here is olive water footprint analysis is that it focuses almost exclusive- cultivation, which is intensely fragmented, with olive ly on volume, which is a very good tool to raise awareness, farming being done either in intensified traditional (low- but probably not sufficient to represent the impact on a

water resource. Thus, while the volume of water footprint [3] Liu, J., L. You, M. Amini, M. Obersteiner, M. Herrero, maybe the same for a crop or a product regardless where A.J.B. Zehnder and H. Yang (2010). A high-resolution assessment on global nitrogen flows in cropland, PNAS it is produced, the impact on a given water resource — 107(17): 8035-8040. reservoir, lake, river, etc., maybe completely different, [4] Gialis, S.E.; Loukas, A. and C. Laspidou (2012). Regional given the local water scarcity, pollution-stress and water Development and World Trade: A Review of Virtual Water quality. Vegetable crops grown in water-stressed areas and Water Footprint Concepts in Agriculture, Proceedings: with significant reliance on irrigation will have a multiple Protection and Restoration of the Environment XI:243-252. impact on water resources than a different crop with the [5] Laspidou, C.S., Kakoulidis, I. and Loukas, A. (2011). Eco- same or even higher volumetric water footprint that is system simulation modeling of nitrogen dynamics in the con- grown in areas with abundant freshwater supplies. The structed wetland Carla in Greece. Desalination and Water Treatment 33:61-67 water footprint concept should be expanded from the volumetric analysis towards including the aspect of im- [6] Hoekstra, A.Y. and Chapagain, A.K. (2007). The water footprints of Morocco and the Netherlands: Global water use as a pact on local water resources, taking into account local- result of domestic consumption of agricultural commodities. ized water quality and scarcity/abundance. Ecological Economics, 64:143-151. [7] Hoekstra, A.Y. and Hung, P.Q. (2005). Globalisation of water resources: international virtual water flows in relation to crop 4. CONCLUSIONS trade. Global Environmental Change Part A, 15(1):45-56. [8] Mekonnen, M.M. and Hoekstra, A.Y. (2010). The green, blue An analogue to the ecological footprint that denotes and grey water footprint of crops and derived crop products, Value of Water Research Report Series No. 47, UNESCO- the area needed to sustain a population, the water foot- IHE, Delft, the Netherlands. print, represents the water volume required and can be [9] Liu, C., Kroeze, C., Hoekstra, A.Y. and Gerbens-Leenes, W. defined for a nation, a product, a process, a group of con- (2012). Past and future trends in grey water footprints of an- sumers, a business, etc. It is a new and important tool that thropogenic nitrogen and phosphorus inputs to major world has been in use for only about a decade, but has already rivers, Ecological Indicators 18, 42-49. done a great deal in raising awareness about the critical [10] Chapagain, A.K.; Hoekstra, A.Y.; Savenije, H.H.G. and Gau- importance of water use and the “hidden” water associat- tam, R. (2006). The water footprint of cotton consumption: ed with production in modern society. Grey water foot- an assessment of the impact of worldwide consumption of cotton products on the water resources in the cotton produc- print specifically places the necessary emphasis on water ing countries, Ecological Economics, 60(1):186-203. pollution and the limited pollutant-assimilative capacity of our water resources. Lack of data and crude assump- [11] Mekonnen, M.M. and Hoekstra, A.Y. (2011). The green, blue and grey water footprint of crops and derived crop products, tions are some of the biggest problems of these calcula- Hydrology and Earth System Sciences, 15:1577-1600. tions. In this paper, we bring out some of the weaknesses [12] Terres, J.M., Campling,.P., Vandewalle, S., Van Orshoven, J. of the grey water footprint calculation method. Nutrient (2002). Calculation of Agricultural Nitrogen Quantity for EU leaching fraction f and natural concentration cnat are two River Basins. Report EUR 20256 EN. 87p. of the factors included in the calculation that are being [13] Niaounakis M. and Halvadakis, C.P. (2006). Olive Processing studied herein, and are proven to be important in predict- Waste Management: Literature Review and Patent Survey, ing grey water footprint accurately. The analysis shows Waste Management Series 5, 2nd ed., Elsevier, Oxford. that research should be directed towards better defining [14] Salmoral, G.; Aldaya, M.M.; Chico, D.; Garrido, A. and these quantities, since they greatly affect grey water foot- Llamas, M.R. (2010). The water footprint of olive oil in print volumes for crops and crop-related products. If wa- Spain. Fundacion Marcelino Botin, Madrid. ter footprint analysis is going to be used as a tool for water resources management, it should probably be expanded to include the localized impact on water re- Received: November 11, 2013 sources, taking into account quantity and quality of local Revised: January 12, 2014 water resources. Accepted: April 02, 2014

The author has declared no conflict of interest. CORRESPONDING AUTHOR

Chrysi S. Laspidou REFERENCES Civil Engineering Department University of Thessaly [1] Laspidou, C.S.; Nydreos-Sakouelos, P. and Kungolos, A. Pedion Areos (2012). Carbon footprint calculation of desalination units in 38334 Volos Greece. Fresenius Environmental Bulletin 21(8b), 2344- GREECE 2349. [2] Chatzimpiros, P. and Barles, S. (2010) Nitrogen, land and E-mail: [email protected] water inputs in changing cattle farming systems and Sci Total Environ 408(20):10, PMID 20624629 FEB/ Vol 23/ No 11a/ 2014 – pages 2899 – 2903

EVALUATION OF WATER QUALITY OF A MOUNTAINOUS STREAM (PELION, CENTRAL GREECE) USING BENTHIC MACROINVERTEBRATES

Christos I. Rumbos1, 2,* and Athanassios Kungolos1

1 Department of Planning and Regional Development, School of Engineering, University of Thessaly, 38334 Volos, Greece 2 Department of Ichthyology and Aquatic Environment, School of Agricultural Sciences, University of Thessaly, 38446, Volos, Greece

Presented at the 4th International Conference on Environmental Management, Engineering, Planning and Economics (CEMEPE), June 24 to 28, 2013, Mykonos, Greece

ABSTRACT nagement strategies [1]. According to the 2000/60 Water Framework Directive of the European Parliament, each Biotic indices are commonly used in biomonitoring member state should establish national freshwater moni- programs to assess the quality of surface waters. In the toring networks and assess the ecological status of fresh- present study, a number of biotic indices based on ben- water ecosystems, in order to achieve the objective of good thic macroinvertebrates have been applied in assessing ecological status in all surface waters by 2015 [2]. In an the water quality of a mountainous stream (Pelion, Cen- ideal situation, an accurate assessment of water quality tral Greece). Samplings were done in summer and au- requires the use of physical, chemical and biological meth- tumn 2012 in two sampling sites (Karamani, Servanates) ods, in order to have the whole spectrum of information using two sampling methods: (a) a standard pond net and under consideration [3]. However, such studies are labo- (b) the Surber method. Invertebrate identification was per- rious and expensive and, therefore, not suitable for routine formed to the family level. The Biological Monitoring water quality monitoring. Biological assessment of water Working Party (BMWP) score, the Average Score Per Taxa quality, using biological communities as bioindicators, is (ASPT), the Lincoln Quality Index, as well as the Iberian widely accepted, relatively inexpensive, and can give BMWP, and the Iberian ASPT were applied to the benthic reliable results as it reflects the cumulative effect of differ- macroinvertebrate samples to assess the water quality of ent stressors [4]. A number of organisms have been used as each site. According to ASPT, IASPT and LQI, the bio- biological indicators of the water quality in freshwater logical water quality of both sampling sites was very good, systems, including diatoms, protozoa, fish as well as vege- regardless of the season and the sampling method used, tation [5]. Benthic macroinvertebrates are among the best with the exceptions of the summer samplings in Servanates bioindicators of running water quality, and have been used with both sampling methods, and in Karamani with in numerous biotic indices and studies [3, 5]. Surber, for which a good water quality was assessed with However, there is a lack of information concerning the LQI and IASPT, respectively. In contrast, the BMWP and use of benthic macroinvertebrates for water quality assess- IBMWP scores for the sampling site Servanates were low, ments in Mediterranean mountainous streams, which are corresponding to fair water quality, but this effect was characterized by extreme hydrological fluctuations and attributed to the particular hydrological conditions of this flow irregularities [6, 7]. Therefore, the objective of the site. The sampling method did not significantly affect the present study was to assess the water quality of a moun- sampling efficiency and, consequently, the results of the tainous stream using biological indices based on benthic water quality evaluation. macroinvertebrates. Moreover, the results obtained with two commonly used sampling methods, the Surber sampler

KEYWORDS: Benthic macroinvertebrates, biological index, and the standard pond net, were compared. biomonitoring, water quality.

2. MATERIALS AND METHODS 1. INTRODUCTION 2.1 Study area The evaluation of water quality in aquatic ecosystems The study was conducted in Vrychonas stream, which is essential for the development of sustainable water ma- is located on mountain Pelion (Central Greece) and discharges into Pagasitikos Gulf (Fig. 1). Platanus orientalis * Corresponding author L. (Platanaceae) is the dominant species of the riparian

FIGURE 1 - Study area and sampling sites (Karamani, Servanates) in Vrychonas stream (black line) (Pelion, Central Greece).

vegetation along the stream while, 20 m away from the 3. RESULTS AND DISCUSSION stream banks, natural vegetation is replaced by olive or- chards. In the two sampling sites of the study area, 461 individuals belonging to 32 families of benthic macroinverte- 2.2 Benthic macroinvertebrate sampling and identification brates were collected (Table 1). Most of them belong to Benthic macroinvertebrates were sampled in July and the class of Insecta (27 families), while among Insecta, October 2012 at two sampling sites (Karamani, Serv- the most abundant are Diptera (7 families), followed by anates). Samples were taken with two sampling tech- Trichoptera (6 families). niques at each site: a) a Surber sampler (900 μm aper- The results of the application of the bioindices at the ture net, 2500 cm2 sampling area) and b) a standard pond two sampling sites, based on the results obtained with the net (500 μm aperture net, 625 cm2 sampling area). At first, two sampling methods, are presented in Table 2. Accord- the benthic macroinvertebrate samples were taken from each ing to ASPT, IASPT and LQI, the biological water quality site using the Surber sampler. Large stones (> 100 mm) were of both sampling sites was “Very Good”, regardless of the removed by hand from the sampling area and scrubbed into season and the sampling method used, with the exceptions the Surber net to collect any attached invertebrates [8]. of the summer samplings in Servanates with both sam- Afterwards, the substrate in the sampling area was dis- pling methods and in Karamani with Surber, for which a turbed and present macroinvertebrates in the sediment were “Good” water quality was assessed with LQI and IASPT, swept by the stream current into the net [8]. The 3-min respectively. In the case of Karamani, the IBMWP scores kick-sweep method was followed for the sampling with the were, in most cases, in agreement with the IASPT scores, pond net. The sampling procedure consisted of placing with the exception of the autumn sampling with pond net, the net in the stream and moving up the stream for 3 min, for which a “Good” water quality was assessed. In contrast, while simultaneously the substrate was disturbed up- in the case of Servanates, the biological water quality of the stream by kicking [8]. Two samples of benthic macroin- site varied from “Bad” to “Good” according to IBMWP vertebrates were taken from each site for each sampling and from “Very Bad” to “Bad” according to BMWP, being session. In both cases, macroinvertebrate samples were in disagreement with IASPT and ASPT, respectively. Gen- removed from the nets, fixed immediately in 70% ethanol erally, the BMWP values were, in most cases, lower than and placed in labeled glass bottles. Back in the laboratory, the respective IBMWP values, mainly because BMWP individuals were sorted and stored in 70% ethanol. The index does not include several characteristic families that macroinvertebrates were identified to the family level were present in our samples, and are taken into considera- using a Leica MZ12 stereoscope (Leica Microsystems, tion by IBMWP (Athericidae, Tabanidae, Calopterygidae, Wetzlar, Germany) and the appropriate dichotomous keys Glossosomatidae). [9-13]. Finally, the Biological Monitoring Working Party The conflicting results, obtained from the BMWP and (BMWP) score [14], the Average Score Per Taxa (ASPT) IBMWP bioindices at Servanates and, consequently, the [15], the Lincoln Quality Index [16], the Iberian BMWP disagreement in the classification of the water quality ac- and the Iberian ASPT [17] were applied to the benthic cording to these two bioindices may be attributed to the macroinvertebrate samples to assess the water quality of particular fluvial conditions at this site. The low water each site. level, especially in summer, may be the reason for the

TABLE 1 - The overall composition and distribution of macroinvertebrates in the Vrychonas stream sampling sites.

Karamani Servanates Summer Autumn Summer Autumn Surber Pond net Surber Pond net Surber Pond net Surber Pond net Phylum Class Order Family Number of individuals Annelida Oligochaeta Lumbriculida Lumbriculidae 40 3 11 3 3 - 1 4 Hirudinea Rhynchobdellida Glossiphoniidae - - 3 - - - - - Mollusca Gastropoda Hygrophila Lymnaeidae 6 4 1 12 - - - - Basommatophora Ancylidae - - 8 4 - - - - Arthropoda Crustacea Amphipoda Gammaridae - 1 - - - - 12 4 Arthropoda Insecta Diptera Athericidae 9 4 11 6 - - 1 1 Simuliidae 1 ------Tipulidae 2 1 ------Culicidae 1 ------Chironomidae - 5 1 3 - - - - Empididae - - 1 - - - - - Tabanidae - - 2 1 - - - - Ephemeroptera Ephemerillidae 1 ------Ephemeridae - - 4 3 - - 3 5 Baetidae 4 1 1 2 - - - - Leptophlebiidae - 1 ------Heptageniidae - - 1 - - - - - Coleoptera Dryopidae 1 ------Dytiscidae 1 ------Gyrinidae - 1 ------Plecoptera Leuctridae - 1 1 9 1 - 93 13 Trichoptera Limnephilidae 1 5 ------Hydropsychidae 18 4 ------Philopotamidae 1 ------Glossosomatidae 2 3 18 5 - - - - Lepidostomatidae - - 1 - - - - - Rhyacophilidae - - - 1 - - - - Odonata Gomphidae 1 2 5 1 2 3 6 8 Aeshnidae - 1 7 3 - 1 - 1 Calopterygidae - 1 2 - - 2 3 - Corduliidae - - 2 4 - - - - Hemiptera Gerridae - 1 - - 15 2 3 9

TABLE 2 - Results of the application of five bioindices for the assessment of water quality of the mountainous stream Vrychonas, using two sampling methods.

Summer Autumn Surber Pond net Surber Pond net Karamani Value 61 79 83 67 BMWP a Water quality Moderate Good Good Moderate Value 5.5 5.6 6.4 6.1 ASPT b Water quality Very Good Very Good Very Good Very Good Value A+ A+ A++ A+ LQI c Water quality Very Good Very Good Very Good Very Good Value 81 101 117 89 IBMWP d Water quality Good Very Good Very Good Good Value 5.4 5.9 6.5 6.4 IASPT e Water quality Good Very Good Very Good Very Good Servanates Value 24 23 48 48 BMWP a Water quality Very Bad Very Bad Bad Bad Value 6.0 7.7 6.9 6.9 ASPT b Water quality Very Good Very Good Very Good Very Good Value B B A A LQI c Water quality Good Good Very Good Very Good Value 22 27 64 56 IBMWP d Water quality Bad Bad Good Moderate Value 5.5 6.8 7.1 7.0 IASPT e Water quality Very Good Very Good Very Good Very Good a Biological Monitoring Working Party; b Average Score Per Taxa; c Lincoln Quality Index; d Iberian Biological Monitoring Working Party; e Iberian Average Score Per Taxa

faunistic scarcity [18], which has led to the low BMWP REFERENCES and IBMWP scores. The macroinvertebrate families present in Servanates (Leuctridae, Gomphidae, Aeshnidae, [1] WWF, (2003) Water and Wetland Index. Critical issues in Calopterygidae) have high ecological value [14-17], indi- water policy across Europe. World Wide Fund for Nature, Madrid, p. 57. cate good water quality, and give high ASPT and IASPT scores, supporting the hypothesis that the small biogenic [2] European Union Council, (2000) Directive of the European capacity of the particular site due to the low water flow Parliament and the Council establishing a framework for Community action in the field of water policy, 2000/60/EC. and not pollution is responsible for the low BMWP and Official Journal of the European Communities 43 (L327), IBMWP scores. Analogous effects have been described Brussels, p. 72. previously, when the impact of a small scale hydropower [3] Metcalfe, J.L. (1989) Biological water quality assessment of in an upland stream in Wales was investigated, and the running waters based on macroinvertebrate communities: low BMWP scores have been attributed to the low num- History and present status in Europe. Environmental Pollu- ber of scoring taxa, due to hydrological variation rather tion 60, 101-139. than differences in water quality [18]. Similarly, other [4] Iliopoulou-Georgudaki, J., Kantzaris, V., Katharios, P., researchers propose that ASPT should be used to correct Kaspiris, P., Georgiadis, Th. and Montesantou, B. (2003) An BMWP scores obtained from sites, where the hydrologi- application of different bioindicators for assessing water quality: A case study in the rivers Alfeios and Pineios (Pelo- cal conditions affect the benthic macroinvertebrate com- ponnisos, Greece). Ecological Indicators 2, 345-360. munity [19]. [5] Hellawell, J.M. (1986) Biological indicators of freshwater Previous studies have shown that the sampling tech- pollution and environmental management. In Pollution Moni- nique significantly affects the composition of the benthic toring Series. Elsevier Applied Science, Amsterdam, p. 546. macroinvertebrate samples, and have a clear impact on [6] Argyroudi, A., Chatzinikolaou, Y., Poirazidis, K. and Lazari- the final water quality assessment, as each sampling tech- dou, M. (2009) Do intermittent and ephemeral Mediterranean nique is selective for specific components of the benthic rivers belong to the same river type? Aquatic Ecology 43, community [8, 20]. In an earlier study, significant differ- 465–476. ences were reported in the composition of samples col- [7] Feio, M.J., Reynoldson, T.B. and Graça, M.A. (2006). Effect lected with Surber and kick-sampling from eleven sam- of seasonal changes on predictive model assessments of streams water quality with macroinvertebrates. International pling sites in streams of south-western Australia [8]. In Review of Hydrobiology 91, 509–520. the present study, the two sampling methods used (Surber, pond net) gave similar results, regarding the [8] Storey, A.W., Edward, D.H.D. and Gazey, P. (1991) Surber and kick sampling: a comparison for the assessment of ma- bioindices scores and the water quality classification. As croinvertebrate community structure in streams of south- far as it concerns the number of taxa collected, Surber western Australia, Hydrobiologia 211, 111-121. samples were characterized by higher species richness in [9] Merrit, R.W. and Cummins, K.W. (1996) An introduction to comparison to the samples collected by pond net, regard- the aquatic insects of north America. Kendall/Hunt Publish- less of the season of sampling. These results are in ing Company, Iowa, p. 862. agreement with previous studies, which attributed the [10] Nilsson, A. (1996) Aquatic insects of north Europe: A taxo- higher species richness of the Surber samples to the great- nomic handbook, Vol. 1 and 2, Apollo Books Aps., Den- er and more in depth effort per unit area expended in mark, p. 714. Surber sampling than pond net [8]. [11] McCafferty, W.P., (1998) Aquatic entomology: The fisher- men’s and ecologists illustrated guide to insects and their rel- atives. Jones and Bartlett Publishers, Massachusetts, p. 448. 4. CONCLUSIONS [12] Askew, R.R. (2004) The dragonflies of Europe (revised edition). Harley Books, Colchester, England, p. 308. The present study illustrates that the use of benthic [13] Malicky, H. (2004) Atlas of European Trichoptera (second macroinvertebrates for the biological assessment of water edition). Springer, Dordrecht, The Netherlands, p. 359. quality can be easily applied in mountainous streams and [14] Chesters, R.K. (1980) Biological Monitoring Working Party. could be used for routine monitoring of mountainous The 1978 national testing exercise. In Water Data Unit Tech- water-bodies. However, the careful selection of the ap- nical Memorandum 19, Department of the Environment, propriate bioindices is necessary, in order to get reliable London, UK, p. 37. results. BMWP and, in a lesser extent, IBMWP proved to [15] Armitage, P.D., Moss, D., Wright, J.F. and Furse, M.T. be relatively ineffective in describing the water quality of (1983) The performance of a new biological water quality Vrychonas stream. Further application of biotic indices in score system based on macroinvertebrates over a wide range of unpolluted running-water sites. Water Research 17, 333- more Greek mountainous streams is essential to evaluate 347. their applicability and reliability. Finally, the present study demonstrated that the sampling method does not [16] Extence, C.A., Bates, A.J., Forbes, W.J. and Barham, P.J. (1987) Biologically based water quality management. Envi- significantly affect sampling efficiency and, consequent- ronmental Pollution 45, 221-236. ly, the results of water quality evaluation. [17] Alba-Tercedor, J. and Sanchez-Ortega, A. (1988) A simple and quick method to evaluate biological quality of running The authors have declared no conflict of interest. freshwater based on Hellawell, 1978. Limnetica 4, 51-56.

[18] Copeman, V.A. (1997) The impact of micro-hydropower on the aquatic environment. Water and Environment Journal 11, 431-435.

[19] Rico, E., Rallo, A., Sevillano, M.A. and Arretxe, M.L. (1992) Comparison of several biological indices based on river macroinvertebrate benthic community for assessment of running water quality. Annals of Limnology 28, 147–156.

[20] Barton, D.R. and Metcalfe-Smith, J.L. (1992) A comparison of sampling techniques and summary indices for assessment of water quality in the Yamaska river, Quèbec, based on benthic macroinvertebrates. Environmental Monitoring and As- sessment 21, 225-244.

Received: October 22, 2013 Revised: March 11, 2014; March 19, 2014 Accepted: April 02, 2014

CORRESPONDING AUTHOR

Christos I. Rumbos Department of Planning and Regional Development School of Engineering University of Thessaly 38334 Volos GREECE

Phone: +30 6972451834 E-mail: [email protected] [email protected]

FEB/ Vol 23/ No 11a/ 2014 – pages 2904 - 2908

ΟIL SPILL DETECTION FROM RADARSAT-1 SYNTHETIC APERTURE RADAR IMAGERY AT NORTHERN ENTRY OF BOSPORUS STRAIT, TURKEY

Burcu Ozsoy-Cicek1,*

1Istanbul Technical University, Faculty of Maritime, Tuzla 34940, Turkey

Presented at the 4th International Conference on Environmental Management, Engineering, Planning and Economics (CEMEPE), June 24 to 28, 2013, Mykonos, Greece

ABSTRACT KEYWORDS: Oil spill, Marine Pollution, Remote Sensing, SAR, Image processing

Shipping and maritime activities may cause a great amount of oil pollution on the sea surface; such as ship 1. INTRODUCTION transportation, oil platforms and ports. There are some de- terrent applications to provide cleaner and safer seas. Oil pollution seriously affects the marine environ- Timely identification of extent and source of oil pollution ment, fisheries, social and economic activities. The main is mandatory for mitigation of its impact and prosecution of sources of marine oil pollution are discharges coming the originator. One of the preferred instruments to detect from ships or offshore platforms, which can be accidental and monitor oil pollution by maritime activities is space- or deliberate. Tanker accidents cause great environmental borne active microwave remote sensing at a few 10 m spatial problems to the marine ecosystem. Well-known examples resolution as enabled independent of cloud cover and day- of accidents are, for instance: Amoco Cadiz (France, on light conditions by Synthetic Aperture Radar (SAR). It March 16, 1978, about 51 million gallons), Exxon Valdez allows us to detect oil pollution at the sea surface over a (Prince William Sound, Alaska, on March 24, 1989, about region of a few 100 kilometers at a glance. It is possible 11 million gallons), M/T Prestige (Spain on November to monitor near-real time, continuously or periodically for 14, 2002, about 20 million gallons), and recently the off- the region of interest. This and its cost effectiveness is a shore oil-platform accident Deep Water Horizon. Major great advantage over air-borne surveillance activities. Pre- shipping routes have potential oil pollution threats due to requisites for successful oil-spill identification by SAR crude oil tanker traffic. Accordingly, high potential for imagery is a sufficiently high wind speed (above 2 m/s), tanker accidents and thus a devastating pollution of the proper application of SAR image processing techniques, marine environment does not only exist for difficult to and a careful consideration of SAR imaging geometry and navigate waters and weather conditions but particularly hydrodynamic modulation of the sea surface by oceanic also for areas with high ship traffic like, for instance, the phenomena. This study provides evidence about the wide- Strait of Malacca, Strait of Dover, Strait of Hormuz, Strait spread occurrence of oil spills in the Black Sea north of of Gibraltar or the Turkish Straits and especially the Is- Bosporus Strait, Turkey. The preliminary results of an tanbul Strait or Bosporus, which is the narrowest strait analysis of 12 RADARSAT-1 SAR ScanSAR Narrow around the world. The strait connects the Black Sea with images acquired during late summer and fall 2010 and the Mediterranean. In order to provide safer ship and mari- 2012 underline the usefulness of SAR imagery for oil spill time safety more effectively, very strict rules are applied detection. Simple image enhancement techniques can perfor Turkish Straits to protect the marine environment. mit to obtain information about the character of the oil Another, often underestimated source of oil pollution of spill. Application of such techniques is however limited by the marine environment is given by the illegal discharges the environmental conditions and in particular the discrimi- of oily residues and oily ballast water intentionally re- nation of oil spills from natural surface slicks requires leased by ships [1]. Even though strict rules are applied, additional data and/or more advanced techniques in ac- as for instance for the Turkish Straits, there is observa- cordance with published knowledge. tional evidence that ships start to illegally discharge oily residues from tank cleaning and oily ballast or bilge water once these have left the inland waters of a country. With- out a routine observation of the seas such pollution remains unreported and the actual input of oily substances * Corresponding author into the marine environment remains unknown.

Ideally such intentional pollution shall not occur. The The goal of this study is to show first results of a fea- ‘International Convention for the Prevention of Pollution sibility study to identify oil spills in the Black Sea in the from Ships’, shortly MARPOL 73/78 includes regulations vicinity of the Bosporus in RADARSAT-1 SAR imagery to aimed at preventing and minimizing pollution from ships— aid human interpretation of SAR images at Turkish gov- both accidental pollution and that from routine opera- ernment. The paper is organized as follows: After this intro- tions—and currently includes six technical Annexes [2]. duction we describe our area of interest and the data ac- This convention includes a concept for especially sensitive quired. Then we briefly review the main aspects of SAR sea areas which are termed as a Special Area in MARPOL based oil spill detection and present the methodology used 73/78. A Special Area is deﬁned by technical reasons in in this paper. This is followed by our results and their relation to its oceanographic and ecological conditions as discussion before we give a short outlook. well as the particular characteristics of the traffic. Adop- tion of special mandatory methods for the prevention of sea pollution by oil is required for Special Areas [2]. The 2. STUDY AREA AND DATA Black Sea is such a Special Area. The MARPOL 73/78 Annex–I Prevention states: “In Special Areas, oil dis- Research study area was the Black Sea. It is one of charges from ships have been completely prohibited, with the largest inland seas in the World with a surface of minor and well-deﬁned exceptions” [3]. There are also 423,000 km² [8]. The Black Sea shoreline is 4,340 km some regional conventions to protect sea from oil pollu- long with Bulgaria, Georgia, Romania, Russia, Turkey tion as the Bucharest Convention for Black Sea and the and Ukraine (http://www.blacksea-commission.org, ac- Bonn Agreement for North Sea. Using the Bonn Agree- cessed October 02, 2013). The Black Sea is the most ment, aerial observation for detection and combating oil isolated sea in the World. It is connected to the World pollution has been carried out for North Sea States since Oceans via the Mediterranean Sea through the Bosporus, three decades [3]. The “Convention on the Protection of Dardanelle and Gibraltar straits and with the Sea of Azov the Black Sea against Pollution” (Bucharest Convention) in the northeast through the Kerch Strait [9]. This isola- is a regional convention to combat oil pollution and foster tion makes any oil pollution particularly severe because joint actions in case of accidents. The document was signed the water mass exchange with the Mediterranean Sea is of in 1992 by the six Black Sea States. The Bucharest Con- the order of 0.1% of the total Black Sea water volume as vention, Article VIII, states: “The Contracting Parties shall is the amount of water input to the Black Sea by river take individually or, when necessary, jointly, all appropri- discharge and precipitation [10]. The level of potential ate measures to prevent, reduce and control pollution of operational and accidental risk of oil spills with conse- the marine environment of the Black Sea from vessels in quent oil pollution of marine and coastal environment is accordance with generally accepted international rules and very high in the Black Sea as shown, e.g. by the Volga- standards” (http://ec.europa.eu/ environment/ water/marine/ neft Tanker accident near the Kerch Strait where about bucharest.htm, accessed October 02 2013). On the other 1.3 million gallons of heavy fuel oil was spilled [11]. A hand for the Black Sea, there is no regular aerial observa- recent study illustrated the results of Russian oil spill tion monitoring to prevent oil pollution or controlling oil detection activities in the northern part of the Black Sea pollution like North Sea. using SAR imagery [12]. Main sources of oil spills in the Black Sea are: oil pipelines, operational oil spills during Both, the coordination of oil pollution combat activi- reloading, trans-shipment, oil transportation by sea/rail, ties in case that an accident has happened already, and the oil refineries, and illegal discharges [10]. observation of the marine environment to detect oil spills from illegal discharges requires a surveillance of the area In this study 12 SAR images acquired by the Canadi- subject to the oil spill. In case of an accident the position an RADARSAT-1 SAR at C-Band (5.3 GHz), HH- is usually well-known; in case of an illegal discharge the polarization (horizontal polarization in transmitting and location is unknown. Doing such surveillance with ships receiving mode) were obtained from Istanbul Technical or via airplanes remains ineffective and expensive and can University - Center for Satellite Communications and Re- be – in case of bad weather conditions – even dangerous. mote Sensing (ITU-CSCRS) summarized in Table 1. All Satellite imagery provides an excellent affordable tool for images were acquired in ScanSAR Narrow (SCN) mode this task [4, 5]. In order to be independent of daylight and and offer a nominal grid resolution of 50 m x 50 m and a clouds on the one hand and to be able to resolve the often pixel spacing of 25 m x 25 m over a 300 km wide swath. small-scale, several ten meters to a few 100 meters wide RADARSAT-1 SAR images in SCN mode have the ad- patches of oil spilled onto the sea surface, synthetic aper- vantage of a finer spatial resolution compared to the stand- ture radar (SAR), i.e. active microwave satellite imagery ard ScanSAR Wide mode images with 100 m x 100 m grid has proven extremely useful. [6,7] provide excellent re- resolution; this is often too coarse to reliably identify an oil views of the satellite techniques and oil spill discrimination spill [6]. All images were taken in the late summer/fall methods and we kindly refer the reader to these two publi- period because at this time of the year tanker traffic is cations, and references therein, to learn more about oil highest. spill detection in satellite imagery.

3. DATA PROCESSING Processing starts with ingesting the SAR image into a software package which allows carrying out the pre- pro- SAR images require a number of processing steps be- cessing steps mentioned in Figure 1, e.g. ENVI or NEST. fore they can be interpreted properly. Figure 1 illustrates The image is geo-referenced, and radiometric and geo- the steps carried out in this study. metric corrections are carried out, leading to a position

TABLE 1 - List of the SAR image parameters obtained

Acquisition Date Acquisition Time Over flight direction Incidence angle at scene center Orbit 05/08/2010 15:42 UTC ascending 29.8° 77005 25/08/2010 15:59 UTC ascending 38.9° 77291 01/09/2010 15:55 UTC descending 38.2° 77391 05/10/2010 16:03 UTC ascending 38.9° 77877 12/10/2010 15:59 UTC ascending 38.9° 77977 19/10/2010 15:55 UTC ascending 38.9° 78077 05/08/2012 4:04 UTC descending 39.3° 87445 22/08/2012 4:08 UTC descending 30.4° 87688 01/09/2012 4:16 UTC descending 30.4° 87831 08/10/2012 15:56 UTC ascending 38.9° 88367 15/10/2012 15:52 U’TC ascending 29.8° 88467 19/10/2012 4:16 UTC descending 30.3° 88517

INPUT SAR SCN IMAGE

PRE-PROCESSING • Geo-Reference

• Radiometric Correction

• Geometric Correction

IMAGE ENHANCEMENT

IMAGE FILTERING

• Gamma • Enhanced Lee (5x5)

DENSITY SLICE

FIGURE 1 - Flow chart of the processing steps used in this study which lead from the raw SAR image to the automatic delineation of an oil spill.

FIGURE 2 - Raw RADARSAT-1 SAR SCN image (amplitude) focusing on the northern entry of the Bosporus Strait on August 5, 2012, 4:04 UTC, before the pre-processing

Look direction

Flight direction

FIGURE 3 - Same image as in Figure 2 but after the pre-processing (e.g. geo-referencing) indicated in the flow chart in Figure 1. The white and black boxes denote the area of the sub-images shown in Figures 4 and 5. The white arrow points towards North.

accuracy of about 50 m. Examples of these initial steps image usually results in some unwanted noise, the so- are shown in Figures 2 and 3: while Figure 2 shows the called speckle noise which needs to be reduced in the raw RADARSAT-1 SAR image, Figure 3 shows the im- image before further analysis. Speckle noise reduction has age after the above-mentioned pre-processing steps. been carried out by applying a Gamma Filter and an enhanced Lee Filter using a 5 x 5 pixel box, which is 125 m Next the image quality needs to be enhanced by fil- x 125 m in case of the RADARSAT SCN images used in tering. The viewing technique used to generate a SAR this study. The usage of a 5 x 5 pixel box is a compromise

between being able to also properly detect small spills and Note that the strength of the radar backscatter is a ships in the image and an optimal speckle noise reduction. function of the incidence angle of the radar – the transi- For ScanSAR wide mode imagery a smaller box, i.e. 3 x 3 tion from bright (right) to dark (left) seen, for instance, at pixel should be used. the top part of Figure 3 along the look direction is most likely caused by this incidence angle dependency – and of the viewing angle with respect to the wind direction. Wind- 4. OIL SPILL DETECTION induced sea surface roughness causes highest radar FROM RADAR IMAGE backscatter when wind direction and viewing direction are opposite to each other; lowest radar backscatter is obtained After the pre-processing steps the main task is to de- when looking perpendicular to the wind direction [14]. lineate features in the SAR images which are potentially These geometrically effects need to be taken into account caused by an oil spill. How can this be done? To give an during selection and interpretation of SAR images. While answer here it is first required to understand how an oil this last issue just can be considered in the interpretation spill manifests itself in a SAR image. SAR images, or in the incidence angle dependency can be either corrected general, images acquired with active microwave instru- for [16] or, as is done in this study, the image is segment- ments (short: radar) over an ocean surface are an imprint ed into slices of one degree incidence angle variation. of the ocean surface roughness. This roughness is caused It should also be noted that gravity waves, swell and a primarily by the surface wind which first causes capillary number of other oceanic phenomena like currents, fronts, waves. Depending on the duration and strength with which internal waves, atmospheric convection, rain cells, and un- the wind blows and the fetch over which the wind blows, derwater bottom topography all can modulate the sea surface gravity waves develop [13, 14]. These have a different in a way that enhances or decreases the impact of the sur- roughness scale but still contribute to the surface rough- face wind on the sea surface roughness [17]. ness. The coupling between the radar waves and the Oil is lighter than sea water. Therefore it spreads at surface roughness causes the so-called “Bragg- the sea surface and effectively damps the capillary waves Scattering”. The Bragg Scattering wavelength matches as well as the centimeter-scale gravity waves [18, 19]. As quite good with the C-Band wavelength which makes C- a result the radar backscatter decreases over the patches Band SAR a particularly good candidate to observe wind- covered by an oil spill and the patches appear dark in a induced sea surface roughness and thus to derive the near- SAR image. However, as pointed out in Brekke and Sol- surface wind speed [13]. As a rule of thumb one can say: berg [6] and in Topouzelis [7], there are a number of fac- The higher the wind speed the higher is the C-Band radar tors which hamper oil spill detection in general. These are backscatter and the brighter is the SAR image. However, listed as follows: there is a threshold wind speed below which no capillary waves are induced and thus the radar backscatter remains 1) Wind speed is below the threshold wind speed for low. SAR images remain dark in this case. This threshold capillary waves to form. Then the radar backscatter wind speed is of the order of 2 m/s at C-Band frequencies. remains low and the respective part of the SAR image In Figure 4 the white dashed, white solid and black circles remains dark. This can also happen during otherwise denote areas where the wind speed is below the threshold, high enough wind speeds due to shadowing effects by where the wind speed is low but above the threshold, and land or islands. This has been illustrated in Figure 4. where the wind speed is high. The quite high radar 2) Above a certain wind speed oil spills are dispersed by backscatter in the red circle is presumably caused by rela- turbulent mixing at the sea surface and simply disap- tively strong catabatic winds descending down the coast pear; this occurs at wind speeds above about 10-14 m/s from the interior northern Turkey and which can be iden- [20]. Only very thick oil spills can withstand such turbu- tified along the entire coastline visible in Figure 4 [15]. lent mixing longer and also at higher wind speeds [6].

FIGURE 4 - Subset of the black box of the SAR image shown in Figure 3 after applying Gamma and Enhanced Lee filters (5x5 pixel). See text for meaning of circles. North is up.

3) Every hydrodynamic modulation of the sea surface, spills which have undergone dispersion and which have e.g., by rain cells [21], internal waves [22], currents been moved by surface currents might look pretty similar [23], etc., can cause radar backscatter to drop to similar- to a natural surface slick. ly low values as observed for wind speeds below the A very good categorization of oil spills, their shapes above-mentioned threshold. and typical appearance in a SAR image is given by Pavla- 4) Young sea ice like frazil and grease ice as well as dark kis et al. [27], where more than 1,600 oil spills are classi- nilas damp the capillary waves and short gravity fied in five categories. As a rule of thumb a low radar waves in a similar manner as an oil spill and appear as backscatter patch in a SAR image is most likely an oil areas of similarly low radar backscatter in SAR im- spill when: agery [24]. • it is a uniformly homogeneous low radar backscatter All these factors are valid for the Black Sea and have area in a uniform high backscatter, i.e., windy area; to be taken into account. In addition one needs to discrim- • it is a linear area of low radar backscatter, not ex- inate between oil spills which are man-made and oily tremely large, with occasional abrupt turns due to substances which occur naturally. Micro-organisms and wind direction or surface (tidal) current changes; natu- plankton can release substances which – similarly to a ral slicks in these conditions tend to disappear while man-made oil spill – damp the surface waves and look oil spills have a higher viscosity and tend to change like an oil spill. These are called surface slicks or surfac- their shape rather than being easily dispersed. tants and complicate reliable oil spill detection in SAR Figure 5 gives a number of excellent examples of oil images considerably [6, 25, 26]. spills. Each of the dark patches seen are clearly defined Any oily substance at the surface – no matter whether and exhibit shapes and sizes typical for oil spills from a it is an oil spill or a natural surface slick – is subject to ship. Comparison with Figure 3 reveals that basically all dispersion over time, e.g. by turbulent mixing or bio-geo- these dark patches are aligned along the major shipping chemical processes, and displacement. The substance is routes across the Black Sea, starting from the Bosporus. moved by tidal, barotropic or eddy currents or simply by In contrast a low radar backscatter patch in a SAR the wind. As a consequence oil spills and slicks form a image is most likely not an oil spill when: wide variety of patterns at the sea surface. While a trained analyst can manually detect most oil spills and slicks and • wind speeds are generally low and the SAR image also discriminate between those any automatic approach shows more or less irregularly distributed patches of needs to be trained with certain parameters with which it low and medium high radar backscatter; can learn and distinguish between slicks and oil spills. • it extends along coastal zones where topography to- Such parameters include total area, perimeter, roundness, gether with the actual wind direction suggests shelter- general shape, degree of connectivity between patches, ing effects; and more shape parameters as well as across spill / slick edge radar backscatter contrast, mean radar backscatter of • it covers elongated dark areas with smooth turnings in the spill / slick, etc. While oil spills are usually very well spiral shape, often several parallel to each other and of- defined, with a sharp step in radar backscatter between the ten not just in one region but spread over a larger area. oil spill area and the surrounding region natural surface Figures 6 illustrates the co-existence of oil spills, see slicks often are less clearly defined. However, the degree the solid white box with the dark patch at the top left and with which an oil spill is very well defined a function of the a few ten kilometers long dark patch stretching wind speed and time elapsed since the spill event. Old oil through almost the entire solid white box, and natural

1 4

FIGURE 5 - Subset of the white box of the SAR image shown in Figure 3. Numbered boxes denote areas used in Figure 7. North is up.

FIGURE 6 - Subset of the SAR images of August 22 2012 (see Table 1) in the southeastern Black Sea. North is up.

slicks, see the dashed white box with the very slim, fore, even though quite advanced methods could be used to aligned parallel to each other, curved dark filaments extract oil spill information from SAR imagery [7] the which are in addition hard to see because of low wind quality of that information remains uncertain. Therefore, speeds. A few of such filaments are also seen at the botin this study, it was decided to use simple image en- tom left part of the solid white box. Note that if the dark hancement techniques to visualize oil spills in RADAR- area due to low wind speed south of the solid white box SAT-1 SAR imagery in an easy-to-understand way to would have been further to north the long oil spill would train and support human interpreters working for Turkish not have been identified. government in oil spill identification and discrimination on SAR images. Figure 7 shows a series of sub-images which have 5. PRELIMINARY RESULTS AND OUTLOOK been selected from Figure 5 such that radar backscatter variation due to incidence angle variation is minimal. In According to Brekke and Solberg [6] and Topouzelis order to enhance the location and the extent of the dark [7] there is a wealth of different methods to delineate patches caused by oil spills the density slice technique is spills and slicks and to discriminate between them in SAR applied. This is done by identifying the Normalized Radar images automatically. However, Topouzelis [7] noted that Cross Section (NRCS) values associated with the dark currently there is a lack of contemporary information patches in a histogram and carefully selecting two NRCS about oil spill and/or natural slick location, timing, extent, ranges (slices) which correspond to the lowest NRCS life-time, type, and many more on from the ground-based values, i.e. the darkest part of these patches and to the observation side and about oil spill and/or natural slick transition regions where the NRCS values start to gradual- location, timing, and extent from the remote sensing per- ly take the values of the background. These two slices are spective – which can be both air- and space-borne. There- displayed in red and green, respectively, in Figure 7 and

FIGURE 7 - From left to right: images of boxes 1 to 4 in Figure 5 with the density slices applied. The white circles in the leftmost image denote location of ships which show up as bright, often star-like spots in SAR images.

FIGURE 8 - Same as Figure 5 but only density slices of the identified oil spills are shown (see also Figure 7 and text).

also in Figure 8, which shows a summary of all dark classification are visible in Figure 8 at the bottom left and patches investigated in Figure 5 with this technique. A bottom right corners. proper selection of the NRCS values used requires experi- Note that the case illustrated in Figure 5 is an easy ence in SAR data analysis and needs to be tuned for each example where even an un-experienced human interpreter SAR image. This will be discussed further below. Note will provide a reasonable result. Figure 6 is more difficult that the background NRCS has been set to the maximum and requires some experience because unlike in Figure 5 NRCS of the image and is thus displayed by white color. there is not a uniform surface wind speed signature but The different slices used give information about the the wind speed, the wind speed induced surface roughness status of the oil spill. A thin oil spill or an oil spill which and thus the SAR image brightness varies considerably. has been subject to dispersion due to winds, currents and In addition surface slicks are present. Therefore a human biogeochemical reactions can expected to be less dense interpreter has to be experienced and has to be able to dis- and might also have a smaller viscosity than a fresh and/ criminate between dark features which are caused by an oil or thick oil spill. Consequently, its damping effect on the spill and dark features which are caused by natural slicks. surface waves is reduced and surface roughness induced Figures 7 and 8 demonstrate that simple image enhance- by the surface wind starts to build up. Dispersion by winds ment techniques can improve the identification of oil spills and currents starts from the edges of an oil spill. Therefore in SAR imagery of the Black Sea and that these even allow almost all oil spills with a dark grey center in Figures 7 obtaining information about the character of the oil spill and 8 are bordered by a light grey area. Based on the area detected. However, already in this SAR image wind speed occupied by these two slices one could deduce more in- variations cause radar backscatter variations which could formation about the respective oil spill. If we assume that mislead an automatic classification method (see Figure 8). environmental conditions have remained constant over a From this experience it is clear that for the SAR image of sufficiently long time period, the distribution of dark grey August 22 2012 shown in Figure 6 a much more careful and light grey areas might indicate the distribution of thin selection of areas for the application of the density slice and/or old oil spills versus thick and/or new oil spills. One approach is required and even then it cannot be excluded has to take into account here also, however, that different that not only oil spills but also natural slicks would be types of oil have different densities and viscosities and classified as oil spill. This would require human interven- therefore will spread differently on the sea surface. tion. The same is true for most of the 12 SAR images While Figure 8 illustrates the location and character (Table 1) acquired for this first preliminary study. of the oil spills visible in Figures 3 and 5 quite nicely it We note therefore, that the used density slice technique also shows the limitations of the density slice technique. has to be used very carefully. SAR images acquired at Since this is a technique relying on simple thresholds, different frequencies, such as by COSMO-SkyMed in X- those areas in the SAR image which exhibit low backscat- Band will most likely need different thresholds for the slicing ter values because of other phenomena than an oil spill technique. But also SAR images acquired by the same sensor are mis-classified as an oil spill as well. This applies for but at a different polarization, a different incidence angle or a instance to box 1 in Figure 5 (leftmost image in Figure 7) different wind speed might require using a different thresh- where the wind speed is partly so low that the associated old for the density slicing technique. For a sound automatic radar backscatter is as low as the radar backscatter used for analysis one needs a look-up table which specifies thresh- the green slice. Consequently areas with low radar olds as function of wind speed and incidence angle ranges backscatter due to low surface wind speed are mis- for different polarization and frequency. classified as belonging to an oil spill. Such areas of mis-

FIGURE 9 - Near surface current (2.5 m depth) in the Black Sea for August 5, 2012. The area displayed by the SAR image in Figure 3 and the SAR subimage in Figure 5 are superposed by black boxes (figure source: MyOcean 2 Project – Ocean Monitoring and Forecasting).

If the SAR image alone does not allow for an unam- is felt that an optimal oil spill detection in SAR imagery biguous detection of dark features as being a spill or a and an optimal assessment of its impact on the marine slick and discrimination between those, then additional environment is requiring additional measurements of the information might help such as a SAR image acquired environmental conditions prior, at and after the anticipat- earlier, information about the environmental conditions ed spilling event, such as by satellites, buoys, ships and like surface currents, e.g. from a numerical model, sea coastal stations and close cooperation between govern- surface temperature, e.g. from Moderate Resolution Spec- mental, observing and modeling groups. troradiometer (MODIS), near surface air temperature and atmospheric stability, near surface wind speed and direction, precipitation records, etc., and information about ship traffic and off-shore activities. ACKNOWLEDGEMENTS Such additional information is not only useful for en- We thank Istanbul Technical University financial hanced interpretation of the SAR images but it is also support during our project. We also thank Istanbul Tech- required for an estimation of the further development of nical University - Center for Satellite Communications detected oil spill as well as for an assessment of its impact and Remote Sensing (ITU-CSCRS) for the additional on the marine environment. Figure 9 illustrates the near SAR image support. We appreciate that MyOcean project surface current situation of the Black Sea on August 5 provides the oceanographic parameters which had great 2012. According to this map the oil spills identified potential to provide additional information to our study. I would be transported southeastward and eastward without appreciate valuable discussion by the reviewers and the threatening the coastal areas immediately. If such spills editor. would have occurred further to the West, however, then these would likely have had an impact at the Turkish The author has declared no conflict of interest. coast close to the Bosporus.

Summary, RADARSAT-1 SAR image analysis is an excellent tool to identify oil spills in the Black Sea by means of their low radar backscatter – provided that envi- REFERENCES ronmental conditions, namely the wind speed, are suitable. Simple image enhancement techniques can be used to [1] Ferraro, G.; Bernardini, A., David, M., Meyer-Roux, S., gain information about the character of the oil spill. How- Muellenhoff, O., Perkovic, M., Tarchi, D. and Topouzelis, K. (2007) Towards an operational use of space imagery for oil ever, such techniques are likely to fail and/or to require pollution monitoring in the Mediterranean Basin: a demon- human intervention when wind speeds are too low, when stration in the Adriatic Sea. Mar. Pollut. Bull. 54, 403-422 surface slick coexist with oil spills and when other hydro- [2] Mattson, G. (2006). MARPOL 73/78 and Annex I: an as- dynamic modulations of the sea surface cause radar sessment of its effectiveness. Journal of International Wild- backscatter values similarly low as those of an oil spill. It life Law and Policy 9, 175–194.

[3] Ferraro, G., Baschek, B., De Montpellier, G., Njoten, O., [20] Girard-Ardhuin, F., G. Mercier, F. Collard, and R. Garello. Perkovic, M. and Vespe, M. (2010). On the SAR derived 2005. Operational oil slick characterization by SAR imagery alert in the detection of oil spills according to the analysis of and synergistic data. IEEE Journal of Oceanic Engineering, the EGEMP. Mar. Pollut. Bull. 60, 91–102. 30(3), 487-495.

[4] Ferraro, G., S. Meyer-Roux, O. Muellenhoff, M. Pavliha, J. [21] Melsheimer, C., W. Alpers, and M. Gade. 2001. Simultane- Svetak, D. Tarchi, and K. Topouzelis. 2009. Long term moni- ous observations of rain cells over the ocean by the synthetic toring of oil spills in European seas. Int. J. Rem. Sens., 30(3), aperture radar aboard the ERS satellites and by surface-based 627-645. weather radars, J. Geophys. Res., 106, 4556-4677.

[5] Gade, M., and W. Alpers. 1999. Using ERS-2 SAR images [22] Alpers, W. 1985. Theory of radar imaging of internal waves, for routine observation of marine pollution in European Nature, 314, 245-247. coastal waters, Sci. Total Environ., 237/238, 441-448. [23] Romeiser, R., S. Ufermann, and W. Alpers. 2001. Remote [6] Brekke, C., and A. H. S. Solberg. 2005. Oil spill detection by sensing of oceanic current features by synthetic aperture ra- satellite remote sensing. Rem. Sens. Environ., 95(1), 1−13. dar - achievements and perspectives, Ann. Telecommun., 56, 661-671. [7] Topouzelis, K. N. 2008. Oil Spill Detection by SAR Images: Dark Formation Detection, Feature Extraction and Classifica- [24] Onstott, R. G. 1992. SAR and scatterometer signatures of sea tion Algorithms. Sensors, 8, 6642-6659. ice, in F. D. Carsey (ed.) Microwave Remote Sensing of Sea Ice, Amer. Geophys. Union Res. Ser., vol. 68, Washington, [8] Çokacar, T. R. 2009. Oil Spills on the Eastern Mediterranean DC., chapter 5. and Black Seas and the Bosphorus and Dardanelles Straits. [25] Alpers, W., and H. Espedal. 2004. Oils and Surfactants, Exploration and Production, 7(1), 120-124. Chapter 11 in "Synthetic Aperture Radar Marine User's Man- [9] Shapiro, G. I. 2009. Black Sea Circulation, University of ual", National Oceanic and Atmospheric Administration, Plymouth, Plymouth, UK, Elsevier Ltd. Center for Satellite Application and Research, NO- AA/NESDIS, Ch.R. Jackson, J.R. Apel, editors, Washington, [10] http://www.blacksea-commission.org, accessed October 02, D.C., USA, ISBN 0-16-073214-X, 263-275. 2013 [26] Huehnerfuss, H., W. Alpers, and K. Richter. 1986. On the [11] http://www.oiledwildlife.eu, accessed October 02, 2013 discrimination between crude oil spills and monomolecular sea slicks by airborne radar and infrared radiometer - possi- [12] Ivanov, A., A. Kucheiko, N. Filimonova, N. Evtushenko, A. bilities and limitations, Int. J. Remote Sensing, 7, 1001-1013. Antonyuk, and N. Terleeva. 2013. Large oil spill detection and polluter identification in the Black Sea using SAR imag- [27] Pavlakis, P., D. Tarchi, A. J. Sieber, G. Ferraro, and G. Vin- es and AIS data. ESA Living Planet Symposium, September cent. 2001. On the Monitoring of Illicit Vessel Discharges, A 9-13 2013, Edinburgh, UK. Reconnaissance Study in the Mediterranean Sea, EC DG Joint Research Center, Institute for the Protection and Securi- [13] Alpers, W. 1983. Imaging ocean surface waves by synthetic ty of the Citizen Humanitarian Security Unit, EU Commis- aperture radar - a review, in T.D. Allan, ed., Satellite Micro- sion Report, EUR 19906 EN, 21 pp. wave Remote Sensing, Ellis Horwood Ltd., England, Chapter 6.

[14] Snoeij, P., P. J. F. Swart, and C. M. H. Unal. 1992. Study on the response of the radar echo from the ocean surface wind vector at frequencies between 1 and 18 H GHz, Final report for the European Space Agency, Dept. of Electr. Eng. Lab. for Telecommunication and Remote Sensing Technology, Delft Univ. of Technology, Delft, The Netherlands. [15] Alpers, W., U. Pahl, and G. Gross. 1998. Katabatic wind- fields in coastal areas studied by ERS-1 synthetic aperture radar imagery and numerical modeling, J. Geophys. Res., 103, 7875-7886. [16] Singha, S., K. Topouzelis, M. Vespe, and O. Trieschmann. 2013. Radiometric normalization on SAR images for oil spill detection. ESA Living Planet Symposium, September 9-13 Received: March 19, 2014 2013, Edinburgh, U.K. Accepted: April 02, 2014 [17] Johannessen, J. A., R. Garello, B. Chapron, R. Romeiser, P. Pavlakis, I. Robinson, N. Connolly, K. Nittis, T. Hamre, S. Ufermann, W. Alpers, H. Espedal, B. Furevik, V. Cummins, CORRESPONDING AUTHOR and D. Tarchi. 2001. Marine SAR analysis and interpretation system - MARSAIS, Ann. Telecommun., 56, 655-660, 2001. Burcu Ozsoy-Cicek [18] Gade, M., W. Alpers, H. Huehnerfuss, V. R. Wismann, and Istanbul Technical University P. A. Lange. 1998. On the reduction of the radar backscatter Faculty of Maritime by oceanic surfac films: Scatterometer measurements and their theoretical interpretation, Remote Sens. Environ., 66, Tuzla 34940 52-70. TURKEY

[19] Huehnerfuss, H., W. Alpers, W. D. Garrett, P. A. Lange, and S. Stolte. 1983. Attenuation of capillary and gravity waves at E-mail: [email protected] sea by monomolecular organic surface films, J. Geophys. Res., 88, 9809-9816. FEB/ Vol 23/ No 11a/ 2014 – pages 2909 – 2918

CORRELATION BETWEEN CHLOROPHYLL A CONCENTRATION AND SEA SURFACE TEMPERATURE IN THE EASTERN MEDITERRANEAN SEA USING GIS AND SATELLITE DATA

Dimitra Kitsiou* and Konstantinos Topouzelis

Department of Marine Sciences, School of the Environment, University of the Aegean, University Hill, 81100 Mytilene, Greece

Presented at the 4th International Conference on Environmental Management, Engineering, Planning and Economics (CEMEPE), June 24 to 28, 2013, Mykonos, Greece

ABSTRACT fish aggregation for mating, feeding and fish production and they are characterized by high values of chlorophyll α Areas of increased sea surface chlorophyll α in the concentration and low values of SST [1]. Several studies aquatic environment are considered as marine productivi- have assessed the relationship between chlorophyll α con- ty hot spots resulting from oceanographic processes, such centrations and SST [2], since it is strongly related to the as upwelling, cyclonic gyres or fronts. Productivity hot spots total abundance of marine life, as well as between fish provide a favorable habitat to marine species for living distribution and physical or biological variables such as and reproduction which can also be characterized by low bathymetry, SST or chlorophyll α [3, 4]. Nevertheless, values of sea surface temperature (SST). Though several there is continuous research interest in this scientific field at studies have revealed this negative correlation between both local and global scale, since the location of hot spots is chlorophyll a and SST, there is ongoing research interest strongly related to the total abundance of marine life. in this scientific field at both local and global scale. In this Remote sensing has been widely applied to monitor marine paper, the correlation between chlorophyll a concentra- phytoplankton seasonality [5] and hot spots of primary tion and SST in the Mediterranean Sea and in particular productivity [6] since it offers the ability to study variabil- the North Aegean Sea is investigated. Monthly satellite ity at both local and global scales. data for year 2012 were acquired and processed in the In this paper, the correlation between chlorophyll a framework of a Geographical Information System (GIS). concentration and SST in the Mediterranean Sea and in Time series of thematic maps for both chlorophyll a con- particular the North Aegean Sea is investigated for year centrations and SST were produced and calculation of 2012. Datasets were acquired from satellite data, since their correlation was performed. The result was the pro- ocean color, represented by chlorophyll α concentrations, duction of a new time series of thematic maps where the and SST are two basic marine parameters measured from correlation between chlorophyll a concentrations and SST several satellite sensors. Calculation of the correlation is illustrated at spatial scale for each month. Therefore, between these two parameters was performed pixel-by- areas with negative correlation, characterized by low SST pixel for every month and representative thematic maps and high chlorophyll a concentrations, were identified in were produced. The latter allowed the identification and the study area. mapping of marine productivity hot spot areas. The detec- ,

tion of the appearance of such correlation to specific loca- KEYWORDS: spatial analysis; marine productivity; geo-database; tions and time periods is important and further discussed mapping; hot spots; North Aegean Sea in this paper.

1. INTRODUCTION 2. MATERIALS AND METHODS

Chlorophyll α concentrations on water masses indi- 2.1 Study area cate the presence of phytoplankton species and therefore Study area is the Eastern Mediterranean Sea and es- high chlorophyll α concentration values appear during algal pecially the North Aegean Sea (Fig. 1). The Aegean Sea blooms. Marine productivity hot spots are critical areas for constitutes the north-eastern part of the Eastern Mediter- ranean Sea with a complicated hydrographical and ecologi- * Corresponding author cal structure due to its geographical position. The North

FIGURE 1 - Study area.

Aegean Sea is connected with the Sea of Marmara through 2.3 Methodology the Dardanelles Straits, nearly 62 km long, 1.3-7 km wide Monthly satellite data of SST and chlorophyll a con- and 55 m in average deep. In this area, the waters from centrations for year 2012 were available after the pre- Levantine and south-central Aegean (relatively warm and processing of the datasets. The correlation between SST highly saline) are diluted by the less saline water inflow and chlorophyll a was calculated pixel-by-pixel for each from the Dardanelles as well as by freshwater nutrient month following the procedure described below. As a re- rich outflows from various rivers discharging along the sult, a correlation raster was produced for each month. Greek and Turkish coastlines [7, 8]. For each month, [X1] is the raster with the satellite data of chlorophyll α and [Y1] the corresponding raster of 2.2 Datasets SST. At this point, there is need for an indicator raster in For this study, datasets of satellite data were retrieved order to identify the pixels where neither of the rasters from myOcean project (www.myocean.eu) granted by the [X1] and [Y1] is empty. The indicator raster, called [IND] European Commission within the Copernicus program. is calculated as Chlorophyll α concentrations for Case-1 waters are the product of the merging of MERIS, MODIS and SeaWiFS [IND] = [X1]*[Y1] - [X1]*[Y1] + 1 satellite data. SST data were obtained from infra-red meas- This raster will have the value 1 in all pixels where urements collected by satellite radiometers (AVHRR on [X1]*[Y1] is not null. Following, the original rasters [X1] board on NOAA and MetOp-A satellites and SEVIRI on and [Y1] are replaced by [X] = [X1]*[IND] and [Y] = board MSG satellite). Both datasets were or converted at 1 [Y1]*[IND] and the following three rasters [XX], [YY] km spatial resolution. Monthly satellite data were ac- and [XY] are calculated: quired for chlorophyll α and daily satellite data for SST [XX] = [X]*[X], [YY] = [Y]*[Y], [XY] = [X]*[Y] for year 2012. The median value of SST daily observations for each month was calculated pixel-by-pixel and a The required correlation raster will represent a mov- timeseries of monthly SST data was produced. It should ing-window correlation raster and is computed from focal be noted here that the observed difference between day- means after definition of a window size and shape. The light and night satellite SST values was not considered window for application of focal means was selected to 3x3 significant in the present study. pixels. Therefore, the focal means of [X], [Y], [XX], [YY]

and [XY] were calculated and the resulting rasters were ary, areas with high values of chlorophyll α concentra- called [Xm], [Ym], [XXm], [Y] and [XYm], respectively. tions are mainly observed in the northern Adriatic Sea and The final correlation raster was calculated by the the eastern coasts of Italy. In the North Aegean Sea high formula: concentrations are observed only in coastal areas and inner gulfs. For May, areas with high values of chloro- Correlation_raster = ([XYm] - [Xm]*[Ym]) / phyll α concentration are mainly observed in the northern Sqrt(([XXm] - [Xm]*[Xm])*([YYm] - [Ym]*[Ym])) Adriatic Sea - in a more extended area than in February, The correlation raster will have non-null values at all the eastern coasts of Italy and the Gulf of Lions in the cells which have neighborhoods where (a) two or more southern coast of France. In the North Aegean Sea high pixels have non-null values for both [X] and [Y] and (b) concentrations are observed not only in coastal areas and not all values of [X] and [Y] in those neighborhoods are inner gulfs, but also in offshore areas in the Thracian Sea. constant. In Fig. 3, SST satellite values in the Mediterranean Sea for February and May 2012 are illustrated. It should be noted that the colorbar is the same for both illustra- 3. RESULTS AND DISCUSSION tions; however the range of values is different. Lower SST values are observed in the Western Mediterranean Pre-processing of the satellite datasets of SST and Sea as well as in the North Aegean Sea compared to the chlorophyll α concentrations for year 2012 resulted in two Eastern Mediterranean. For February 2012 the SST values monthly timeseries, respectively. In this paper, the results were about 1oC lower than in May 2012. only for February and May 2012 are shown. They are representative of winter and spring conditions in the ma- In Fig. 4(a) the hot spot areas which are characterized rine environment in terms of SST and illustrate clearly the by negative correlation between chlorophyll α and SST and seasonal variation concerning the location and extent of values of SST lower than 11oC, are presented for February marine productivity hot spots. In Fig. 2 monthly chloro- 2012. These areas are located in the northern part of the phyll α satellite data for (a) February 2012 and (b) May Adriatic Sea and the North Aegean Sea. In Fig. 4(b) the hot 2012 are presented for the Mediterranean Sea. For Febru- spot areas which are characterized by negative correlation

FIGURE 2 - Chlorophyll α concentrations (mg/m3) for (a) February 2012 and (b) May 2012 in the Mediterranean Sea.

FIGURE 3 - SST values for (a) February 2012 and (b) May 2012 in the Mediterranean Sea.

FIGURE 4 - Hot spot areas with negative correlation between chlorophyll α and SST and values of SST (a) lower than 11oC for February 2012 and (b) lower than 13οC for May 2012 in the Mediterranean Sea (marked dark on the map).

FIGURE 5 - Hot spot areas with negative correlation between chlorophyll α and SST and values of SST (a) lower than 11oC for February 2012 and (b) lower than 13oC for May 2012 in the North Aegean Sea (marked dark on the map).

between chlorophyll α and SST and values of SST lower Thracian Sea up to the island of Lemnos. It is clear that than 13oC, are presented for May 2012. These areas are during spring the marine productivity hot spot areas are located in the northern part of the Adriatic Sea, the North extended to a larger geographic area covering almost the Aegean Sea and the Gulf of Lions in the southern coast of whole North Aegean Sea, since it is highly influenced by France. Compared to the hot spot areas of February 2012, river plume and inflow of Black Sea Water as well as by those of May 2012 are extended to a larger geographic coastal upwelling [9]. This result was also observed by [1]. area. The Aegean Sea is the third major sea of the Eastern In Fig. 5(a), it is observed that for February 2012 in the Mediterranean and is characterized by high spatiotem- North Aegean Sea, hot spot areas are detected in Ther- poral variability. Though it is considered to be an oligo- maikos Gulf and especially in the eastern part of the Thra- trophic sea, large gradients of primary productivity along cian Sea up to the Dardanelles Straits. In Fig. 5(b), it is with highly productive areas have been detected [10-12]. observed that for May 2012 in the North Aegean Sea, hot The general scheme of annual phytoplankton blooms is spot areas are detected in Thermaikos Gulf and the whole represented by a short bloom during spring and a smaller

one in autumn [13]; however, the major phytoplankton importance for monitoring and management of fisheries blooms in the Eastern Mediterranean occur in winter [14]. resources as well as for the design and determination of The latter has been attributed to short periods of cool marine protected areas. weather followed by periods of warmer conditions which allow phytoplankton to bloom [15]. ACKNOWLEDGEMENT In particular, in the Aegean Sea, the dominated water masses are cold waters of low salinity coming from the This work is part of the research project entitled: Black Sea (BSW) and warm waters of high salinity origi- “ThalASAR - Automatic detection and classification of nated from the Levantine region; both create a strong front meso-scale ocean phenomena on satellite Advanced Syn- which characterizes the circulation pattern in the North thetic Aperture Radar (ASAR) data” implemented within Aegean. The area influenced by BSW is considered as the the framework of the Action «Supporting Postdoctoral most productive area of the Eastern Mediterranean. The Researchers» of the Operational Program "Education and main water mass of the Northern Aegean is characterized Lifelong Learning" (Action’s Beneficiary: General Secre- by low nutrient concentrations [16]. Three rivers, Axios, tariat for Research and Technology), and is co-financed Loudias and Aliakmon supply the Thermaikos bay with by the European Social Fund (ESF) and the Greek State. freshwater depending on the season, dam supply and the water extraction for irrigation, contributing to the increase The authors have declared no conflict of interest. of nutrient loads, supporting therefore marine productivity. Consequently, primary productivity in the areas influenced by the rivers is closely related to the variability of the rivers’ load [17]. REFERENCES

In the Northern Aegean, higher values of chlorophyll [1] Valavanis, V., Kapantagakis, A., Katara, A. and Palialexis, concentration have been associated to lower temperatures A. (2004) Critical regions: A GIS-based model of marine [18], as in the present work. It was also found that the sur- productivity hotspots. Aquatic Sciences, 66, 139–148. face chlorophyll maximum happens in late winter due to [2] Volpe, G., Nardelli, B.B., Cipollini, P., Santoleri, R. and the intense mixing of the water column allowing the trans- Robinson, I.S. (2012) Seasonal to interannual phytoplankton response to physical processes in the Mediterranean Sea from portation of nutrients to the upper layers [18]. During win- satellite observations. Remote Sensing of Environment, 117, ter, due to strong cooling events over the shelf areas of the 223–235. Northern Aegean, SST has a considerable decrease lead- [3] Wingfield, D.K., Peckham, S.H., Foley, D.G., Palacios, ing to dense water formation [19]. Recently, a significant D.M., Lavaniegos, B.E., Durazo, R., Nichols, W.J., Croll, relationship between the seasonal cycles of chlorophyll and D.A. and Bograd, S.J. (2011) The Making of a Productivity SST was detected [18]. Finally, the SST and chlorophyll Hotspot in the Coastal Ocean. PlosOne, 6(11), e27874. seasonal cycles amplitudes were found to be much larger in [4] Georgakarakos, S. and Kitsiou, D. (2008) Mapping abun- the Northern Aegean than in the Southern part. dance distribution of small pelagic species applying hydroa- coustics and Co-Kriging techniques. Hydrobiologia, 612, 155–169.

4. CONCLUSIONS [5] Demarcq, H., Reygondeau, G., Alvain, S. and Vantrepotte, V. (2012) Monitoring marine phytoplankton seasonality from space. Remote Sensing of Environment, 117, 211–222. In the Northern Aegean Sea marine productivity [6] van Ruth, P.D., Ganf, G.G. and Ward T.M. (2010) Hot-spots hotspots seem to follow the annual cycle of mixing and of primary productivity: An Alternative interpretation to stratification of surface waters. Tsiaras et al. [17] investi- Conventional upwelling models. Estuarine, Coastal and Shelf gated the inter-annual productivity in the Northern Aege- Science, 90, 142-158. an Sea and their results indicated that the open sea [7] Zervakis, V. and Georgopoulos, D. (2002) Hydrology and productivity variability is related to both vertical mixing circulation in the North Aegean (eastern Mediterranean) and nutrient fluxes, induced by the variability of thermo- throughout 1997 and 1998. Mediterranean Marine Science, 3(1), 5-19. haline circulation. In addition, the impact from the Black Sea Water inflow variability on the Northern Aegean Sea [8] Krasakopoulou, E., Zervakis, V., Souvermezoglou, E. and Georgopoulos, D. (2002) North-eastern Aegean sea: an effort is significant. Negative correlations between high chl-a to estimate steady-state N & P budgets during September concentrations and SST have also been detected in other 1998. Mediterranean Marine Science, 3(1), 43-54. parts of the world [20] and related to potential fishery [9] Valavanis, V.D., Drakopoulos, P. and Georgakarakos, S. locations. In conclusion, further study in this field could (1999) A Study of Upwellings using GIS. In: Proceedings of be carried out by acquiring and processing longer CoastGIS ’99: International Conference GIS and New Ad- timeseries of satellite data. Furthermore, the geographic vances in Integrated Coastal Zone Management, IFREMER, distribution of marine productivity hotspots should be Brest, France. further assessed in relation to fishery resources in order to [10] Ignatiades, L., Psarra, S., Zervakis, V., Pagou, K., Souverme- zoglou, E., Assimakopoulou, G. and Gotsis-Skretas, O. (2002) reveal the relation between marine productivity and major Phytoplankton size-based dynamics in the Aegean Sea (Eastern species catch areas. Mapping of these areas is of great Mediterranean). Journal of Marine Systems, 36, 11–28.

[11] Lykousis, V., Chronis, G., Tselepidis, A., Price, N.B., Theo- charis, A., Siokou-Frangou, I., Van Wambeke, F., Danovaro, R., Stavrakakis, S., Duineveld, G., Georgopoulos, D., Ignati- ades, L., Souvermezoglou, A. and Voutsinou-Taliadouri, F. (2002) Major outputs of the recent multidisciplinary biogeochemical researches undertaken in the Aegean Sea. Journal of Marine Systems, 33–34, 313–334.

[12] Siokou-Frangou, I., Bianchi, M., Christaki, U., Christou, E., Giannakourou, A., Gotsis, O., Ignatiades, L., Pagou, K., Pitta, P., Psarra, S., Souvermezoglou, E., Van Wambeke, F. and Zervakis, V. (2002) Organic carbon partitioning and carbon flow along a gradient of oligotrophy in the Aegean Sea (Mediterranean Sea). Journal of Marine Systems, 33–34, 335–353. [13] EEA (1999) State and pressures of the marine and coastal Mediterranean environment. Environmental Assessment Se- ries, No 5, Copenhagen: European Environment Agency.

[14] Karydis, M. and Kitsiou, D. (2012) Eutrophication and environmental policy in the Mediterranean Sea: a review. Envi- ronmental Monitoring and Assessment, 184, 4931–4984.

[15] Krom, M.D.,Woodwardb, E.M.S., Herut, B., Kress, N., Car- boa, P., Mantourad, R.F.C., et al. (2005) Nutrient cycling in the south east Levantine basin of the eastern Mediterranean: results from a phosphorus starved system. Deep-Sea Re- search PART II-Topical Studies in oceanography, 52(22–23), 2879–2896. [16] Ignatiades, L. (2005) Scaling the trophic status of the Aegean Sea, eastern Mediterranean. Journal of Sea Research, 54, 51– 57.

[17] Tsiaras, K.P., Kourafalou, V.H., Raitsos, D.E., Triantafyllou, G., Petihakis, G. and Korres G. (2012) Inter-annual productivity variability in the North Aegean Sea: Influence of ther- mohaline circulation during the Eastern Mediterranean Tran- sient. Journal of Marine Systems, 96–97, 72–81.

[18] Skliris, N., Mantziafou, A., Sofianos, S. and Gkanasos A.

(2010). Satellite-derived variability of the Aegean Sea eco-

hydrodynamics. Continental Shelf Research, 30, 403–418.

[19] Zervakis, V., Georgopoulos, D. and Drakopoulos, P.G. (2000). The role of the North Aegean in triggering the recent Eastern Mediterranean climatic changes. Journal of Geophys- ical Research, 105(C11), 26103–26116. [20] Boyce, D.G., Lewis, M.R. and Worm, B. (2010) Global phytoplankton decline over the past century. Nature, 466, 591- 596. Received: November 20, 2013 Revised: April 10, 2014 Accepted: May 10, 2014

CORRESPONDING AUTHOR

Dimitra Kitsiou

Department of Marine Sciences School of the Environment University of the Aegean University Hill

Mytilene, Lesvos, 81100 GREECE

Phone/ Fax: +30 22510 36819

E-mail: [email protected]

FEB/ Vol 23/ No 11a/ 2014 – pages 2919 – 2925

TOXICOLOGICAL ASSESSMENT OF THE EFFECTS OF CLOSED LANDFILL ON NEIGHBOURING HYDROECOSYSTEM

Gintaras Svecevičius*, Nijolė Kazlauskienė, Asta Slučkaitė and Tomas Makaras

Institute of Ecology of Nature Research Centre, Akademijos 2, 08412 Vilnius 21, Lithuania

Presented at the 4th International Conference on Environmental Management, Engineering, Planning and Economics (CEMEPE), June 24 to 28, 2013, Mykonos, Greece

ABSTRACT cleaning processes, which can lead to unpredictable environmental changes in the ecosystem [10]. Moreover, Urban waste landfills can remain long-term pollution landfills could remain long-term pollution sources even sources far beyond the termination of their operation. Land- after the termination of their operation [4, 11]. Therefore, fill leachates entering water bodies are becoming one of the it is very important to establish and evaluate the ecologi- main factors of pollution of the aquatic environment. We cal significance of outgoing pollution from closed land- evaluated the effects of the closed Kairiai landfill (Šiauliai fills. For this purpose, physico-chemical and biological- District, Lithuania) on the neighbouring hydroecosystem toxicological parameters should be assessed. (a drainage channel, pond and creek). Water sampling was performed at increasing distances from the leachate The objectives of this study were: (1) to estimate the holding reservoirs in the direction of water flow. Water toxicity of the water of a possibly-affected hydro-eco- physico-chemical parameters and its toxicity level were system neighbouring a closed landfill through a complex of determined. Bioassay testing was carried out on the fol- bioassay tests on aquatic animals (medicinal leeches, lowing aquatic animals: medicinal leech, daphnia and rain- daphnia and fish) and to assess the extent of decline in the bow trout (larvae and juveniles), in each case evaluating toxicity recorded with increasing distance from the landfill their mortality and behavioural responses. An evident cor- in the direction of water; (2) to carry out physico-chemical relation between water quality parameters and its toxicity analysis of the water samples; (3) to perform correlation were found related to the distance from pollution source. analysis between analytical and toxicological data; (4) to predict possible consequences of persistent pollutant migra-

tion on the ecotoxicological state of aquatic ecosystems. KEYWORDS: Landfill; leachate; physico-chemical analysis; bioassay testing; toxicity. 2. MATERIALS AND METHODS

1. INTRODUCTION The Kairiai landfill is located 5 km east of the Ši- auliai City (55°55'42.7", 23°23'42.81", WGS). The land- Urban waste landfills are one of the main environ- fill began operation in 1960 and was closed in 2007. Dur- mental point source polluters posing a hazard to the envi- ing its operation, household and industrial waste contain- ronment [1-4]. Landfills are associated with negative ef- ing toxic substances was deposited in the landfill, includ- fects on the environment and human health, causing soil ing from bicycle and television factories, leather pro- and groundwater pollution [2-8]. Landfill leachates usual- cessing plants and furniture and food industries. In 2002, ly contain persistent (stable) inorganic (heavy metals) and during maintenance work on a retaining dam, an accident organic pollutants and are non-biodegradable. These per- resulted in the dam being destroyed and about 44,000 m3 sistent pollutants are more dangerous as they migrate of leachate leaking out into a drainage channel. It took a from one biological system to another and accumulate in long time to localize this emergency. Still today, the land- aquatic organisms and ecosystems [9]. fill continues to seep leachate and the flow is channelled to pass through local treatment facilities before entering Leachates entering water bodies are becoming one of two isolated holding reservoirs, maintained under open- the main factors in the pollution of aquatic environments, air conditions, and from time to time transported to treat- as they change the chemical composition of the water and ment plants. However, the question arises whether a for- sediment and disturb the biological balance of the self- mer landfill, even one closed quite long ago, could affect water bodies located nearby. Possibly, the landfill leachate * Corresponding author could penetrate through permeable soils from holding

reservoirs and pollute neighbouring water bodies. The Bioassay testing was carried out under controlled la- hydro-ecosystem incorporated in the landfill area under boratory conditions. The following test-organisms and their study consists of a nameless drainage channel surround- test-functions were investigated: medicinal leech Hirudo ing the landfill which flows for 1.5 km into the Ginkūnai verbena, bred under laboratory conditions (cluster for- pond (of 1.1 km2 area). In turn, Švedė Creek flows out of mation, mobility); daphnia – Daphnia magna (mortality, the pond. In total, six sampling sites (No. 0, 1, 2, 3, 4, 5) mobility inhibition); rainbow trout – Oncorhynchus were set at increasing distances from the leachate-holding mykiss, hatchery reared (mortality, behavioural responses reservoirs in the drainage channel, the pond and the creek, in larvae and respiratory–locomotor activity in juveniles). the distances being about 10, 800, 1300, 2200, 2900 and Deep-well water was used as the maintenance and con- 3200 meters along the water flow from the reservoirs trol water. The average hardness of the water was approxi- respectively (Figure 1). mately 284 (271–296) mg/l as CaCO , the alkalinity was 3 200 (190–210) mg/l as CaCO3, the pH ranged from 7.9 to 8.1, the temperature was maintained at approximately 10oC for rainbow trout, and 20oC for daphnia and leech. The dissolved oxygen concentration was no less than 8 to 10 mg/l, and the dissolved organic carbon (DOC) was less than the detection limit (< 0.3 mg/l) [20]. The tests on adult medicinal leech were performed according to the method described in detail by Petrauskienė [21]. The worms were placed individually into 0.5-litre glass vessels filled with test or control water. Seven individuals were used for each sample. Mobility (in %), avoidance response (in %) and the cluster formation (gath- erings of animals in one cluster - in one knot) were estimated. Every test lasted 96 hours and had three replications. The tests on daphnia were performed using ten adult individuals for each water sample placed into 0.5–litre glass vessels filled with test or control water. Mortality and mobility inhibition in the crustaceans was recorded according to standardized procedures [22]. Every test lasted 96 hours and had two replications. The tests on rainbow trout larvae were performed using ten individuals (immediately after hatching) for each water sample placed in 1–litre plastic vessels filled with

FIGURE 1 - The scheme of the study area and sampling sites: land- test or control water. Larvae were incubated in a cold dark fill leachate reservoir (F), drainage channel (site No. 0 and 1), room at a temperature of 10  1.5oC. Mortality, breathing Ginkūnai pond (site No. 2, 3, and 4) and Švedė Creek flowing out of rate (counts/min), nest formation (in %) and response to the pond (site No. 5). external stimuli (in %) in larvae were recorded [23]. Eve-

ry test lasted 96 hours and had two replications. All water samples were collected in June, August, and October, 2012 and underwent complete hydrochemi- All these above–motioned tests were performed un- cal analysis. Multiple measurements of dissolved oxygen der semi-static conditions. (O2 mg/l), pH, temperature (ºC), salinity (‰) and conduc- Juvenile rainbow trout were examined in a flow-through tivity (µS/cm) were performed during the whole test- test apparatus consisting of six isolated test boxes connected season with a hand-held multi-meter (WTW Multi 340i/SET, to one battery with exposure duration of < 1 hour [24]. Res- Germany). The other parameters were measured once in piratory–locomotor responses were measured during 1- summer. The following physico-chemical characteristics min periods for each test fish individually, the mean value of water have been established: permanganate index (mg being calculated for 12 individuals. O/l), steady–state CO2 (mg/l), total hardness as CaCO3 + + 2+ 2+ + Obtained data were analysed statistically using one- (mg/l); cations: Na , K Ca , Mg , NH4 (mg/l, respec- ¯ 2¯ ¯ 2¯ ¯ ¯ tively); anions: Cl , SO4 , HCO3 , CO3 , NO2 , NO3 way ANOVA followed by Dunnett's multiple comparison (mg/l, respectively), as well as total priority heavy metal test at p < 0.05 through STATISTICA Version 6.0 (StatSoft (Cu, Zn, Ni, Cr, Pb, Cd, Hg) concentrations in the water Inc., Tulsa, Oklahoma, USA) software. (mg/l, respectively) and bottom sediments (mg/kg of dry The most sensitive and reliable toxicity indexes were mass, respectively) were determined according to stand- selected for further integral water sample toxicity assess- ardized procedures [12–19]. ment.

According to Sprague [25], the “safe” concentration 20 (DL) to 66, Ni from < 4 (DL) to 21, Cr from 5 to 66, means the concentration of a pollutant which does not Pb from 4 to 8, Cd from 0.15 to 0.40 and Hg from < 0.05 have an adverse sublethal or chronic effect on the fish, and (DL) to 0.13 mg/kg of dry mass, respectively (Figure 4). estimates for some pollutants, especially cumulative and Meanwhile, in neighbouring natural waters, the in- persistent poisons, are as low as 0.01 toxic units (LC50), crease in heavy metal (Cu, Ni, and Cr) concentration was i.e. an application factor (the ratio between chronic maxi- found only in the drainage channel (site No. 0 and 1) and mum-acceptable-concentration and median lethal concen- varied from 1 to 4 µg/l (Figure 4). All these above– tration: MATC/LC50) of 0.01 should be applied. Further- mentioned parameters significantly improved with in- more, many researchers also documented that the “safe” creasing distance from the pollution source (leachate concentration could be successfully estimated using fish holding reservoirs) (Multiple R = 0.84, F = 9.70 and p = behavioural responses, for example, gill-cleaning reflex 0.017). (coughing), etc. [26, 27]. Comparatively high heavy metal (Cu, Ni, Zn and Cr) In this study, we developed a water toxicity level concentrations were found in the Kairiai landfill leachate (Cu classification based on Sprague’s suggestion and our own – 2, Zn – 76, Ni – 100, Cr – 620, Pb – 3 and Hg – 0.2 mg/l, many-year toxicity studies on fish using either mortality respectively) and were close to those medians commonly or behavioural endpoints; the latter comparing with stand- occurring in landfill leachates [1]. However, chromium ard (reference) toxicity test results [as a fraction of 96- concentration was significantly higher, perhaps because hour LC50 using Cu and Zn binary mixture for medicinal the Kairiai landfill was used for many years for industrial leeches and seven metal (Cu, Zn, Ni, Cr, Pb, Cd and Mn ) wastes containing a lot of chromium from leather pro- model mixture for rainbow trout juveniles as reference cessing plants. substances] through regression analysis [21, 28] as fol- The data obtained here conforms to the generally as- lows: sumed statement that due to the high concentration of many  >1 – extremely high toxicity (acutely lethal effects are pollutants, municipal landfill leachates are one of the types observed within 96 hours. Mortality exceeds 50 % of of wastewaters with the highest environmental impact. individuals); Municipal landfill leachates generally contain four main  >0.5<1 – high toxicity (acutely lethal effects are ob- groups of pollutants: I) dissolved organic matter, II) inor- served within 96 hours. Mortality amounts 25-50 % of ganic compounds such as ammonium and chlorides, III) individuals); metals and IV) xenobiotic organic substances [6, 29].  >0.1<0.5 – moderately toxic (acutely lethal effects The obtained data showed that water samples from within 96 hours. Mortality amounts less than 25 % of the drainage channel (sites No. 0 and 1) were the more individuals); toxic for the most of the test–organisms studied. High  >0.01<0.1 – low toxicity (mortality is not observed – mortality in daphnia and fish larvae, as well as significant sub-lethal effects – chronic toxicity zone); changes in behavioural responses of test–animals, has been recorded (Table 1).  <0.01 – non-toxic (“safe” concentration range). Test-organisms demonstrated different sensitivity and responded quite differently to the water samples. Test 3. RESULTS AND DISCUSSION water toxicity level according to the suggested classification could be characterized as follows: Analytical data showed that there was a significant  Drainage channel water (site No. 0) was of “extremely variation in a number of physico-chemical parameters high toxicity” to rainbow trout larvae, “highly toxic” between the site water samples. to daphnia and rainbow trout juveniles, and of "low toxicity" to medicinal leech; Average dissolved oxygen concentration during whole test-season (spring, summer and autumn) ranged from 2.7  Drainage channel water (site No. 1) was "moderately toxic" to daphnia, rainbow trout larvae and juveniles, to 10.0 O2 mg/l, temperature from 13.5 to 17.2ºC, pH from 7.21 to 8.18, conductivity from 706 to 1203 µS/cm, salini- and of "low toxicity" to medicinal leech; ty from 0.03 to 0.35 ‰, permanganate index from 7.92 to  Ginkūnai pond water (site No. 2) was of “extremely high toxicity” to daphnia, "moderately toxic" to rain- 14.9 mg O/l, total hardness from 308 to 438 as CaCO3 bow trout larvae and juveniles, and of "low toxicity" to mg/l, steady-state CO2 from 3.01 to 39.9 mg/l (Figure 2); cations: Na+ from 25.2 to 93.2, K+ from 3.2 to 11.4, Ca2+ medicinal leech; 2+ + from 79.6 to 135.0, Mg from 24.6 to 28.7, NH4 from  Ginkūnai pond water (site No. 3) was “highly toxic” to 0.08 to 5.73 mg/l, respectively; anions: Cl¯ from 52.6 to daphnia, of "low toxicity" to rainbow trout juveniles 2¯ ¯ 172, SO4 from 51.1 to 77.6, HCO3 from 256 to 440, and "non-toxic" to rainbow trout larvae and to medici- 2¯ ¯ CO3 from 0.14 to 0.66, NO2 from 0.01 [less than the nal leech; ¯ detection limit (DL)] to 0.46, NO3 from < 0.05 (DL) to  Ginkūnai pond water (site No. 4) was of "low toxicity" 47.4 mg/l, respectively (Figure 3) as well as total heavy to rainbow trout juveniles and medicinal leech, and metals in bottom sediments: Cu from 7 to 23, Zn from < "non-toxic" to rainbow trout larvae and daphnia;

FIGURE 2 - Physico-chemical parameters of sampling site water. Diagrams A–E represent average parameter values for whole test-season (spring-summer-autumn). Vertical bars denote ± 1 standard deviation.

FIGURE 3 - Ion concentration in sampling site water.

FIGURE 4 - Heavy metal concentration in bottom sediments and sampling site water.

TABLE 1 - Toxicity of Kairiai landfill regional aquatic ecosystem water samples to aquatic animals and multiple correlations with water physico-chemical characteristics (including the distance from pollution source)

Test-organism Medicinal leeches Daphnia Rainbow trout larvae Rainbow trout juveniles Sampling Toxicity index site (No) Mobility (fraction of refer- Mortality (%) Mortality (%) Coughing rate (fraction of reference ence 96-hr LC50 value) 96-hr LC50 value) 0 0.09 45 85 0.86 1 0.07 25 20 0.28 2 0.016 65 15 0.26 3 0.009 45 0.0 0.015 4 0.085 0.0 0.0 0.016 5 0.012 5 0.0 0.023 Multiple R* 0.44 0.64 0.83 0.88 F 0.98 2.76 8.60 14.25 p 0.047 0.021 0.036 0.017 * – in coefficient calculating values of physico-chemical parameters below the detection limit were equalled to zero.

 Švedė Creek water (site No. 5) was of "low toxicity" 4. CONCLUSION to rainbow trout juveniles, daphnia and medicinal leech, and "non-toxic" to and rainbow trout larvae. Summarizing the results of bioassay testing and physico-chemical measurements, it should be pointed out that In general, the data confirmed the existence of a strict the Kairiai landfill remains a serious source of permanent relationship between the integral toxicity level of test pollution, which continues to affect the neighbouring water water and the distance from the leachate holding reser- bodies. Although the landfill is closed, it still continues to voirs, and could be arranged into the following toxicity seep leachate. Despite the former landfill having treatment order (according to the site No.): 0> 1 = 2 > 3> 4 = 5. facilities and the periodic removal of the wastewater to According to Kalčíková et al. [29], the results of eco- municipal facilities, it is evident that part of leachate toxicity testing also showed that various organisms re- penetrates through permeable soils from the holding res- spond differently to landfill leachate and therefore, the ervoirs and pollutes neighbouring water bodies. It is obvi- use of a battery of bioassay tests is essential for a reliable ous that the landfill leachate is still extremely toxic assessment of the effects of landfill leachates on the envi- wastewater, which flows from the holding reservoirs to the ronment. Sensitivity of all studied species to leachate toxic drainage channel and further into Ginkūnai pond and Švedė effects, as obtained by the average LC50 or IC50 mean Creek. Sampling site water toxicity is significantly related values from each species tested, was ranked as: Thamno- to the distance from the pollution source. Water toxicity cephalus platyurus> Dunaliella tertiolecta> Brachionus decreases when the distance from the pollution source plicatilis≥ Artemia franciscana [8]. increases, and in turn the distance from the pollution Therefore, the assessment of water toxicity by biolog- source is significantly related to water physico-chemical ical methods suggests the need to use at least three organ- parameters: water quality increases when the distance isms of different phylogenetic level. At low concentra- from the pollution source also increases. tions, in the absence of mortality, we recommend the use When the landfill leachate enters water bodies, it be- of aquatic organism behavioural responses that are signif- comes the primary and secondary source of pollution and icantly more sensitive than other biological indicators. can cause irreversible changes in the aquatic ecosystem.

Complex physico-chemical and bioassay testing in- [11] Afolayan, O.S., Ogundele, F.O. and Odewumi, S.G. (2012) vestigation could enable to the extrapolation of the ob- Hydrological implication of solid waste disposal on groundwater quality in urbanized area of Lagos state, Nigeria. Inter- tained experimental results in the natural environment in national Journal of Applied Science and Technology, 2(5): order to predict possible consequences of persistent pollu- 74-82 tant migration and the effect on the organism-population- [12] ISO 10304-1:2007. Water quality – Determination of dis- community set for integrated ecotoxicological evaluation of solved anions by liquid chromatography of ions – Part 1: De- the state of aquatic ecosystems in general. termination of bromide, chloride, fluoride, nitrate, nitrite, phosphate and sulfate. ISO, the International Organization for Standardization.

[13] ISO 9963-1:1994. Water quality – Determination of alkalini- ACKNOWLEDGEMENTS ty – Part 1: Determination of total and composite alkalinity. ISO, the International Organization for Standardization.

This work is funded by the Research Council of Lith- [14] 14 ISO 14911:1998. Water quality – Determination of dis- + + 2+ 2+ 2+ 2+ 2+ uania, Project No. MIP-038/2012. solved Li+, Na+, NH4 , K , Mn , Ca , Mg , Sr and Ba using ion chromatography – Method for water and waste wa- The authors have declared no conflict of interest. ter. ISO, the International Organization for Standardization. [15] ISO 8467:1993. Water quality – Determination of permanganate index. ISO, the International Organization for Standardi- zation. REFERENCES [16] ISO 15586:2003. Water quality – Determination of trace elements using atomic absorption spectrometry with graphite [1] Qasim, S.R. and Chiang, W. (1994) Sanitary Landfill Leach- furnace. ISO, the International Organization for Standardiza- ate, Generation, Treatment and Control. Technomic Publish- tion. ISO, the International Organization for Standardization. ing Co. Inc., Lancaster, Pennsylvania, 338 p. [17] ISO 1483:2007. Water quality – Determination of mercury – [2] Slack, R.J., Gronow, J.R., and Voulvoulis, N. (2005) House- Method using atomic absorption spectrometry. ISO, the In- hold hazardous waste in municipal landfills: contaminants in ternational Organization for Standardization. leachate. Science of the Total Environment, 337:119-137 [18] ISO 11047:1998. Soil quality – Determination of cadmium, [3] Vasarevičius, S., Čegariova, J., and Sližytė, D. (2005) Inves- chromium, cobalt, copper, lead, manganese, nickel and zinc – tigation and evaluation of landfill leachate permeability in the Flame and electrothermal atomic absorption spectrometric soil. Journal of Environmental Engineering and Landscape methods. ISO, the International Organization for Standardi- Management, XIII (3):108-115 zation.

[4] Pablos, M.V., Martini, F., Fernandez, C., et al. (2011) Corre- [19] ISO 16772:2004. Soil quality – Determination of mercury in lation between physicochemical and ecotoxicological ap- aqua regia soil extracts with cold-vapour atomic spectrometry proaches to estimate landfill leachates toxicity. Waste Man- or cold–vapour atomic fluorescence spectrometry. ISO, the agement, 31:1841-1847 International Organization for Standardization.

[5] Alkassasbeh, J.Y.M., Heng, L.Y. and Surif S. (2009) Toxici- [20] ISO 8245:1999 Water quality – Guidelines for the determina- ty testing and the effect of landfill leachate in Malaysia on tion of total organic carbon (TOC) and dissolved organic car- behavior of common carp (Cyprinus carpio L., 1758; Pisces, bon (DOC). ISO, the International Organization for Standard- Cyprinidae). American Journal of Environmental Sciences, ization. 5(3):209-217 [21] Petrauskienė, L. (2008) Lethal effects of Zn, Cu and their [6] Singh, V. and Mittal, A. K. (2009) Toxicity analysis and pub- mixture on the medicinal leech (Hirudo verbana). Ekologija, lic health aspects of municipal landfill leachate: A case study 54(2):77-80 of Okhla landfill, Delhi. 8th World Wide Workshop for Young Environmental Scientists WWW-YES 2009: Urban [22] ISO 6341 (1996) Water quality – Determination of the inhi- waters: resource or risks? Arcueil, France. bition of the mobility of Daphnia magna Straus (Cladocera, Crustacea) – Acute toxicity test. ISO, the International Or- [7] Aderemi, A.O., Adewumi, G.A. and Otitoloju, A.A. (2012) ganization for Standardization. Municipal landfill leachate characterization and its induction [23] Kazlauskienė, N., Svecevičius, G., Petrauskienė, L. and Vo- of glycogen vacuolation in the liver of Clarias gariepinus. sylienė, M.Z. (2010) Behavioural responses of medicinal International Journal of Environmental Protection, 2(4):20-24 leech and rainbow trout exposed to crude oil and heavy fuel [8] Tsarpali, V. and Dailianis, S. (2012) Landfill leachate com- oil in ontogenesis. Polish Journal of Environmental Studies, position and toxic potency in semi-arid areas: an integrated 19(2): 429-433 approach with the use of physicochemical and toxicological [24] Svecevičius, G. (2005) Behavioral responses of rainbow trout data. Third International Symposium on Green chemistry for Oncorhynchus mykiss to sublethal toxicity of a model mix- environment, health and development, 3-5 October 2012, ture of heavy metals. Bulletin of Environmental Contamina- Skiathos Island, Greece. tion and Toxicology, 74: 845-852

[9] Bernes, C. (1998) Persistent organic pollutants. A Swedish [25] Sprague, J.B. 1971 Measurement of pollutant toxicity to fish. view of an international problem. III. Sublethal effects and safe concentrations. Water Reseach, [10] Słomczyńska, B. and Słomczyński, T. (2004) Physico- 3: 793-821 chemical and toxicological characteristics of leachates from [26] Atchison, G.J, Henry, M.G., Sandheinrich, M.B. (1987) Ef- MSW landfills. Polish Journal of Environmental Studies, fects of metals on fish behavior: A review. Environmental 13(6): 627-37 Biolology of Fish, 18: 11-25

[27] Scherer, E. (1992) Behavioural responses as indicators of environmental alterations: Approches, results, developments. Journal of Applied Ichthyology, 8: 122-131 [28] Svecevičius, G. (2004) Estimation of toxicity of the Kairiai landfill leachate and polluted by it surrounding region water- body water due to accident spill by use of rapid behavioral responses of juvenile rainbow trout Oncorhynchus mykiss. In: Protection and Control of Water Bodies. Proceedings of In- ternational Scientific Conference, Kaunas, Lithuania, 2004, p. 208-213. (in Lithuanian) [29] Kalčíková, G., Zagorc-Končan, J. and Gotvajn, A.Ž. (2011) Evaluation of landfill leachate quality with battery of bi- otests. Acta Environmentalica Universitatis Comenianae (Bratislava) 19:145-150

Received: June 05, 2014 Accepted: June 18, 2014

CORRESPONDING AUTHOR

Gintaras Svecevičius Institute of Ecology Nature Research Centre Akademijos 2 08412 Vilnius 21 LITHUANIA

Phone: +370 52796312 Fax: +370 52729352 E-mail: [email protected]

FEB/ Vol 23/ No 11a/ 2014 – pages 2926 - 2932

ASSESSING TOOLS AND METHODS FOR WATER QUALITY MONITORING IN MEDIUM SIZED RIVERS. THE CASE OF ARCADIKOS RIVER (MESSINIA, GREECE)

Konstantinos C. Gritzalis*, Evangelia Th. Anastasopoulou, Vasiliki Markogianni and Nikolaos A. Georgiopoulos

Hellenic Centre for Marine Research. Institute of Marine Biological Resources & Inland Waters, 46.7 km Athens - Sounion Ave., 19013 Anavyssos, Greece

Presented at the 4th International Conference on Environmental Management, Engineering, Planning and Economics (CEMEPE), June 24 to 28, 2013, Mykonos, Greece

ABSTRACT 1. INTRODUCTION

Various quality assessment methods have been devel- Arkadikos river basin is located at the western part of oped over the past decades in order to estimate the pollu- Messinia prefecture, covers an area of 183.7 km2 and its tion loads and their impact on river water quality, by taking altitude ranges from 0-1225m, with an average of 409.3m. into account among other things the analysis of pressures, Hilly (63.1%) and semi-mountainous (18.2%) terrain pre- including the discharges coming from small industries and dominate, while plain terrain exhibits a percentage of 13.8%. the leaching from adjacent croplands. In the present research From a geological point of view it is a neotectonic basin we considered such methods for assessing quality with filled with post alpine sediments and surrounded by alpine a special emphasis on monitoring, as the European Di- configurations of the unity Olonos- Pindos [1]. Abundant rective of 2000/60/EU requires. The aim of the study is to flora and fauna occurs, significant habitats and two determine both the physicochemical and biological status NATURA biotopes inside the catchment area. The basin of the river and link it to the current pressures. As an indic- exhibits an asymmetry with the development of a thicker ative ecosystem, River Arkadikos was chosen, located in and longer network located at its northern part, while Messinia Prefecture which flows into Kyparissiakos Gulf from the hydrographic network analysis after Strahler, it and receives great quantities of olive oil wastewater annual- is indicated that Arkadikos River belongs to class 4 [2,3]. ly. This constitutes the major issue concerning river Also an extended abundance of common reeds is present pollution in the prefecture since no treatment method is at the downstream part of the river while maquis vegeta- currently applied and the wastes result in nearby streams. tion prevails at the upper part (Figure 1). A characteristic Several chemical parameters were recorded such as total feature of the basin is its rural character, where olive phenols, pesticide residues of 90 active substances, trace groves occupy the 16.8% and complex culture systems metals, nutrients and particulate organic carbon. Subse- are found at 3.2% [4]. The basic industrial activity in the quently, the assessment method of river quality, River prefecture is the operation of 250 olive oil industries which Habitat Survey was applied and the STAR_ICMi biological consist one of the most important pollution pressures to- indicator was calculated. The downstream part of Arkadikos wards the aquatic system along with the outflow from the R., as it is evident from the physicochemical and biologi- cultivation crops. The olive oil production in the prefecture cal results, was found to be severely downgraded during is estimated to be 50.000 tn annually, most of which takes the winter period of 2011 which can be partly attributed to place at the western part of the prefecture in which the the impact of olive oil waste waters. The system seems to studied river basin belongs. This entails that 250.000 tn of recover during spring, since the values of the parameters olive mill waste waters (OMWW) are being generated were drastically decreased but the pollution burden is annually [5]. The majority of the olive mills (78%) operate shifted towards pesticides whose values were recorded at a three phase system producing great quantities of elevated. OMWW and the current practice is to discharge them in nearby streams [5]. The aim of the present research is to KEYWORDS: classify the river in quality classes according to the physi- river, pollution, habitat survey, nutrients, pesticides. cochemical as well as the biological data obtained, connect the results to the current pressures and attribute which * Corresponding author parameters should be monitored frequently.

and p-coumaric acid = 0,05mg/L.The determination of Gal- lic, Caffeic and p-Coumaric acid was performed using High Purified Liquid Chromatography (HPLC) [6]. Specifical- ly, ion Exchange Chromatography was used, which is a method that allows the separation of ions based on their affinity to the ion exchanger and was performed with the use of Metrohm’s Ion Chromatography instrument [7]. The analysis of the pesticides was performed using liquid chromatography/mass spectrometry (LC-MS and LC-MS/MS), a separation method that combines the physical separation capabilities of liquid chromatography with the mass analysis capabilities of mass spectrometry (MS) [8]. The physicochemical parameters of pH, temperature, conductivity, salinity, total solids and dissolved oxygen were determined in situ with the portable multiparameter instrument of HANNA company HI9828 [9]. Alkalinity and total hardness were calculated by the use of Radiometer TIM900 automatic titrator [10]. The main ions, Ca+2, Mg+2, K+1, +1 -1 -1 - Na , SO4 , Cl and NO3 , were determined by ion chromatography using Metrohm’s ion chromatographer based - –3 + on their retention times [11]. Nutrients (NO2 , PO4 , NH4 ), Total Phosphorus (TP) and Total Nitrogen (TN) were measured photometrically by the MERCK- VEGA 400 spectrophotometer. The method was used for concentrations of P> - + 0,010 mg/l, ΝΟ2 > 0,016 mg/l and ΝΗ4 >0,013 mg/l [12]. Nitrites were determined according to Griess’s Reaction, ammonium according to Berthelot’s reaction and total nitrogen using Koroleff’s method for concentrations Ν FIGURE 1 - Geographical Information System (GIS) morphological >1,0 mg/l [13]. Heavy metals in the soluble phase were map of Arcadikos River Basin in Messinia, Greece determined using Inductively Coupled Plasma Mass Spec- trometry (ICP-MS), Thermo-XSeries II after being fil- tered and acidified. The accuracy of the method was con- 2. MATERIALS AND METHODS firmed by measuring a standard reference river water sample. This method is capable of detecting metals at concen- Two sampling stations were examined, station A, lo- trations as low as one part in 1012 and it includes atomiza- cated close to R. Arkadikos headwaters and station B tion, ionization and ion separation [14]. Particulate Or- located a few meters before the estuaries of the river into ganic Carbon was measured via an elemental analyzer Kyparissiakos Gulf (3.1 km from the coastal zone) (Fig- (CHNS Flash 2000 Thermo scientific) [15]. ure 1). Several chemical and biological parameters were determined during the wet and the dry period of 2.2 Materials and Methods for Sampling Benthic Macroinver- 2011while phenols were measured during the wet period tebrates of 2012 as well. The biological indicator of STAR ICMi The aquatic macroinvertebrate organisms in this study was calculated and the river was classified according to were collected with the RIVPACS method which gives, an the physicochemical data obtained in quality classes. overall image of the landscape, including hydrogeomor- phology, riparian condition, anthropogenic pressures and 2.1 Methods for Chemical parameters land uses using a rectangle framed net (0.25 x 0.25 m) Total phenols were measured using the Folin Ciocal- based on a long arm [16,17]. The macroinvertebrate data, teau method which is based on the oxidation of hydroxy after being identified with appropriate keys, was regis- phenyl groups in an alkaline environment. The absorb- tered together with the data from the AQEM protocol in ance was measured with a UV-Vis spectrophotometer at the AQEMdip 2.6 database and subsequently in the AS- 765nm and the results were expressed as mg of Gallic TERICS program [18]. Finally, the multimetric index acid based on relevant calibration curve [5]. The limit of STAR_ICMi was calculated from the analysis of the ma- detection (LOD) was 0,33mg GAE/L and the limit of quan- croinvertebrates and the biological status of the stations tification (LOQ) 1,09mg GAE/L regarding total phenols was recorded. STAR_ICMi index results from the combi- while the corresponding values for each one of the phenolic nation of six indicators which fully meet the requirements acids were the following: LOD Gallic acid =0,02mg/L, of the WFD [19]. In order to assess the physical character Caffeic acid = 0,02 mg/L and p-coumaric acid = 0,01mg/L; and the habitat quality of Arkadikos R the hydromorphologi- LOQ Gallic acid =0,10mg/L, Caffeic acid = 0,10 mg/L cal method of River Habitat Survey (RHS) was applied

[20]. This method was developed in U.K. and classifies River Nutrient classification system [21,22]. The classifi- the site examined in one of the six different classes from cation system of the chemical- physicochemical status of “Pristine” to “Severally Modified”. High quality is deter- the rivers applied constitutes a development of the classi- mined by habitat features occurring at sites in a predomi- fication system developed by Skoulikidis 2006 [21] and nantly unmodified physical state. The Habitat Modifica- refers to geographical zone 3 to which the study area tion Score and the Habitat Quality Assessment score are belongs [22]. The chemical–physicochemical properties calculated. Data collection is based on a standard 500 m and the dissolved oxygen of the stations were classified in length of river channel. Map information is collected for quality classes according to Tables 1 and 2, respectively. each site and includes grid reference (or latitude, longi- In the case in which the nutrient values were below the tude), altitude, slope, geology, height of source and dis- detection limit of the analytical method, the value that tance from source. During the field survey, features of the was recorded corresponded to the half of the detection channel (both in-stream and banks) and adjacent river limit according to the relevant legislation [23]. For the corridor are recorded. In all, more than 200 compulsory classification of the dissolved oxygen the Norwegian data entries are made at each site, in the form of the pres- system was applied which is the most severe among the ence, absence and (in some case) extent of specific fea- familiar classification systems applied in Europe. For the tures, collectively building a comprehensive picture of qualitative classification of each nutrient and dissolved oxy- habitat diversity and character. gen a value from the scoring system is applied (Table 3). Each quality element, according to the quality class that it 2.3 Classification according to the Greek River Nutrient Clas- belongs, receives a value from column C. For the calcula- sification System tion of the chemical/physico-chemical status of each sta- Classification of the Chemical–Physicochemical station the average value is obtained and categorized accord- tus of Arkadikos R. took place according to the Greek ing to column A.

TABLE 1 - River Classification System based on nutrient concentrations in the geographical zone 3 (Greek River Nutrient Classification System_GR-NCS) [13].

Quality classes towards nutrients High Good Moderate Poor Bad - N-NO3 mg/l ≤0,19 ≤0,89 ≤1,28 ≤1,40 >1,40 - N- NO2 mg/l ≤0,002 ≤0,016 ≤0,021 ≤0,066 > 0,066 + N- NH4 mg/l ≤0,016 ≤0,036 ≤0,063 ≤0,555 >0,555 3- P-PO4 mg/l ≤0,017 ≤0,028 ≤0,058 ≤0,064 >0,064 TP mg/l ≤0,023 ≤0,084 ≤0,092 ≤0,138 >0,138 TN mg/l ≤0,41 ≤2,27 ≤2,90 ≤4,09 > 4,09

TABLE 2 - Classification of dissolved oxygen concentration according to the Norwegian system

Quality classes towards Dissolved Oxygen High Good Moderate Poor Bad Norway mg/l > 9 9–6.4 6.4-4 4-2 < 2

TABLE 3 - Scoring system of quality classes

Scoring System Indicators A B C Classes Score boundaries Average H (High) >4-5 (4.1+5)/2 4.55 G (Good) >3-4 (3.1+4)/2 3.55 M (Moderate) >2-3 (2.1+3)/2 2.55 P (Poor) >1-2 (1.1+2)/2 1.55 B (Bad) <1 1/2 0.5

3. RESULTS AND DISCUSSION villages, olive oil presses etc) factors. The concentration of dissolved oxygen exhibited a rather decreased value Total phenols, the most characteristic compound of (6,35 mg/L) during the wet period of 2011 at station B at olive oil mill waste waters, were recorded at a very high the downstream part of Arkadikos R compared to the dry rate (7.91mg/L) during January of 2012 at station B. This one. Contrary to what one would expect the reduced val- value is far above the limit value (0.5mg/L) according to ues of dissolved oxygen during the winter period are a the most recent European and Greek legislation concern- common phenomenon before the estuaries of certain riv- ing surface waters [24-26]. The HPLC analysis of specific ers in the prefecture of Messinia according to local moni- phenolic acids revealed the existence of p-coumaric acid toring programs [4]. This is mostly due to the pollution (1.39mg/L) during the same period at station B. However, load of untreated OMWW that these rivers carry, as it is there were not detected at either period of 2011 probably proven by the elevated values of phenols, Mn and Cu and because they are drifting away rapidly in running waters. the elevated concentration of POC all of which are compo- Concerning the physicochemical parameters, pH ranged nents of OMWW [5]. Additionally at the downstream part at normal levels (pH ca 8-9), the average temperature was of Arkadikos R. there is almost no vegetation, the abiotic recorded higher during the dry period (ca. 19 °C), com- substrate mainly consists of sand, gravel and pebbles and pared to the wet (ca. 11°C) and salinity remained constant therefore we do not expect to observe a decrease in the and at low levels for both stations during the wet and the dry dissolved oxygen concentration during summer due to periods of 2011 and 2012. Conductivity was recorded at a photosynthetic procedures. The concentration of dissolved slightly higher level during winter of 2011 (ca. 450 μS/cm) oxygen in water is an undeniable indicator of the condition at station A compared to the summer period (<300 μS/cm) and sustainability of the aquatic ecosystem. The minimum while at the downstream part of the river it remained con- concentrations that are required for a number of healthy stant (ca. 540μS/cm). fish is 5-8mg/L [16] and thus special attention should be given at the studied aquatic system since the recorded The values of trace metals, pesticides and dissolved value during winter are close to the lower limit and the oxygen recorded during 2011 are presented at Tables 4 area seems to be affected by discontinuous pollutants. and 5. Low concentrations of all the heavy metals examined (Cr, Mn, Fe, Co, Ni, Cu, Zn, Cd, Pb) were recorded The stations also contained a wide range of pesticide in the water body of Arkadikos river in both sampling residues (Table 5). During the winter period of 2011 in periods. The water basin of Arkadikos has the lowest station A all pesticides were found below the detection pollution load towards heavy metals compared with the limit, however during summer the following five pesti- other rivers examined in the area [3]. A slight increase in cides were recorded in the amounts shown at Table 5. the majority of trace metals during the wet period con- Downstream of Arkadikos River, at station B, three pesti- firms that the basic heavy metal input comes from the cides were detected during winter and seven during sum- terrestrial environment through soil leaching. However, a mer in rather small concentrations. Thus, the pesticides major increase was recorded regarding Mn and a slight residues exhibit an increase during the dry period mostly increase in the concentration of Cu at station B during near the river’s estuaries. As far as the biological classifi- winter. The concentrations recorded in both stations were cation of the stations is concerned the upstream part of the found to be much lower than the limit values that apply river was classified as “Moderate” during the wet period for surface inland waters according to the European legis- and as “Good” during the dry period subject to STAR_ICMi lation [24]. At station B the value of POC was recorded biological indicator (Table 6). This differentiation is proba- elevated during the wet period of 2011 (884,6 mg/l) probably due to a random event like rainfall. The downstream bly due to the input of natural and anthropogenic organic part of the river though was classified as “Poor” at both load (olive oil tree cultivation, various cultivations, small periods since a limited number of taxa were present at a

TABLE 4 - Concentrations of Trace Metals and Dissolved Oxygen at Arkadikos R.

Cr Mn Fe Co Ni Cu Zn Cd Pb D.O.

Station μg/L μg/L μg/L μg/L μg/L μg/L μg/L μg/L μg/L mg/L

A_ Jan 2011 0,153 3,740 3,407 0,152 0,475 0,590 n.d. n.d. 0,007 6,950

A_ June 2011 0,106 7,444 6,876 0,144 0,427 0,510 n.d. n.d. 0,021 9,160

Average value 0,130 5,592 5,142 0,148 0,451 0,550 - - 0,014 8,055

B_January 2011 0,166 25,980 6,438 0,180 0,527 0,926 n.d. 0,002 0,072 6,350

B_June 2011 0,193 4,959 10,560 0,132 0,445 0,320 0,659 n.d. 0,057 8,990

Average value 0,180 15,470 8,499 0,156 0,486 0,623 0,659 - 0,065 7,670

TABLE 5 - Concentrations of Pesticides at Arkadikos R.

Pesticides A B Pesticides A B January 2011 μg/L μg/L June 2011 μg/L μg/L Abamectine n.d. n.d. Fenbuconazole n.d. 0,006 Bifentrine n.d. 0,025 Formetanate 0,007 n.d. Carbedazim n.d. n.d. Imidaclorpid 0,005 n.d. Diazinon n.d. n.d. Methoxyfenozide n.d. 0,006 Dimethoate n.d. 0,008 Dimethoate n.d. 0,009 Diphenylamine n.d. n.d. Propamocarb 0,006 0,005 Metalaxyl n.d. 0,014 Metalaxyl 0,066 0,015 Methamidophos n.d. n.d. Methamidophos 0,011 n.d. Omethoate n.d. n.d. Omethoate n.d. 0,02 n.d. not detected Tebuconazol n.d. 0,007

TABLE 6 - Classification of Arkadikos River in reference to STAR_ICMi index Biological Indicator, RHS and the Greek River Nutrient Classification System (NCS_GR) [21, 22].

RHS RHS NCS_GR NCS_GR STAR_ICMi STAR_ICMi River Station Wet period Dry period Wet period Dry period Wet Period 2011 Dry Period 2011 2011 2011 2011 2011 Arkadikos A Moderate Good Good Good 3.84/Good 4.12/High Arkadikos B Poor Poor Poor Poor 3.84/Good 3.84/Good

relatively low abundance, something that can be partly sessment showed that the same station is in a “Poor” attributed to the pressures which mainly involve the oper- condition due to the severe lack of habitats especially ation of numerous olive oil industries and untreated sew- during winter. Heavy metals, pesticides and phenols were age from small settlements to a much lesser extent. Ac- not detected at levels of concern during 2011. The in- cording to the results of the River Habitat Survey applica- creased concentration of phenols during the next wet tion the Habitat Quality Assessment (HQA) score and the period and the high levels of particulate organic carbon Habitat Modification score (HMS) were low, at the up- during winter indicate that the pollution is seasonal. Ra- stream part of the river and elevated at the downstream ther low concentrations of several pesticides were detect- part classifying the station as “Poor”, which means that ed with the highest values to be recorded during the dry station B was found to be “modified” while station A period. In general, we can conclude there are no severe exhibited a much more natural character (Table 6). Con- pollution incidents in the area examined and the aquatic cerning the physicochemical classification based on nutri- ecosystem can thoroughly recover completely if the dis- ents, the springs of the river (Station A) were found to be continues sources of pollution are totally absent. The at a “High” status during summer and at a “Good” status pollution burden arises mainly from phenols, particulate during winter. Station B was also found to be at a “Good” organic carbon, ammonia and Mn during winter while the status at both samplings and the small increase in the only parameter that exhibited an increase during summer concentration of N-NH4 (0.04 mg/l) during summer did was the pesticides residues. Also a most complete picture not influence its overall classification. Taking both sam- arises by taking into account all the above tools and pling campaigns into account the stations were classified methods and not one by one. From all the parameters as “Good” during 2011. measured in the present study an emphasis should be given to the biological results and the abundance of the macroinvertebrates families as well as to the nutrient, 4. CONCLUSIONS phenol and pesticides levels in order to determine more clearly the pollution sources for each season and poten- Overall Arkadikos River seems to be at a rather good tially develop a multimetric index that combines all the ecological status despite the numerous active olive oil above data. industries, the large agricultural areas and the untreated domestic sewage that derives from small settlements. The authors have declared no conflict of interest. However, the quality of the river should be continually reviewed due to the majority of pressures and the “Poor” biological quality that was recorded before the estuaries of the river. According to the biological results, which REFERENCES capture at a greater depth the situation that prevails in a [1] Institute of Geological and Mineral Exploration and Surveys, station, station B was classified as “Poor” both the wet I.G.M.E. 1:50.000 Sheets of Meligalas, Filiatra Kiparissia and the dry period of 2011. Also the Habitat Quality As- and K.Figaleia. http://www.igme.gr retrieved on 2 OCT 2013

[2] Ministry of the Environment, http://www.minenv.gr Data- [20] Raven, J., Holmes, N., Dawson, F., Fox, P., Everard, M., base of NATURA regions 2000 retrieved on 2 OCT 2013 Fozzard, R.and Rouen, K. (1998) River Habitat Quality: the physical character of rivers and streams in the UK and isle of [3] Gritzalis, K.C. (2011) Editor-Coordinator of the Program: Man. River Habitat Survey Report No 2. Environment Agen- Monitoring of the Quality and Assessment of the Ecological cy, UK. Status of the Messinian Rivers in South Peloponnese, Greece (Pamisos, Aris, Ligdou, Epis, Karias, Tzanes-Polylimnio, [21] Skoulikidis, N.Th., Amaxidis, Y., Bertahas, I., Laschou, S., Maurozoumena, Despotis, Mourtia, Arkadikos, Nedas & Ve- and Gritzalis K.C. (2006). Analysis of factors driving stream likas), Hellenic Centre for Marine Research, Institute of In- water composition and synthesis of management tools – A lands Waters and Biological Resources, Anavyssos, Greece. case study on small/medium Greek catchments. The Science of the Total Environment 362: 205-241. [4] Anastasopoulou E.T. (2012) Effects of oil olive wastewaters in coastal environments: The case of Messinia. PhD Thesis, [22] Laschou, S. (2010) Identifying Reference Conditions and University of Athens, pp.46, 91-96, 99-108. Classification System Development of Nutrient for Assess- ment of Chemical-physico-chemical state of the Greek rivers. [5] Folin, O. and Ciocalteu, V. (1927). On tyrosine and trypto- Master Thesis, University of Athens and Hellenic Centre for phan determination in protein. The Journal of Biological Marine Research, Athens. Chemistry 73: 627-50 [23] Directive 2009/90/EC (2009) Laying down technical specifi- [6] Vinha A.F., Silva B.M., Andrade P.B., Seabra R.M., Pereira cations for chemical analysis and monitoring of water status J.A. and Oliveira M.B., (2002) Development and evaluation in accordance with Directive 2000/60/EC of the European of an HPLC/DAD method for the analysis of phenolic com- Parliament and of the Council of 31 July 2009. pounds from olive fruits Liq. Chrom. and Rel. Technology 25 (1), 151-160. [24] Directive 2008/105/EC (2008) of the European Parliament and of the Council of 16 December 2008 on environmental [7] Metrohm Ion Analysis, I.C. (2011) Cap 2.2. quality standards in the field of water policy, amending and http://www.metrohm.com subsequently repealing Council Directives 82/176/EEC, [8] Arpino, P. (1992) Combined liquid chromatography mass 83/513/EEC, 84/156/EEC, 84/491/EEC, 86/280/EEC and spectrometry. Part III. Applications of thermospray. Mass amending Directive 2000/60/EC of the European Parliament Spectrometry Reviews 11: 3. and of the Council. [9] www.hannainst.com retrieved on 2 May 2012. [25] Gov. Sheet No. 1909, Issue B, Greece (2010). Determining Environmental Quality Standards (EQS) for concentrations [10] Standard Methods for the Examination of Water and of certain pollutants and priority substances in surface waters th Wastewater (1971), 13 Ed. American Public Health Associ- in compliance with the provisions of Directive 2008/105/EC ation, Washington, 306-309. of the European Parliament and of the Council of 16 Decem- [11] Jackson, P. and Haddad, Paul R. (1990). Ion chromatog- ber 2008 on environmental quality standards (EQS) in the raphy: principles and applications. Amsterdam: Elsevier. field of water policy and on the concentrations of specific ISBN 0-444-88232-4. pollutants into inland surface waters and other devices, pp. 29265-29280. [12] ΜΕRCΚ, Methods of Chemical Analysis (2009), Method No 14752 (NH4), No 14729 (PO4), No 14537 (TN), No 14776 [26] Karaouzas I., Cotou E., Albanis, T., Kamarianos, A., Skoulikidis, N.T. and Giannakou, U. (2010) Bioassays and (NO2), No 14794 (SiO2) Manual Spectroquant Vega 400. Biochemical Biomarkers for Assessing Olive Mill and Citrus [13] Koroleff, F. (1970). Revised version of Direct determination Processing Wastewater Toxicity, Environ. Toxicol., 669-676. of ammonia in natural waters as indophenol blue. Interna- tional Council for the Exploration of the Sea C. M. 1969/ C:9 ICES information on techniques and methods for sea water analysis. [14] Foerstner, U. and Wittmann, G.T.W. (1983). Metal Pollution in the Aquatic Environment, Springer-Verlag, Berlin. [15] Cutter, G.A. and Radford–Knoery J., (1991). Determination Received: June 05, 2014 of carbon, nitrogen sulfur and inorganic sulfur species in ma- Accepted: June 18, 2014 rine particles, Analysis and Characterization. Am. Geophys. Union, Geophys. Monogr. 63: 57-63 [16] Armitage, P.D., Moss, D., Wright, J.F. and Furse, M.T. (1983) CORRESPONDING AUTHOR The performance of a new biological water quality score system based on macroinvertebrates over a wide range of unpol- Dr. Konstantinos Gritzalis luted running-water sites. Water Resources 17: 333-347. Hellenic Centre for Marine Researc [17] Alba-Tercedor, J. and Pujante, A.M. (2000) Running water Institute of Marine Biological Resources & Inland biomonitoring in Spain: opportunities for a predictive ap- Waters proach. In assessing the biological quality of fresh waters: RIVPACS and other techniques. Wright, J.F., Sutcliffe, D.W. 46.7 km Athens - Sounion Ave. and Furse, M.T. Freshwater Biology Association, Ambleside 19013 Anavyssos 207-216. GREECE [18] http://www.aqem.de and www.eu-star.at visited on 13 Sep 2012. Phone: +302291076394 [19] Buffagni, A., Erba, S. and Furse, M. (2007) A simple proce- Fax: +30 2291076419 dure to harmonize class boundaries of assessment systems at E-mail: [email protected] the pan-European scale, Environmental Science and Policy, 10: 709-724. FEB/ Vol 23/ No 11a/ 2014 – pages 2933 - 2938

SUBJECT INDEX

A L accidents 2840 leachate 2926 Agia Sofia axis 2826 agricultural products 2890 M aquatic invasive species 2895 mapping systems 2814 mapping 2919 B marine pollution 2909 ballast water 2895 marine productivity 2919 benthic macroinvertebrates 2904 market research 2870 bioassay testing 2926 materials exchange 2876 biological index 2904 microbial biomass 2852 biomonitoring 2904 microbial respiration 2852 mortality 2859 C chromate reduction 2865 N collection 2884 natural environment 2814 comparative study 2890 near well flow 2847 cost-benefit analysis 2890 nitrogen pollution 2899 noise 2803 D North Aegean Sea 2919 decision making 2884 nutrient leaching fraction 2899 nutrients 2933 E earthworm 2859 O eco label 2870 Oil spill 2909 environmental impacts 2826 oily waste 2852 environmental monitoring 2809 operational efficiency 2847 exchange 2890 P F pavement 2803 frequency 2840 pedestrian infrastructure 2819 fuzzy logic 2840 pesticides 2933 physico-chemical analysis 2926 G pollution 2933 geo-database 2814 port waste reception facility 2895 geo-database 2919 Port 2840 GIS 2814 GIS 2884 Q green products 2870 Questionnaire survey 2819 grey water footprint 2899 growth 2859 R rational well spacing 2847 H recyclable products 2890 habitat survey 2933 recycling 2884 hexavalent chromium 2865 remote Sensing 2809 home composting 2876 remote sensing 2814 hot spots 2919 remote Sensing 2909 reproduction 2859 I river 2933 image processing 2909 road construction 2803 iron sulphate 2865 Romania 2870

L S landfill 2926 SAR 2909 Larissa city 2819 ship-generated waste 2895

AUTHOR INDEX

S A shoreline position 2809 Abeliotis, Costas 2876 soil column 2852 Anastasopoulou, Evangelia Th. 2933 soil microbial biomass 2865 Aravossis, Konstantinos 2890 soil remediation 2847 space-time model 2832 B spatial analysis 2919 Barsan, Narcis 2870 Bikaki, Nantia 2876 T technologically enhanced naturally C occurring radioactive materials (TENORM) 2852 Caraman, Iuliana 2870 Thessaloniki’s urban zone 2832 Chroni, Christina 2876 toxicity 2859 Coutsikos, Philipos 2870 toxicity 2926 traffic calming 2826 D traffic prediction per time period 2832 Dagdilelis, Emmanouil 2884 triclosan 2859 Darbra, Rosa Mari 2840 two-phase flow in porous media 2847 Doğan-Sağlamtimur, Neslihan 2895 Drakopoulos, Panos 2809 U uncertainty 2840 F urban regeneration 2826 Feleki, Eleni 2884 Fountzoula, Christina 2890 W walking conditions 2819 G waste bins 2884 Galitskaya, Polina 2852 waste campaign 2876 Gavanas, Nikolaos 2819 waste prevention 2876 Gavanas, Nikolaos 2826 water footprint 2899 Georgiopoulos, Nikolaos A. 2933 water quality 2904 Georgiou, Anastasios 2876 Ghionis, George 2809 González Dan, José R. 2840 Grigoriadis, Nikolaos 2884 Gritzalis, Konstantinos C. 2933 Gumerova, Raushaniya 2852

I Inglezakis, Vassilis 2870

K Karkanias, Christos 2884

Kazlauskienė, Nijolė 2926 Kehagia, Fotini 2803 Kitsiou, Dimitra 2919

Kungolos, Athanassios 2904 Kyriakou, Kalliopi 2832

Lakakis, Konstantinos 2832 Lasaridi, Katia 2876 Laspidou, Chrysi S. 2899 Lazar, Gabriel 2870 Lazogiannis, Kostas 2809

Leita, Liviana 2865 Loizia, Pantelitsa 2876

M Makaras, Tomas 2926 Manthos, Evangelos 2803 Margon, Alja 2865 Markogianni, Vasiliki 2933 Miškelytė, Diana 2859 Mondini, Claudio 2865 Moussiopoulos, Nicolas 2884 Moysiadis, Athanasios 2814

N Nedeff, Valentin 2870

O Ozsoy-Cicek, Burcu 2909

P Panainte, Mirela 2870 Perakis, Konstantinos 2814 Perkoulidis, George 2884 Phanou, Korina 2876 Pitsiava-Latinopoulou, Magda 2826 Poulos, Serafim 2809

R Rumbos, Christos I. 2904

S Satır, Tanzer 2895 Savvaidis, Paraskevas 2832 Sdoukopoulos, Alexandros 2819 Sdoukopoulos, Alexandros 2826 Selivanovskaya, Svetlana 2852 Sequi, Paolo 2865 Sinicco, Tania 2865 Skouras, Eugene D. 2847 Slučkaitė, Asta 2926 Stafylas, Stavros 2884 Svecevičius, Gintaras 2926

T Topouzelis, Konstantinos 2919 Tsakalidis, Anastasios 2819

V Valavanides, Marios S. 2847 Valentini, Massimiliano 2865 Venetis, Christos 2870 Voukkali, Irene 2876 Z Žaltauskaitė, Jūratė 2859 Zorpas, Antonis A. 2876

2941