Dairy Farm Effluents and Groundwater
Contamination – Insights from the
Vadose Zone
Thesis submitted in partial fulfillment
of the requirements for the degree of
"DOCTOR OF PHILOSOPHY"
by
Shahar Baram
Submitted to
The Senate of Ben-Gurion University of the Negev
10 April 2013
Beer-Sheva Dairy Farm Effluents and Groundwater
Contamination – Insights from the
Vadose Zone
Thesis submitted in partial fulfillment
of the requirements for the degree of
"DOCTOR OF PHILOSOPHY"
by
Shahar Baram
Submitted to
The Senate of Ben-Gurion University of the Negev
Approved by the advisors:
______
______
______
Approved by the Dean of the Kreitman School of Advanced Graduate Studies:
______
10 April 2013
Beer-Sheva This work was carried out under the supervision of Dr. Ofer Dahan and Prof. Zeev Ronen Department of Environmental Hydrology and Microbiology, The Zuckerberg Institute for Water Research (ZIWR) Faculty: Jacob Blaustein Institutes for Desert Research, and Dr. Daniel Kurtzman Institute of Soil, Water and Environmental Sciences, Agricultural Research Organization, The Volcani Center, Bet Dagan 50250, Israel.
i Research-Student's Affidavit
Research-Student's Affidavit when Submitting the Doctoral Thesis for Judgment
I Shahar Baram, whose signature appears below, hereby declare that (Please mark the appropriate statements):
_X_ I have written this Thesis by myself, except for the help and guidance offered by my Thesis Advisors.
_X_ The scientific materials included in this Thesis are products of my own research, culled from the period during which I was a research student.
___ This Thesis incorporates research materials produced in cooperation with others, excluding the technical help commonly received during experimental work. Therefore, I am attaching another affidavit stating the contributions made by myself and the other participants in this research, which has been approved by them and submitted with their approval.
Date:______Student's name:______Signature:______
ii Acknowledgments
Acknowledgements
"Water is H2O, hydrogen two parts, oxygen one, but there is also a third thing, that makes water and nobody knows what that is": D.H. Lawrence (1885-1930), Pansies, 1929
First of all, I thank my wife and best friend, Ruth, for accompanying me in this adventure! Together, we found a little piece of heaven on earth where we formed our family and established our home. Many thanks to my supervisors, Ofer, Zeev and Dani, for being wonderful human beings and for their gracious guidance and kind help throughout the time of my studies and research at the Zuckerberg Institute for Water Research. In their unique way, they made this work an enjoyable quest for answers to questions. To all the students, researchers and staff of the Jacob Blaustein Institutes for Desert Studies – too many to name – who guided, assisted, and in innumerable ways, helped me throughout my research. Special thanks to Yuval Shani and Michael Kugel for their extensive help in all the technical aspects of my research. Special thanks to the dairy farm owner for allowing me to conduct my field research at his farm. This research was made possible through funding by The Israel Water Authority.
iii Table of contents
Table of Contents
Research-Student's Affidavit ...... ii Acknowledgements ...... iii Table of Contents ...... iv List of Figures ...... v List of Abbreviations ...... vi Abstract ...... vii 1. Introduction ...... 1 1.1 Scientific background ...... 1 1.2 Objectives ...... 4 1.3 Significance of the research ...... 5 1.4 Geohydrological background ...... 5 1.5 The degradation in the groundwater quality of Israel’s Coastal Aquifer ...... 7 1.6 Study site ...... 8 1.7 The dissertation structure ...... 9 2. Published Papers ...... 11 3. Discussion and Summary ...... 55 3.1 Key observations and findings ...... 55 3.2 Conclusions ...... 61 3.3 Future perspectives ...... 62 4. Additional References (not included in published papers) ...... 63 5. Appendix A ...... 67
iv Figures
List of Figures
Figure 1. Map of Israel with its coastal aquifer...... 6
Figure 2. Schematic cross section of the coastal aquifer in the study area...... 6
- - Figure 3. Average nitrate (NO3 ) and chloride (Cl ) concentrations with time in
operational cell No. 82 of Israel’s coastal aquifer...... 8
v Abbreviations
List of Abbreviations
AMX Anammox-Bacteria AOA Ammonia-Oxidizing Archaea AOB Ammonia-Oxidizing Bacteria BLS Below Land Surface BOD Biological Oxygen Demand CAFO Concentrated Animal Feeding Operations CEC Cation Exchange Capacity CND Coupled Nitrification-Denitrification CWS Crack-Water Sampler DCIS Desiccation Crack Induced Salinization GIS Geographic Information System GWR Geographically Weighted Regression IDW Inverse Distance Weighted LPI Local Polynomial Interpolation N Nitrogen NOB Nitrite-Oxidizing Bacteria RBF Radial Basis Function TN Total Nitrogen VMS Vadose-zone Monitoring System WHO World Health Organization
vi Abstract
Abstract Earthen waste lagoons are commonly used to store liquid wastes from concentrated animal feeding operations (CAFOs), such as dairy farms. To reduce leakage from such lagoons, soil liners are normally constructed from heavy clay soils, which are regarded as less permeable due to their low saturated hydraulic conductivities. Unfortunately, it has repeatedly been shown that this practice fails to prevent substantial leaching of pollutants into the subsurface and leads to groundwater contamination. One of the main concerns with seepage from dairy earthen lagoons is the high concentrations of nitrogen-species [N-species: organic-nitrogen (organic-N), + ammonia (NH3) and ammonium (NH4 )] in the lagoon leachates. Studies on the fate of N-species in the vadose zone and groundwater underlying dairy earthen waste lagoon environments have shown that their fate can vary, both spatially and temporally between lagoons and within a single lagoon. In this study, water flow, contaminant transport and the fate of N-species in the vadose zone and groundwater underlying a dairy waste lagoon were studied using long-term in-situ measurements and supporting lab experiments. The study initially investigated the physical characteristics of the desiccation-crack networks that form naturally in the unsaturated clay sediments underlying the waste sources and their role in water infiltration and percolation mechanisms in the vadose zone. In-situ monitoring of temporal changes in the water content profile throughout the unsaturated zone indicated fast (m h-1) and deep (>12 m) propagation of water during intensive rain events and during fluctuations in wastewater level. The water percolation pattern indicated that a substantial amount of water crosses the clay soil and recharges the underlying calcareous sandstone. It was further suggested that desiccation cracks cross the entire clay layer, and subsist on land and in the subsurface year-round, even during the wet winter and at near saturation conditions. In addition to temporal preferential infiltration events, continuous slow (mm d-1) infiltration of wastewater was observed underlying the permanently flooded sections of the waste lagoon (away from its margins). Wastewater percolation from the lagoon bed was controlled by the hydraulic conductivity of fine organic material that settled on it and decreased the permeability of the clay matrix. As a result of low flux from the permanently flooded zones, unsaturated conditions prevailed immediately below the subsurface year round.
vii Abstract
Frequent sampling of the sediment pore water using in-situ monitoring devices, along with sediment core samples, indicated that all the N-species (organic- + N, NH3 and NH4 ) that leached into the subsurface underlying the lagoon surroundings were oxidized within the upper 0.5 m of the vadose zone. The oxidation - was followed by NO3 accumulation. Ammonia-oxidizing bacteria (AOB) and 7 -1 anammox-bacteria (AMX) were most abundant (10 gene copies g dry sediment) in the top soil (<0.2 m) and decreased significantly with depth (down to 0.5 m). On the other hand, the numbers of ammonia-oxidizing archaea (AOA) were relatively constant 7 -1 throughout the soil profile (10 gene copies g dry sediment). The microbial abundance - and the complete oxidation of the infiltrating N-species indicated that NO3 accumulation in the vadose zone probably resulted from complete aerobic nitrification - (NH3 oxidation to NO3 ) and as a byproduct of anammox activity. The microbial - abundance and the NO3 accumulation also indicated that the formation of desiccation cracks in the clay sediment enhanced the aeration of the subsurface even at near saturation conditions. Coupled nitrification denitrification (CND) reactions were - found to be the key factor regulating the fate of NO3 in the vadose zone underlying earthen waste lagoons. Under low water content (45% saturation in this study), CND leads to minor removal of the propagating N mass; under intermediate water contents (70% saturation in this study), CND leads to 90% removal of the propagating N mass; and under high water contents (90% saturation in this study), CND leads to nearly complete removal of the propagating N mass. Even though CND led to substantial - removal of NO3 from the vadose zone and groundwater, observations made in the region underlying the lagoon indicated that the dairy waste lagoons serve as a point - source for groundwater contamination by NO3 . Chloride concentration in pore-water from the subsurface under the waste lagoon and its surroundings exhibited up to a 5.5 fold increase along the clayey vadose zone. Water stable isotope (δ18O and δ2H) values in pore water and in sediment samples showed isotopic enrichment with depth, and indicated subsurface evaporation down to ~3 m below land surface (BLS). Daily oscillations in the thermal gradients between the air temperatures at the land surface and the air temperatures in the crack voids were found to be sufficient to trigger thermally driven convective air flow in the crack voids, and to drive subsurface evaporation. A conceptual model termed: desiccation-crack-induced salinization (DCIS) was suggested to explain the
viii Abstract field observations. The DCIS conceptual model supports previous conceptual models on vadose zone and groundwater salinization in fractured rock in arid environments and extends its validity to clayey soils in semi-arid environments. The infiltration of relatively salty water (1600 mg L-1) from the lagoon surroundings, as well as the relatively high water content underneath it, enhances the subsurface evaporation process and allows substantial accumulation of salts in the deep vadose zone. Accordingly, the waste lagoon serves as a point source for groundwater salinization. The results from the detailed monitoring of the vadose zone underlying the waste lagoon and its surroundings and from the chemical composition of the upper eight meters of the groundwater under the lagoon were combined with regional groundwater surveys to assess the regional impact of dairy waste lagoons on - - groundwater quality. Regional Cl and NO3 mass balances were based on Geographic-Information-System (GIS) interpolations and showed that from the initiation of the dairy industry in the 1960s up until 2010, leachates from dairy - - lagoons have contributed 5.6 % and 14 % of the total mass of Cl and NO3 , respectively, that was added to and pumped from the regional groundwater combined. Close examination of the ratio between the mass that has reached the groundwater from dairy waste lagoon leachates and the mass added to the groundwater under each settlement in the Beer-Tuvia region, from 1960 to 2010, indicated average - - contribution of 19 % Cl and 42 % NO3 . These high values indicated that due to the ongoing pumping in the region, the whole area of the settlement, rather than just the - - area underlying the lagoons, should be addressed as a point source for Cl and NO3 contamination. This assumption was strengthened by the low coefficient of
2 - - determination (R ) found between the Cl and NO3 concentrations in the groundwater and the distance to dairy farms, despite the fact that the dairy waste lagoon is a point
- - source for groundwater contamination by Cl and NO3 . Calculations also indicated - - that 5.7 % and 14.1 % of the Cl and NO3 masses respectively, that are stored in the vadose zone of the region today are located in the vadose zone under dairy lagoons. Accordingly, drying of the lagoons would have a relatively minor effect on the regional groundwater salinization process and a considerable effect on the - groundwater contamination by NO3 . Nonetheless, the drying up of lagoons would stop the point source salinization and contamination of the groundwater.
ix Abstract
Keywords: clay sediment, desiccation cracks, preferential flow, subsurface salinization, subsurface evaporation, waste lagoons, groundwater contamination, dairy farms.
x Introduction
1. Introduction
1.1 Scientific background
Groundwater is the most important fresh water source in large parts of the world. While in the majority of arid and semiarid environments, groundwater forms the only available natural water resource, in humid environments, groundwater exploitation is preferred over surface water. For example, in most European countries, groundwater use exceeds 70% of the total water consumption (Zektser and Everett, 2004). Groundwater usage has a number of essential advantages when compared with surface water: it is usually of higher quality, better protected from possible pollution including infectious agents less subject to seasonal and perennial fluctuations, and much more uniformly spread over large regions than surface water (Zektser and Everett, 2004). Despite its advantages, there is an increase in the number of cases in which the quality of groundwater resources has been found to deteriorate over time. In Israel, 58–70% of the fresh water consumed annually (domestic, industrial and agricultural uses) is drawn from groundwater resources, of which ~20% comes from the Coastal Aquifer, which is the country's largest fresh-water storage volume (Shavit and Furman, 2001; Weinberger, 2007). The quality of the groundwater in Israel’s Coastal Aquifer is affected by several hydrological processes: salinization due to over-exploitation over the years (resulting from limited flushing of salts to the sea, and from sea water penetration from the west) (Weinberger, 2007), salinization from natural in-land sources (Melloul and Wollman, 2003; Vengosh and Ben-Zvi, 1994), salinization due to the flushing down of salts from the vadose zone during the initial cultivation of virgin soils and reclamation of swamps (Kanfi et al., 1983; Kurtzman, 2011; Kurtzman and Scanlon, 2011), a decrease in the recharge due to urbanization (Weinberger, 2007) and continuous pollution from anthropogenic activity (from industry, agriculture, and urbanization) (Bernstein et al., 2011; Kass et al., 2005; Melloul and Wollman, 2003; Ronen et al., 1983; Vengosh and Ben-Zvi, 1994; Wilkison and Blevins, 1999). The Coastal Aquifer underlies the most populated, and until recently, the most intensively cultivated and industrialized area of the state of Israel.
1 Introduction
Intense agricultural activity and concentrated animal feeding operations (CAFO) are known to have a negative impact on groundwater quality (Almasri and Kaluarachchi, 2004b; Baker and Hawke, 2007; Burkart and Stoner, 2007; Burkholder et al., 2007; Harter et al., 2002; Melo et al., 2012; van der Schans et al., 2009). Such facilities have centralized feeding systems where animals are concentrated in small spaces and most of their food is supplied from external sources. The direct unavoidable result is large amounts of solid and liquid waste that cannot be treated in domestic waste treatment plants due to its very high biological oxygen demand (BOD). Hence, it is either being stored in unlined/soil-lined anaerobic waste lagoons in the dairy farm area, or being spread as a readily available fertilizer to nearby agricultural fields. These practices are considered to have potential polluting impacts on the underlying ground water quality (Almasri and Kaluarachchi, 2004a; Bobier et al., 1993; Chang and Entz, 1996; DeSutter et al., 2005; Harter et al., 2002). Previous studies have indicated that earthen waste lagoons leak to the subsurface, and that the infiltration rate is controlled by the hydraulic properties of the natural soil or the constructed soil-liner (DeSutter and Pierzynski, 2005; DeSutter et al., 2005; Ham and Baum, 2009; Ham and DeSutter, 2000), the lagoon’s structure and the waste management regime (Gooddy et al., 2002; Gooddy et al., 1998; Korom and Jeppson, 1994; Parker et al., 1999a; Parker et al., 1999b) and by physical, chemical, and biological processes that reduce the hydraulic conductivity of the lagoon-bed, commonly termed seal formation (Cihan et al., 2006; Tyner et al., 2006). To minimize downward or lateral seepage of the wastewater from earthen waste lagoons and to prevent groundwater contamination, clay is typically mixed with local sediment and compacted to form an earthen liner along the lagoon bottom and banks (Brown and Associates, 2003; SCS, 1997). The use of some of the clay soils (especially those with a high composition of minerals from the smectite group) as soil liners might increase rather than decrease the downward or lateral seepage of the wastewater from earthen waste lagoons, due to the tendency of such soils to form cracks upon desiccation. It is well known that desiccation cracks can preferentially serve as water conduits and preferentially transport water and solutes (movement through a fraction of the total cross-sectional area available for flow) into deep sections of the vadose zone during the generation of local runoff or occasional flooding (Bronswijk et al., 1995; Brown et al., 1999; Gerke et al., 2010; Kelly and Pomes, 1998; Kurtzman and
2 Introduction
Scanlon, 2011; Novák et al., 2000; Oostindie and Bronswijk, 1995). The dimensions and the consequent hydraulic properties of the desiccation cracks vary dynamically in time and space following the wetting conditions on the surface and in the subsurface. An increasing number of observations from field-scale experiments in the vadose zone and large-scale lysimeter experiments suggest that, even though desiccation cracks may disappear on the land surface under wet conditions, the cracks will not completely disappear from the subsurface and may still serve as preferential flow paths for air and water and enhance evaporation in the presence of surface winds (Acworth and Timms, 2009; Adams and Hanks, 1964; Adams et al., 1969; Gerke, 2006; Graham, 2004; Greve et al., 2010; Kishne et al., 2010; Mermut et al., 1996; Nachshon et al., 2012; Selim and Kirkham, 1970). To date, there is still lack of data and understanding of the role that desiccation cracks play in the field-scale dynamics of water infiltration, subsurface aeration and evaporation. The main concern with seepage from earthen waste lagoons is the risk to + groundwater quality. Organic-nitrogen (organic-N) and ammonium (NH4 ) are the dominant N species in CAFO anaerobic waste lagoons, whereas in their subsurface, - - nitrate (NO3 ) is the most common contaminant. High NO3 levels in drinking water have been associated with risks for methemoglobinemia (blue-baby syndrome) in infants and diarrheal and respiratory diseases and have even been suggested to be a risk factor for specific cancers (Ward et al., 2005). Accordingly, many countries and - organizations around the world have provided guideline values for maximum NO3 levels in drinking water [for example: World Health Organization (WHO) – 50 mg L- 1, USA – 45 mg L-1, Israel – 70 mg L-1 (Committee, 2010; Ward et al., 2005; WHO,
+ - 2007)]. Microbial oxidation of ammonia (NH3) and NH4 into NO3 (nitrification) in a porous medium requires the presence of molecular oxygen (O2) (Prosser, 1989), and it can be generated by two processes: (1) complete nitrification (oxidation of ammonia to nitrate) and (2) as a byproduct of anammox bacterial activity that consumes ammonia and nitrite (Mulder et al., 1995). Partial nitrification (oxidation of ammonia to nitrite) can provide nitrite for the activity of anammox bacteria (Sliekers et al., - 2002). NO3 can be further reduced by microorganisms into N gas (N2) through denitrification. In sediments where favorable conditions for both nitrification and - - denitrification are present in neighboring microhabitats, nitrite (NO2 ) or NO3 produced during nitrification can be utilized by denitrifiers. This process is termed
3 Introduction coupled nitrification-denitrification (CND). CND plays an important role in the removal of nitrogenous compounds in sediments (Kremen et al., 2005; Nielsen and Revsbech, 1998). Studies on the fate of N species in the vadose zone and groundwater in dairy earthen waste lagoon environments have shown that their fate varies from place to - place. Monitoring of NO3 concentration in the pore water of the vadose zone under and around earthen lagoons constructed in sand to clay loam sediments demonstrated - that during lagoon operation, in some cases, NO3 concentrations remained similar to background concentrations (Meyer et al., 1972; Parker et al., 1999b), and in others, they reached very high concentrations (>900 mg L-1) (Korom and Jeppson, 1994; - Oliver and Meyer, 1974). Groundwater monitoring for NO3 contamination from earthen lagoons has demonstrated that the impact of the lagoon varies with time from - its initial operation (Sewell, 1978). While in some studies, elevated NO3 concentrations in groundwater were attributed to lagoon seepage (Nordstedt et al., - 1971), in others, elevated NO3 could not be distinguished from the background levels caused by manure application in surrounding fields (Harter et al., 2002; Pettygrove et - al., 2010; Viers et al., 2012). A recent study has even shown that the NO3 concentration in the saturated zone varies with depth due to denitrification processes in the groundwater column (Singleton et al., 2007).
1.2 Objectives The study was triggered by the lack of data about the bio-geo-hydrological processes in the vadose zone underlying dairy waste lagoons and their impact on the groundwater quality. The main objectives of this research were to investigate the water flow and contaminant transport mechanisms in the vadose zone underlying a dairy waste lagoon and its surroundings, through detailed in-situ monitoring of the bio-geo-hydrological processes in the vadose zone. In particular, this study evaluates: (a) the infiltration mechanisms and the resulting sediment water content around a dairy waste lagoon constructed in expansive clay sediment, (b) the impact of the + infiltration mechanisms and sediment water content on NH4 oxidation and denitrification in the clayey vadose zone underlying the dairy waste lagoon and its surroundings, (c) the impact of dairy waste lagoons on subsurface and groundwater salinization, and (d) the regional impact of leachates from dairy waste lagoons on the
4 Introduction
- - Cl and NO3 loads in the vadose zone and groundwater in the Beer-Tuvia region, Israel.
1.3 Significance of the research
To date, studies on the effect of dairy farm waste lagoons on groundwater quality were mainly based on sediment sampling and extraction, on pore water sampling and on groundwater sampling. Results from such studies have indicated high spatial and temporal variability in the fate of leaching N-species in the vadose zone and groundwater underlying dairy earthen waste lagoon environments. In some cases, differences were observed between the water content and the N-species concentrations in the sediment at the lagoon banks and the water content and the N- species concentrations under the rest of the lagoon. Studies have also indicated that physical, chemical, and biological processes reduce the hydraulic conductivity of the lagoon-bed (commonly termed “seal formation”). The effects of the infiltration through the sections of the lagoon bed where a seal has formed and the infiltration through the rest of the lagoon on the bio-hydrological processes in the vadose zone have not yet been studied.
1.4 Geohydrological background
The Coastal Aquifer underlies the Coastal Plain of Israel and stretches parallel to the Mediterranean Sea, extending east between 8 km in the north to about 30 km in the south (Fig. 1) (Gvirtzman, 2002). The Kurkar Group, which forms the aquifer, overlies the impervious marine clays of the Saqiye Group of Pliocene age and is composed of alternations of sandstone, calcareous sandstone (eolianites or ‘Kurkar’), siltstone, red sandstones and loamy soils (‘Hamra‘), and marine clay and shales of Pleistocene age (Issar, 1968; Vengosh and Ben-Zvi, 1994) (Fig. 2). The latter divides the aquifer into a number of sub-units, but the aquifer is generally considered as a single-water system. The main recharge to the aquifer is from rain that falls directly above it (average rainfall of 500 mm y-1 in the central area and less than 300 mm y-1 in the south). The general flow direction in the aquifer is from the east towards the Mediterranean Sea in the west. Morphologically, the Coastal Plain is characterized by a series of parallel north-south ridges, made of calcareous sandstone or partly
5 Introduction cemented sand dunes, stretching parallel to the shore line with up to 20 m of deep clayey longitudinal valleys (‘troughs’) between them (Issar, 1968; Tolmach, 1977; Vengosh and Ben-Zvi, 1994).These troughs were once swamps or riverbeds, and their soil is therefore fertile (Kass et al., 2005).
35o
n a e n a r Nablus r a e e it S d e n M a Tel-Aviv d Jerusalem r o
l J Study area e a r s Gaza I Hebron
31o Beer Sheva E g y p t Negev Desert
0 50km
Figure 1. Map of Israel with its Coastal Aquifer (highlighted in yellow).
Azrikam +80 West Beer-Tuvia 2 East 13/3 Azrikam T1 Bitania 2 Orot B +60 Ashdod 4 13/5 Ashdod 7 +40
+20
0
-20
-40
-60 oup)
Elevation aboveseaElevation level (m) r up iye g -80 Gro (Saq rkar ation Ku form -100 Yafo
-120
0 1 2 3 4 5 6 7 8 Km Clay Sand, calcareous sandstone Pebbles, conglomerate MarineMarin clay clay and and shale shale Sandy clay, loam Figure 2. Schematic cross section of the Coastal Aquifer in the study area (modified from Tolmach (1977)).
6 Introduction
1.5 The degradation in the groundwater quality of Israel’s Coastal Aquifer
Prior to the intense urbanization and cultivation of Israel’s coastal plain (1930s), the underlying Coastal Aquifer was characterized by very high water quality (Cl<100 mg L-1). The high quality of groundwater in the aquifer has since been impaired due to salinization processes along the western and eastern margins, as well as in internal areas of the aquifer in the form of salt plumes (Kass et al., 2005; Mercado, 1985; Shavit and Furman, 2001; Vengosh and Ben-Zvi, 1994). This degradation is manifested by a continuous increase in the average chloride and nitrate concentrations (Fig. 3). Nowadays, the chloride and nitrate concentrations in the groundwater of the Beer-Tuvia region has reached an average concentration of ~800 and ~45 mg L-1, respectively, with an ongoing yearly increase of 8 mg L-1 of chloride (Weinberger, 2007) (Fig. 3). A geochemical study concluded that the formation of salt plumes in the Coastal Aquifer of Israel is not directly related to anthropogenic sources, but to the intrusion of saline water bodies from deeper strata as a result of a drop in the water level (Vengosh and Ben-Zvi, 1994). Shavit and Furman (2001) identified the geographic location of the salinization source described by Vagosh and Ben-Zvi (1994) using a model. However, their model could not explain the increase in the chloride concentration following the recovery of the hydraulic depression. Furthermore, the deep saline water bodies have not been detected in specially drilled boreholes. Kurtzman and Scanlon (2011) and Kurtzman (2011) have suggested that in natural clayey soils (vertisols), a change in land use, from natural to cultivated land, leads to a flushing down of salts that have accumulated in the vadose zone into the groundwater. Kurtzman (2011) suggested that such a mechanism may explain the salt plume in the southern part of the Coastal Aquifer, including the Beer-Tuvia region.
7 Introduction
700 60
600 50
500 40 ) ) -1 -1 400 30 (mg L - - (mg L (mg 3 - 300 Cl 20 NO 200 NO - 3 10 100 Cl-
0 0 1968 1974 1980 1986 1992 1998 2004 Year - - Figure 3. Average nitrate (NO3 ) and chloride (Cl ) concentrations with time in operational cell No. 82 of Israel Coastal Aquifer (modified from Weinberger (2007)).
1.6 Study site
The Beer-Tuvia region is located above the southern section of the Israeli Coastal Aquifer. The area hosts 140 dairy farms (8.5% of the dairy cows in Israel) with an average herd size of 80 animal units (including calves, heifers and milking cows). A dairy farm with approximately 60 dairy cows and 30 calves was chosen for the study of contaminant transport from a dairy farm through the vadose zone to the underlying groundwater. The dairy farm is located in the Beer-Tuvia region (Fig. 1). The facility uses a 200 m3 (20 m long, 10 m wide and 1 m deep) earthen unlined wastewater storage lagoon, which is the common manure-management practice in the area. About 7 m3 of wastewater (manure, feces and corral washing and cooling water) are drained into the lagoon daily. Excess wastewater from the lagoon overflows into a drainage channel year round. No specific maintenance procedures, such as drainage or solid removal, are used at the site. Wastewater is periodically pumped from the lagoon and channel to serve as a readily available fertilizer for nearby agricultural fields. Even so, both the waste lagoon and the channel are flooded with liquid manure year-round. Natural annual shallow-rooted plants (mainly: Malva sylvestris) grow on the waste source banks from November to April.
8 Introduction
1.7 The dissertation structure The insights that were formulated from the data collected during the research period 2008 – 2013 and their implications are presented in this dissertation as five independent papers. The papers contribute to a better understanding of the impact of dairy farms on groundwater contamination by initially addressing the bio-geo- hydrological processes in a clayey vadose zone underlying a dairy farm, and then by discussing the regional impact of these processes on the groundwater quality of the study site. The last part of the dissertation provides an integral discussion of the main results of the study and their implications. The first paper, entitled "Water percolation through a clayey vadose zone" (published in the Journal of Hydrology), describes the infiltration dynamics through the clay soil at the study site. The article focuses on the impact of desiccation cracks on water infiltration and propagation in the vadose zone. The second paper, entitled "Infiltration mechanism controls nitrification and denitrification processes under dairy waste lagoon" (published in the Journal of Environmental Quality), describes the nitrogen transformations under a dairy waste lagoon. The article focuses on the impact of the infiltration mechanisms on subsurface aeration and nitrogen transformation in the vadose zone around the lagoon. The third paper, entitled "Ammonia transformations and abundance of ammonia oxidizers in a clay soil underlying a manure pond" (published in FEMS Microbiology Ecology journal), describes the diversity and abundance of ammonia- transforming microorganisms (AOA, AOB and AMX) under the waste lagoon. This article focuses on the relations between the abundance of AOA, AOB and AMX and the possible nitrogen bio-transformation under the wastewater storage lagoon. The fourth paper, entitled "Desiccation-crack-induced salinization in deep clay sediment" (accepted for publication in the journal Hydrology and Earth System
Sciences Discussions), describes subsurface evaporation and salinization processes, and offers a conceptual model termed desiccation-crack-induced-salinization (DCIS), in which thermally driven convective air flow in the desiccation cracks induces evaporation and salinization in relatively deep sections of the subsurface. The article
9 Introduction also highlights how leachates from dairy waste lagoons enhance the subsurface salinization process.
The fifth paper, entitled "Assessing the impact of dairy waste lagoons on groundwater quality using a spatial analysis of vadose zone and groundwater information: example from the Beer-Tuvia region Israel" (under review, see Appendix A), evaluates the regional impact of leachates from dairy farm waste - - lagoons on groundwater quality. The article shows regional Cl and NO3 mass balances that are based on the observations from the vadose zone underlying the waste lagoon, on groundwater surveys and on the historical development of the dairy farming industry in the Beer-Tuvia region. In the initial stages of the research project, natural hormones (testosterone and estrogen) were observed throughout the entire unsaturated sediment profile under the lagoon. A paper that studied the flow and transport properties of these hormones in the vadose zone underlying the lagoon was published by Arnon et al. [Transport of testosterone and estrogen from dairy-farm waste lagoons to groundwater (Environmental Science and Technology)] (with me as the 8th author). Even though this paper is not presented in the manuscript, it is mentioned here since its conclusions that the hormones were transported throughout the entire profile by a flow mechanism different from the normal advection-dispersion driven flow through the sediment matrix, stimulated my research and scientific insights.
10 Published Papers
2. Published Papers
11 Journal of Hydrology 424–425 (2012) 165–171
Contents lists available at SciVerse ScienceDirect
Journal of Hydrology
journal homepage: www.elsevier.com/locate/jhydrol
Water percolation through a clayey vadose zone ⇑ S. Baram a, D. Kurtzman b, O. Dahan a, a Department of Environmental Hydrology & Microbiology, Zuckerberg Institute for Water Research, Blaustein Institutes for Desert Research, Ben-Gurion University of the Negev, Sede Boqer Campus 84990, Israel b Institute of Soil, Water and Environmental Sciences, Agricultural Research Organization, The Volcani Center, Bet Dagan 50250, Israel article info summary
Article history: Heavy clay soils are regarded as less permeable due to their low saturated hydraulic conductivities, and Received 10 April 2011 are perceived as safe for the construction of unlined or soil-lined waste lagoons. Water percolation dynam- Received in revised form 1 December 2011 ics through a smectite-dominated clayey vadose zone underlying a dairy waste lagoon, waste channel and Accepted 29 December 2011 their margins was investigated using three independent vadose-zone monitoring systems. The monitoring Available online 5 January 2012 systems, hosting 22 TDR sensors, were used for continuous measurements of the temporal variation in This manuscript was handled by Philippe Baveye, Editor-in-Chief vadose zone water-content profiles. Results from 4 years of continuous measurements showed quick rises in sediment water content following rain events and temporal wastewater overflows. The percolation pattern indicated dominance of preferential flow through a desiccation-crack network crossing the entire Keywords: À1 Vadose zone clay sediment layer (depth of 12 m). High water-propagation velocities (0.4–23.6 m h ) were observed, Smectite-dominated clay indicating that the desiccation-crack network remains open and serves as a preferential flow pathway 3 À3 Desiccation crack year-round, even at high sediment water content (0.50 m m ). The natural formation of desiccation- Infiltration crack networks at the margins of waste lagoons induces rapid infiltration of raw waste to deep sections Preferential flow of the vadose zone, bypassing the sediment’s most biogeochemically active parts, and jeopardizing Waste lagoon groundwater quality. Ó 2012 Elsevier B.V. All rights reserved.
1. Introduction especially near the surface (Chertkov and Ravina, 1998). Desiccation cracks can serve as preferential water conduits, and transport water Due to their low saturated hydraulic conductivities (on the order and solutes into the vadose zone below the root zone during the of 1 cm dayÀ1 or less), clay soils are considered safe for the construc- generation of local runoff by intense rain events or floods (Bronswijk tion of soil-lined and unlined waste lagoons. This perception is et al., 1995; Kurtzman and Scanlon, 2011). Little field work was based on the assumption that the low hydraulic conductivity of done on the depths reached by preferential wetting fronts via desic- the clay matrix slows down infiltration, increases runoff and pre- cation cracks, and field data are still lacking especially on the prop- vents deep percolation. Slow propagation of wetting fronts further agation rates of these preferential wetting fronts. exposes infiltrating water to intensive evapotranspiration in shal- Several models have been developed over the years to charac- low soil horizons and prevents deep migration of water and pollu- terize the formation of desiccation cracks in expansive clay soil tants to the underlying groundwater. However, while these (Chertkov, 2008; Chertkov and Ravina, 1998; Cornelis et al., assumptions hold for matrix flow, their validity for clay soil profiles 2006). These models are based on commonly used laboratory mea- under unsaturated natural conditions is questionable. Clay soils, surements, such as the coefficient of linear extensibility (COLE) especially those dominated by clay minerals built from tetrahe- which relates soil shrinkage potential to soil water loss, and the dral–octahedral–tetrahedral sheets (2:1 structure) which have very soil shrinkage characteristic curve. The models incorporate the weak interlayer bonding (such as Montmorillonite and Vermicu- assumption that desiccation cracks start to develop from the min- lite), are subjected to dynamic structural changes during natural ute the soil becomes unsaturated (Chertkov and Ravina, 1998). The drying and wetting cycles (Horn et al., 1994). In these soils, the loss soil profile is normally divided into an upper soil layer of macro- of water from the matrix due to evapotranspiration is partially or shrinkage with cracks of large aperture (>1 mm) and a lower soil completely replaced by shrinkage of the clay matrix (Jarvis, 2007). layer of micro-shrinkage with cracks of small aperture (<1 mm) Consequently, in arid and semiarid environments, desiccation (Chertkov, 2008; Chertkov and Ravina, 1998; Cornelis et al., cracks can make up a substantial proportion of the soil’s volume, 2006). While the models deal mainly with the macro-shrinkage layer, from the point of view of hydraulic conductivity, the thinner cracks that develop in the micro-shrinkage layer may make a ⇑ Corresponding author. Tel.: +972 8 6596917. E-mail addresses: [email protected] (S. Baram), [email protected] significant contribution to the overall hydraulic conductivity of a (D. Kurtzman), [email protected] (O. Dahan). clay layer.
0022-1694/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.jhydrol.2011.12.040 166 S. Baram et al. / Journal of Hydrology 424–425 (2012) 165–171
Field observations have demonstrated the importance of desic- desiccation cracks over time and its direct impact on the observed cation cracks to the overall hydraulic conductivity of a clay layer. deep percolation. These cracks have been found to allow rapid transport of water, nutrients and pesticides to the subsoil where they are inaccessible 2. Materials and methods to shallow rooting plants and can pollute the local groundwater (Bronswijk et al., 1995; Harris et al., 1994; Scorza et al., 2004). Most 2.1. Study area field studies have been performed on the upper section (1m)of clay soils with tile drains. However, Acworth and Timms (2009) The study area is located in the Beer Tuvia region, above the studied an untiled agricultural clay field with a relatively deep southern part of the coastal aquifer in Israel. The phreatic aquifer groundwater table (16 m). Those authors showed evidence of shal- in the area is composed of calcareous sandstone and layers of Pleis- low aquifer (16 m depth) freshening due to downward leakage of tocene age clays (Kurkar Group) (Issar, 1968; Weinberger, 2007). irrigation water during an irrigation season. In addition, during The clay in the area is mostly dominated by 2:1 clay minerals the drilling of a borehole, they observed substantial amounts of (Table 1), and its thickness can vary from several centimeters to drilling fluid being lost to the soil, which resulted in a rise in several meters within a distance of several hundred meters. The groundwater level. More and more field observations and large- climate in the study area is characterized as Mediterranean with scale lysimeter experiments are suggesting that even though desic- average summer and winter temperatures of 24.3 °C and 14.2 °C, cation cracks may disappear at land surface under wet conditions, respectively, and an average annual precipitation of 450 mm that the cracks do not completely disappear from the subsurface, and falls during the winter season (November–March). The main re- may still serve as preferential flow paths (Gerke, 2006; Greve charge to the aquifer is related to percolation of seasonal rainwater et al., 2010). These observations, along with the discussed impor- and agricultural return flow from irrigated fields. tance of desiccation cracks to the overall hydraulic conductivity of Three sites were chosen for the study of infiltration through clay clay layers, emphasize the need for a better understanding of flow sediments near and under waste sources. One site was located un- and transport in these dynamic hydraulic domains. der a single-stage earthen unlined dairy farm waste lagoon, and Soil-lined and unlined liquid waste lagoons on clay soils are fre- the other two sites were located under a waste channel and its quently used, all over the world, to store effluents from concen- margins, 150 m downstream of the lagoon. The dairy farm selected trated animal feeding operations (CAFOs) such as dairy farms. It for this study is representative of the area: it has 60 dairy cows and has been frequently shown that direct leakage from dairy waste la- 30 heifers and calves, and discharges about 7 m3 dÀ1 of wastewater goons and their side walls impacts the quality of the underlying and liquid manure into the lagoon (200 m2 and 1 m deep). Over- groundwater (DeSutter et al., 2005; Gooddy et al., 2002; Ham flow from the lagoon flows into a waste channel (2 m wide, 1 m and Baum, 2009; Harter et al., 2002; Parker et al., 1999). Yet, due deep and 200 m long). The dairy farm has been operating in the to the low saturated hydraulic conductivity of clay soils, in many same way for the past 40 years and no specific maintenance proce- countries dairy waste lagoons are constructed in such soils without dures, such as drainage or solids removal, have been used at the the need for any special procedures (SCS, 1997). It has been sug- site. Both the waste lagoon and the channel are flooded with liquid gested that the side walls of the liquid-waste lagoons, which are manure year-round. The hydraulic conductance of the lagoon bed susceptible to wetting and drying cycles, are likely to form desicca- is controlled by an organic layer which was formed on it (Baram tion cracks regardless of the soil-lining method used (McCurdy and et al., 2010). The stratigraphy at the sites consists of a thick clay McSweeney, 1993; Parker et al., 1999). Not only does the formation layer overlaying sandy loam and calcareous sandstone (Table 2). of cracks reduce the effectiveness of the lagoon’s seal (Parker et al., The water table in the area is located at a depth 46 m below 1999), it also increases the potential for preferential transport of li- the surface. quid waste from the waste lagoons during fluctuations in water level. The objective of this study was to assess water percolation 2.2. Field instrumentation and monitoring concept dynamics in clayey vadose zones underlying dairy waste lagoons and their margins. We performed a detailed analysis of the tempo- The study was carried out using three vadose-zone monitoring ral variations in water-content profiles along deep sections of the systems (VMSs) that were designed to facilitate real-time continu- vadose zone. Special attention was paid to the development of ous measurements of the temporal variation in sediment
Table 1 Selected properties of the clay sediment at the study site.
Particle distributiona Bulk mineralogical composition (%)b Composition of the phyllosilicates (%) b Sand (%) Silt (%) Clay (%) Phyllosilicates Quartz Calcite K-feldspar and plagioclase Illite–smectite Kaolinite 26±14 22±8 52±8 44±2 39±5 9±1 8±2 90±7 10
a Average of six samples from under the lagoon and channel (three each). b Average of three samples, two from under the lagoon and one from a nearby agricultural field.
Table 2 Stratigraphy and probe distribution underlying the waste lagoon, the waste channel and their margins.
Site Probes depth (m) Lithology Clay Sandy loam Calcareous sandstone Lagoon margins 1, 2, 3 0–6.5 m 6.5–8 m 8–52 m Lagoon 6, 10, 15, 20 Channel 1.7, 3.9, 8.1, 10.2, 13.4, 16.6, 20.6 0–12 m 12–25 m Channel margins 1.6, 3.7, 5.8, 8.0, 10.1, 13.3, 16.4, 20.7 S. Baram et al. / Journal of Hydrology 424–425 (2012) 165–171 167
A a b Section A-A front view Waste channel Waste lagoon Waste channel
FTDR probes VMS # II VMS # III
Vadose zone
Groundwater * Not to scale A Fig. 2. Calibration curve for water-content measurements by FTDR probes. The data Fig. 1. Schematic side view illustration of the vadose-zone monitoring system points represent fine to coarse sand and the clay soil from the study area mixed À1 À1 À1 (VMS) installed under the waste lagoon, waste channel and the waste channel with tap water (EC 0.3 mS cm ) and NaCl solution (EC 21 mS cm ). La L is h À1 À À1 margins (a), and front view (A–A) of the VMS installed under channel and its linearly correlated to soil water content ( = 0.2432(La L ) 0.3981). La L margins (b). defines the soil permittivity as the ratio between the apparent lengths of the TDR probe (La) to its real length (L)(Dahan et al., 2007). water-content profiles (Fig. 1). A detailed description of the moni- of the VMSs, the volumetric water content of the sediment was toring system, installation procedure and performance can be measured continuously every hour during the dry period (April– found in previous publications (Dahan et al., 2007, 2008; Rimon October) and every 15 min during the wet season (November– et al., 2007). For the sake of convenience, a general description of March). the monitoring system is provided here. In each site, a VMS com- posed of a flexible sleeve made of thin PVC liner (0.5 mm thick) 2.3. Desiccation crack characteristics hosting a set of flexible time domain reflectometry (FTDR) probes was installed through a 140-mm diameter uncased slanted bore- Desiccation cracks in the upper clay sediment appear on land hole (Fig. 1). Once placed in the borehole, the sleeve was filled with surface over the entire area, including the immediate margins of two-component high-density (1.65 kg LÀ1) liquid urethane that the waste lagoon and channel (Fig. 3). Dimensions and distribution solidifies shortly after its application. The hydrostatic pressure of the desiccation cracks were measured over the course of a year generated by the liquid urethane within the flexible sleeve causes in an undisturbed 14 Â 5.4 m plot located away from the lagoon its expansion, seals the borehole void and ensures proper attach- and channel, using three transects in each direction. The cracks’ ment of the probes to the sediment facing the upper side walls aperture, depth and location were measured at each intersection of the borehole. The volumetric water content of the sediment between a desiccation crack and a transect. The aperture was mea- was measured using a time-domain reflectometry instrument sured perpendicular to the crack axis using a caliper, while the (TDR100 Campbell Scientific, Logan, UT). Calibration of the TDR depth of the macro-cracks was measured using a 6-mm straight measurements accounting for the specific structure of the FTDR probe, instrumentation setup and variety of soil types, including clay soils from the site, in a range of salinities is presented in Fig. 2. The overall accuracy of water-content measurements under 5.4m field conditions using the VMS is ±5% of the absolute value (Rimon et al., 2011). The vadose zone underlying the waste lagoon was monitored using a 35-m long VMS. The system was installed in a slanted bore- hole (35° from the vertical), spanning the sediment from the margins of the waste lagoon to the sediment underlying it. The upper section of the VMS (1–3 m) monitored the shallow clay hori- zons underlying the margins of the lagoon while the deeper section (6–20 m) monitored the deep sections that are located under the 14m lagoon. The probes’ distribution across the vadose zone is pre- sented in Table 2. Two additional 24-m long VMSs were installed in two independent slanted boreholes (30° from the vertical) under the waste channel and its margins. The first monitoring system was installed in the center of the channel, aligned with the channel axis, such that all of the probes were positioned across the vadose zone underlying the channel. The second monitoring system was installed 2 m away from the channel edge, parallel to the channel direction, positioning all of the probes under the channel margins (Table 2). These two independent monitoring systems allowed us to monitor both the vadose zone under a permanently flooded Fig. 3. Undisturbed land plot which was used for continuous survey of desiccation waste source, and the vadose zone under its dry margins. In all crack dimensions. 168 S. Baram et al. / Journal of Hydrology 424–425 (2012) 165–171 metal rod (Chertkov and Ravina, 1998). All dimensions are channel is a narrow line-shaped waste source (2 m wide) whereas reported as mean ± standard deviation, with total number of sam- the waste lagoon is a much wider waste source (20 m wide). ples (n) in brackets. Hence, the measuring points positioned under the waste channel are close to the channel margins on both sides and are more likely to be influenced by the infiltration processes occurring at the mar- 3. Results and discussion gins. Similarly, the probe that was positioned in clay 6 m under the lagoon, close to the lagoon margins, also recorded wetting events 3.1. Water percolation through the clay at the beginning of each winter season (Fig. 4c). In general, the water-content values in the deep sandy sections of the vadose zone Continuous monitoring of the vadose zone underlying the waste at all sites showed no observable fluctuations with time (Fig. 4). An sources and their margins indicated temporal variations in mea- exception can be seen with the probe located 13.3 m under the sured water contents which were mostly associated with signifi- channel margins. This probe is located at the top of the sand forma- cant rain events (Fig. 4). Water percolation through the entire tion, immediately under the clay, and is therefore influenced by the clay cross section underlying the margins of the waste channel rapid water percolation from the clay layer to the sand (demon- was expressed by quick rises in the sediment water content, within strated during October 2009–March 2010 in Fig. 4a). Redistribution 2–15 h of most of the significant rain events (Fig. 4a). During these of the percolating water from the upper section of the sand was not events, a wetting process was observed in the underlying calcare- followed by observable changes in the sediment water content (ar- ous sandstone (13.3 m) within 36–48 h (Fig. 4a), indicating signif- rival of the wetting front) at the deeper sections of the vadose zone icant water penetration into the sandy formation underlying the (such as observed by Rimon et al. (2007)), indicating that steady 12 m clay layer. Similar patterns were observed along the clay pro- flow conditions prevail at the deep vadose zone. file underlying the margins of the waste lagoon throughout the Close inspection of the water-content variation in all three sites winter of 2008/2009 (Fig. 4c). Water-content profiles of the reveals an erratic wetting sequence of the clay sediments. This ob- sediment underlying the waste channel showed that near-satura- served percolation pattern indicates domination of a preferential 3 À3 tion conditions (0.50–0.60 m m ) prevail in the upper part of flow mechanism (Dahan et al., 1999) rather than gradual wetting the clay sediment (Fig. 4b). Nevertheless, fluctuations with time, front propagation in porous matrix (Dahan et al., 2008; Rimon and rapid responses to some of the intensive rain events were ob- et al., 2007). Preferential flow in soils refers to faster than average served at that site down to a depth of 8.1 m (November 2009, water movement which takes place through only a fraction of the December 2010, Fig. 4b). When interpreting the results from the pore space, thereby bypassing most of the matrix. Moreover, it is waste lagoon and channel, it should be remembered that the waste commonly accepted that preferential flow is predominantly grav- ity-driven, thus consisting of rapid water movement from higher to lower depths (Gerke et al., 2010). The observed pattern of sharp and rapid increases in the water-content profiles (Fig. 4) cannot be explained by the slow matrix flow rate (1myÀ1) measured for this clay soil by Arnon et al. (2008). It is suggested that the desic- cation-crack network formed in this clay soil (Fig. 3) serves as a water conduit for preferential infiltration through the entire thick clay layer. Even though the maximal depth of the desiccation cracks and their spatial distribution with depth were not directly measured in the field, the observations of wetting fronts crossing an entire clay layer (12 or 6 m) within hours suggests that the des- iccation cracks cross the entire unsaturated clay layer. Further- more, the observation of sharp increases in water content in the nearly saturated clay under the waste channel (Fig. 4b) suggests that desiccation cracks prevail in clay soils, even at very high water contents that are still lower than saturation. These observations agree with the assumptions made by Chertkov and Ravina (1998) that desiccation cracks start to form from the minute the soil be- comes unsaturated, and imply high connectivity between the formed cracks, even at high water contents. An additional indica- tion for dominant fracture flow in the clay sediments emerges from the relatively quick drainage pattern of the sediments after wetting events (Fig. 4): shortly after most wetting events, the water con- tent in the clay profiles decreased relatively rapidly to lower val- ues. Such a quick reduction in water content may be attributed to both water imbibition by the surrounding drier clay matrix and to deep drainage via preferential pathways.
3.2. Desiccation-crack properties
A survey of the properties and distribution of the major desicca- tion cracks at the end of the dry season (October 2008) yielded an average crack aperture of 5.5 ± 1.7 cm with an average depth Fig. 4. Temporal changes in water content of sediments under the waste channel (where the aperture was greater than 6 mm) of 65 ± 19 cm margins (a), waste channel (b) and waste lagoon (c), along with rain events during (n = 70). It should be noted that the maximal depth of the desicca- 2008–2010 (d). FTDR probes at depths of 1, 2 and 3 m from the waste lagoon represent the lagoon margins, while FTDR probes at depths of 10, 15 and 20 m tion cracks was measured with a stiff 6-mm metal rod that limits represent the area under the flooded region of the waste lagoon. the ability to follow the crack irregular structure or to measure S. Baram et al. / Journal of Hydrology 424–425 (2012) 165–171 169 the depth at smaller apertures. Therefore, these depths represent macro-shrinkage and to the formation of major desiccation cracks only minimal values. Cases in which consecutive measurements (aperture >1 mm) (Chertkov, 2008; Cornelis et al., 2006). According of the same crack, by the metal rod, indicated shallower crack to these models, the deeper section of the clay is only subjected to depths, served as an indication for swelling processes in the crack. micro-shrinkage and to the formation of minor desiccation cracks The intersections among the major desiccation cracks created poly- (aperture <1 mm). However, the observations in this study show gons of 99 ± 48 Â 69 ± 53 cm (n = 70) (Fig. 5a). These polygons the significant contribution of desiccation cracks in the deeper sec- were further dissected by a substantial number of smaller cracks tions to the overall hydraulic conductivity of dispersive clays. (aperture <2 cm) which divided them into smaller separate units. Moreover, they demonstrate the dynamic structure of desicca- Repeat measurements of the same desiccation cracks during the tion-crack networks as preferential flow paths. For example, in following two winters (2008/2009 and 2009/2010) showed that the winter of 2009/2010, significant wetting along the entire clay they were never fully closed during the rainy seasons (Fig. 5b). profile was only observed after the first rain event of that season Note that the cumulative rain precipitation in these years was be- (November 2009), while in the winter of 2008/2009, significant low and around average (280 and 416 mm yÀ1, respectively). In wetting was observed following each intensive rain event (Fig. 4c most cases, swelling of the sediment was not uniform along a crack and d). Furthermore, preferential infiltration was observed even line, leaving closed and unclosed sections (Fig. 5b). Changes in the under near-saturation conditions (see December 2010, probe at depths of the cracks varied from complete closure to no change at 2m; Fig. 4b–d), demonstrating the nontrivial relations between all (Fig. 5c). In places were the cracks were fully closed on the sur- clay water content and the dimensions of the desiccation-crack face, the metal rod was forcefully pushed into the sediment to network. check for closure depth. In many places, when the metal rod Additional insight into the impact of desiccation cracks on reached a depth of 20 cm it plunged into an open void, indicating water flow and contaminant transport in clay sediments near that cracks which healed on the surface remained open in their waste sources is related to observations of erratic wetting which deeper subsurface sections (below 20 cm), sometimes with no did not follow rain events. These were observed at the channel change in total depth (see, for example, desiccation crack 1 in margins in March 2009, under the waste channel in September Fig. 5c and d). The pattern of rapid and deep water penetration fol- 2009, and under the waste lagoon margins in December 2008, Sep- lowing intensive rain events throughout the winter (Fig. 4) pro- tember 2010 and October 2010 (Fig. 4). These observations imply vided an additional indication that the desiccation-crack network that raw wastewater preferentially infiltrated into the deep vadose remains open year-around, and hence may serve as a preferential zone through desiccation cracks formed on the side walls of the flow path for local runoff infiltration. Note that local runoff (to waste lagoon and waste channel during minor fluctuations in the the nearest desiccation crack) may be generated locally on each wastewater level. Parker et al. (1999) and McCurdy and McSwee- clay polygon due to the low infiltration capacity of the top wet clay ney (1993) have suggested that such processes may occur on the (Fig. 5a and b). These observations coincide with other reported side walls of waste lagoons which are susceptible to wetting/dry- field observations and large-scale lysimeter experiments suggest- ing processes. ing that even though desiccation cracks may disappear on land sur- face under wet conditions, the cracks will not completely 3.3. Validation of the percolation mechanism disappear from the subsurface, and may still serve as preferential flow paths (Acworth and Timms, 2009; Gerke, 2006; Greve et al., The VMS and the FTDR technique have been previously used to 2010). monitor propagation of a wetting front through the vadose zone To date, most of the models that predict formation of vertical under a variety of natural hydraulic and sedimentological condi- desiccation cracks in dispersive clays at variable water contents tions including: (a) rainwater infiltration through sand dunes (Ri- have mainly dealt with the upper layer that is subjected to mon et al., 2007), (b) floodwater infiltration through a sandy
Fig. 5. Desiccation cracks in the top soil at the end of summer (a) and in mid winter (b). Non-uniform healing of the desiccation cracks represented by healed (I) and unhealed (II) sections. Changes in depth (c) and aperture (d) of four representative desiccation cracks plotted against cumulative rainfall. 170 S. Baram et al. / Journal of Hydrology 424–425 (2012) 165–171 ephemeral riverbed (Dahan et al., 2008) and (c) floodwater infiltra- and Bronswijk, 1992) and MACRO (Jarvis, 1994) require a determi- tion through coarse alluvial deposits in desert ephemeral rivers nation of parameters that either describe the spatial properties of (Dahan et al., 2007). In those studies, wetting front propagation the crack networks or the properties of the soil itself. At this point, across the vadose zone was monitored using several independent our field observations cannot be simulated by the existing models. VMSs which were technically identical to those used in the present For example the SWAT, FRACTURE and SWAP models all assume study. Fig. 6 represents the wetting fronts propagating depth vs. that the desiccation cracks do not cross the entire clay layer, and time from rain/flood events in the abovementioned sediments. that water accumulates at the bottom of the crack and then redis- The propagation velocity of each wetting front with depth can be tributes into the matrix. Moreover, models which are based on estimated directly from the linear slope between the origin and physical parameters of the clay (e.g. shrinkage curve, etc.) cannot each plotted point. The wetting-front propagation velocity across predict the fact that at a given water content, cracks sometimes the vadose zone, due to either floods or rain-induced infiltration preferentially transport water into the deep vadose zone whereas events, was shown to be 0.06–0.9 m hÀ1 in the alluvial deposits, at other times, they do not. The difficulty in applying dual-perme- 0.07–0.6 m hÀ1 in the sandy river bed and 0.002–0.07 m hÀ1 in ability models (such as MACRO) to cracked soils lies in the fact that the sand dune (Fig. 6). Surprisingly, wetting events in the clay soil flow in both regions is described using the Richards equation, an propagated at much faster rates, ranging from 0.4 to 23.6 m hÀ1 assumption that is likely not valid for flow in large drying cracks (Fig. 6). Such rapid infiltration rates in heavy clay soil (Table 1) (Novák et al., 2000). Furthermore, the erratic behavior of the desic- can only be achieved via preferential flow paths, such as a desicca- cation cracks as water conduits cannot be parameterized to solve tion-crack network. It should be noted that such intensive prefer- expressions such as the quasi-empirical first-order rate expression ential flow was not observed at any other site that was needed for simulations of the exchange between matrix and mac- continuously monitored by VMSs, even in extremely heteroge- ropore regions in the MACRO model. The collected data demon- neous alluvial deposits under flood conditions (Dahan et al., strate the dynamic changes in the desiccation-crack network as 2007, 2008; Rimon et al., 2007). Moreover, during a single rain water conduits for preferential water flow and transport, and stress event, the propagating wetting events were not always observed the need for a determination of field-scale parameters such as rep- by all FTDR probes. For example, during the winter of 2009/2010, resentative elementary volume (REV; Bear, 1972) for desiccation- only the first wetting event was monitored at a depth of 3.7 m un- crack networks in order to improve field-scale flow and transport der the waste channel margins, while the deeper profile responded predictions. to all of the significant rain events during that winter (Fig. 4a and d). Similarly, at the beginning of December 2008, the wetting event 3.4. Implications for pollutant transport was not observed at a depth of 2 m, yet it was observed at a depth of 3 m (Fig. 4c and d). This phenomenon emphasizes the dynamic Observations of deep and rapid wetting processes through thick structure of the desiccation-crack network and its influence on clay sediment indicate that preferential flow through desiccation- the percolation pattern. crack networks actively transports a significant amount of water As a whole, desiccation cracks significantly contribute to the directly into the deep sections of the vadose zone. This observation ability of water to flow and percolate through dispersive clay for- implies that pollutant transported by the flowing water will bypass mations. Moreover, their impact varies dynamically in time and the biogeochemically active parts of the vadose zone and jeopar- space following the wetting conditions. The observed data stress dize groundwater quality. The response of all FTDR probes to rain the difficulty involved in providing an accurate and meaningful events, even under near-saturation conditions, by three indepen- estimation of desiccation crack input parameters for numerical dent VMSs installed at three sites and providing 22 independent models. Models such as SWAT (Arnold et al., 2005), SWAP (Van measurement points of water content, each positioned under an Dam, 2000), FRACTURE (Novák et al., 2000), FLOCR (Oostindie undisturbed sediment column, implies that the preferential perco- lation mechanism through the desiccation-crack networks makes a substantial contribution to groundwater recharge. The pollution potential of such preferential infiltration events at the margins of a waste lagoon is further stressed by work done by Arnon et al. (2008) at the same site. Those authors concluded that normal transport via advection and dispersion could not explain the field observations of natural hormones in the deep vadose zone and in the groundwater. It is reasonable to assume, based on our field observations, that the hormones were transported directly into the deep vadose zone via preferential infiltration, bypassing the absorptive clay. Overall, our observations indicate that the margins of unlined/soil-lined waste lagoons, constructed in dispersive clays in semiarid environments, may substantially contribute to the pol- lution potential of waste lagoons.
4. Conclusions
Rapid rises in water-content profiles of smectite-dominated clay vadose zone following rain events and overflow of waste la- Fig. 6. Propagation of wetting fronts under a sandy ephemeral riverbed during goons indicate an infiltration pattern that is dominated by prefer- flood events (after Dahan et al., 2008), under a sand dune during natural rain events ential flow through desiccation-crack networks. Water percolation (after Rimon et al., 2007), under a coarse alluvial ephemeral riverbed during flood propagated at high velocities (0.4–23.6 m hÀ1) and crossed the en- events (after Dahan et al., 2007), and under dispersive clay soil (from this study). tire clay sediment (>12 m). The network of naturally formed desic- The data points represent the lag time between the initiation of every rain or flood event on land surface and the time of increase in sediment water content at each cation cracks remained open year-round, even at high sediment 3 À3 depth. water contents (0.50 m m ). Natural formation of desiccation S. Baram et al. / Journal of Hydrology 424–425 (2012) 165–171 171 cracks at the margins of waste lagoons induces rapid infiltration of Gerke, H.H., 2006. Preferential flow descriptions for structured soils. J. Plant Nutr. raw waste into deep sections of the vadose zone, bypassing the Soil Sci. 169, 382–400. Gerke, H.H., Germann, P., Nieber, J., 2010. Preferential and unstable flow: from the sediment’s most biogeochemically active parts, and jeopardizing pore to the catchment scale. Vadose Zone J. 9, 207–212. groundwater quality. Gooddy, D.C., Clay, J.W., Bottrell, S.H., 2002. Redox-driven changes in porewater chemistry in the unsaturated zone of the chalk aquifer beneath unlined cattle slurry lagoons. Appl. Geochem. 17, 903–921. Acknowledgements Greve, A., Andersen, M.S., Acworth, R.I., 2010. Investigations of soil cracking and preferential flow in a weighing lysimeter filled with cracking clay soil. J. Hydrol. We thank Michael Kogel and Yuval Shani for their extensive 393, 105–113. Ham, J.M., Baum, K.A., 2009. Measuring seepage from waste lagoons and earthen efforts in constructing and operating the VMS, the dairy farm own- basins with an overnight water balance test. Trans. ASAE 52, 835–844. er for allowing us to conduct the research on his farm, and Ms. Sara Harris, G.L., Nicholls, P.H., Bailey, S.W., Howse, K.R., Mason, D.J., 1994. Factors Elchanani for fruitful discussions. The work was funded by the Is- influencing the loss of pesticides in drainage from a cracking clay soil. J. Hydrol. 159, 235–253. rael Water Authority. Comments and suggestions provided by Harter, T., Davis, H., Mathews, M.C., Meyer, R.D., 2002. Shallow groundwater quality Andrew M. O’Reilly and an anonymous reviewer helped to signifi- on dairy farms with irrigated forage crops. J. Contam. Hydrol. 55, 287–315. cantly improve this manuscript. Horn, R., Taubner, H., Wuttke, M., Baumgartl, T., 1994. Soil physical properties related to soil structure. Soil Till. Res. 30, 187–216. Issar, A., 1968. Geology of central coastal plain of Israel. Israel J. Earth Sci. 17, 16–29. References Jarvis, N.J., 1994. The MACRO model (version 3.1)—technical description and sample simulations. Rep. and Dissertations 19, Dept. Soil Sci., Swedish University of Acworth, R.I., Timms, W.A., 2009. Evidence for connected water processes through Agricultural Sciences, Uppsala, Sweden. smectite-dominated clays at Breeza, New South Wales. Aust. J. Earth Sci. 56, 81– Jarvis, N.J., 2007. A review of non-equilibrium water flow and solute transport in 96. soil macropores: principles, controlling factors and consequences for water Arnold, J.G., Potter, K.N., King, K.W., Allen, P.M., 2005. Estimation of soil cracking and quality. Eur. J. Soil Sci. 58, 523–546. the effect on surface runoff in a Texas Blackland Prairie watershed. Hydrol. Kurtzman, D., Scanlon, B.R., 2010. Groundwater recharge through Vertisols— Process. 19, 589–603. irrigated cropland versus natural land, Israel. Vadose Zone J. 10, 662–674. Arnon, S., Dahan, O., Elhanany, S., Cohen, K., Pankratov, I., Gross, A., Ronen, Z., Baram, McCurdy, M., McSweeney, K., 1993. The origin and identification of macropores in S., Shore, L.S., 2008. Transport of testosterone and estrogen from dairy-farm an earthen-lined dairy manure storage basin. J. Environ. Qual. 22, 148–154. waste lagoons to groundwater. Environ. Sci. Technol. 42, 5521–5526. Novák, V., Šimu˚ nek, J., van Genuchten, M.T., 2000. Infiltration of water into soil with Baram, S., Arnon, S., Ronen, Z., Kurtzman, D. Dahan, O. 2010. The Impact of Dairy cracks. J. Irrig. Drain. Eng. 26, 41–47. Farms on the Groundwater Quality of Israel’s Coastal Aquifer, Toward Oostindie, K., Bronswijk, J.J.B., 1992. FLOCR—a simulation model for the calculation Sustainable Groundwater in Agriculture – Linking Since and Technology, San of water balance, cracking and surface subsidence of clay soils. Report 47, DLO- Francisco, California. Available from:
Ini ltration Mechanism Controls Nitrii cation and Denitrii cation Processes under Dairy Waste Lagoon
S. Baram,* S. Arnon, Z. Ronen, D. Kurtzman, and O. Dahan
arthen waste lagoons are commonly used to Earthen waste lagoons are commonly used to store liquid wastes from store liquid wastes from concentrated animal feeding concentrated animal feeding operations. h e fate of ammonium + − operations (CAFOs) such as dairy farms. h e liquid waste (NH4 ) and nitrate (NO3 ) was studied in the vadose zone below earthen-clay dairy farm waste lagoons using three independent storedE in the lagoons contains high concentrations of nitrogen vadose zone monitoring systems. h e vadose zone was monitored (N), phosphorus (P), salts, organic compounds, and other nutri- from 0.5 to 30 m below land surface through direct sampling of the ents. For economic reasons, most dairy farms store the liquid sediment porewater and continuous measurement of the sediment waste on-site in earthen (soil-lined or unlined) lagoons rather proi le’s water content variations. Four years of monitoring revealed that wastewater ini ltration from the lagoon is controlled by two than in plastic-lined lagoons, steel tanks, or concrete tanks (Ham mechanisms: slow (mm d−1), constant ini ltration from the lagoon and DeSutter, 2000). However, this practice fails to prevent the bed; and rapid (m h−1) ini ltration of wastewater and rainwater via substantial leaching of organic-nitrogen (organic-N) and ammo- preferential l ow in desiccation cracks formed in the unsaturated + nium nitrogen (NH4 –N) into the subsurface (Gooddy et al., clay sediment surrounding the lagoon banks. h e preferential l ow 1998, 2002; Ham and DeSutter, 2000; Ham, 2002; Harter et al., mechanism is active mainly during wastewater-level l uctuations and intensive rain events. h e vadose zone below the waste 2002; DeSutter et al., 2005). sources remained unsaturated throughout the monitoring period, To minimize downward or lateral seepage of the wastewater + and all ini ltrating NH4 was oxidized in the upper 0.5 m. h e from earthen waste lagoons and prevent groundwater + − NH4 oxidation (nitrii cation) was coupled with NO3 reduction contamination, it is recommended that the saturated hydraulic (denitrii cation) and depended on the sediment water content, conductivity coei cient (K ) of the earthen lining would be less which was controlled by the ini ltration mechanism. Coupled sat −6 −1 nitrii cation–denitrii cation (CND) resulted in 90 to 100% than a prescribed level (Ksat < 1 × 10 cm s , the level in many reduction in the total nitrogen mass in the vadose zone, with higher U.S. states) (SCS, 1997). Hence, during the construction of the removal under high water content (~0.55 m3 m−3). Mass balance of waste lagoon, clay is typically mixed with local sediment and − nitrogen and isotopic composition of NO3 indicated that CND, compacted to form an earth liner along its bottom and banks. rather than cation exchange capacity, is the key factor regulating Following the introduction of wastewater into the lagoon, nitrogen’s fate in the vadose zone underlying earthen waste lagoons. the hydraulic conductivity of the earth lining will most likely be reduced by at least an order of magnitude due to physical, chemical, and biological processes, commonly termed seal formation (SCS, 1997; Cihan et al., 2006; Tyner et al., 2006). Laboratory-scale experiments on dairy waste ini ltration into clay, loam, and sand sediments have shown that all sediment types have similar ini ltration l uxes (4.6 to 6.9 × 10−7 cm s−1) within 10 d of manure application (Culley and Phillips, 1982). h e ef ect of the hydraulic parameters of the organic seal and the underlying sediment on the ini ltration rate from animal waste lagoons was analyzed by Tyner and Lee (2004) using a steady- state two-layer model. h eir model predicted that sediment thickness has no ef ect on the ini ltration rate, and sensitivity analysis predicted that ini ltration rate is highly dependent on seal hydraulic conductivity and minimally dependent on
Copyright © 2012 by the American Society of Agronomy, Crop Science Society S. Baram, S. Arnon, Z. Ronen, and O. Dahan, Dep. of Environmental Hydrology & of America, and Soil Science Society of America. All rights reserved. No part of Microbiology, Zuckerberg Institute for Water Research, Blaustein Institutes for this periodical may be reproduced or transmitted in any form or by any means, Desert Research, Ben-Gurion Univ. of the Negev, Sede Boqer Campus, Israel 84990; electronic or mechanical, including photocopying, recording, or any information D. Kurtzman, Institute of Soil, Water and Environmental Sciences, Agricultural storage and retrieval system, without permission in writing from the publisher. Research Organization, The Volcani Center, Bet Dagan 50250, Israel. Assigned to Associate Editor Jim Miller. J. Environ. Qual. 41 doi:10.2134/jeq2012.0015 Abbreviations: BLS, below land surface; CAFO, concentrated animal feeding Received 8 Jan. 2012. operation; CEC, cation exchange capacity; CND, coupled nitrii cation– *Corresponding author ([email protected]). denitrii cation; FTDR, l exible time domain rel ectometry; USS, ultraviolet © ASA, CSSA, SSSA spectrophotometric screening; VMS, vadose zone monitoring system; VSP, vadose 5585 Guilford Rd., Madison, WI 53711 USA zone sampling port.
1623 − − sediment hydraulic conductivity. Cihan et al. (2006) developed microhabitats, nitrite (NO2 ) or NO3 produced during a model for describing the sealing process in dairy and swine nitrii cation can be utilized by denitrii ers. h is process is waste lagoons with time. At er a stable seal develops, the model termed coupled nitrii cation–denitrii cation (CND). Coupled predicts that the seal properties, and not the sediment properties, nitrii cation–denitrii cation plays an important role in the are responsible for limiting ini ltration. Furthermore, Tyner et removal of nitrogenous compounds in sediments (Nielsen and al. (2006) showed, in a laboratory experiment, that within 85 Revsbech, 1998; Kremen et al., 2005). d of manure application to undisturbed columns of silt loam, Studies on the fate of N species in the vadose zone and the sediment proi le becomes unsaturated. h ey indicated groundwater in dairy earthen waste lagoon environments have that unsaturated conditions appeared immediately below the shown that their fate varies from place to place. Monitoring of − sediment–liquid interface. “In situ” i eld measurements showed NO3 concentration in the porewater of the vadose zone under that seepage rates from earthen lagoons are typically on the order and around earthen lagoons constructed in sand to clay loam of several millimeters per day or less (Miller et al., 1985; Parker sediments demonstrated that during lagoon operation, in some − et al., 1999c; Ham and Baum, 2009). h ese seepage rates were cases NO3 concentrations remained similar to background observed in various sediment types (sand to clay) and validated concentrations (Meyer et al., 1972; Parker et al., 1999c), and in the i ndings of Cihan et al. (2006) and Tyner et al. (2006) that others they reach very high concentrations (>900 mg L−1) (Oliver the hydraulic conductivity of the seal, rather than that of the and Meyer, 1974; Korom and Jeppson, 1994). Groundwater − sediment, controls the ini ltration rate. While the potential monitoring for NO3 contamination from earthen lagoons has seepage loss from the bottom of waste lagoons has been well demonstrated that the impact of the lagoon varies with time studied, that from the lagoon banks, which are subjected to from its initial operation (Sewell, 1978). While in some studies, − wetting/drying cycles during frequent water-level l uctuations, elevated NO3 concentrations in groundwater were attributed warrants further research (Parker et al., 1999c). to lagoon seepage (Nordstedt et al., 1971; Withers et al., 1998), − Clayey earth lining is very attractive for use in waste lagoons in others, elevated NO3 could not be distinguished against due to its low hydraulic conductivity. However, clay sediments the background levels from manure application in surrounding are highly sensitive to water content and tend to form desiccation i elds (Harter et al., 2002). A recent study has even shown that − cracks under unsaturated conditions (Chertkov and Ravina, the NO3 concentration in the saturated zone varies with depth 1998; Parker et al., 2001). h ese desiccation cracks serve as due to denitrii cation processes in the groundwater column preferential l ow paths for water and contaminants. Quick rises (Singleton et al., 2007). in sediment water content following rain events and temporal The objective of this research was to study the fate of − wastewater overl ows were observed across a clay vadose zone NO3 in the clay vadose zone underlying dairy earthen exceeding 10-m depth (Baram et al., 2012). h e ini ltration waste lagoons and their margins, by implementing a unique velocities calculated from these wetting events (0.4 to 23.6 m h−1) array of vadose zone monitoring systems (VMSs). More indicated that the desiccation-crack network in the unsaturated specifically, this study evaluates the impact of infiltration + clay sediment remains open and serves as a preferential l ow mechanism and sediment water content on NH4 oxidation pathway year-round, even at high sediment water contents and denitrification in the subsurface. (~0.50 m3 m−3). Kurtzman and Scanlon (2011) concluded from the chemical composition of the vadose zone sediment Materials and Methods and groundwater that preferential l ow is the major mechanism Study Area dominating the recharge processes in the noncultivated vertisol h e study area was located in the Beer Tuvia region (40 km2), of the region in which this research was performed. above the southern part of the Coastal Aquifer in Israel. h e h e main concern with seepage from earthen waste lagoons phreatic aquifer in the area is composed of calcareous sandstone is the risk to groundwater quality. Organic-N and NH + are 4 (Kurkar Group) overlaid by Pleistocene-age clays (Issar, 1968; the dominant N species in CAFO anaerobic waste lagoons, Weinberger, 2007). h e land surface area is covered by a clay whereas in their subsurface, nitrate (NO −) is the most common 3 layer, dominated by smectite minerals, with a typical thickness contaminant. High NO − levels in drinking water have been 3 spanning 3 to 12 m. h e climate is characterized as Mediterranean associated with risks for methemoglobinemia (blue baby with average summer and winter temperatures of 24.3°C and syndrome) in infants and diarrheal and respiratory diseases, 14.2°C, respectively. h e average annual precipitation is ~450 and have even been suggested to be a risk factor for specii c mm, falling during the winter season (November to March) cancers (Ward et al., 2005). Accordingly, many countries and mostly in i ve to eight rainy episodes. h e main recharge to organizations around the world have provided guideline values the aquifer is related to percolation of seasonal rainwater and for maximum NO − levels in drinking water (for example: 3 agricultural return l ow from irrigated and rain-fed i elds that World Health Organization—50 mg L−1, United States— have been intensively cultivated in the area for >60 yr. Over the 45 mg L−1, Israel—70 mg L−1 [Ward et al., 2005; WHO, 2007; past four decades, the quality of the groundwater in the Beer Israel Internal Af airs and Environment Committee, 2010]). Tuvia region has been continuously deteriorating, as rel ected by Microbial oxidation of ammonia (NH ) and NH + into 3 4 a gradual increase in chloride concentrations (>600 mg L−1) and NO − (nitrii cation) in porous medium requires the presence 3 local spots of high NO − concentration (Weinberger, 2007). of molecular oxygen (O ) (Prosser, 1989). Nitrate can be 3 2 h e Beer Tuvia region currently hosts ~10,000 cows in 140 further reduced by microorganisms into N gas (N ) through 2 dairy farms. h e dairy farm selected for this study has 60 dairy denitrii cation. In sediments where favorable conditions for cows and 30 heifers and calves, which is representative of the both nitrii cation and denitrii cation are present in neighboring 1624 Journal of Environmental Quality region. Wastewater and liquid manure from the farm (7 m3 d−1) is borehole axis, while ensuring proper attachment of the probes constantly discharged into a single-stage earthen unlined lagoon to the sediment facing the upper side walls of the borehole. (also commonly termed a waste storage pond) without treatment Assuming that the general l ow direction in the vadose zone is and remains untreated. h e lagoon is roughly 20 m long, 10 m rather vertical, each point on the upper side of a slanted borehole wide, and 1 m deep. Wastewater overl ow is constantly drained faces an undisturbed sediment column (Fig. 1). into a waste channel roughly 200 m long, 2 m wide, and 1 m h e volumetric water content of the sediment was measured deep. h e dairy farm has been operating in the same way for the using the FTDR probes operated with a TDR100 instrument past 50 yr and no specii c maintenance procedures, such as solids (Campbell Scientii c, Logan, UT). Time domain rel ectometry removal, have been used at the site. Both the waste lagoon and measurements were calibrated by accounting for the specii c the channel are l ooded with liquid manure year-round. Natural structure of the FTDR probe, instrumentation setup, variety of annual shallow-rooted plants (mainly Mala sylestris L.) grow soil types (including the clay at the site), and salinities (Baram et on the waste source banks from November to April. al., 2012). h e overall accuracy of water content measurements h e vadose zone underlying the waste lagoon and drainage under i eld conditions using the VMS is ±5% of the measured channel is composed of clay sediments extending from land surface value (Rimon et al., 2011). to depths of 6 m and 12 m, respectively. h e clay overlies sand h e sediment porewater from the unsaturated zone was and calcareous sandstone down to the water table at ~46 m below sampled using VSPs. h e VSPs are essentially suction lysimeters land surface (BLS). Particle size distribution of the clay sediment that have been modii ed for operation with the VMS in deep in the area is composed of 26% sand, 22% silt, and 52% clay, 90% sections of the vadose zone. Porewater sampling from unsaturated of which consists of illite–smectite minerals. Desiccation cracks sediments is achieved by creating hydraulic continuity between form regularly in the clay sediment and create polygons with the sediment and the sampling cell using a l exible porous average dimensions of 1.0 × 0.70 m2, separated by cracks with an interface (Dahan, 2005, 2008, 2009a,b; Dahan et al., 2009). average aperture of 0.055 ± 0.017 m. h e cracks extend to a depth h e vadose zone at the study site was monitored using three of 0.65 ± 0.19 m (cracks are dei ned when the aperture is >0.006 independent VMSs that were installed under the waste lagoon m). h e desiccation-crack network remains active year-round, and its drainage channel. One VMS with i ve monitoring units even at er rain events and under high water contents (>0.50 m3 was installed in a 35-m-long borehole extending from the lagoon m−3). A detailed discussion of the role of the desiccation cracks margins (1.5 m from the lagoon bank) toward the center of the in the water-ini ltration process through the clay sediment at the waste lagoon to a vertical depth of 30.5 m (I in Fig. 1a and Table studied site can be found in Baram et al. (2012). 1). Two additional VMSs (II and III in Fig. 1b), with seven monitoring units each, were installed ~150 m downstream of Monitoring Approach and Field Instrumentation the waste lagoon in 24-m-long boreholes under the center of the h e hydraulic and chemical properties of the sediment in the waste channel and 2 m away from the channel bank to a vertical vadose zone underlying the waste lagoon and its margins were depth of 21.3 m (Table 1). In addition to the monitoring units evaluated by VMS. h e VMS is designed to collect continuous real-time measurements of the sediment water content, and to allow frequent sampling of the sediment porewater across the vadose zone (Dahan et al., 2007, 2008; Rimon et al., 2007, 2011). h ese latter publications present a detailed description of the monitoring system and its installation procedure. h erefore, in this article, only a brief description is provided of the monitoring concept and the technical aspects that were important for evaluating the presented data. h e VMS is composed of a l exible sleeve made of a thin, l exible liner hosting a set of l exible time domain rel ectometry (FTDR) probes and vadose zone sampling ports (VSPs). h e VMS is installed in a 140-mm-diameter, uncased, slanted borehole (30–35° from the vertical). During the installation, the sleeve is inserted into the borehole and i lled with two-component high-density liquid urethane (~1.65 kg L−1) that solidii es shortly at er its application. Fig. 1. Schematic illustration of the vadose zone monitoring systems (VMS) and observation h e hydrostatic pressure generated by the well installed in the sediment proi le under (a) the waste lagoon and its margins, and (b) the waste channel and its margins. The monitoring units (c) include (1) a l exible time domain liquid urethane inside the sleeve expands the rel ectometry (FTDR) probe, and (2) a vadose zone porewater sampling port (VSP). (a) shows sleeve and seals the borehole void to prevent a side view of the waste lagoon while (b) shows a front view of the waste channel, and thus generation of preferential pathways along the VMSs II and III appear as vertical lines in (b) although they were installed in slanted boreholes. www.agronomy.org • www.crops.org • www.soils.org 1625 that were installed via the VMS, three conventional suction Chemical Analysis lysimeters (0.06 m long, 0.02 m in diameter; A.M.I., Ashdod, Concentrations of N species (NO −–N, NO −–N, NH +–N, Israel) were installed in the sediments underlying the central part 3 2 4 and total N) were determined within 48 h at er sampling. h e of the waste lagoon (0.5 and 1.4 m below the lagoon bottom), NO −–N concentration was measured using two methods: (i) and at the interface between the bottom of the lagoon and 3 ultraviolet spectrophotometric screening (USS) (APHA 4500- the underlying sediments (Fig. 1a and Table 1). h e suction NO3 B), and (ii) ion chromatography (Dionex-4500i, Sunnyvale, lysimeters were installed in 0.05-m holes that were backi lled CA). Ion chromatography was also used to measure NO −–N, with a few centimeters of i ne quartz sand (<75 μm) and sealed 2 and the phenate method was used for NH +–N (APHA 4500- to the surface with bentonite clay. A groundwater observation 4 NH3 F). Total N was determined using the persulfate digestion well was drilled a few meters away from the lagoon bank to a method (APHA 4500-N C) (APHA, 1998). Accuracy of the depth of 54 m with a screen interval from the water table at 46 m measured concentrations was validated using 10% duplicates and BLS to the bottom of the well. spike recovery tests. Matrix interference was not observed in the h e volumetric water content of the sediment in the vadose USS method, and high recovery (91–98%) was observed in all zone was measured every 60 min during the dry season (April– cases. During the i rst 2 yr (2007 and 2008), the USS method October) and every 15 min during the wet season (November– was used to determine the NO −–N concentration; later, ion March). Water samples from the lagoon, vadose zone, and 3 chromatography was used. groundwater were collected every 6 to 12 wk. h e wastewater δ18O and δ15N in NO − of porewater samples from the vadose samples were collected using suction lysimeters. Groundwater 3 zone were measured at the Laboratory of the Israel Geological was sampled from the observation well using a submersible Survey, using an isotope ratio mass spectrometer (Delta Plus pump (Model MP1, Grundfos, Denmark). All water samples XP h ermo Scientii c, Braunschweig, Germany), at er chemical were stored in polypropylene bottles and kept on ice until they conversion of NO − to nitrous oxide (McIlvin and Altabet, reached the laboratory (<12 h), where they were i ltered (45-μm 3 2005). Values are given as mean ± standard deviation, with total glass i ber i lter) and kept at 4°C until analysis (<2 wk). number of samples (n) in parentheses.
Table 1. The monitoring setup. Ini ltration Rates
Monitoring unit depth (m)† Sediment No. of porewater Conductance (K/Δx) of the waste lagoon bed and VSP‡ FTDR type samples waste channel bed was measured “in situ” using pipe 0 Lagoon sludge 14 ini ltrometers that were hammered into the top ~15 Vadose zone underlying the waste lagoon cm of the clay soil (where K [L t−1] is the bulk hydraulic 1.5 Clay 16 conductivity of the unknown thickness Δx [L] of 2.4 Clay 14 the soil through which ini ltration takes place, under 6.5 6 Clay 15 positive pressures, before reaching the point of drainage 10.5 10 Calcareous sandstone 8 in unsaturated conditions [L]). h e ini ltrometers 15.5 15 Calcareous sandstone 18 (2 and 1 m long with diameters of 0.35 and 0.2 m, 20.5 20 Calcareous sandstone respectively) were i lled with wastewater i ltered to 30.5 Calcareous sandstone 15 remove coarse l oating particles (using sieve ASTM Vadose zone underlying the waste lagoon margins No. 40; Ari. J. Levi, Petach Tikva, Israel) and covered 2.5 2 Clay 14 at the top to eliminate water loss due to evaporation. 3.5 3 Clay 14 h e l ux was evaluated by measuring the drop-in water Vadose zone underlying the waste channel level inside the ini ltrometer at nine dif erent locations 2.4 1.7 Clay 8 around the lagoon and channel. Ini ltration was 4.5 3.9 Clay 9 measured under a range of hydraulic heads between 0.1 6.5 Clay 8 m below and 1.2 m above the wastewater level. In each 8.7 8.1 Clay 7 experiment, the water level was allowed to drop by up 10.7 10.2 Clay 8 to 5% of the initial level to measure ini ltration under 14.0 13.5 Calcareous sandstone 3 relatively constant water head. 21.3 19.8 Calcareous sandstone 2 Vadose zone 2 m away from the overl ow waste channel bank Sediment Sampling 2.3 1.6 Clay 4 Detailed sampling of the water content and organic 4.3 3.7 Clay 7 matter content in the upper 0.1 m of the sediment under 6.5 6.0 Clay 8 the waste lagoon and channel was performed by exposing 8.4 7.9 Clay 4 small sections of the lagoon and channel bottom in the 10.5 10.0 Clay 4 following manner. A metal cylinder (0.9 m long and 0.6 13.8 13.3 Calcareous sandstone 4 m in diameter) was i rst pushed through the wastewater 21.1 19.8 Calcareous sandstone into the waste lagoon and channel bottom. h en, before Groundwater 12 sampling, the sediment, wastewater, and 0.1 to 0.2 m † Vertical depth measured relative to land surface at each site. of i ne organic matter that had settled on the bottom ‡ VSP, vadose zone sampling port; FTDR, l exible time domain rel ectometry probe. were carefully removed from the cylinder, exposing
1626 Journal of Environmental Quality the bottom. Sediment cores were then collected Table 2. Seepage rates under dairy and cattle waste lagoons. from the exposed surface using plastic cylinders Seepage rate Reference Sediment type (0.11 m long and 0.028 m in diameter) and kept Range Avg. on ice until they were brought to the laboratory. —— mm d−1 —— h e core samples were immediately dissected to Full-scale seepage studies 0.005-m-thick slices that were analyzed for water Meyer et al. (1972) Sand–clay loam NA† ~1 content (oven-dried at 105°C for 72 h) and total Parker et al. (1999c) and references therein Sandy loam–clay loam 0.26–0.87 0.48 organic matter content through combustion Ham (2002) 18–30% clay 0.2–2.4 1.1 (450°C for 4 h) (Nelson and Sommers, 1996). This study Clay 0.64–7.6 2.4 Altogether, six dif erent locations were sampled Small-scale laboratory studies under the center of the channel and the center of Parker et al. (1999c) and references therein Sand–clay 0.24–36 3.12 the lagoon (three each). Cihan et al. (2006) Sand–clay NA ~1 + Sampling of the water content and NH4 –N Tyner et al. (2006) Silt loam NA 0.7 distribution in the upper 0.5 m of the sediment † NA, data not available. under the waste lagoon was performed once a year from three random locations by hammering laboratory studies with various sediment types (sand to clay) a metal pipe (0.05 m in diameter and 0.5 m long) into the lagoon (Table 2). h e similarities between the ini ltration rates through bed. Two cores were taken from each location. Sections of 0.1 m dif erent types of sediment further support conclusions of from one core were used to measure the volumetric water content Tyner et al. (2006) and Cihan et al. (2006) that the hydraulic of the sediment, while 0.1-m sections from the other core were conductivity of the organic seal rather than the hydraulic + conductivity of the underlying sediment matrix controls the extracted for NH4 –N using 2 M KCl (Maynard et al., 2008). ini ltration rate from the bottom of waste lagoons. Saturation of the clay sediment (θsat) was evaluated using 10 undisturbed core samples (0.02 m long and 0.01 m in diameter) h e organic matter content and the water content in the clay taken from 10 to 20 cm BLS. h e samples were oven-dried (at sediment below the lagoon and channel rapidly decreased from the surface down to a depth of 2 cm, while they remained relatively 105°C for 72 h), l ushed with CO2 gas (5 min), and then placed in degassed water. At er 10 h the samples were weighed and constant below 2-cm depth (Fig. 2). h e water content in the their volume was remeasured to obtain their volumetric and upper 1.5 cm resembled the laboratory-measured water contents −1 gravimetric saturated water contents. at saturation (θsat 0.64 ± 0.06 g g ), indicating that unsaturated conditions appeared immediately below that depth (1.5 cm). Results and Discussion h is water content proi le indicates that the bulk hydraulic While measurements of the ini ltration pattern from the conductivity of the upper few centimeters of the clay layer lagoon and channel bottom enabled assessment of water losses (matrix and desiccation cracks under unsaturated conditions) is through the sediment matrix, changes in sediment water content lower than that of the rest of the clay proi le to depths of 6 to 12 enabled detection of water losses following preferential l ow m. Reduction of the hydraulic conductivity in the top section from the lagoon and channel banks. Combining results of of sediments below the wastewater lagoon has been attributed both methods showed that ini ltration of polluted wastewater to seal formation (Rowsell et al., 1985; SCS, 1997; Cihan et al., from the dairy farms to the subsurface can occur via four major 2006; Tyner et al., 2006). h e linear correlation between the 2 regions: (i) the area under the waste lagoon bottom, which is water content and organic matter content in the upper 6 cm (R 2 permanently l ooded; (ii) the area along the waste lagoon banks, = 0.66 for the lagoon, R = 0.99 for the channel) suggests that which is subjected to l uctuations in wastewater level; (iii) the seal formation is driven by clogging of the clay sediment pores by area under the waste channel, which is permanently l ooded; and (iv) the area underlying the waste channel margins, which is subjected to occasional l ooding by wastewater overl ows and rain events. Monitoring of the waste channel area, a narrow waste source proximal to two hydraulically active banks, represents the process occurring at the lagoon banks, while monitoring of the lagoon represents the whole lagoon area including the area far from the banks (Fig. 1). Ini ltration from the Bottom of the Waste Sources and Seal Formation Average ini ltration l uxes from the waste lagoon and channel bottom, estimated from the pipe ini ltrometers, was 2.4 mm d−1 (n = 20) (Table 2) with an average conductance value of 4.4 × 10−8 s−1. No signii cant dif erences were observed between the ini ltration l uxes measured in the lagoon and channel (p = 0.22). h e ini ltration rates measured in this study were very similar to Fig. 2. Organic matter content and water content in the sediments under the waste channel and the waste lagoon. Results are presented measurements conducted previously in earthen lagoons and in as averages, and the horizontal bars are the standard deviation. www.agronomy.org • www.crops.org • www.soils.org 1627 organic matter. h is organic seal is produced by either biological preferential l ow paths being generated. h e crack intensity in activity (SCS, 1997; Parker et al., 1999c) or physical settling of clayey sediments, such as in our study site, is more pronounced i ne organic particles from the wastewater (SCS, 1997; Parker et at the waste source (lagoon or channel) margins and banks. As al., 1999c; Tyner et al., 2006). a result, at the lagoon banks there are interfaces between the water source, the wet sediments (under the source), and the dry, Deep Percolation cracked hydraulically active sediments. h erefore, the dimensions Percolation of water through the vadose zone was evaluated by of the waste source on the surface have a strong inl uence on the spatial and temporal variations in the sediment water content water content distribution in the subsurface, dependent on the across the unsaturated zone beneath the waste sources and their extent of the wet–dry interfaces at the banks. For example, the margins (Fig. 3). Water content of the clay sediment underlying vadose zone under a narrow lagoon is near two hydraulically the central parts of the lagoon was signii cantly lower (~0.40 m3 active banks. Preferential water ini ltration from the banks and m−3) than that under the channel (0.55–0.62 m3 m−3). Temporal redistribution into the sediment matrix over such a narrow area variations in sediment water content across the unsaturated sustains high water content in the vadose zone (i.e., the channel, proi le at both sites showed dramatic variations throughout the with water content of 0.50–0.60 m3 m−3; Fig. 3). On the other year: although these variations were more pronounced during hand, the sediment under the central part of a wide lagoon is the rainy season (November–April) following intensive rain relatively more distant from the hydraulically active banks. h is events, they were also observed throughout the dry summer area will be less af ected by the ini ltration processes at the banks season (May–October) when no rain was recorded (see vertical and will sustain a lower water content (i.e., the lagoon, ~0.40 dashed arrows on Fig. 3). Baram et al. (2012) attributed these m3 m−3; Fig. 3). h e described mechanism explains the elevated rapid changes in water content to quick ini ltration and drainage water contents observed under the lagoon banks (0.40–0.60 m3 of percolating water through preferential l ow paths formed by m−3) (Baram et al., 2012). Similar observations have been made desiccation-crack networks in the unsaturated clay sediment. by Parker et al. (1999a, 1999c), who suggested that drying/ h ey showed that the desiccation-crack network remains open freezing processes enhance seepage from the lagoon banks, and and hydraulically active year-round, and serves as a fast conduit by Gooddy et al. (2002), who suggested that a large component for water movement, even under conditions of near saturation of the wastewater migrates through the lagoon banks.
(laboratory-measured water contents at saturation θsat were 0.63 ± 0.04 m3 m−3). h e repeated observations of rapid increase in Ini ltration Mechanism and Nitrogen Transformations sediment water content during the dry season indicated that h e dominant N species in the dairy farm wastewater was + −1 preferential ini ltration events occur regularly from the lagoon NH4 –N, with an average concentration of 2012 ± 482 mg L (n − and channel banks during l uctuations in the wastewater level = 14). h e NO3 –N concentration in the wastewater was always −1 − and overl ows: the wastewater l ows through the desiccation- below 6.8 mg L (n = 14), and NO2 –N was below detection + crack network into the clay matrix and redistributes in the (n = 14). h e NH4 –N concentrations in the sediment (KCl subsurface, sustaining elevated water contents in the clayey extract) under the lagoon and the channel decreased dramatically vadose zone around the lagoon and channel banks. in the upper shallow cross section, from 2700 to 4200 mg kg−1 h is mechanism suggests that the water content in the bulk dry sediment at 0.05 m to ~10 mg kg−1 dry sediment at a depth of + sediment is highly dependent on the intensity of the crack 0.45 m (Fig. 4). Moreover, NH4 –N concentration in porewater network and on proximity to the cracks. Crack intensity and samples collected from deeper parts of the vadose zone (>0.5 m) connectivity increase as the sediment dries (Chertkov and remained very low (<5 mg L−1) throughout the entire sampling Ravina, 1998), increasing the likelihood of interconnected period (over 200 samples in 26 sampling campaigns from January 2007 and January 2011). + h e disappearance of NH4 –N from the vadose − zone was accompanied by NO3 –N formation (Fig. 5), + suggesting intensive microbial oxidation of NH4 –N to − NO3 –N (Fig. 4 and 5), which requires molecular O2 (Francis et al., 2007). Gooddy et al. (1998) showed a clear + − inverse relationship between NH4 and NO3 , which was demonstrated to be driven by oxygen availability. We therefore assume that aerobic conditions exist within the clay vadose zone under the l ooded waste lagoon and channel. h e development of aerobic conditions might be explained by the development of unsaturated conditions in the vadose zone (Fig. 2), and the well- developed desiccation-crack networks at the banks which further enhance aeration of the vadose zone. It is Fig. 3. Temporal changes in sediment water content measured by l exible time likely that aerobic conditions exist in regions close to the domain rel ectometry probes under the waste channel and waste lagoon desiccation cracks, while anaerobic conditions prevail in (continuous lines) and by gravimetric methods on sediment cores (squares), sediment microstructures within the clay matrix which along with daily cumulative rain events during 2008–2011 (bars). Dashed arrows indicate events in which sharp increases in sediment water content were not are not well aerated by these cracks. h e constant supply + associated with rainstorms. of organic-N and NH4 –N and the proximity of aerobic 1628 Journal of Environmental Quality − and anaerobic zones suggest that the NO3 –N derived from nitrii cation is immediately available for denitrii cation, resulting in CND. Coupled nitrii cation–denitrii cation can occur in all sediment types where favorable conditions for both nitrii cation and denitrii cation are present in neighboring microhabitats (Kremen et al., 2005). h e abundance of both nitrifying and denitrifying bacteria and the nitrii cation and denitrii cation potentials were determined in sediment samples collected from the lagoon bed down to a depth of 0.5 m. A detailed description of the methods used and the dif erent microbial groups can be found in Sher et al. (2012). h e high number of nitrifying and denitrifying bacteria (~108 gene copies g dry sediment−1 of each) and the high nitrii cation and denitrii cation potentials found in the uppermost sediment suggest that CND is likely to occur under the lagoon and channel. + Examinations of NH4 –N proi les from other studies on Fig. 4. Vertical distribution of NH +–N concentrations in the sediment seepage from earthen waste lagoons have shown similar rapid 4 + at three dif erent locations underneath the waste lagoon and three attenuation of NH4 –N with depth (Ham and DeSutter, 2000; dif erent locations underneath the waste channel. Ham, 2002; DeSutter et al., 2005). h ese studies were conducted in soils with clay contents between 18 and 46%. In those studies, composition in the shallow clay proi le under the waste channel, + since the NO −–N concentrations in that area were mostly close the rapid attenuation of NH4 –N was attributed to the cation 3 exchange capacity (CEC) of the sediment under the lagoon. To to zero (Fig. 5 and 6). Nevertheless, the isotopic composition compare our results with those of Ham and DeSutter (2000), in one of the porewater samples from 4 m below the channel − Ham (2002), and DeSutter et al. (2005), we calculated the total (a sample with relatively high NO3 –N concentration), along − N mass that had ini ltrated into the subsurface by multiplying with the low NO3 –N concentrations in that area, suggest that the inl ux of wastewater by the operation period and by the denitrii cation was more prominent in locations with relatively + − high water contents (0.50–0.60 m3 m−3). h is observation is concentrations of organic-N, NH4 –N, and NO3 –N in it. h e total N mass ini ltrated into the subsurface was then compared in agreement with other studies, for example, Arah and Smith to the total N mass stored in the subsurface (integrating the N-species concentration proi les reported in those papers). Our calculations showed that in all of those studies, regardless of the clay content, >90% of the N mass was removed during transport in the subsurface. In most cases, the removal occurred within the upper 2 m of the sediment proi le. h e results from the current study showed that 85 to 100% of the N mass is removed from the subsurface (above 1.5 m) under the lagoon and channel through CND (Fig. 4 and 5). h erefore, while high CEC values can explain some attenuation + of NH4 –N concentration in the sediment proi le, it cannot explain the extent of N removal from the subsurface environments under waste lagoons. − Clear dif erences between NO3 –N concentrations in the propagating porewater underlying the waste lagoon and waste channel − were observed. Whereas NO3 –N concentration in the clay sediments underlying the lagoon was very high (183–524 mg L−1), the concentrations under the waste channel were very low (0–29 mg L−1) (Fig. 5). h e isotopic composition − 15 18 Fig. 5. Vertical distribution of (a) NO −–N concentrations in porewater from the vadose zone of NO (δ N and δ O) in the propagating 3 3 beneath the waste channel and the waste channel margins, and (c) beneath the waste porewater had a distinct denitrii cation signature lagoon and the waste lagoon margins. Vertical sediment water content proi les represent (Kendall, 1998) (Fig. 6), further supporting the the average value measured by the l exible time domain rel ectometry probes (b) under assumption of CND reactions occurring in the the channel and its margins and (d) under the lagoon and its margins. Concentrations in the groundwater presented at 47 m. Results are presented as averages and the horizontal vadose zone under the waste lagoon and channel. bars are the standard deviation. Transition from clay to sand/loam occurs at 6 m below − We were unable to evaluate the NO3 isotopic land surface under the lagoon and at 12 m under the channel (Fig. 1). www.agronomy.org • www.crops.org • www.soils.org 1629 no seal has formed. h ey suggested that wastewater ini ltration would generate saturated anaerobic conditions in the sediment underlying the lagoon, while the lagoon banks would remain unsaturated + and aerobic. Accordingly, NH4 –N would be the dominant N species in the anaerobic saturated − environment under the lagoon, while NO3 –N would be the dominant N species in the aerobic unsaturated environment under the lagoon banks. In this work, the inl uence of the percolation mechanisms on the sediment water content and the consequent fate of the N species in the vadose zone beneath an unlined dairy waste lagoon where a seal has formed are illustrated in Fig. 7. h ree dif erent zones should be considered when studying the fate of redox-related compounds in seepage from clayey earthen waste lagoons. h e i rst zone is the vadose zone under the center of − 15 18 the waste lagoon (away from the banks). Water in Fig. 6. Isotopic composition of NO3 (δ N and δ O) in the propagating porewater from the vadose zone beneath the waste lagoon, waste channel, and their margins. The isotopic this zone continuously ini ltrates from the bottom composition of dairy manure is based on Kendall (1998). l ooded area at relatively low l uxes (mm d−1). Clogging of the clay matrix at the bottom of the (1989) and Schurgers et al. (2006). Bakken and Dörsch (2006) l ooded area by i ne organic matter (seal formation) leads to the further suggested that sediment water content acts almost like an development of unsaturated conditions (70% saturation) in the on/of switch with respect to denitrii cation. Because sediment underlying sediment of the vadose zone. h e desiccation-crack aeration is dependent on water content, it is reasonable to assume network developing in the unsaturated clay allows air penetration that as the conditions in the vadose zone approach saturation, and formation of aerobic conditions with anaerobic niches where more anaerobic niches will be generated and denitrii cation CND occurs. Coupled nitrii cation–denitrii cation leads to will become more prominent. Despite the importance of water substantial reduction in N mass. content on the fate of N under CAFOs, a number of key studies h e second zone consists of the banks of the waste lagoon. on seepage from waste lagoons did not report the values of water h is zone is subjected to continuous slow ini ltration under its content in the sediments (Oliver and Meyer, 1974; Korom and l ooded section as well as to repeated, rapid (m h−1) preferential Jeppson, 1994; Ham and DeSutter, 2000; Gooddy et al., 2001; ini ltration via the desiccation-crack network formed in the Harter et al., 2002; DeSutter et al., 2005). On the other hand, dryer bank areas during l uctuations in wastewater level. the work of Parker et al. (1999b) revealed that NO −–N is not 3 Redistribution of the preferentially ini ltrating wastewater found beneath the lagoon bottom, but is found in isolated areas generates higher sediment water content (90% saturation) and beneath the lagoon side-slopes where higher water contents are reduces its aeration, such that aerobic conditions exist in niches. observed in the sediment. In this section of the vadose zone, CND leads to nearly complete h e results of the study by Parker et al. (1999b) and those removal of the propagating N mass. reported here strongly suggest that the water content and h e third zone is the area underlying the waste lagoon consequent fate of N in the vadose zone are impacted by proximity margins. h is area is subjected to rare l ooding events from to the waste source banks. h e mechanism governing N-species wastewater overl ow and to deep percolation during intensive transport under the center of large waste lagoons may dif er rain events. h e vadose zone in this section is characterized signii cantly from that at the banks of the l ooded area. While by low water content (45% saturation) and contains a well- minor l uctuations in water level in the waste lagoon or channel developed desiccation-crack network. Aerobic conditions have a negligible ef ect on the ini ltration rate from the lagoon prevail in the vadose zone and wastewater overl ow preferentially bed, these l uctuations might cause seepage of wastewater through + ini ltrates into its deep sections, where NH4 –N is oxidized to desiccation cracks formed at and near the l ooded area’s banks, − − NO3 –N. h e NO3 –N propagates deeper into the vadose zone leading to wetting and drying cycles in the subsurface. When with minor transformation. preferential ini ltration occurs regularly, the water content of the Nitrate-N was the only N form found in the groundwater sediment will be higher and closer to saturation. Consequently, under the lagoon (71 ± 19 mg L−1 [n = 12]) (Fig. 5). h e NO −–N removal from the vadose zone will be much more 3 average concentration under the lagoon was 3.5 times higher ef ective, as observed here under the waste channel (Fig. 5c and than the average concentration in the regional groundwater 5d). On the other hand, overl ows and isolated preferential −1 − (~20.2 mg L ; Weinberger, 2007). Moreover, NO3 –N ini ltration events from the banks will not sustain high water concentration in porewater sampled across the entire (~40 m) contents and will lead to very high NO −–N concentrations, as 3 vadose zone under the waste lagoon was similar to concentrations observed under the waste channel margins (Fig. 5c and 5d). measured in the upper groundwater, indicating that leachates Gooddy et al. (1998) summarized the primary physical and from the waste lagoon have reached the groundwater (Fig. 5c). chemical processes beneath an unlined dairy waste lagoon where − On the other hand, NO3 –N concentrations in the deep section 1630 Journal of Environmental Quality (>10 m) of the vadose zone below the channel and its margins were lower than the concentrations in the groundwater under the lagoon (Fig. 5). Under the channel margins, preferential ini ltration during rainstorms transports substantial − amounts of water with low NO3 –N concentration (22 ± 25 mg L−1 [n = 55]; Baram, unpublished data, 2012) into the deep parts of the vadose zone (>10 m) (Baram et al., 2012). Subsurface mixing between the channel leachates and the preferentially ini ltrating rainwater could − lead to such a decrease in NO3 –N concentrations. h is study strengthens the generally accepted idea that the hydraulic conductivity of the organic seal formed at the bottom of waste lagoons controls the ini ltration rate from the lagoon bed. h e results indicate that the formation of an organic seal in the bed of a waste lagoon leads to the development of Fig. 7. Conceptual model (not to scale) of the three zones of a waste lagoon constructed in clay sediment, the governing seepage mechanisms, and the resulting fate of the N species in the unsaturated conditions under the lagoon. subsurface. Zone I represents the vadose zone under the center of the waste lagoon (away from Consequently, the unsaturated sediment the banks); Zone II represents the banks of the waste lagoon; and Zone III represents the lagoon under the lagoon is sui ciently aerated margin area, which is subjected to rare l ooding events by wastewater overl ow and rain events. Small arrows represent continuous slow (mm d−1) ini ltration l ux and long arrows represent to support the microbial oxidation of preferential ini ltration via the desiccation-crack network. Relative saturation (%) is the ratio + − between the measured water content and the water content at saturation. NH4 –N to NO3 –N. An examination of the published literature has shown that this process is likely to occur in completely oxidized in the upper 0.5 m of the sediment below lagoons where a seal has formed, regardless of the sediment in the lagoon, channel, and their banks. Ammonium-N oxidation − which the lagoon was constructed. For example, in the work was coupled with NO3 –N reduction, removing >90% of the of Gooddy et al. (2002) on lagoons constructed in chalk, high leached N, with up to 100% N removal under regions with NH +–N concentrations were observed under lagoons where no higher water contents. Nitrogen removal suggests that neither 4 + + − NH4 –N nor NO3 –N can serve alone as indicators of lagoon seal had formed, while almost complete oxidation of NH4 –N − leakage and that CND, rather than the CEC of the sediment, to NO3 –N was observed within the upper vadose zone under lagoons where a seal had formed. In the works of Ham and regulates the fate of N in the vadose zone under permanently DeSutter (2000), Ham (2002), and DeSutter et al. (2005) on l ooded waste lagoons. lagoons constructed in sediment with clay contents of 18 to 46% + Acknowledgments where a seal had formed, almost complete oxidation of NH4 –N to NO −–N was observed within the upper vadose zone. We thank Michael Kogel and Yuval Shani for their extensive ef orts in 3 the construction and operation of the VMS, the dairy farm owner for allowing us to conduct the research on his farm, and Ms. Sara Elchanani Conclusions for fruitful discussions. h e work was funded by Israel’s Water Authority. h e subsurface under a dairy waste lagoon, waste channel, Comments and suggestions provided by three anonymous reviewers and their margins can be divided into two main zones: (i) that helped to signii cantly improve this manuscript. under the permanently l ooded area and (ii) that under the banks; this division is based on the ini ltration mechanism, the References sediment water content, and their impact on N transformations. APHA (American Public Health Association). 1998. Standard methods for the h e sediment water content under the permanently l ooded examination of water and wastewater. United Book Press, Baltimore, MD. 3 −3 Arah, J.R.M., and K.A. Smith. 1989. Steady-state denitrii cation in area was below saturation (~0.40 m m ) due to formation of aggregated soils—A mathematical model. J. Soil Sci. 40:139–149. an organic seal at the bottom and continuous slow (mm d−1) doi:10.1111/j.1365-2389.1989.tb01262.x ini ltration through it. At the same time, the sediment water Bakken, L., and P. Dörsch. 2006. Nitrous oxide emission and global changes: 3 −3 Modeling approaches. In: H. Bothe et al., editors, Biology of the nitrogen content under the banks was closer to saturation (~0.50 m m ) cycle. Elsevier, Amsterdam. p. 453. −1 due to rapid (m h ) and deep (<8 m) preferential ini ltration Baram, S., D. Kurtzman, and O. Dahan. 2012. Water percolation through a and redistribution of wastewater through the desiccation-crack clayey vadose zone. J. Hydrol. 424–425:165–171. network to the sediment matrix, which occurs regularly during Chertkov, V.Y., and I. Ravina. 1998. Modeling the crack network of swelling clay soils. Soil Sci. Soc. Am. J. 62:1162–1171. doi:10.2136/ lagoon-level l uctuations and rainstorms. h e formation of sssaj1998.03615995006200050002x desiccation cracks enhanced vadose zone aeration and hence Cihan, A., J.S. Tyner, and W.C. Wright. 2006. Seal formation beneath animal + waste holding ponds. Trans. ASAE 49:1539–1544. aerobic processes. Consequently, organic-N and NH4 –N were www.agronomy.org • www.crops.org • www.soils.org 1631 Culley, J.L.B., and P.A. Phillips. 1982. Sealing of soils by liquid cattle manure. Miller, M.H., J.B. Robinson, and R.W. Gillham. 1985. Self-sealing of earthen Can. Agric. Eng. 24:87–89. liquid manure storage ponds. 1. A case-study. J. Environ. Qual. 14:533– Dahan, O. 2005. Flexible probe for measuring moisture content in soil. U.S. 538. doi:10.2134/jeq1985.00472425001400040013x Patent 6956381. Date issued: 18 October. Nelson, D.W., and L.E. Sommers. 1996. Total carbon, organic carbon, and Dahan, O. 2008. Method and system for monitoring soil properties. European organic matter. In: D.L. Sparks, editor, Methods of soil analysis. Part 3. Patent Application Publication EP 1982047 A1. Application EP Chemical methods. SSSA and ASA, Madison, WI. p. 961–1010. 07706061.4; i led 31 Jan. 2007. Date application published: 22 October. Nielsen, T.H., and N.P. Revsbech. 1998. Nitrii cation, denitrii cation, and Dahan, O. 2009a. Method and system for monitoring soil properties. U.S. Patent N-liberation associated with two types of organic hot-spots in soil. Soil Application Publication US 2009/0038390 A1. Application 12/222,069; Biol. Biochem. 30:611–619. doi:10.1016/S0038-0717(97)00211-3 i led 31 July 2008. Date application published: 12 February. Nordstedt, R.A., L.B. Baldwin, and C.C. Hortenstine. 1971. Multistage Dahan, O. 2009b. Method and system for monitoring soil properties. Israel lagoon systems for treatment of dairy farm waste. In: Proceedings of the Patent Application Publication 193126; i led 31 Jan. 2007. Date International Symposium of Livestock Waste Management and Pollution application published: 11 February. Abatement. ASAE, St. Joseph, MI. p. 77–80. Dahan, O., Y. Shani, Y. Enzel, Y. Yechieli, and A. Yakirevich. 2007. Direct Oliver, J.C., and J.L. Meyer. 1974. Subl oor monitoring of Shady Grove dairy measurements of l oodwater ini ltration into shallow alluvial aquifers. J. liquid manure holding pond. Calif. Agric. 28:6–7. Hydrol. 344:157–170. doi:10.1016/j.jhydrol.2007.06.033 Parker, D.B., D.E. Eisenhauer, D.D. Schulte, and D.L. Martin. 1999a. Modeling Dahan, O., R. Talby, Y. Yechieli, E. Adar, and Y. Enzel. 2009. In situ monitoring seepage from an unlined beef cattle feedlot runof storage pond. Trans. of water percolation and solute transport using a vadose zone monitoring ASAE 42:1437–1445. system. Vadose Zone J. 8:916–925. doi:10.2136/vzj2008.0134 Parker, D.B., D.E. Eisenhauer, D.D. Schulte, and J.A. Nienaber. 1999b. Seepage Dahan, O., B. Tatarsky, Y. Enzel, C. Kulls, M. Seely, and G. Benito. 2008. Dynamics characteristics and hydraulic properties of a feedlot runof storage pond. of l ood water ini ltration and ground water recharge in hyperarid desert. Trans. ASAE 42:369–380. Ground Water 46:450–461. doi:10.1111/j.1745-6584.2007.00414.x Parker, D.B., W.J. Rogers, M.C. McCullough, J.E. Cahoon, M.B. Rhoades, and C. DeSutter, T.M., G.M. Pierzynski, and J.M. Ham. 2005. Movement of lagoon- Robinson. 2001. Ini ltration characteristics of cracked clay soils in bottoms of liquor constituents below four animal-waste lagoons. J. Environ. Qual. feedyard playa catchments. Paper presented at: ASAE Annual International 34:1234–1242. doi:10.2134/jeq2004.0296 Meeting, Sacramento, CA. 29 July–1 Aug. 2001. Paper 01-2281. Francis, C.A., J.M. Beman, and M.M.M. Kuypers. 2007. New processes and Parker, D.B., D.D. Schulte, and D.E. Eisenhauer. 1999c. Seepage from earthen players in the nitrogen cycle: h e microbial ecology of anaerobic and animal waste ponds and lagoons—An overview of research results and archaeal ammonia oxidation. ISME J. 1:19–27. doi:10.1038/ismej.2007.8 state regulations. Trans. ASAE 42:485–493. Gooddy, D.C., J.W. Clay, and S.H. Bottrell. 2002. Redox-driven changes in Prosser, J.I. 1989. Autotrophic nitrii cation in bacteria. Adv. Microb. Physiol. porewater chemistry in the unsaturated zone of the chalk aquifer beneath 30:125–181. doi:10.1016/S0065-2911(08)60112-5 unlined cattle slurry lagoons. Appl. Geochem. 17:903–921. doi:10.1016/ Rimon, Y., O. Dahan, R. Nativ, and S. Geyer. 2007. Water percolation through S0883-2927(02)00055-0 the deep vadose zone and groundwater recharge: Preliminary results based Gooddy, D.C., A.G. Hughes, A.T. Williams, A.C. Armstrong, R.J. Nicholson, on a new vadose-zone monitoring system. Water Resour. Res. 43:W05402. and J.R. Williams. 2001. Field and modeling studies to assess the risk to doi:10.1029/2006WR004855 UK groundwater from earth-based stores for livestock manure. Soil Use Rimon, Y., R. Nativ, and O. Dahan. 2011. Physical and chemical evidence for Manage. 17:128–137. doi:10.1111/j.1475-2743.2001.tb00018.x pore-scale dual-domain l ow in the vadose zone. Vadose Zone J. 10:322– Gooddy, D.C., P.J.A. Withers, H.G. McDonald, and P.J. Chilton. 1998. 331. doi:10.2136/vzj2009.0113 Behaviour and impact of cow slurry beneath a storage lagoon: II. Chemical Rowsell, J.G., M.H. Miller, and P.H. Groenevelt. 1985. Self-sealing of earthen composition of chalk porewater at er 18 years. Water Air Soil Pollut. liquid manure storage ponds. 2. Rate and mechanism of sealing. J. Environ. 107:51–72. doi:10.1023/A:1004975615804 Qual. 14:539–543. doi:10.2134/jeq1985.00472425001400040014x Ham, J.M. 2002. Seepage losses from animal waste lagoons: A summary of a four- Schurgers, G., P. Dörsch, L. Bakken, P. Lef elaar, and L.E. Haugen. 2006. Modelling year investigation in Kansas. Trans. ASAE 45:983–992. soil anaerobiosis from water retention characteristics and soil respiration. Ham, J.M., and K.A. Baum. 2009. Measuring seepage from waste lagoons and Soil Biol. Biochem. 38:2637–2644. doi:10.1016/j.soilbio.2006.04.016 earthen basins with an overnight water balance test. Trans. ASAE 52:835–844. SCS. 1997. National engineering handbook: Part 651. Agricultural waste Ham, J.M., and T.M. DeSutter. 2000. Toward site-specii c design standards for management i eld handbook: 651.1080. Appendix 10D: Geotechnical, animal-waste lagoons: Protecting ground water quality. J. Environ. Qual. design, and construction guidelines. USDA–SCS, Washington, DC. 29:1721–1732. doi:10.2134/jeq2000.00472425002900060001x Sewell, J.I. 1978. Dairy lagoon ef ects on groundwater quality. Trans. ASABE Harter, T., H. Davis, M.C. Mathews, and R.D. Meyer. 2002. Shallow groundwater 21:948–952. quality on dairy farms with irrigated forage crops. J. Contam. Hydrol. Sher, Y., S. Baram, O. Dahan, Z. Ronen, and A. Nejidat. 2012. Nitrogen 55:287–315. doi:10.1016/S0169-7722(01)00189-9 transformations and abundance of nitrii ers and denitrii ers in a clay Israel Internal Af airs and Environment Committee. 2010 Health regulations— soil underlying a manure pond. FEMS Microbiol. Ecol. (in press). Sanitary quality of drinking water. (In Hebrew.) http://portal.knesset.gov. doi:10.1111/j.1574-6941.2012.01347.x il/NR/rdonlyres/065AE099-4333-44EA-963B-49AAE9561F5E/646/ Singleton, M.J., B.K. Esser, J.E. Moran, G.B. Hudson, W.W. McNab, and water.pdf (accessed 10 Mar. 2007). T. Harter. 2007. Saturated zone denitrii cation: Potential for natural Issar, A. 1968. Geology of central coastal plain of Israel. Isr. J. Earth Sci. 17:16–29. attenuation of nitrate contamination in shallow groundwater under dairy Kendall, C. 1998. Tracing nitrogen sources and cycling in catchments. In: operations. Environ. Sci. Technol. 41:759–765. doi:10.1021/es061253g C. Kendall and J.J. McDonnell, editors, Isotope tracers in catchment Tyner, J.S., and J. Lee. 2004. Inl uence of seal and liner hydraulic properties on hydrology. Elsevier, Amsterdam. p. 519–576. the seepage rate from animal waste holding ponds and lagoons. Trans. Korom, S.F., and R.W. Jeppson. 1994. Nitrate contamination from dairy lagoons ASAE 47:1739–1745. constructed in coarse alluvial deposits. J. Environ. Qual. 23:973–976. Tyner, J.S., W.C. Wright, and J. Lee. 2006. Lagoon sealing and i lter cakes. Trans. doi:10.2134/jeq1994.00472425002300050018x ASAE 49:527–531. Kremen, A., J. Bear, U. Shavit, and A. Shaviv. 2005. Model demonstrating the Ward, M.H., T.M. deKok, P. Levallois, J. Brender, G. Gulis, B.T. Nolan, and potential for coupled nitrii cation denitrii cation in soil aggregates. J. VanDerslice. 2005. Workgroup report: Drinking-water nitrate and Environ. Sci. Technol. 39:4180–4188. doi:10.1021/es048304z health—Recent i ndings and research needs. Environ. Health Perspect. Kurtzman, D., and B.R. Scanlon. 2011. Groundwater recharge through Vertisols— 113:1607–1614. doi:10.1289/ehp.8043 Irrigated cropland versus natural land, Israel. Vadose Zone J. 10:662–674. Weinberger, G. 2007. h e development and use of water resources in Israel until Maynard, D.G., Y.P. Kalra, and J.A. Crumbaugh. 2008. Nitrate and exchangeable autumn 2006. Ministry of Natl. Infrastructure, Jerusalem. ammonium nitrogen. In: M.R. Carter and E.G. Gregorich, editors, Soil WHO. 2007. Nitrates and nitrites in drinking-water. WHO/SDE/WSH/07.01/16. sampling and methods of analysis. 2nd ed. CRC Press, Taylor and Francis Nitrate and nitrite in drinking-water. Background document for development Group, Boca Raton, FL. p. 71–80. of WHO guidelines for drinking-water quality. World Health Organization, McIlvin, M.R., and M.A. Altabet. 2005. Chemical conversion of nitrate and nitrite Geneva. http://www.who.int/water_sanitation_health/dwq/chemicals/ to nitrous oxide for nitrogen and oxygen isotopic analysis in freshwater and nitratenitrite2ndadd.pdf (accessed 10 Mar. 2007). seawater. Anal. Chem. 77:5589–5595. doi:10.1021/ac050528s Withers, P.J.A., H.G. McDonald, K.A. Smith, and C.G. Chumbley. 1998. Meyer, J.L., E. Olson, and D. Baier. 1972. Manure holding ponds found self- Behaviour and impact of cow slurry beneath a storage lagoon: I. sealing. Calif. Agric. 26:14–15. Groundwater contamination 1975–1982. Water Air Soil Pollut. 107:35– 49. doi:10.1023/A:1004923631734 1632 Journal of Environmental Quality RESEARCH ARTICLE Ammonia transformations and abundance of ammonia oxidizers in a clay soil underlying a manure pond Yonatan Sher, Shahar Baram, Ofer Dahan, Zeev Ronen & Ali Nejidat
Department of Environmental Hydrology & Microbiology, Zuckerberg Institute for Water Research, Blaustein Institutes for Desert Research, Ben-Gurion University of the Negev, Midreshet Sede Boqer, Israel
Correspondence: Ali Nejidat, Department of Abstract Environmental Hydrology & Microbiology, The Jacob Blaustein Institutes for Desert Unlined manure ponds are constructed on clay soil worldwide to manage farm Research, Ben-Gurion University of the waste. Seepage of ammonia-rich liquor into underlying soil layers contributes Negev, Sede Boqer Campus, Midreshet Sede to groundwater contamination by nitrate. To identify the possible processes Boqer 84990, Israel. Tel.: +972 8 6596832; that lead to the production of nitrate from ammonia in this oxygen-limited fax: +972 8 6596831; e-mail: [email protected] environment, we studied the diversity and abundance of ammonia-transform- ing microorganisms under an unlined manure pond. The numbers of ammo- Received 20 October 2011; revised 15 February 2012; accepted 21 February 2012. nia-oxidizing bacteria and anammox bacteria were most abundant in the top Final version published online 27 March of the soil profile and decreased significantly with depth (0.5 m), correlating 2012. with soil pore-water ammonia concentrations and soil ammonia concentra- tions, respectively. On the other hand, the numbers of ammonia-oxidizing DOI: 10.1111/j.1574-6941.2012.01347.x archaea were relatively constant throughout the soil profile (107 amoA copies
per gsoil). Nitrite-oxidizing bacteria were detected mainly in the top 0.2 m. The Editor: Tillmann Lueders results suggest that nitrate accumulation in the vadose zone under the manure pond could be the result of complete aerobic nitrification (ammonia oxidation Keywords to nitrate) and could exist as a byproduct of anammox activity. While the manure ponds; ammonia-oxidizing bacteria; ammonia-oxidizing archaea; anammox majority of the nitrogen was removed within the 0.5-m soil section, possibly bacteria. by combined anammox and heterotrophic denitrification, a fraction of the pro- duced nitrate leached into the groundwater.
lagoons to control pollution through water seepage Introduction (Sweeten, 1998). Agricultural facilities known as concentrated animal feed- The gradual accumulation of organic matter and the ing operations extract large quantities of manure waste development of microbial biofilms at the bottom of (Burkholder et al., 2007), and different management earthen manure ponds and in the underlying soils (Tyner practices have been developed to control and mitigate & Lee, 2004) reduce hydraulic conductivity and inhibit their impact on environmental quality (Day & Funk, infiltration of manure liquor into the groundwater 1998). Manure storage in anaerobic ponds is widely used (Maule´ et al., 2000). Nitrogen in manure ponds, originat- owing to the ponds’ low construction costs (Bernet & ing from cattle feces and urine, appears mainly in the Be´line, 2009). However, the operation of these ponds form of ammonia and organic nitrogen (Safley et al., may involve severe environmental risks as a source of 1986). The manure ponds are highly anoxic, and oxida- contaminants to the air (Amon et al., 2006), as well as tion of ammonia to nitrate can be achieved only by to ground and surface water bodies (Arnon et al., 2008). intensive aeration (McGarvey et al., 2007). The construc- One of the major concerns of animal manure’s environ- tion of manure ponds on clay soil is based on the
MICROBIOLOGY ECOLOGY MICROBIOLOGY mental impact is its contamination of ground and assumption that its low hydraulic conductivity and high surface water bodies with nutrients such as phosphorus cation exchange capacity would limit the downward and nitrogen (Mallin & Cahoon, 2003). Therefore, strict leaching of ammonia and its confinement to the anaero- regulations standardize the construction of manure bic upper soil layers without further transformations
FEMS Microbiol Ecol 81 (2012) 145–155 ª 2012 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved 146 Y. Sher et al.
(DeSutter & Pierzynski, 2005; DeSutter et al., 2005). core samples were immediately covered with aluminum However, nitrate has been detected in the vadose zone foil and kept on ice until they reached the laboratory and the groundwater under dairy manure ponds (Korom (< 12 h). The 0.5- and 0.1-m cylinders were cut into sec- & Jeppson, 1994; Parker et al., 1999). tions every 10 and 2 cm, respectively. Samples were taken Based on the generation of nitrate from ammonia in from each section of the core for further analyses, after these soil layers, it can be hypothesized that ammonia- removing the layers that had been in contact with either oxidizing microorganisms do colonize this oxygen-limited the core barrel walls or the cutting tools. All samples were environment. Nitrate can be generated by two processes: kept at 4 °C until use. Core sections are reported as: (1) (1) complete nitrification (oxidation of ammonia to sediment – representing the sediment in the manure nitrate) and (2) as a byproduct of anammox bacteria pond bed, (2) interface – representing the interface layer activity that consumes ammonia and nitrite (Mulder between the pond sediment and the underlying soil, and et al., 1995). Partial nitrification (oxidation of ammonia (3) soil profile – representing soil samples at different to nitrite) can provide nitrite for the activity of anammox depths below the interface. bacteria (Sliekers et al., 2002). However, to the best of To continuously sample the propagating pore water in our knowledge, the microbial groups that may contribute the unsaturated zone, a custom-made suction cup (6 cm to nitrate generation in the soils underlying anaerobic long and 2 cm diameter) was installed in the soil under- unlined manure ponds have not been studied before. In lying the manure pond, at a vertical depth of 0.5 m. this study, we report upon the diversity and abundance of ammonia-transforming microorganisms [ammonia- Chemical analyses oxidizing bacteria (AOB), ammonia-oxidizing archaea (AOA), and anammox bacteria] in a soil profile below an Water content was determined after 72 h of air-drying at unlined anaerobic manure pond. 105 °C, and total organic matter was determined by the combustion of these air-dried soil samples at 450 °C (Nelson & Sommers, 1996). Total ammonia nitrogen was Materials and methods determined following extraction with 1 M KCl. Nitrate, nitrite, and pore-water ammonia were extracted (~ 910 Study site dilution) with double distilled water (DDW). Ammonia Samples analyzed in this study were collected from a was determined by the phenate method, nitrite by the dairy farm, located in the lowlands (Shfelat Yehuda) of sulfanilamide colorimetric method, and nitrate by cad- southern central Israel, containing 60 dairy cows and 30 mium reduction of nitrate and subsequent analysis as for À heifers and calves. It discharges ~ 7m3 day 1 of liquid the nitrite (Clesceri et al., 1998). Pore-water ammonia waste (manure, feces, water from washing shades, and concentrations were assessed with the assumption of no cooling water) into an unlined earthen manure pond, desorption from the solids during the DDW extractions; with dimensions of ~ 16 9 12 m and depth ~ 0.8 m. hence, the concentrations in the DDW extractions repre- The manure pond is constructed in clayey soil (51% clay, sented solely the dilution of the pore water. Oxidation– 85% of which consists of illite/smectite minerals), and the reduction potentials (ORPs) of the manure pond slurry sediment in its bed consists of organic sludge and coarse and propagating pore water in the unsaturated zone were sand derived from the wear of the dairy farm’s concrete monitored continuously for 4 months using an ORP elec- structures. trode (Cole-Parmer KH27300-19, Vernon Hills, IL).
Sampling of soil and pore water under the Assessment of nitrification potential manure pond Aerobic ammonia oxidation potential was assessed in 0.5- To sample the sediment underlying the pond, a metal L flasks containing a 200 mL medium of 25 mM ring (1 m high and 0.6 m diameter) was placed in the K2HPO4 (pH 7.8) and 2.86 mM (NH4)2SO4, covered with manure pond to allow slurry removal and exposure of air-permeable paper stoppers. Following amendment of the bottom of the pond’s top sediments. Sediment and 5 g soil, flasks were incubated for 24 h in the dark with soil samples were then collected from the bottom of the continuous shaking (200 r.p.m.). Samples were with- pond using two types of core barrels: a 0.5-m long steel drawn and frozen at À80 °C. Nitrogen species were ana- corer with a diameter of 0.15 m to collect a large undis- lyzed as described in the previous section. Nitrification turbed 0.5-m long sediment sample, and a 0.1-m long rates were calculated based on the linear regression sterilized PVC cylinder with diameter of 0.075 m for (R2 = 0.76–0.94, P-value < 0.006) of nitrate accumulation higher sampling resolution of the topsoil. In the field, the vs. time.
ª 2012 Federation of European Microbiological Societies FEMS Microbiol Ecol 81 (2012) 145–155 Published by Blackwell Publishing Ltd. All rights reserved Ammonia transformations under a manure pond 147
ume of 50 µl, containing 5 µlof109 PCR buffer DNA extraction and PCR amplification (Sigma), 250 lM of each deoxynucleoside triphosphate, À1 Genomic DNA was extracted with a PowerSoilTM DNA 2.5 mM MgCl2, 0.1 mg mL BSA, 0.5 µM of each of Isolation kit (MO BIO Laboratory Inc., Solana Beach, the relevant primers, and 2 µl of DNA template. Amplifi- CA). The column of the kit was rinsed twice to ensure cation reactions were carried out in a TGradient thermo- maximal DNA extraction. The abundance of AOA, AOB, cycler (Biometra, Gottingen, Germany). nitrite-oxidizing bacteria (NOB), and anammox was esti- mated via a SYBR green chemistry quantitative PCR DGGE analysis (qPCR) of the following marker genes: putative archaeal amoA gene, bacterial amoA, 16S rRNA gene, and anam- DGGE analysis was performed with a DcodeTM Universal mox 16S rRNA gene, respectively. The primers and PCR Mutation Detection System (Bio-Rad, Hercules, CA) in a conditions are given in Table 1. qPCR contained 12.5 µl 1-mm thick 8% (w/v) polyacrylamide gel at 60 °C. PCR reaction mix (DyNAmoTM Flash SYBR® Green qPCR products of the putative archaeal amoA gene and 16S kit; Finnzymes, Espoo, Finland), 2.5 µl of each of the rel- rRNA gene fragments were analyzed with denaturing gra- evant primers, 5 µl of DNA template or standard, and dients of 15–50% and 35–50% (Nejidat, 2005; Nicol 2.5 µl DDW in a total volume of 25 µl. Melting curves et al., 2008) urea/formamide, for 1160 min at 80 V and (72–95 °C) showed only one peak for all qPCR reactions. 980 min at 70 V for AOA and AOB, respectively. Poly- Calibration curves were created according to a 10-fold acrylamide gels and all DGGE solutions were prepared dilution series (103–109 copies) of plasmids containing according to the manufacturer’s instructions (Bio-Rad). environmental copies of the relevant genes. Calibration Ethidium bromide-stained gels were visualized on a Gel curves had R2 > 0.975, and the slope was between À3.0 Doc XR gel-imaging system (Bio-Rad), and DNA bands and À3.9, corresponding to PCR efficiencies of 90–111%. were excised on a UV transilluminator table using a scal- Amplification reactions were carried out in a Rotor- pel. The DNA was eluted from the gel and used as a tem- GeneTM 6000 (Corbett Life Science, Concorde, NSW, plate for reamplification using the same sets of PCR Australia). primers (Table 1) except CTO primers without the For denaturing gradient gel electrophoresis (DGGE) GC-clamp. analysis, the following genes were amplified: putative archaeal amoA gene and 16S rRNA gene fragments of the Cloning, sequencing, and phylogenetic analysis AOB (Table 1). The latter were amplified by CTO prim- ers using a nested PCR approach, with initial amplifica- Reamplified DGGE bands were cloned in a pTZ57R tion using 27f-1492r primers, as indicated in Table 1 plasmid using an InsTAcloneTM PCR cloning kit (MBI (Mahmood et al., 2006). PCRs were carried out in a vol- Fermentas, Hanover, MD). Cloned DNA was then sent
Table 1. PCR and qPCR primers and reaction conditions applied in this study
Target gene Application Primers Conditions Size (bp) Reference
AOA (amoA) DGGE and Arch amoAF-5′-TTATGGTCTGGCTTAGACG-3′ 95 °C 5 min; 35 cycles: 95 °C35s, 635 Francis et al. qPCR Arch amoAR-5′-GCGGCCATCCATCTGTATG 54 °C45s,72°C50s (2005) T-3′ AOB (amoA) qPCR amoAF-5′-GGGGHTTYTACTGGTGGT-3′ 95 °C 5 min; 35 cycles: 95 °C35s, 491 Rotthauwe amoA2R-5′-CCCCTCKGSAAAGCCTT 54 °C45s,72°C50s et al. (1997) CTTC-3′ AOB DGGE CTO189F-5′-(GC clamp)-GAGRAAAGCA 95 °C 5 min; 30 cycles: 94 °C45s, 460 Kowalchuk (16S rRNA GGGATC G-3′ 58 °C 45 s,72 °C 1 min; 10 min at et al. (1997) gene) CTO649R-5′-CTAGCTTGTAGTTTCAAA 72 °C CGC-3′ Bacteria Nested PCR 27F-5′-AGAGTTTGATCCTGGCTCAG-3′ 95 °C 5 min; 30 cycles: 94 °C45s, 1465 Lane (1991) (16S rRNA 1492R-5′-GGTTACCTTGTTACGACTT-3′ 55 °C45s,72°C 1 min; 10 min at gene) 72 °C NOB qPCR FGPS 872f-5′-CTAAAACTCAAAGGAAT 95 °C 15 min; 35 cycles: 95 °C60s, 397 Degrange & (16S rRNA TGA-3′ 50 °C60s,72°C60s Bardin (1995) gene) FGPS 1269r-5′-TTTTTTGAGATTTGCTAG-3′ Anammox qPCR AMX 368F-5′-TTCGCAATGCCCGAAAGG-3′ 95 °C 15 min; 35 cycles: 95 °C45s, 452 Schmid et al. (16S rRNA AMX 820F-5′-AAAACCCCTCTACTTAGTGC 59 °C50s,72°C60s (2003) gene) CC-3′
FEMS Microbiol Ecol 81 (2012) 145–155 ª 2012 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved 148 Y. Sher et al. for sequencing (Macrogen Inc., Seoul, Korea). To iden- The decrease in pore-water ammonia concentrations tify the anammox bacteria in the soil, corresponding showed the same trend as the KCl-extracted ammonia 16S rRNA gene fragments were PCR-amplified (Table 1) concentrations. However, in the sediment of the manure and a cloned library was constructed. Ten clones were pond bed, the latter (442 mg-N per kgsoil) was signifi- randomly selected and sequenced. Phylogenetic trees, cantly lower than its concentrations in the underlying soil based on 16S rRNA gene sequences of AOB and anam- profile. Nitrate was not detected in the sediment or in the mox bacteria, were constructed using the neighbor- top 0.1 m of the soil profile, whereas detectable nitrate joining method with evolutionary distances computed concentrations were measured at a depth of about 0.3 m, using the maximum composite likelihood method. Phy- reaching up to 6 mg-N per kgsoil at 0.5 m (Fig. 1e). logenetic analyses were conducted with MEGA4 software Nitrite concentrations throughout the soil profile were in
(Tamura et al., 2007). Sequences obtained from this the range of zero to 0.2 mg-N per kgsoil. Sampling of the study were deposited in GenBank and assigned accession vadose zone pore water under the manure pond, at a numbers HQ652082–HQ652103, HQ652105–HQ652107, depth of 0.5 m, showed high nitrate concentrations of
HQ407496, and JF313147. 513 mg-N per Lpore water, coinciding with low ammonia concentrations of 0.3 mg-N per Lpore water. Manure pond slurry exhibited a highly reduced ORP of À0.44 V, while Results the underlying pore water, collected from a depth of 0.5 m below the pond bed, showed an ORP of 0.18 V. Chemical parameters along a soil profile under the manure pond Nitrification activity throughout the soil Soil chemical parameters under the manure pond under- profile went notable changes throughout the soil profile (Fig. 1). The gravimetric water content and the organic matter The decrease in ammonia concentrations and the accu- content in the soil decreased with depth from values of mulation of nitrate throughout the soil profile indicated 38% and 5%, respectively, at the interface between the nitrification activity. Nitrification potential was, therefore, pond sediments and the underlying clay, to the values of assessed throughout the soil profile (Fig. 2). The highest À 24% and 3%, respectively, at a depth of 0.5 m (Fig. 1a nitrification activity (30 mg-N (NO3 ) per kgsoil per h) and b). Ammonia concentrations followed the same trend was measured at a depth of 0.1 m in the soil layer and À (Fig. 1c and d). Soil ammonia concentration (KCl decreased steadily with depth, reaching 0.7 mg-N (NO3 ) extracts) decreased from 3442 mg-N per kgsoil at the per kgsoil per h at 0.5 m. interface to 2 mg-N per kgsoil at a depth of 0.5 m under the manure pond. In the first 10 cm of the profile, Abundance of ammonia-transforming ammonia concentration remained relatively constant and microorganisms then decreased below this depth. Pore-water ammonia concentrations, extracted with DDW (concentration cal- The abundance of the aerobic ammonia oxidizers culated per liter of pore water), also decreased from throughout the soil profile was estimated based on the
1098 mg-N per Lpore water at the interface to 32 mg-N per copy number of the amoA genes of the bacterial and Lpore water at a depth of 0.5 m under the manure pond. archaeal ammonia oxidizers using qPCR (Fig. 3). Copy
Fig. 1. Physiochemical properties of the soil-depth profile under the manure pond: (a) water content, (b) organic matter, (c) soil ammonia, (d) pore-water ammonia, and (e) nitrate. Dashed line indicates the interface between the manure pond sediment and the underlying soil. Water content and organic matter values are the averages of two measurements – one from each core. Soil ammonia, pore-water ammonia, and nitrate values are averages of four measurements – two from each core.
ª 2012 Federation of European Microbiological Societies FEMS Microbiol Ecol 81 (2012) 145–155 Published by Blackwell Publishing Ltd. All rights reserved Ammonia transformations under a manure pond 149
Fig. 2. Nitrification potential, measured by the accumulation of nitrate along a large-scale soil core (0.5 m). Error bars represent standard error of the slope, calculated by linear regression of four to six time points. numbers of the putative archaeal amoA were, in most cases, higher than those of the AOB, with an average 6 copy number of 5 9 10 per gsoil, with no significant change with depth and with no correlation to ammonia concentration (Fig. 1) or nitrification potential (Fig. 2). On the other hand, the highest copy number of the 7 bacterial amoA (1 9 10 copies per gsoil) was recorded Fig. 3. Abundance of ammonia-transforming microorganisms in the top 10 cm, followed by a steep decrease in three throughout the soil profile under the manure pond, represented as orders of magnitude at a depth of 0.5 m (Fig. 3), copy numbers of marker genes per gram dry soil. Abundance of AOB correlated with pore-water ammonia concentration (●), AOA (○), and anammox [AMX (▼)]. Each point represents 2 2 (R = 0.83, P = 0.0001) and nitrification potential (R = average of four qPCR runs – two from each core, and error bars 0.80, P = 0.042). indicate standard deviation. The anammox 16S rRNA gene showed an increase in 4 copy number from 9 9 10 copies per gsoil in the sedi- the soil profile than in those from its lower parts. The 7 ment of the manure pond to 2 9 10 copies per gsoil at major DNA bands revealed by the DGGE analysis the interface, followed by a sharp decrease with depth in (Fig. 4a) were sequenced, and the DNA sequences were the soil profile, to a concentration of 2 9 103 copies per used to construct a phylogenetic tree (Fig. 5). AOB gsoil at 0.5 m (Fig. 3). The abundance pattern of anam- sequences obtained from soil samples at a depth of 0.1– mox 16S rRNA gene copies correlated with soil ammonia 0.5 m were related to both Nitrosomonas and Nitrosospira concentration (R2 = 0.74, P = 0.0006), largely due to the lineages. Within the latter, manure pond soil sequences low soil ammonia concentration and anammox abun- clustered mainly near Nitrosospira briensis and Nitrosospir- dance in the manure pond sediment (Figs 1c and 3). a multiformis sequences within cluster 3 of the Nitroso- Nitrobacter species were detected in 7 out of 18 soil samples spira genus (Purkhold et al., 2000). Nitrosomonas species tested and spanning the 0.5 m soil profile. The positive were distributed in several clusters, including those of samples were mainly in the upper parts of the soil profile Nitrosomonas europea and Nitrosomonas communis (top 0.2 m), and their16S rRNA gene copy numbers were (Fig. 5). DGGE analysis of AOA amoA gene fragments 6 6 in the range of 2 9 10 to 8 9 10 copies per gsoil. throughout the soil profile did not show major changes (Fig. 4b), and a dominant DNA band (4; Fig. 4b) was evident throughout the soil profile. Its sequence Community structure of ammonia oxidizers (JF313147) showed 95% similarity to corresponding The dominant species of AOB along the soil profile were sequences (DQ148902, DQ148904, and DQ148891) of identified by DGGE (Fig. 4a). Higher numbers of DNA AOA retrieved from saline estuaries (Francis et al., 2005). bands were found in the samples from the top 0.1 m of In addition, the sequence (HQ407496) of the DNA bands
FEMS Microbiol Ecol 81 (2012) 145–155 ª 2012 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved 150 Y. Sher et al.
Abundance of ammonia-transforming microorganisms The generation of nitrate from ammonia is an indicator of aerobic nitrification. Aerobic AOB and AOA, which oxidize ammonia to nitrite, were highly abundant at the interface between the manure pond sediment and the underlying soil (Fig. 3). The AOB 16S rRNA gene sequences belonged to both the Nitrospira and Nitroso- monas genera (Fig. 5). All Nitrospira-like 16S rRNA gene sequences were associated with cluster 3 (Fig.5). Nitroso- monas (Koops et al., 2003) and Nitrosospira cluster 3 (Kowalchuk et al., 2000; Webster et al., 2005) species may have been selected for by the high ammonia concen- tration throughout the soil profile (Fig. 1). The number of AOB amoA gene copies (Fig. 3) decreased with soil depth (Fig. 4a), possibly driven by variations in ammonia concentration (Princic et al., 1998; Avrahami et al., 2002) and heterogeneity of the environ- Fig. 4. DGGE profiles of two soil cores sampled under the manure mental conditions within the soil matrix (Brune et al., pond: (a) AOB and (b) AOA communities. Arrows indicate DNA bands 2000). In contrast, the gene copy number of the putative that were sequenced; AOB identification numbers correspond to archaeal amoA and AOA diversity did not change signifi- those in the phylogenetic tree in Fig. 5. cantly with soil depth (Figs 3 and 4b, respectively), in accordance with the suggestion that they can also grow at (6–1; Fig. 4b), which showed a significant reduction in its low levels of ammonia (Erguder et al., 2009; Martens- intensity along the soil profile, has a 99% similarity to Habbena et al., 2009), as found at the lower soil layers. sequences (AB542171 and AB542169) of AOA retrieved In addition, the distribution patterns of the AOB and the from composted cattle manure (Yamamoto et al., 2011). AOA were possibly affected by the availability of oxygen Ten randomly selected anammox 16S rRNA gene throughout the soil profile as AOA were reported to be clones that form a library were sequenced, and a phyloge- able to occupy environments of very low dissolved oxygen netic tree was constructed. The sequences were highly concentrations (Coolen et al., 2007; Erguder et al., 2009). similar and clustered with the sequence of Candidatus The relative contribution of the two ammonia-oxidizing Jettenia asiatica (Fig. 6), suggesting selection for limited groups to the nitrification activity that is measured in diversity in the manure pond soil environment. environmental samples is still under debate with conclu- sions being based mainly on the relative abundance of the Discussion respective amoA gene copy numbers (Prosser & Nicol, 2008). Whereas in some environments, AOB have been Nitrate, in contrast to ammonia that adsorbs to clay par- found to be the dominant ammonia oxidizer, in others, ticles, is easily leached and contaminates groundwater. AOA numbers surpass those of AOB (e.g., Nicol et al., Therefore, this research aimed at identifying the ammo- 2008; De Corte et al., 2009; and Jia & Conrad, 2009). It nia-consuming microbial communities that are involved is difficult to determine the relative contribution of the in the generation of nitrate from ammonia in a reduced two groups in the studied system. The correlation of environment of clay soil that underlies anaerobic manure ammonia concentrations in the soil layers (Fig. 1) and ponds. We have detected and studied the abundance and ammonia oxidation activity in batch experiments (Fig. 2) diversity of AOB, AOA, and anammox bacteria in a soil with AOB abundance (Fig. 3) suggests their significant profile beneath a dairy farm manure pond. The detection role. However, it should be mentioned that the measured of these microbial groups in the reduced vadose zone activity in the batch experiments do not necessarily repre- suggested the subsistence of heterogeneous environmental sent the in situ activity because the used medium can be conditions that allowed aerobic (AOA and AOB) and preferable to one of the ammonia-oxidizing groups (AOB anaerobic (anammox) microbial activity. This can be vs. AOA). Although ammonia concentration in the nitri- attributed to the complexity of the soil environment in fication medium was relatively high (5.6 mM), it was still which aerobic and anaerobic niches can subsist in very lower than its concentration in the pore water in most close proximity (Tiedje et al., 1984; Holden & Fierer, sections of the soil profile (Fig. 1), which further supports 2005).
ª 2012 Federation of European Microbiological Societies FEMS Microbiol Ecol 81 (2012) 145–155 Published by Blackwell Publishing Ltd. All rights reserved Ammonia transformations under a manure pond 151
Fig. 5. AOB phylogenetic tree based on 16S rRNA gene fragments (440 bp) and inferred using the neighbor-joining method. Sequences obtained in this study are indicated in bold. The bootstrap consensus tree inferred from 1000 replicates is taken to represent the evolutionary history of the taxa analyzed. Branches corresponding to partitions reproduced in < 50% bootstrap replicates are collapsed. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates) is shown to the left of the branches. The tree is drawn to scale; evolutionary distances were computed using the maximum composite likelihood method and are in the units of the number of base substitutions per site. the role of AOB. However, the effects of other factors, in part, owing to transport processes and organic matter such as oxygen concentrations (Lam et al., 2007) that contents. Nevertheless, the high copy number of the might differentiate between the in situ and the batch amoA gene of the AOB and AOA, in particular, in the experiments’ measured activities cannot be ruled out. In upper soil samples, indicates an active mixture of ammo- addition, the measured ammonia concentrations can be, nia-oxidizing population.
FEMS Microbiol Ecol 81 (2012) 145–155 ª 2012 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved 152 Y. Sher et al.
Fig. 6. Phylogenetic tree, based on 16S rRNA gene fragments (462 bp) that are unique to anammox bacteria, inferred using the neighbor- joining method. Sequences obtained in this study are indicated in bold. The bootstrap consensus tree inferred from 1000 replicates is taken to represent the evolutionary history of the taxa analyzed. Branches corresponding to partitions reproduced in less than 50% bootstrap replicates are collapsed. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates) is shown to the left of the branches. The tree is drawn to scale; evolutionary distances were computed using the maximum composite likelihood method and are in the units of the number of base substitutions per site.
tions in the vadose zone caused by daily atmospheric Possible routes for oxygen supply pressure fluctuations (Rimon et al., 2011), and (3) wind Although measurements showed that the environment is gusts (Auer et al., 1996; Neeper, 2001). reduced, the generation of nitrate from ammonia indi- cates that considerable amounts of oxygen were reaching Cooperation between nitrogen-transforming the environment and supporting aerobic nitrification. The microorganisms oxygen supply for aerobic microbial activity can be facili- tated by the sediment and soil structure under the man- The lack of nitrite accumulation (only residual levels were ure pond. Deposition of fine organic particles on the bed detected) in the soil samples may stem from either effi- of the manure pond reduces its hydraulic conductivity cient activity of the NOB, nitrite consumption by the (Parker et al., 1999; Cihan et al., 2006). In return, the anammox bacteria and heterotrophic denitrifiers, or as a underlying clay layer becomes unsaturated (Fig. 1a) and result of their combined activity. Detection of autotrophic susceptible to the formation of desiccation cracks (Chert- anammox bacteria (Fig. 3) indicates the existence of kov & Ravina, 1999; Chertkov, 2002). Indeed, desiccation anaerobic niches in the soil layers because of the con- cracks with average aperture of 5.5 ± 1.7 cm and average sumption of the limited amounts of oxygen by aerobic depth of 65 ± 19 cm (with average aperture larger than microbial activity. Carbon dioxide originated from the 6 mm) were observed over the entire land surface at our crack-perfused air or was generated by the heterotrophic site, including the margins of the manure pond (Baram microbial activity that can support the autotrophic et al., 2012). Cracks to the depths of 12 m were also mea- growth of both the ammonia oxidizers and the anammox sured. The cracks network was found to remain opened bacteria. Anammox bacteria have mostly been detected in and serves as a preferential flow pathway year-round aquatic environments and wastewater treatment plants (Baram et al., 2012). It is highly possible that desiccation (Dalsgaard et al., 2005; Kuenen, 2008). However, anam- cracks from the margins create a network of horizontal mox bacteria were also recently detected in terrestrial eco- connectivity under the pond, as has been observed in systems (Humbert et al., 2010; Hu et al., 2011), and the large-scale lysimeter experiments and modeled by numeri- anammox 16S rRNA gene sequences obtained from the cal models (Chertkov, 2002; Greve et al., 2010). It is sug- soil profile were related to C. Jettenia asiatica (Fig. 6), gested that the formation of desiccation cracks around which has been detected in terrestrial ecosystems (Quan and under the manure pond can accelerate the aeration et al., 2008; Humbert et al., 2010). The number of the of the unsaturated vadose zone, via a variety of mecha- anammox bacteria was relatively low in the manure pond nisms: (1) thermal-induced air convection in the cracks sediment, and the highest numbers were measured in the (Nachshon et al., 2008), (2) barometric pressure fluctua- interface between the sediment and the soil (Fig. 3). The
ª 2012 Federation of European Microbiological Societies FEMS Microbiol Ecol 81 (2012) 145–155 Published by Blackwell Publishing Ltd. All rights reserved Ammonia transformations under a manure pond 153 low anammox numbers in the sediment may be related Israel’s Water Authority and by a grant from the Israel to their specific ecological requirements of nitrite concen- Science Foundation (734/05). tration and C/N ratio and their preference to colonize microscopic particles, such as clay particles, in the soil profile, rather than sand particles (Woebken et al., 2007; Authors’ contribution Dang et al., 2010). On the other hand, the decreasing Yonatan Sher and Shahar Baram contributed equally to number of anammox bacteria throughout soil profile this work. (Fig. 3) can be attributed to the decreasing concentration of ammonia (Fig. 1). Nitrifiers have been found to interact with anammox References bacteria under extremely low oxygen systems, such as the Amon B, Kryvoruchko V, Amon T & Zechmeister-Boltenstern S water column of oxygen minimum zones in marine eco- (2006) Methane, nitrous oxide and ammonia emissions during systems (Lam et al., 2007; Yan et al., 2010). The activity of storage and after application of dairy cattle slurry and influence anammox bacteria consumes ammonia and nitrite while of slurry treatment. Agric Ecosyst Environ 112:153–162. producing mainly gaseous N2 and nitrate as a byproduct Arnon S, Dahan O, Elhanany S, Cohen K, Pankratov I, Gross (Kuenen, 2008). In the studied system, nitrite can be pro- A, Ronen Z, Baram S & Shore LS (2008) Transport of duced as an intermediate of nitrification (ammonia oxida- testosterone and estrogen from dairy-farm waste lagoons to tion to nitrite) and/or denitrification (reduction of nitrate groundwater. Environ Sci Technol 42: 5521–5526. to nitrite), as well. Nitrate was found to be the dominant Auer LH, Rosenberg ND, Birdsell KH & Whitney EM (1996) À1 À nitrogen form (166 ± 100 mg L NO3 -N) in the propa- The effects of barometric pumping on contaminant gating pore water at the depth of 0.5 below the pond bed, transport. J Contam Hydrol 24: 145–166. which is significantly lower than the expected concentra- Avrahami S, Conrad R & Braker G (2002) Effect of soil tion that may result from the oxidation of the ammonia ammonium concentration on N2O release and on the pore water in the upper soil layers (Fig. 1), indicating community structure of ammonia oxidizers and denitrifiers. 68 – nitrogen removal processes. Nitrogen removal from the Appl Environ Microbiol : 5685 5692. Baram S, Kurtzman D & Dahan O (2012) Water percolation vadose zone underlying the manure pond can be the result through a clayey vadose zone. J Hydrol 424–425: 165–167. of anammox activity (consuming ammonia and nitrite Bernet N & Be´line F (2009) Challenges and innovations on resulting from partial nitrification), although the occur- biological treatment of livestock effluents. Bioresour Technol rence of canonical denitrification cannot ruled out. The 100: 5431–5436. relative contribution of each pathway to nitrogen removal Brune A, Frenzel P & Cypionka H (2000) Life at the oxic- in the studied system is not yet clear. However, the resid- anoxic interface: microbial activities and adaptations. FEMS ual nitrate can leach down into the deeper soil layers (Paul Microbiol Rev 24: 691–710. & Zebarth, 1997) and can potentially contaminate the Burkholder J, Libra B, Weyer P, Heathcote S, Kolpin D, groundwater. Thorne P & Wichman M (2007) Impacts of waste from concentrated animal feeding operations on water quality. Environ Health Perspect 115: 308–312. Conclusions Chertkov VY (2002) The horizontal hydraulic conductivity of This study showed that aerobic and anaerobic ammonia- vertical interaggregate capillary cracks in clay soils. Physical transforming microbial species inhabit the soil layers that Methods in Agriculture: Approach to Precision and Quality ´ – underlie an anaerobic pond receiving ammonia-rich cow (Blahovec J & Kutılek M, eds), pp. 39 50. Kluwer Academic/Plenum Publishers, New York. manure. Therefore, leaked ammonia from manure ponds Chertkov VY & Ravina I (1999) Morphology of horizontal can be readily transformed, and nitrate is produced. cracks in swelling soils. Theor Appl Fract Mech 31:19–29. While most of the nitrogen is removed within a 0.5-m Cihan A, Tyner JS & Wright WC (2006) Seal formation soil layer, significant levels of nitrate were detected in the beneath animal waste holding ponds. Trans ASAE 49: 1539– vadose zone. Therefore, the results indicate that manure 1544. ponds that are underlined with clay soil should be con- Clesceri LS, Greenberg AE & Eaton AD (1998) Standard sidered as a potential point source for the contamination Methods for the Examination of Water and Wastewater. of groundwater with nitrogenous compounds. American Public Health Association, Washington, DC. Dalsgaard T, Thamdrup B & Canfield DE (2005) Anaerobic ammonium oxidation (anammox) in the marine Acknowledgements environment. Res Microbiol 156: 457–464. We thank the dairy farm owner for allowing us to con- Coolen MJL, Abbas B, van Bleijswijk J, Hopmans EC, Kuypers duct the research at his farm. The work was funded by MMM, wakeham SG & Sinninghe Damste JS (2007)
FEMS Microbiol Ecol 81 (2012) 145–155 ª 2012 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved 154 Y. Sher et al.
Putative ammonia-oxidizing crenarchaeota in suboxic waters Korom SF & Jeppson RW (1994) Nitrate contamination from of the Black see: a basin-wide ecological study using 16S dairy lagoons constructed in coarse alluvial deposits. J ribosomal and functional genes and membrane lipids. Environ Qual 23: 973–976. Environ Microbiol 9: 1001–1016. Kowalchuk GA, Stephen JR, DeBoer W, Prosser J, Embley T & Dang H, Chen R, Wang L, Guo L, Chen P, Tang Z, Tian F, Li Woldendorp JW (1997) Analysis of ammonia-oxidizing S & Klotz MG (2010) Environmental factors shape sediment bacteria of the beta subdivision of the class Proteobacteria anammox bacterial communities in hypernutrified Jiaozhou in coastal sand dunes by denaturing gradient gel Bay, China. Appl Environ Microbiol 76: 7036–7047. electrophoresis and sequencing of PCR-amplified 16S Day D & Funk T (1998) Processing manure: physical, chemical ribosomal DNA fragments. Appl Environ Microbiol 63: and biological treatment. Animal Waste Utilization: Effective 1489–1497. Use of Manure as a Soil Resource (Hatfield JL & Stewart BA, Kowalchuk GA, Stienstra AW, Heilig GHJ, Stephen JR & eds), pp. 243–282. CRC Press, Boca Raton, FL. Woldendorp JW (2000) Changes in community structure of De Corte D, Yokokawa T, Varela M, Agogue H & Herndl G ammonia-oxidizing bacteria during secondly succession of (2009) Spatial distribution of Bacteria and Archaea and calcareous grassland. Environ Microbiol 2:99–110. amoA gene copy numbers throughout the water column of Kuenen JG (2008) Anammox bacteria: from discovery to the Eastern Mediterranean Sea. ISME J 3: 147–158. application. Nat Rev Microbiol 6: 320–326. Degrange V & Bardin R (1995) Detection and counting of Lam P, Jensen MM, Lavik G, McGinnis DF, Muller B, Schubert Nitrobacter populations in soil by PCR. Appl Environ M, Amann R, Thamdrup B & Kuypers MMM (2007) Linking Microbiol 61: 2093–2098. crenarchaeal and bacterial nitrification to anammox in the DeSutter TM & Pierzynski GM (2005) Evaluation of soils for Black Sea. P Natl Acad Sci USA 104: 7104–7109. use as liner materials: asoil chemistry approach. J Environ Lane DJ (1991) 16S/23S rRNA sequencing. Nucleic Acid Qual 34: 951–962. Techniques in Bacterial Systematics (Stackebrandt E & DeSutter TM, Pierzynski GM & Ham JM (2005) Movement of Goodfellow M, eds), pp. 115–175. John Wiley & Sons Inc., lagoon-liquor constituents below four animal waste lagoons. Chichester, UK. J Environ Qual 34: 1234–1242. Mahmood S, Freitag TE & Prosser JI (2006) Comparison of Erguder TH, Boon N, Wittebolle L, Marzorati M & Verstraete PCR primer-based strategies for characterization of W (2009) Environmental factors shaping the ecological ammonia oxidizer communities in environmental samples. niches of ammonia-oxidizing archaea. FEMS Microbiol Rev FEMS Microbiol Ecol 56: 482–493. 33: 855–869. Mallin MA & Cahoon LB (2003) Industrialized animal Francis CA, Roberts KJ, Beman JM, Santoro AE & Oakley BB production—a major source of nutrient and microbial (2005) Ubiquity and diversity of ammonia-oxidizing archaea pollution to aquatic ecosystems. Popul Environ 24: 369– in water columns and sediments of the ocean. P Natl Acad 385. Sci USA 102: 14683–14688. Martens-Habbena W, Berube PM, Urakawa H, de la Torre JR Greve A, Andersen MS & Acworth RI (2010) Investigations of & Stahl DA (2009) Ammonia oxidation kinetics determine soil cracking and preferential flow in a weighing lysimeter niche separation of nitrifying Archaea and Bacteria. Nature filled with cracking clay soil. J Hydrol 393: 105–113. 461: 976–979. Holden PA & Fierer N (2005) Microbial processes in the Maule´ CP, Fonstad TA, Vanapalli SK & Majumdar G (2000) vadose zone. Vadose Zone J 4:1–21. Hydraulic conductivity reduction due to ponded hog Hu B, Rush D, van der Biezen E, Zheng P, van Mullekom M, manure. Can J Agric Eng 42: 157–163. Schouten S, Damste′ JSS, Smolders AJP, Jetten MSM & McGarvey JA, Miller WG, Zhang RH, Ma YG & Mitloehner F Kartal B (2011) New anaerobic ammonium-Oxidizing (2007) Bacterial population dynamics in dairy waste during community enriched from peat soil. Appl Environ Microbiol aerobic and anaerobic treatment and subsequent storage. 77: 966–971. Appl Environ Microbiol 73: 193–202. Humbert S, Tarnawski S, Fromin N, Mallet M-P, Aragno M & Mulder A, van de Graaf AA, Robertson LA & Kuenen JG Zopfi J (2010) Molecular detection of anammox bacteria in (1995) Anaerobic ammonium oxidation discovered in a terrestrial ecosystems: distribution and diversity. ISME J 4: denitrifying fluidized bed reactor. FEMS Microbiol Ecol 16: 450–454. 177–184. Jia Z & Conrad R (2009) Bacteria rather than Archaea Nachshon U, Weisbrod N & Dragila MI (2008) Quantifying dominate microbial ammonia oxidation in an agricultural air convection through surface-exposed fractures: a soil. Environ Microbiol 11: 1658–1671. laboratory study. Vadose Zone J 7: 948–956. Koops H-P, Purkhold U, Pommerening-Roser A, Neeper DA (2001) A model of oscillatory transport in granular Timmermann G, Wagner M (2003) The lithoautotrophic soils, with application to barometric pumping and earth ammonia-oxidizing bacteria. The Prokaryotes: An Evolving tides. J Contam Hydrol 48: 237–252. Electronic Resource for the Microbiological Community, 3rd Nejidat A (2005) Nitrification and occurrence of salt-tolerant edn, release 3, 13 (Dworkin M et al., eds). Springer, New- nitrifying bacteria in the Negev desert soils. FEMS Microbiol York. http://link.springerny.com/link/service/books/10125/. Ecol 52:21–29.
ª 2012 Federation of European Microbiological Societies FEMS Microbiol Ecol 81 (2012) 145–155 Published by Blackwell Publishing Ltd. All rights reserved Ammonia transformations under a manure pond 155
Nelson DW & Sommers LE (1996) Total carbon, organic Safley LM, Westerman PW & Barker JC (1986) Fresh dairy carbon, and organic matter. Methods of Soil Analysis. Part 3. manure characteristics and barnlot nutrient losses. Agric Chemical Methods (Sparks DL, ed) , pp. 961–1010. Soil Wastes 17: 203–215. Science Society of America & American Society of Schmid M, Walsh K, Webb R et al. (2003) Candidatus Agronomy, Madison, WI. “Scalindua brodae,” sp. nov., Candidatus “Scalindua Nicol G, Leininger S, Schleper C & Prosser J (2008) The wagneri”, sp. nov., two new species of anaerobic ammonium influence of soil pH on the diversity, abundance and oxidizing bacteria. Syst Appl Microbiol 26: 529–538. transcriptional activity of ammonia oxidizing archaea and Sliekers AO, Derwort N, Campos Gomez JL, Strous M, bacteria. Environ Microbiol 10: 2966–2978. Kuenen JG & Jetten MSM (2002) Completely autotrophic Parker DB, Eisenhauer DE, Schulte DD & Nienaber JA (1999) nitrogen removal over nitrite in one single reactor. Water Seepage characteristics and hydraulic properties of a feedlot Res 36: 2475–2482. runoff storage pond. Trans ASAE 42: 369–380. Sweeten JM (1998) Cattle feedlot manure and wastewater Paul J & Zebarth B (1997) Denitrification and nitrate leaching management practices. Animal Waste Utilization: Effective during the fall and winter following dairy cattle slurry Use of Manure as a Soil Resource (Hatfield JL & Stewart BA, application. Can J Soil Sci 77: 231–240. eds), pp. 125–156. CRC Press, Boca Raton, Florida. Princic A, Mahne I, Megusar F, Paul EA & Tiedje JM Tamura K, Dudley J, Nei M & Kumar S (2007) MEGA4: (1998) Effects of pH and oxygen and ammonium molecular evolutionary genetics analysis (MEGA) software concentrations on the community structure of nitrifying version 4.0. Mol Biol Evol 24: 1596–1599. bacteria from wastewater. Appl Environ Microbiol 64: Tiedje JM, Sexstone AJ, Parkin TB, Revsbech NP & Shelton DR 3584–3590. (1984) Anaerobic processes in soil. Plant Soil 76: 197–212. Prosser JI & Nicol GW (2008) Relative contributions of Tyner JS & Lee J (2004) Influence of seal and liner hydraulic archaea and bacteria to aerobic ammonia oxidation in the properties on the seepage rate from animal waste holding environment. Environ Microbiol 10: 2931–2941. ponds and lagoons. Trans ASAE 47: 1739–1745. Purkhold U, Pommerening-Roser A, Juretschko S, Schmid Webster G, Embley TM, Freitag TE, Smith Z & Prosser JI MC, Koops H-P & Wagner M (2000) Phylogeny of all (2005) Links between ammonia oxidizers species recognized species of ammonia oxidizers based on composition, functional diversity and nitrification kinetics comparative 16S rRNA and amoA sequence analysis: in grassland soils. Environ Microbiol 7: 676–684. implications for molecular diversity surveys. Appl Environ Woebken D, Fuchs BA, Kuypers MAA & Amann R (2007) Microbiol 66: 5368–5382. Potential interactions of particle-associated anammox Quan ZX, Rhee SK, Zuo JE, Yang Y, Bae JW, Park JR, Lee ST bacteria with bacterial and archaeal partners in the & Park YH (2008) Diversity of ammonium-oxidizing Namibian upwelling system. Appl Environ Microbiol 73: bacteria in a granular sludge anaerobic ammonium- 4648–4657. oxidizing (anammox) reactor. Environ Microbiol 10: Yamamoto N, Asano R, Yoshii H, Otawa K & Nakai Y (2011) 3130–3139. Archaeal community dynamics and detection of ammonia- Rimon Y, Nativ R & Dahan O (2011) Vadose zone water oxidizing archaea during composting of cattle manure using pressure variation during infiltration events. Vadose Zone J culture-independent DNA analysis. Appl Microbiol 10: 1105–1112. Biotechnol 90: 1501–1510. Rotthauwe J, Witzel K & Liesack W (1997) The ammonia Yan J, Op en Camp HJM, Jetten MSM, Hu YY & Haaijer monooxygenase structural gene amoA as a functional SCM (2010) Induced cooperation between marine marker: molecular fine-scale analysis of natural ammonia- nitrifiers and anaerobic ammonium-oxidizing bacteria by oxidizing populations. Appl Environ Microbiol 63: incremental exposure to oxygen. Sys Appl Microbiol 33: 4704–4712. 407–415.
FEMS Microbiol Ecol 81 (2012) 145–155 ª 2012 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved Geoscientiic Geoscientiic
Geoscientiic Geoscientiic ss
Hydrol. Earth Syst. Sci., 17, 1533–1545, 2013 Hydrology and OpenAccess www.hydrol-earth-syst-sci.net/17/1533/2013/ doi:10.5194/hess-17-1533-2013 Earth System © Author(s) 2013. CC Attribution 3.0 License. Sciences
Desiccation-crack-induced salinization in deep clay sediment
S. Baram1, Z. Ronen1, D. Kurtzman2, C. Kulls¨ 3, and O. Dahan1 1Department of Environmental Hydrology & Microbiology, Zuckerberg Institute for Water Research, Albert Katz International School for Desert Studies, Blaustein Institutes for Desert Research, Ben-Gurion University of the Negev, Sede Boqer Campus, 84990, Israel 2Institute of Soil, Water and Environmental Sciences, Agricultural Research Organization, The Volcani Center, Bet Dagan 50250, Israel 3Institute for Hydrology, Albert-Ludwigs-University Freiburg, Fahnenbergplatz, 79098 Freiburg, Germany
Correspondence to: S. Baram ([email protected])
Received: 22 October 2012 – Published in Hydrol. Earth Syst. Sci. Discuss.: 21 November 2012 Revised: 7 March 2013 – Accepted: 28 March 2013 – Published: 22 April 2013
Abstract. A study on water infiltration and solute transport 1 Introduction in a clayey vadose zone underlying a dairy farm waste source was conducted to assess the impact of desiccation cracks on Expansive clay soils, which are dominated by 2 : 1 clay min- subsurface evaporation and salinization. The study is based erals, crack during desiccation when substantial matric suc- on five years of continuous measurements of the temporal tion in the pore structure of the fine-grained clays develops variation in the vadose zone water content and on the chem- and the tensile strength exceeds the cohesive strength of the ical and isotopic composition of the sediment and pore wa- soil matrix (Mermut et al., 1996; Nahlawi and Kodikara, ter in it. The isotopic composition of water stable isotopes 2006; Costa et al., 2013). Desiccation cracks can serve as (δ18O and δ2H) in water and sediment samples, from the area water conduits and preferentially transport water and solutes where desiccation crack networks prevail, indicated subsur- into deep sections of the vadose zone during the generation face evaporation down to ∼ 3.5 m below land surface, and of local runoff or occasional flooding (Bronswijk et al., 1995; vertical and lateral preferential transport of water, following Kurtzman and Scanlon, 2011; Baram et al., 2012a). Several erratic preferential infiltration events. Chloride (Cl−) con- field-scale and large-scale lysimeter experiments have shown centrations in the vadose zone pore water substantially in- that, even though desiccation cracks may disappear on the creased with depth, evidence of deep subsurface evaporation land surface under wet conditions, the cracks will not com- and down flushing of concentrated solutions from the evapo- pletely disappear from the subsurface and may still serve as ration zones during preferential infiltration events. These ob- preferential flow paths (Mermut et al., 1996; Gerke, 2006; servations led to development of a desiccation-crack-induced Acworth and Timms, 2009; Greve et al., 2010; Kishne et al., salinization (DCIS) conceptual model. DCIS suggests that 2010; Baram et al., 2012a). Mermut et al. (1996) stated that thermally driven convective air flow in the desiccation cracks while cultivation and wetting of a Vertisol land surface may induces evaporation and salinization in relatively deep sec- result in the removal of observed desiccation cracks from the tions of the subsurface. This conceptual model supports pre- plow zone, the cracks below this zone will continue to exist. vious conceptual models on vadose zone and groundwater Acworth and Timms (2009) showed that groundwater at 16 salinization in fractured rock in arid environments and ex- and 34 m below a Vertisol field was recharged by irrigation tends its validity to clayey soils in semi-arid environments. water during a single irrigation season. Baram et al. (2012a) showed rapid (within hours) increases in sediment water con- tent in deep sections (> 12 m) of a clayey vadose zone, fol- lowing intensive rain events and short duration flooding on the land surface.
Published by Copernicus Publications on behalf of the European Geosciences Union. 1534 S. Baram et al.: Desiccation-crack-induced salinization in deep clay sediment
Evaporation of water from sediment profiles has been rel- (Weisbrod et al., 2000; Weisbrod and Dragila, 2006). In the atively well studied over the years. Salt buildup and iso- fracture void, the air replacement rate by convection and its topic enrichment of water stable isotopes (deuterium (2H), consequent water vapor flux potential out of the fracture void and oxygen-18 (18O)) in the sediments near the land surface were found to be orders of magnitude faster than the wa- due to evaporation have been observed in many studies, in- ter replacement rate by water flow in the sediment matrix cluding column experiments and field observations of unsat- surrounding the fracture void. Due to the fast air exchange urated soils (Allison et al., 1983, 1985; Barnes and Allison, rate, resulting from nighttime thermal convection, air humid- 1988; Nativ et al., 1995, 1997; Scanlon et al., 1997; Scanlon ity within the fracture void is likely to be similar to the hu- and Goldsmith 1997; Gazis and Feng, 2004, and citations midity in the atmospheric air during the night (Nachshon within; Kurtzman and Scanlon, 2011). In arid and semi-arid et al., 2008; Weisbrod et al., 2009). During the day, wind- environments, the annual evapotranspiration potentials are induced convection can also result in enhanced fracture void- higher than the annual precipitation levels (Scanlon et al., atmosphere air circulation (Nachshon et al., 2012). The re- 1997). In such environments, the high evapotranspiration po- moval of water vapor (as humidity) from the fracture void tentials minimize the deep percolation and lead to the accu- dries the sediment surrounding crack walls which leads to the mulation of salts in the upper unsaturated zone (Allison et al., development of a capillary gradient and advective pore-water 1985; Sharma and Hughes, 1985; Scanlon, 1991). Accord- flow from the wetter zones of the sediment matrix toward the ingly, chloride (Cl−) concentrations have been used to quan- drier fracture surfaces. This study aims to assess the hypoth- tify moisture fluxes in the field (Sharma and Hughes, 1985; esis that a similar phenomenon can occur in clay soils due to Scanlon, 1991; Scanlon and Goldsmith, 1997). Kurtzman the formation of desiccation cracks. and Scanlon (2011) compared Cl− concentration profiles The objective of this study is to asses the impact of desic- taken from the clayey vadose zone beneath both cultivated cation cracks in clay sediment in a semi-arid environment on and uncultivated Vertisol. They observed a significant accu- subsurface evaporation and salinization processes. The study mulation of Cl− in the vadose zone under the uncultivated was conducted through a detailed investigation of the interre- plot. Using flow simulations, they concluded that the ma- lations between (a) water infiltration mechanisms, (b) distri- trix infiltration rates under the uncultivated plots were much bution of solute concentration across the vadose zone profile, lower than the ones under the cultivated plots and that the and (c) stable isotope composition (δ18O, δ2H) of water and uncultivated plots enabled rapid freshwater recharge through sediment samples across a clayey vadose zone in the vicin- discrete preferential paths. Nativ et al. (1995, 1997) studied ity of a wastewater source. The study was conducted in the groundwater recharge under fractured chalk in an arid area. framework of a research project on the impact of dairy farms’ They proposed, based on chemical and isotopic data, that a waste storage ponds on groundwater recharge. small portion of the rainwater percolates downward through the matrix, while a larger percentage of the percolating wa- ter moves through preferential pathways in fractures. They 2 Materials and methods found that the water flowing through the fractures was typ- ically not exposed to evaporation, and that it penetrated the 2.1 Study area matrix across the fracture walls, depleting the stable isotopic composition and diluting the salt concentrations across the The study area is located in the Beer Tuvia region (40 km2), unsaturated zone. above the southern part of the Coastal Aquifer in Israel. While fractures in the vadose zone are usually consid- This phreatic aquifer is composed mainly of sand and cal- ered as preferential flow paths that serve mainly as water careous sandstone of the Pleistocene age, and is overlaid by conduits for deep and fast water percolation (Dahan et al., clay that originated from swamps or riverbeds. The thick- 1999, 2000; Zhou et al., 2006; Ireson et al., 2009), recent ness of the clay layer can vary from several centimeters studies have shown that they can serve as air conduits that to several meters (up to 20 m) within a distance of sev- enhance aeration and water evaporation in deep sections of eral hundred meters (Issar, 1968). The climate is Mediter- the vadose zone (Nachshon et al., 2008; Kamai et al., 2009; ranean with a hot (24.3 ◦C on average) and dry summer and Weisbrod et al., 2009). It has been shown that, in arid en- a cooler (14.2 ◦C on average) rainy winter (November to vironments, the density gradients between the air within a March). The mean annual precipitation is ∼ 450 mm, which fracture void (light, wet and warm air) and the atmospheric falls mostly through 5–8 rainy episodes. The average yearly air (cooler, drier and heavier (denser)) trigger air convection evaporation potential is 1725 mm (using a class-A pan; data within the void, which enhances water evaporation and salt available at http://www.ims.gov.il/IMSEng/Meteorologika/ accumulation in the sediment surrounding the fracture walls evaporation+Tub/monthly+data/). Groundwater is mainly (Weisbrod and Dragila, 2006; Kamai et al., 2009). It was fur- recharged by percolation of seasonal rainwater and agricul- ther suggested that preferential infiltration of water during tural return flow from irrigated fields. Over the past 60 yr the rain storms flushes down the solutes that have accumulated area has been intensively cultivated, and since the 1990s, on the fracture walls and induces groundwater salinization it has hosted approximately 12 500 lactating cows in 140
Hydrol. Earth Syst. Sci., 17, 1533–1545, 2013 www.hydrol-earth-syst-sci.net/17/1533/2013/ S. Baram et al.: Desiccation-crack-induced salinization in deep clay sediment 1535 dairy farms. Throughout the past five decades, the aver- age Cl− concentration in the groundwater in the Beer Tu- via region has increased from ∼ 200 mg L−1 in the 1930s to > 600 mg L−1 today (Weinberger, 2007). This research was conducted in a dairy farm representa- tive of the area that has 60 dairy cows and 30 heifers and calves. Discharges of wastewater and liquid manure from the farm (about 7 m3 d−1) are stored on site in a single-stage un- lined earthen storage pond (∼ 200 m2 and 1 m deep) without specific maintenance procedures, such as drainage or solids removal. Overflow from the pond continuously flows into a waste channel (∼ 2 m wide, 1 m deep and 200 m long). Both the waste storage pond and the channel are flooded with liq- uid manure year-round. The dairy farm has operated under very similar conditions from the late 1960s until today. The stratigraphy underlying the waste pond is clay (0–6 m) overlaying sandy loam (6–8 m), and then calcareous sand- stone (> 8 m). The stratigraphy underlying the liquid waste channel is clay (0–12 m) overlaying calcareous sandstone (Fig. 1). The water table at the study site is located at ∼ 47 m below the surface (b.l.s.). The particle-size distribution of the local top clay is 26 % sand, 22 % silt, and 52 % clay, of which 90 % are illite-smectite minerals (Baram et al., 2012a). Fig. 1. Water content profiles under the waste channel and its mar- Major and minor desiccation cracks (aperture > 1 mm and gins and under the waste storage pond and its banks. Results are < 1 mm, respectively) form a desiccation-crack network that presented as averages, and the horizontal bars are the standard de- crosses the entire clay layer and remains open and hydrauli- viation between all the measurements (2007–2011) made by each cally active year-round, even under relatively high water con- FDTR probe. tent (> 0.50 m3 m−3) (the method in which the major desic- cation cracks were surveyed can be found in Baram et al., 2012a) (Fig. 2). The major desiccation cracks create poly- 193126). Detailed descriptions of the VMSs’ setup and the gons with average dimensions of 1.0 × 0.7 m2, separated by monitoring probes used in this study can be found in previous cracks with an average aperture of 0.055 ± 0.017 m and an publications by Baram et al. (2012a, b). Generally, one VMS average depth, in which the aperture of the crack is greater was installed from the pond margins towards its center, and than 0.006 m, of 0.65 ± 0.19 m. two additional VMSs were installed under the waste channel and its margins (Fig. 3 and Table 1). Pore-water samples, vol- 2.2 Monitoring, sampling and analysis umetric water content and temperature measurements were taken down to a depth of 30.5 m below land surface (b.l.s.) The hydrological processes in the unsaturated zone in the en- under the waste storage pond, to a depth of 3.5 m b.l.s. un- tire waste pond area (center, banks and margins) were stud- der the pond margins and to a depth of 21 m b.l.s. under both ied using three independent vadose-zone monitoring systems the waste channel and its margins. The infiltrating wastew- (VMSs), as well as sediment sampling, using various drilling ater was sampled at the interface between the bottom of the techniques (bucket-augers and direct push coring methods). waste storage pond and the underlying sediments using con- The VMS is designed to collect in situ real-time continuous ventional suction lysimeters. measurements of the water contents and the temperatures in The volumetric water content of the sediment in the va- the vadose zone, using flexible time domain reflectometry dose zone was measured every 60 min during the dry sea- probes (FTDRs) and thermocouples, respectively, and to al- son (April–October) and every 15 min during the wet season low frequent sampling of the sediment pore water, using va- (November–March). The volumetric water content in the up- dose zone sampling ports (VSPs) (Dahan et al., 2007, 2008, per clay sediment under the waste storage pond was evalu- 2009; Rimon et al., 2007, 2011). The VSPs are essentially ated once a year using sediment core samples as described in suction lysimeters that have been modified for operation with Baram et al. (2012b). The temperature profile along the des- the VMS in deep sections of the vadose zone. Pore-water iccation crack void and along the sediment profile was mea- sampling from unsaturated sediments is achieved by creating sured every 10 and 60 min, respectively. Rainwater was col- hydraulic continuity between the sediment and the sampling lected in a PVC cylinder (4 cm i.d. × 30 cm) filled with 2 cm cell using a flexible porous interface (Dahan et al., 2009; of liquid paraffin (Floris, Israel), and sampled once a month Patents # US 6,956,381; US 12/222,069; EP 07706061.4; IL to represent the average isotopic (δ18O, δ2H) values. Water www.hydrol-earth-syst-sci.net/17/1533/2013/ Hydrol. Earth Syst. Sci., 17, 1533–1545, 2013 1536 S. Baram et al.: Desiccation-crack-induced salinization in deep clay sediment
A a b Section A-A front view Waste margins Waste storage pond Waste channel channel
V a d o 4 s e V z o M n S e # m VM 1 I o I
n a S i n t 2 d o #
VMS II # r I I # VMS III i I n 3 I g
s y s t e m Vadose zone
Groundwater A * Not to scale 1. FTDR probe 2. Pore-water Sampling port (VSP) 3. Thermocouple 4. Suction lysimeter Desiccation cracks at the study site. Fig. 2. Fig. 3. Schematic side-view illustration of the vadose-zone moni- toring system (VMS) installed under the waste storage pond, waste samples from the waste sources (pond and channel) and from channel and the waste channel margins (a), and front view (A-A) of the vadose zone were collected every 6 to 12 weeks. Water the VMS installed under the channel and its margins (b) (modified samples for major ions were stored in polypropylene bottles from Baram et al., 2012a). and kept on ice until they reached the laboratory (< 12 h), 18 2 where they were filtered (45 µm glass fiber filter) and kept at (δ O-H2O, δ H-H2O) in the rain, wastewater, pore water 4 ◦C until analysis (< 2 weeks). Water samples for isotopic and sediment samples were measured in the laboratory of the analysis were taken between January 2011 and May 2011, Albert Ludwigs University of Freiburg, by the H2O(liquid)– and placed in 2 mL (12×32 mm) screw-top borosilicate glass H2O(vapor) equilibration and the tunable laser-spectrometry vials cupped by closures with silicone septa until analysis methodology (detailed description of the method can be (Fisher Scientific, Catalog No. 03-391-8 and No. 03-391-14, found in Garvelmann et al., 2012). A two-point calibration respectively). for the δ18O and δ2H values in the samples was preformed Sediment sampling for the determination of the isotopic with standards in clay matrix from the same site, in loam and composition of the bulk pore water in the vadose zone was in sand. The overall analytical error, including uncertainties performed using a direct push coring machine (PowerProbe from the sampling, transport and sample preparation during 9700-VTR AMS, American Falls, ID, USA) equipped with equilibration, is 0.5 ‰ for δ18O and 2.5 ‰ for δ2H. a 0.06 m-i.d. dual tube piston sampler (5005.74; AMS) and To evaluate the potential for thermally triggered air con- a single-use 0.035 m-i.d. PVC liner (5006.427; AMS). Sedi- vection within the desiccation cracks, the air temperatures ment core samples were taken from three locations: (a) under within the voids of three neighboring desiccation cracks were an undisturbed plot which is unaffected by the waste sources, measured using thermocouples (copper-constantan thermo- representing the background; (b) 3 m away from the waste couples; PP-T-24 Omega, Stamford, CT, USA), similar to source, representing an area that is seldom subjected to flood- the ones installed on the VMSs. The thermocouples were ing by wastewater overflows; and (c) under the center of the installed in the desiccation crack voids from the land sur- waste channel. The core samples were taken at the end of face to a depth of 0.7 m b.l.s. in 0.1 m intervals. All the mon- the dry summer (October 2010, following the last precipita- itoring sensors (thermocouples and FTDRs) were operated tion event in March 2010) and at the end of the wet winter with Campbell Scientific (Logan, UT, USA) data acquisition (April 2011), in which 0.325 m of rain had precipitated, and and logging instruments, including a CR10X data-logger, eight days after a 0.03 m rain event. At each sampling event, TDR100 and AM16/32 and SDM50 multiplexers. cores were taken to a depth of 8–10 m b.l.s. At the end of Potential evaporation due to thermally driven air convec- the winter, sediment samples were also taken from the upper tion in the desiccation cracks was calculated on the basis of a 0.3 m of the sediment profile using a garden hand shovel. All theory that was previously developed by Weisbrod and Drag- cores were immediately dissected into 0.1 m slices, placed in ila (2006) and Nachshon et al. (2008). In general, the calcu- aluminum-coated bags and sealed to avoid water losses. Due lations simplify the complex natural systems and assume that to compaction during direct push coring, the upper sample evaporation occurs from the desiccation cracks’ walls and (0–1.2 m) was not dissected. the land surface. As such, the theory simplifies the natural Chloride (Cl−) concentrations were determined within system and does not take into account a reduction in veloc- two weeks after sampling using an ion-chromatograph ity due to thermal diffusion, the effect of fracture aperture (Dionex-4500i, USA). Oxygen and hydrogen isotopes variability, the fracture surface texture or limiting factors
Hydrol. Earth Syst. Sci., 17, 1533–1545, 2013 www.hydrol-earth-syst-sci.net/17/1533/2013/ S. Baram et al.: Desiccation-crack-induced salinization in deep clay sediment 1537
Table 1. The monitoring setup (modified from Baram et al., 2012a). 3 Results and discussion
Number of For the convenience of readers and proper interpretation of Monitoring unit depth (m) pore-water the results described here, a brief description of the main VSP a,b FTDRa,c TCa,d Sediment samples findings from previous stages of this study are presented 0 Pond sludge 14 here. Baram et al. (2012a) showed that water percolation through the clay sediments in the vicinity of a waste source Vadose zone underlying the waste storage pond is controlled by two main processes: (a) continuous slow in- 1.5 Clay 22 2.4 Clay 21 filtration from the bottom of the waste source through the 6.5 6 5.9 Clay 9 clay matrix (flux rate of a few millimeters per day), and (b) 10.5 9.9 Calcareous sandstone fast percolation of rain and wastewater through desiccation 15.5 15 14.9 Calcareous sandstone 23 cracks in velocities exceeding several meters per hour, to a 30.5 30 29.9 Calcareous sandstone 19 depth exceeding 10 m. Following these hydrological findings Vadose zone underlying the waste storage pond margins and observations on the bio-chemical characteristics of the 1 0.9 Clay subsurface (Baram et al., 2012b; Sher et al., 2012), the sub- Vadose zone underlying the waste storage pond banks surface around the waste source was divided into three main 2.5 2 1.9 Clay 13 zones: (a) the center of the permanently flooded waste pond 3.5 3 2.9 Clay 14 (away from the banks). This area is subjected mainly to slow Vadose zone underlying the waste channel water percolation from the pond bottom. In this zone, sedi- 2.4 1.7 2.3 Clay 9 ment water content is relatively constant and low, reaching 4.5 3.9 4.4 Clay 9 only 0.40 m3 m−3 (70 % saturation) (Fig. 1 waste pond). 6.5 6.4 Clay 7 ∼ 8.7 8.1 8.6 Clay 7 (b) The waste source banks. This area is located under the 10.7 10.2 10.6 Clay 7 edges of the waste source where the dominant flow mech- 14.0 13.5 13.9 Calcareous sandstone 7 anism is controlled by preferential flow through desiccation 21.3 19.8 21.2 Calcareous sandstone 4 cracks. In this respect, the waste channel is considered as Vadose zone underlying the waste channel margins two neighboring waste source banks. The vadose zone in this 2.3 1.6 2.2 Clay 4 region is subjected to frequent infiltration events of wastew- 4.3 3.7 4.2 Clay 7 ater due to fluctuations in the wastewater level, as well as to 6.5 6.0 6.4 Clay 8 8.4 7.9 8.3 Clay 4 rainwater infiltration during the winter time. The clay sedi- 10.5 10.0 10.4 Clay 4 ment profile underlying the waste source banks was found to 13.8 13.3 13.7 Calcareous sandstone 4 be consistently wetter than all the other regions of the pond, 21.1 20.6 21.0 Calcareous sandstone reaching ∼ 0.55 m3 m−3 (90 % saturation) (Fig. 1 waste pond a Depth measured relative to the land surface at each site; b vadose zone pore-water banks and waste channel). (c) The waste pond margins. This sampling port; c flexible time domain reflectometry probe; d thermocouple. area is located several meters away from the waste source. It is subjected mainly to rainwater, but also to occasional occur- such as the transport of the pore water from the sediment ma- rences of flooding by wastewater. Infiltration in this area is trix to the crack walls. We assumed that the atmospheric air predominantly controlled by preferential flow via desiccation becomes water saturated (relative humidity of 100 %) imme- crack networks, and the sediment water content in this area diately as it enters the crack void, and that once convection is relatively low, reaching only ∼ 0.30 m3 m−3 (45 % satura- is initiated, the entry of cool dry fingers of atmospheric air tion) (Fig. 1 waste channel margins). and the venting of the crack will continue until atmospheric warming would reverse the thermal gradient in the crack 3.1 Stable isotope composition (typically right after sunrise; Weisbrod et al., 2009). Detailed descriptions of the equations we used can be found in the 3.1.1 Vadose zone under the background area supporting information. It should be clarified that actual wa- ter vapor loss, due to convection, depends on the water vapor Profiles of water stable isotopes (δ18O and δ2H) in sedi- pressure differences between the atmospheric and the frac- ment samples from the background site (an area that is not ture air. As the thermal gradient increases and, subsequently, influenced by wastewater) during the end of the dry sum- convective fluxes, the limiting factor becomes the ability of mer exhibited uniform values, from near land surface to a the matrix to provide water vapor to the fracture air (Kamai et depth of 8 m b.l.s. (δ18O −2 and δ2H −5 ‰) (Fig. 4a and al., 2009; Weisbrod et al., 2009). Kamai et al. (2009) showed b). The δ values of this profile are significantly enriched that at a temperature difference (delta T ) of 10 ◦C, the water compared to the rainwater values (δ18O −4.6 ± 0.7 and δ2H vapor loss due to thermal convection is at a maximum, and −19.1±4.8 ‰, n = 7, each-representing an average value for increasing delta T will not result in increased water vapor the 0.225 m of rain sampled, out of the 0.325 m of precipita- loss. tion during the 2010–2011 winter). Such enrichment in the www.hydrol-earth-syst-sci.net/17/1533/2013/ Hydrol. Earth Syst. Sci., 17, 1533–1545, 2013 1538 S. Baram et al.: Desiccation-crack-induced salinization in deep clay sediment
δ-values suggests that the infiltrating rainwater is subjected to evaporation from the soil profile (Allison et al., 1983). To- wards the end of the wet winter, the isotopic profile showed a clear trend of shifting from depleted rainwater values, near the land surface, to enriched/evaporated values at a depth of ∼ 3.5 m b.l.s. (δ18O −1.2‰ and δ2H −5.7 ‰) (Fig. 4a and b). This trend of near-surface depletion implies that a signif- icant amount of rainwater percolated to the upper layers of the unsaturated zone (< 1 m) during the wet winter. Below 3.5 m, the δ18O and δ2H values fluctuated between −1.7 and −2.9 ‰ and between −5.1 to −7.2 ‰, respectively. These fluctuations in the δ-values, at the deeper parts of the sedi- ment profile, suggest subsurface mixture between depleted rainwater that percolated deep into the vadose zone via a preferential flow path and the ambient enriched/evaporated water (Nativ et al., 1995, 1997). The assumption that the fluctuations in the δ-values resulted from deep water infil- tration through a preferential flow path is supported by the study of Baram et al. (2012a), which demonstrated that a nat- urally formed desiccation crack network in the clay serves as deep (> 12 m) preferential flow paths for rainwater infil- tration. Fluctuations in the isotopic profile during the win- ter, which then evened out during the summer, could have resulted from both slow water redistribution into the matrix (smoothing effect) of the preferentially propagating rainwa- ter, and deep (> 3 m) subsurface evaporation (enrichment). 18 2 These differences between the summer and winter profiles at Fig. 4. δ O and δ H in rainwater, wastewater, pore water and in undisturbed sediment samples from the vadose zone underlying the the background site differ from the normally observed saw- undisturbed background site ( , ), permanently flooded waste pond tooth pattern of evaporated (enriched) summer values and un- a b and channel (c, d) and waste source margins (e, f). The top sample evaporated (depleted) winter values along the vadose zone represents an average value for the 0–1.2 m depth, section. Values (DePaolo et al., 2004). Unlike in a sawtooth pattern where for the depths of 0–0.3 m were obtained only at the end of the winter, the differences in the δ-values can be used to estimate yearly from sediment samples taken by a garden hand shovel. recharge (Garvelmann et al., 2012), here the isotopic profile from the end of the winter represents mixing between the ambient enriched water (from the end of the summer) and Hughes, 1985; Fontes et al., 1986; DePaolo et al., 2004), the fresh depleted rainwater that preferentially propagates which fits the relative humidity during the nighttime at the through the desiccation cracks into deeper parts of the va- study site (http://www.ims.gov.il/IMSEng/CLIMATE). The dose zone. vadose zone water line also indicates that the evaporation Presenting the relation between δ2H and δ18O (in a δ2H- was not limited by the diffusive movement of water and va- δ18O plane) of all the sediment samples from the background por through a dry soil layer, since in such cases, the slope is site, during the winter and summer, indicates a clear shift normally smaller than 3.5–3.2 (Allison et al., 1983). from the average rainwater values on the local meteoric water line (slope of 7.06) to subsurface values on the vadose zone 3.1.2 Vadose zone under the waste pond area water line (slope of 4.45) (Fig. 5). While the samples from the winter demonstrate a shift with depth from the meteoric The water stable isotope composition in sediment samples water line towards enriched and evaporated values, which are and in pore water collected from the vadose zone under located further below the meteoric water line (winter values the waste source area (waste pond and channel) exhibited depths of 0 to 3 m vs. depths > 3 m, Fig. 5), the entire sum- a trend of gradual depletion with depth, from the wastewa- mer profile presents enriched and evaporated values com- ter values near the land surface (δ18O 2.5 ± 0.8 ‰ and δ2H pared to both the rain and the meteoric water line (summer 20.1±2.7 ‰, n = 3, each) to depleted values (δ18O ∼ −2 ‰ values, Fig. 5). Such an isotopic shift from the rain values and δ2H ∼ −5 ‰) at depths of 3.5 m and deeper (Fig. 4c and the meteoric water line is indicative of subsurface evap- and d). Unlike isotopic enrichment that may be attributed oration. The slope of the vadose zone water line indicates to evaporation, isotopic depletion of a water source is usu- that the evaporation occurs at relatively high humidity (70 %) ally attributed to mixture with a more depleted water source (Allison et al., 1983; Barnes and Allison, 1983; Sharma and (Nativ et al., 1995, 1997). Accordingly, the observed trend
Hydrol. Earth Syst. Sci., 17, 1533–1545, 2013 www.hydrol-earth-syst-sci.net/17/1533/2013/ S. Baram et al.: Desiccation-crack-induced salinization in deep clay sediment 1539
through the desiccation cracks formed at the banks and mar- gins of the pond, and the enriched wastewater that perco- lates through the matrix. This assumption is supported by Baram et al. (2012a), which demonstrated the dominance of deep preferential rainwater infiltration through desicca- tion crack networks. Nativ et al. (1995, 1997) proposed a similar mechanism to explain subsurface depletion of en- riched (surface-evaporated) matrix water with depleted rain- water that preferentially infiltrated through rock (chalk) frac- tures. In this study, the subsurface mixing process seems to be significant down to ∼ 3 m b.l.s. (Fig. 4c, d); below that depth, the δ-values reach a steady state and resemble the ones at the margins and background (Fig. 4). Other mechanisms, such as subsurface mixing between enriched infiltrating wa- ter and the capillary rise of depleted shallow groundwater or near/on-surface evaporation of a subsurface-depleted wa- ter source, may lead to a similar trend of gradual depletion in δ-values with depth. Yet, these mechanisms are consid- ered irrelevant for this site where the water table is located 18O vs. 2H in rainwater and sediment samples from the Fig. 5. δ δ at 47 m b.l.s. and the land surface is permanently covered by vadose zone under the background site. Local meteoric water line is based on the data presented in the work of Asaf et al. (2004). wastewater. An additional example of preferential infiltration of rain- water into deep sections of the vadose zone can be seen in Fig. 4c and e, where the δ18O values of pore-water sam- of depletion with depth under the waste source represents ples (collected by the VMS) from the margins and from a mixing process between the two water sources that in- the deep profile under the banks (4.3 and 6.3 m b.l.s.) are filtrate into the vadose zone in the area: (a) wastewater, more depleted and closer to the rainwater values. One should which slowly infiltrates from the permanently flooded waste note that the VSPs at the margins yielded pore-water sam- source’s bed, characterized by enriched (heavier) isotopic ples mainly following intense rain events, which led to deep composition (Fig. 4c and d); and (b) rainwater, which infil- rainwater infiltration and increased the water content of the trates down the desiccation cracks from the banks and mar- clay sediment (> 0.35 m3 m−3) (Baram et al., 2012a). This gins of the waste source, characterized by depleted (lighter) relationship between the ability to sample pore water from isotopic composition (see Sect. 3.1.1 background) (Fig. 4a the relatively dry clay sediments and the deep infiltration and b). The δ-values and the trend of the profile from the of rainwater, along with the observation made by Landon margins of the waste source resemble those obtained from et al. (1999) on the higher proportion of more mobile water the background (Fig. 4e and f vs. Fig. 4a and b), implying (15–95 %) in suction lysimeter samples compared with sam- that the dominant infiltration and evaporation processes oc- ples from sediment cores (5–80 %) at the same depth, can curring at the margins are similar to those observed at the explain the consistently lighter composition of the “mobile” background. Nevertheless, closer examination of the pro- pore-water samples compared to the clay matrix water sam- files from the margins reveals higher fluctuations for both ples. These differences, along with the depleted δ18O values enriched and depleted values, suggesting some influence of observed in sediment of the vadose zone (5.5 and 7.2 m un- occasional preferential wastewater infiltrations during pond der the background plot and 5–6 m under the pond margins, overflows (Baram et al., 2012b). An outlier, of highly de- Fig. 4a and e, respectively), demonstrate a substantial deep pleted δ2H values, was observed between 0.1 to 0.4 m under (> 7 m) recharge of rainwater during the wet winter and a re- the pond (Fig. 4d). Such highly depleted values may have re- distribution into the drier matrix (summer values, Fig. 4a, b, sulted from the presence of isotopically different molecules e and f). (such as hydrogen sulfide (H2S), methane (CH4) and hydro- gen gas (H2)), which were detected in the sediment sam- 3.2 Soluble salt accumulation in the vadose zone ples (unpublished lab measurements using gas analyzer), and are known byproducts of the microbial anaerobic respiration Chloride concentration profiles in sediment pore water (col- of proteins found in the manure (Gerardi, 2003). Accord- lected using the VMS) under all sites surrounding the waste ingly, theses values were not taken into account in the ob- source exhibited a dramatic increase with depth (Fig. 6). served trends and ratios. We suggest that the trend of deple- The concentrations ranged from ∼ 1600 mg L−1 in the in- tion with depth (Fig. 4c, d) is an outcome of subsurface mix- filtrating wastewater to a very high concentration reaching ing between preferential infiltrations of depleted rainwater, 11 500 mg L−1 at a depth of 6.4 m under the channel margins. www.hydrol-earth-syst-sci.net/17/1533/2013/ Hydrol. Earth Syst. Sci., 17, 1533–1545, 2013 1540 S. Baram et al.: Desiccation-crack-induced salinization in deep clay sediment
root zone, the chloride concentrations remained unchanged, indicating piston flow of a single water source under steady state conditions (Gardner, 1967). However, in the work of Kurtzman and Scanlon (2011) and in this work, increasing salinization was observed to a depth of > 4 m, while signif- icant root water uptake by the natural annual shallow-rooted vegetation (maximum 1 m below the surface, mainly: Malva sylvestris and Malva nicaeensis) could only account for in- creasing salinization down to 1 m. To overcome this prob- lem, Kurtzman and Scanlon (2011) attributed water loss be- low 1 m to deep subsurface evaporation (simulated with the root-water-uptake sink term in the Richards equation using the Hydrus1D code). However, they did not provide a mech- anism that can explain such deep evaporation or isotopic val- ues to support this assumption. The Cl− profiles (Fig. 6), along with the isotopic composition (Fig. 4) and the slope Fig. 6. Vertical distribution of Cl− concentrations in pore water from the vadose zone beneath the waste channel margins, the waste of the vadose zone water line (Fig. 5), indicate deep subsur- channel and storage pond banks, and the storage pond. Results are face evaporation, especially since root water uptake does not presented as averages, and the horizontal bars are the standard de- affect the isotopic ratio of the infiltrating water (Allison et al., viation between all the water samples collected by each VSP (Ta- 1983; Dawson and Ehleringer, 1991; Thorburn et al., 1993). ble 1). 3.3 Subsurface evaporation and salinization of clay sediments Under the waste pond, a relatively large surface area that is permanently covered by wastewater, the Cl− concentration Temperature measurements in the desiccation crack void to increased in the clay profile to ∼ 4000 mg L−1, which then a depth of 0.6 m at the site indicated daily oscillations of the maintained a relatively constant concentration up to the wa- thermal gradient between the land surface and the subsurface ter table at 47 m b.l.s. (Fig. 6). Further, under the pond banks (Fig. 7). The thermal gradient was created by the daily tem- and under the channel (an area that is considered as two prox- perature fluctuations. The air inside the deeper sections of the imate banks), the Cl− concentration increased down through cracks is warmer than the air at the land surface (1 > 5 ◦C) the clay profile up to ∼ 8000 mg L−1 at a depth of 9 m. The during the nighttime (6 p.m. to 6 a.m. LT), and the gradient most substantial increase in chloride concentration occurred reverses during the daytime (6 a.m. to 6 p.m. LT). Measure- within the upper 7 m of the clayey vadose zone at the channel ments of the temperature along the sediment profile (matrix) margins, where the concentration was 5.5 fold higher than from the land surface down to 6 m b.l.s. (using the thermo- the concentration in the infiltrating wastewater (Fig. 6). In couples on the VMSs; Table 1) showed very small daily os- this study, the increase in Cl− concentration, with respect to cillations and clear seasonal trends (unpublished field obser- the known water source (wastewater), could not be attributed vations). The differences between the temperature of the at- to the dissolution of soluble salts from the vadose zone, due mospheric air and the temperature of the matrix (down to to their low concentrations in the natural undisturbed sedi- 6 m b.l.s.) were most significant during the winter (> 10 ◦C, ment in the area. A similar trend of increasing salinization extending down to 6 m b.l.s.) and were smaller, but still sig- with depth was observed to depths of 1 to 4 m in natural en- nificant, during the summer (July–September) (> 2 ◦C, ex- vironments composed of different sediment types, such as tending down to 1.5–2.5 m b.l.s.). Weisbrod et al. (2000), calcrete plains, sands, and sandy to sandy loam mallee (Eu- Weisbrod and Dragila (2006) and Kamai et al. (2009) demon- calyptus spp.) woodland (Allison et al., 1985; Cook et al., strated evaporation and salt buildup in fractured rocks, due to 1989; Scanlon et al., 2007), silty to gravelly loam shrub land thermally driven convective air flow in fracture voids. More (Scanlon, 1991), and in natural uncultivated cracking clays research is needed to quantitatively link between the ther- (Radford et al., 2009). In all cases, the Cl− buildup along mal gradient in the field and the depth and magnitude of the the profile reflected the enrichment of atmospherically de- thermal convection. Baram et al. (2012a, b) showed that the rived Cl− due to evaporation and root water uptake (transpi- desiccation crack network in the clay sediment at the vicin- ration) by the natural vegetation, which resulted in minimal ity of the waste source is hydraulically active and open year- downward matrix flux (several mm yr−1) and a long passage round. Accordingly, the daily oscillation in the thermal gra- time in the vadose zone. In these studies, the Cl− concen- dient across the desiccation network may well promote sub- trations increased monotonically throughout the root zone, surface evaporation processes deep into the vadose zone. It until reaching a peak at the maximal depth reached by the is possible that the higher variability in the oxygen signa- roots of the natural vegetation. Beneath the bottom of the ture near the end of the winter, compared to the signature at
Hydrol. Earth Syst. Sci., 17, 1533–1545, 2013 www.hydrol-earth-syst-sci.net/17/1533/2013/ S. Baram et al.: Desiccation-crack-induced salinization in deep clay sediment 1541
of deep preferential infiltration of rainwater and wastewater (Baram et al., 2012a), enhance the capillary gradient between the sediment in the polygon and the drier sediment near the walls of the cracks, and enable substantial water loss and salt accumulation due to thermally driven air convection (evap- oration). On the other hand, under the pond banks, there is a continuous supply of saltier water (wastewater) from the wet sediment to the drier sediment near the desiccation crack surfaces (mainly at the margins). This continuous water sup- ply prevents the drying of the sediment and the transition to diffusive evaporation through a dry soil layer, maintains the potential for convective evaporation in the cracks, and allows substantial accumulation of salts in the deep vadose zone. Water infiltration through preferential flow paths created by the desiccation cracks during intensive rain or flooding Fig. 7. Daily fluctuations in the temperature profile inside desicca- tion cracks at the study site. Each line represents a different tem- events transports and redistributes salt that has been accumu- perature sensor and its depth, relative to the land surface. Note the lated in the fractured clay during the evaporative phases deep unstable conditions (cold air above warm air) that develop in the into the subsurface. Dissolved salts may be transported by cracks every night. the percolating water, both vertically and laterally, follow- ing the cracks’ orientation and the governing flow mecha- nism. While preferential flow along the crack walls may dis- the end of the summer (Fig. 4a, c and e), also reflects deeper solve and flush down salts directly into the deep vadose zone, evaporation during the winter, when the thermal gradients are redistribution of the infiltrating water into the clay matrix steeper and deeper. Such processes are more significant at may flush down salts from the matrix surrounding the cracks the pond margins where the desiccation crack networks are to the deeper sections of the vadose zone, via a gravity- fully developed and the wet sediment under the pond or chan- driven, piston-like matrix flow. Solute redistribution in the nel provides a continuous source of moisture for evaporation. vadose zone due to vertical and lateral preferential trans- The continuous exposure of the clay sediment at the subsur- port of fresher water has been previously reported in frac- face to evaporation explains the observed gradual increase in tured rocks (Nativ et al., 1995; Weisbrod et al., 2000) and in Cl− concentration and the enrichment in the isotopic signa- soils (Allison et al., 1985; Cook et al., 1989; DePaolo et al., ture of δ18O and δ2H with depth. Moreover, subsurface evap- 2004). Accordingly, the increases in salinity with the depth oration provides an explanation as to why the slope of the of the sediments underlying the waste pond (from ∼ 1600 vadose zone water line is not indicative for diffusive evapo- to ∼ 4000 mg L−1), the pond banks and waste channel (from ration. ∼1600 to ∼ 8000 mg L−1), and the channel margins (from A comparison between the Cl− concentrations under the ∼ 1600 to ∼ 9000 mg L−1) (Fig. 6) can be attributed to the margins (∼ 10 000 mg L−1; Fig. 6) and the Cl− concentra- lateral transport of saline solution from the evaporative zones tions under natural undisturbed clays in the near region at the margins to the less evaporative zones underlying the (4000–5500 mg L−1, Kurtzman and Scanlon, 2011) showed permanently flooded regions. Evidence for the existence of that the concentrations under the pond banks and margins active desiccation crack networks under the permanently are up to twofold higher. We suggest that these differences flooded pond and channel at this site were previously re- reflect the difference in both the Cl− concentration of the ported by Baram et al. (2012a, b) and Sher et al. (2012). infiltrating water source (i.e., rainwater (∼ 20 mg L−1; Asaf et al., 2004) vs. wastewater (∼ 1600 mg L−1)) and the dif- 3.4 Evaporation and deep drainage fluxes ferences in the water supply to the sediment–air interfaces along the major desiccation crack voids. Under natural con- The annual water loss from the clayey subsurface under ditions, throughout most of the year, the unsaturated clay ma- the channel and at its banks due to evaporation was esti- trix is relatively dry (Kurtzman and Scanlon, 2011; Baram et mated using chloride mass balance equations. Accordingly, al., 2012a), and only a small capillary gradient may exist be- we assumed that the flow and transport processes in the sub- tween the “dry” sediment matrix away from the cracks (i.e., surface are in a steady state: Jrain + Jwaste = Jev + Jdr and in the center of the polygon) and the drier sediment near the JrainCrain + JwasteCwaste = JdrCdr, where Jwaste and Cwaste walls of the cracks. Accordingly, only a limited amount of are the yearly wastewater infiltration flux through the chan- water can flow towards the sediment–air interfaces, along the nel bed (0.875 m yr−1, Baram et al., 2012b) and the Cl− major desiccation crack voids where evaporation occurs. In- concentration in the infiltrating wastewater (1600 mg L−1 creases in the sediment water content of the clay matrix in the (Fig. 6), respectively; Jrain and Crain are the average yearly polygon (away from the cracks), following the redistribution precipitation (0.45 m yr−1; assuming no runoff) and the Cl− www.hydrol-earth-syst-sci.net/17/1533/2013/ Hydrol. Earth Syst. Sci., 17, 1533–1545, 2013 1542 S. Baram et al.: Desiccation-crack-induced salinization in deep clay sediment concentration in the rainwater (20 mg L−1, Asaf et al., 2004), respectively; Jev is the yearly evaporation flux from the clay layer; and Jdr and Cdr are the yearly water flux from the base of the clay layer to the underlying calcareous sand forma- tion. Cl− concentration in the draining water was measured as ∼ 8000 mg L−1; Fig. 6). Based on these assumptions, the yearly evaporation flux from the clay sediment under the channel was found to be 1.1 m yr−1, indicating that 85 % of the infiltrating water is lost yearly from the clayey vadose zone due to subsurface evaporation (root water uptake is con- sidered negligible in the vicinity of the waste source). Simul- taneously, the annual drainage flux from the clay layer to the underlying calcareous sand formation (Jdr) was found to be 0.20 m yr−1. The calculated fluxes were validated using the HYDRUS 1-D version 3 computer program (Simunek et al., 1998). We performed forward modeling of over 40 yr of con- tinuous infiltration of water to a 30 m deep vadose zone com- posed of homogeneous sandstone layer (using the values for sand predicted by the Rosetta lite, v 1.1 embedded in HY- DRUS). A range of constant infiltration fluxes of wastewater were applied at the top boundary (0.05–0.50 m yr−1) while the lower boundary was the water table with zero pressure. In the simulations, the water content distribution in the sandy vadose zone ranged from 0.075 to 0.12 m3 m−3 (high wa- ter contents were in agreement with high infiltration fluxes). Fig. 8. Desiccation-crack-induced-salinization (DCIS) conceptual The simulations indicated that under an infiltration flux of model. 0.2 m yr−1, the water content in the sandy vadose zone pro- file resembled the field observations in the upper part of the sand layer (0.097 m3 m−3, Fig. 1), in agreement with the flux a major fraction (85 %) of the propagating pore water evapo- calculated from the chloride mass balance. rates from the upper 4 m of the vadose zone at the study site. Theoretical calculations of water evaporation potential The depth to which thermally driven evaporation can occur from the vadose zone due to thermally driven convective air in conical-shaped fractures, like desiccation cracks, is still flow in the desiccation crack voids were performed using the unknown and warrants further research. equations described in Nachshon et al. (2008) and Weisbrod Evaporation from a water body may be estimated through and Dragila (2006) (detailed descriptions of the equations are the relationship between the initial isotopic values of the wa- presented in the supporting information). Using the cracks’ ter source and the isotopic values of the same water body dimensions at the site (apertures of 0.003–0.05 m; polygon following evaporation using the Rayleigh distillation equa- size 1.0 × 0.7 m2; depth 2 m), the daily temperature gradi- tion. However, in order to properly estimate the evaporation ents (1–5 ◦C) and the relative humidity (RH) of the air in based on Rayleigh’s distillation equation, no water can be the atmosphere and in the cracks (70 % and 100 %, respec- added to the system throughout the evaporation process. In tively) at the site, indicated high yearly evaporation poten- this site, we know that this assumption is not valid since: (a) tials of 0.644–14 900 m yr−1. These calculated values repre- preferential infiltration of depleted rainwater reaches the sub- sent the maximal evaporation potentials for a crack with two surface and decreases the enriched evaporated values of the parallel walls when the temperature gradient prevails down ambient water and (b) the subsurface values represent a mix- to 2 m below the surface. The calculated values indicate an ture between the ambient enriched water (from the end of the evaporation potential that is up to four orders of magnitude summer) and the preferentially infiltrating water. greater than the yearly infiltration. It is likely that just like with fractures in chalk, the loss of water from the cracked clayey subsurface is limited by (1) water transport through 4 Conclusions the clay sediment matrix to the crack surfaces, and (2) the water vapor concentration of the atmosphere (relative hu- Our observations on increased salinization and isotopic en- midity) during the nighttime (Weisbrod and Dragila, 2006; richment to depths exceeding 4 m in a clay formation where Kamai et al., 2009). Accordingly, the slow matrix flow in the infiltration is predominantly controlled by preferential flow clay sediment and nights with high relative humidity proba- in desiccation cracks shows that thermally driven evaporation bly limit the evaporation from the subsurface, such that only in the deep sections of the vadose zone is a likely mechanism.
Hydrol. Earth Syst. Sci., 17, 1533–1545, 2013 www.hydrol-earth-syst-sci.net/17/1533/2013/ S. Baram et al.: Desiccation-crack-induced salinization in deep clay sediment 1543
Furthermore, this work expands on previous studies that have clayey soils in semi-arid environments. The long-term perva- explored evaporation from cracked soil (Adams and Hanks, siveness of the suggested invading atmospheric plumes and 1964; Adams et al., 1969; Selim and Kirkham, 1970; Ritchie their repeatability from winter to winter may have significant and Adams, 1974) and provides comprehensive subsurface implications for the physical and chemical evolution of frac- observations and a theoretical framework to explain why tured vadose zones and need to be further explored. the presence of desiccation cracks increases the total evap- oration from clay sediment and consequent sediment salin- ization. Combining observations on (1) rapid and deep wa- Supplementary material related to this article is ter infiltration through desiccation cracks in clay sediments available online at: http://www.hydrol-earth-syst-sci.net/ (Baram et al., 2012a), (2) increased sediment salinization 17/1533/2013/hess-17-1533-2013-supplement.pdf. with depth under the waste sources and their margins, (3) isotopic signature (δ2H and δ18O) of the vadose zone pore water that indicates subsurface evaporation, (4) transport of sorptive contaminants to deep sections of the vadose zone (Arnon et al., 2008), and (5) extensive aeration of the va- Acknowledgements. We thank Michael Kogel for his extensive dose zone, which supports nitrification under anaerobic water efforts in the construction and operation of the VMS, the dairy farm owner for allowing us to conduct this research on his farm, sources with a very high organic load (Baram et al., 2012b; Irena Pankratov for her lab work and Ms. Sara Elchanani for fruitful Sher et al., 2012), has led to the development of a conceptual discussions. We express special gratitude to Noam Weisbrod for model, hereafter termed as desiccation-crack-induced salin- his helpful suggestions and constructive comments. The work ization (DCIS) (Fig. 8). was funded by Israel’s Water Authority and by the Israel Science The DCIS conceptual model suggests that thermally Foundation (Grant # 141412). Comments and suggestions provided driven convective air flow in the desiccation cracks induces by M. Dragila, I. Jolly and an anonymous reviewer helped to water evaporation in relatively deep sections of the subsur- significantly improve this manuscript. face (Fig. 8). Under these conditions, capillary flow between the wet sediment (0.40–0.60 m3 m−3) (high hydraulic poten- Edited by: T. Blume tial) under the waste sources and the drier sediment (0.25– 0.40 m3 m−3) (low hydraulic potential) under their margins is likely to be maintained, allowing continuous water va- References por flux from the desiccation crack network to the surface Acworth, R. I. and Timms, W. A.: Evidence for connected (Fig. 8). As a result, high solute concentrations are built water processes through smectite-dominated clays at up in the deep sediment surrounding the crack network. Er- Breeza, New South Wales, Aust. J. Earth Sci., 56, 81–96, ratic preferential infiltration events that follow intensive rains doi:10.1080/08120090802541952, 2009. and occasional flooding by waste of the waste source banks Adams, J. E. and Hanks, S.: Evaporation from soil shrinkage cracks, and margins transport solutes from the saline sediments sur- Soil Sci. Soc. Am. Pro., 28, 281–284, 1964. rounding the cracked zone, both vertically and laterally into Adams, J. E., Ritchie, J. T., Burnett, E., and Fryrear, D. W.: Evapo- the clayey cross section. These flushing events promote a ration from a simulated soil shrinkage crack, Soil Sci. Soc. Am. gradual increase with depth of the solute concentration un- Pro., 33, 609–613, 1969. der the waste source and its margins. Allison, G. B., Barnes, C. J., and Hughes, M. W.: The distribution of deuterium and 18O in dry soils 2. Experimental, J. Hydrol., The DCIS conceptual model was based on a set of sys- 64, 377–397, 1983. tematic field observations presenting a complex relationship Allison, G. B., Stone, W. J., and Hughes, M. W.: Recharge in karst between gas flow, water flow and solute transport in an un- and dune elements of a semi-arid landscape as indicated by nat- saturated clay vadose zone. Assessing the DCIS characteris- ural isotopes and chloride, J. Hydrol., 76, 1–25, 1985. tics requires further work on simulations that would account Arnon, S., Dahan, O., Elhanany, S., Cohen, K., Pankratov, I., for the (1) water flow and solute transport in clay sediment Gross, A., Ronen, Z., Baram, S., and Shore, L. S.: Trans- where the desiccation cracks’ geometry, as well as the hy- port of testosterone and estrogen from dairy-farm waste la- draulic properties of the clay matrix, changes dynamically; goons to groundwater, Environ. Sci. Technol., 42, 5521–5526, (2) the depth and rate to which water losses as vapor flux doi:10.1021/es800784m, 2008. from desiccation cracks due to thermally driven evaporation Asaf, L., Nativ, R., Shain, D., Hassan, M., and Geyer, S.: Controls can occur; and (3) solute accumulation and redistribution in a on the chemical and isotopic compositions of urban stormwater in a semiarid zone, J. Hydrol., 294, 270–293, 2004. flow domain that is controlled by multiple flow mechanisms, Baram, S., Kurtzman, D., and Dahan, O.: Water percolation through such as preferential and matrix flow, as well as multiple flow a clayey vadose zone, J. Hydrol., 424–425, 165–171, 2012a. directions and water sources. The DCIS conceptual model Baram, S., Arnon, S., Ronen, Z., Kurtzman, D., and Dahan, O.: supports previous conceptual models on vadose zone and Infiltration mechanism controls nitrification and denitrification groundwater salinization in fractured rock in an arid environ- processes under dairy waste lagoons, J. Environ. Qual., 5, 1623– ment (Weisbrod and Dragila, 2006) and extends its validity to 1632, doi:10.2134/jeq2012.0015, 2012b. www.hydrol-earth-syst-sci.net/17/1533/2013/ Hydrol. Earth Syst. Sci., 17, 1533–1545, 2013 1544 S. Baram et al.: Desiccation-crack-induced salinization in deep clay sediment
Barnes, C. J. and Allison, G. B.: The distribution of deuterium and Ireson, A. M., Mathias, S. A., Wheater, H. S., Butler, A. P., and 18O in dry soils: 1. Theory, J. Hydrol., 60, 141–156, 1983. Finch, J.: A model for flow in the chalk unsaturated zone incor- Barnes, C. J. and Allison, G. B.: Tracing of water movement in the porating progressive weathering, J. Hydrol., 365, 244–260, 2009. unsaturated zone using stable isotopes of hydrogen and oxygen, Issar, A.: Geology of Central Coastal Plain of Israel, Israel J. Earth J. Hydrol., 100, 143–176, 1988. Sci., 17, 16–29, 1968. Bronswijk, J. J. B., Hamminga, W., and Oostindie, K.: Field-scale Kamai, T., Weisbrod, N., and Dragila, M. I.: Impact of ambient tem- solute transport in a heavy clay soil, Water Resour. Res., 31, 517– perature on evaporation from surface-exposed fractures, Water 526, doi:10.1029/94WR02534, 1995. Resour. Res., 45, W02417, doi:10.1029/2008WR007354, 2009. Cook, P.G., Walker, G.R. and Jolly, I.D.: Spatial variability of Kishne, A. S., Morgan, C. L. S., Ge, Y. F., and Miller, W. L.: An- groundwater recharge in a semiarid region, J. Hydrol., 111, 195- tecedent soil moisture affecting surface cracking of a Vertisol in 212, 1989. field conditions, Geoderma, 157, 109–117, 2010. Costa, S., Kodikara, J., and Shannon, B.: Salient factors controlling Kurtzman, D. and Scanlon, B. R.: Groundwater recharge through desiccation cracking of clay in laboratory experiments, Geotech- Vertisols – irrigated cropland versus natural land, Israel, Vadose nique, 63, 18–29, 2013. Zone J., 10, 662–674, 2011. Dahan, O., Nativ, R., Adar, E. M., Berkowitz, B., and Ronen, Z.: Landon, M. K., Delin, G. N., Komor, S. C., and Regan, C. P.: Com- Field observation of flow in a fracture intersecting unsaturated parison of the stable-isotopic composition of soil water collected chalk, Water Resour. Res., 35, 3315–3326, 1999. from suction lysimeters, wick samplers, and cores in a sandy un- Dahan, O., Nativ, R., Adar, E. M., Berkowitz, B., and Weisbrod, N.: saturated zone, J. Hydrol., 224, 45–54, 1999. On fracture structure and preferential flow in unsaturated chalk, Mermut, A. R., Dasog, G. S., and Dowuona, G. N.: Soil morphol- Ground Water, 38, 444–451, 2000. ogy, in: Vertisols and technologies for their management, edited Dahan, O., Shani, Y., Enzel, Y., Yechieli, Y., and Yakirevich, A.: Di- by: Ahmed, N. and Mermut, A., Amsterdam: Elsevier Science, rect measurements of floodwater infiltration into shallow alluvial 89–110, 1996. aquifers, J. Hydrol., 344, 157–170, 2007. Nachshon, U., Weisbrod, N., and Dragila, M. I.: Quantifying Dahan, O., Tatarsky, B., Enzel, Y., Kulls, C., Seely, M., and Ben- air convection through surface-exposed fractures: A laboratory ito, G.: Dynamics of flood water infiltration and ground water study, Vadose Zone J., 7, 948–956, 2008. recharge in hyperarid desert, Ground Water, 46, 450–461, 2008. Nachshon, U., Dragila, M., and Weisbrod, N.: From atmospheric Dahan, O., Talby, R., Yechieli, Y.,Adar, E., Lazarovitch, N., and En- winds to fracture ventilation: Cause and effect, J. Geophys. Res.- zel, Y.: In situ monitoring of water percolation and solute trans- Biogeosci., 117, G02016, doi:10.1029/2011JG001898, 2012. port using a vadose zone monitoring system, Vadose Zone J., 8, Nahlawi, H. and Kodikara, J.: Laboratory experiments on desicca- 916–925, 2009. tion cracking of thin soil layers, Geotech. Geol. Eng., 24, 1641– Dawson, T. E. and Ehleringer, J. R.: Streamside trees that do not use 1664, 2006. stream water, Nature, 350, 335–337, 1991. Nativ, R., Adar, E., Dahan, O., and Geyh, M.: Water recharge and DePaolo, D. J., Conrad, M. E., Maher, K., and Gee, G. W.: Evap- solute transport through the vadose zone of fractured chalk under oration effects on oxygen and hydrogen isotopes in deep vadose desert conditions, Water Resour. Res., 31, 253–261, 1995. zone pore fluids at Hanford, Washington, Vadose Zone J., 3, 220– Nativ, R., Adar, E., Dahan, O., and Nissim, I.: Water salinization 232, 2004. in arid regions - observations from the Negev desert, Israel, J. Fontes, J. C., Yousfi, M., and Allison, G. B.: Estimation of long- Hydrol., 196, 271–296, 1997. term, diffuse groundwater discharge in the Northern Sahara us- Radford, B. J., Silburn, D. M., and Forster, B. A.: Soil chloride and ing stable isotope profiles in soil-water, J. Hydrol., 86, 315–327, deep drainage responses to land clearing for cropping at seven 1986. sites in central Queensland, northern Australia, J. Hydrol., 379, Gardner, W. R.: Water uptake and salt distribution pattern in saline 20–29, 2009. soils, in: Isotopes and Radiation Techniques in Soil Physics and Rimon, Y., Dahan, O., Nativ, R., and Geyer, S.: Water percolation Irrigation Studies, Istanbul. Int., Energy Agency, Vienna, 335– through the deep vadose zone and groundwater recharg: prelim- 340, 1967. inary results based on a new vadose-zone monitoring system, Garvelmann, J., Kulls,¨ C., and Weiler, M.: A porewater-based sta- Water Resour. Res., 43, W05402, doi:10.1029/2006WR004855, ble isotope approach for the investigation of subsurface hy- 2007. drological processes, Hydrol. Earth Syst. Sci., 16, 631–640, Rimon, Y., Nativ, R., and Dahan, O.: Physical and chemical evi- doi:10.5194/hess-16-631-2012, 2012. dence for pore-scale dual-domain flow in the vadose zone, Va- Gazis, C. and Feng, X.: A stable isotope study of soil water: ev- dose Zone J., 10, 322–331, 2011. idence for mixing and preferential flow paths, Geoderma, 119, Ritchie, J. T. and Adams, J. E.: Field measurement of evaporation 97–111, 2004. from soil shrinkage cracks, Soil Sci. Soc. Am. Pro., 38, 131–134, Gerardi, M. H.: The Microbiology of Anaerobic Digesters, Hobo- 1974. ken, New Jersey, John Wiley & Sons, Inc., 2003. Scanlon, B. R.: Evaluation of moisture flux from chloride data in Gerke, H. H.: Preferential flow descriptions for structured soils, J. desert soils, J. Hydrol., 128, 137–156, 1991. Plant Nutr. Soil Sci., 169, 382–400, 2006. Scanlon, B. R. and Goldsmith, R. S.: Field study of spatial vari- Greve, A., Andersen, M. S., and Acworth, R. I.: Investigations of ability in unsaturated flow beneath and adjacent to playas, Water soil cracking and preferential flow in a weighing lysimeter filled Resour. Res., 33, 2239–2252, 1997. with cracking clay soil, J. Hydrol., 393, 105–113, 2010. Scanlon, B. R., Tyler, S., and Wierenga, P.: Hydrologic issues in arid, unsaturated systems and implications for contaminant trans-
Hydrol. Earth Syst. Sci., 17, 1533–1545, 2013 www.hydrol-earth-syst-sci.net/17/1533/2013/ S. Baram et al.: Desiccation-crack-induced salinization in deep clay sediment 1545
port, Rev. Geophys., 35, 461–490, 1997. Weinberger, G.: The development and use of water resources in Is- Scanlon, B. R., Jolly, I., Sophocleous, M., and Zhang, L.: Global rael until autumn 2006, in: Jerusalem: Ministry of National In- impacts of conversions from natural to agricultural ecosystems frastructure, 405 pp., 2007 (in Hebrew). on water resources: Quantity versus quality, Water Resour. Res., Weisbrod, N. and Dragila, M. I.: Potential impact of convective frac- 43, W03437. doi:10.1029/2006WR005486, 2007. ture venting on salt-crust buildup and ground-water salinization Selim, H. M. and Kirkham, D.: Soil temperature and water content in arid environments, J. Arid. Environ., 65, 386–399, 2006. changes during drying as influenced by cracks: A laboratory ex- Weisbrod, N., Nativ, R., Adar, E. M., and Ronen, D.: Salt accumula- periment, Soil Sci. Soc. Am. Pro., 34, 565–569, 1970. tion and flushing in unsaturated fractures in an arid environment, Sharma, M. L. and Hughes, M. W.: Groundwater recharge estima- Ground Water, 38, 452–461, 2000. tion using chloride, deuterium and oxygen-18 profiles in the deep Weisbrod, N., Dragila, M. I., Nachshon, U., and Pillersdorf, M.: coastal sands of Western Australia, J. Hydrol., 81, 93–109, 1985. Falling through the cracks: The role of fractures in Earth- Sher, Y., Baram, S., Dahan, O., Ronen, Z., and Nejidat, A.: Nitrogen atmosphere gas exchange, Geophys. Res. Lett., 36, L02401, transformations and abundance of nitrifiers and denitrifiers in a doi:10.1029/2008GL036096, 2009. clay soil underlying a manure pond, FEMS Microbiol. Ecol., 81, Zhou, Q. L., Salve, R., Liu, H. H., Wang, J. S. Y., and Hudson, 145–155, 2012. D.: Analysis of a mesoscale infiltration and water seepage test Simunek, J., Sejna, M., and van Genuchten, M. T.: The HYDRUS- in unsaturated fractured rock: Spatial variabilities and discrete 1D software package for simulating the one-dimensional move- fracture patterns, J. Contam. Hydrol., 87, 96–122, 2006. ment of water, heat, and multiple solutes in variably saturated media, in: IGWMC-TPS-70, Golden, CO: International Ground- water Modeling Center: Colorado School of Mines, 1998. Thorburn, P. J., Walker, G. R., and Brunel, J. P.: Extraction of water from Eucalyptus trees for analysis of deuterium and O-18 – Lab- oratory and field techniques, Plant Cell Environ., 16, 269–277, 1993.
www.hydrol-earth-syst-sci.net/17/1533/2013/ Hydrol. Earth Syst. Sci., 17, 1533–1545, 2013 Supporting information – Paper 4
Supporting information to "Desiccation-Crack-Induced
Salinization in Deep Clay Sediment"
Shahar Baram1, Zeev Ronen1, Daniel Kurtzman2, Christoph Külls3 and Ofer
Dahan1
1Department of Environmental Hydrology & Microbiology, Zuckerberg Institute for
Water Research, Blaustein Institutes for Desert Research, Ben-Gurion University of the Negev, Sede Boqer Campus, Israel 84990.
2Institute of Soil, Water and Environmental Sciences, Agricultural Research
Organization, The Volcani Center, Bet Dagan 50250, Israel.
3Institute for Hydrology, Albert-Ludwigs-University Freiburg, Fahnenbergplatz
Freiburg 79098, Germany.
Equations used for estimations of water evaporation
potential from the vadose zone due to thermally driven
convective air flow in the desiccation crack voids
Calculations of evaporation due to thermally driven convection in desiccation cracks were based on theoretical equations that were previously presented by Nachshon et al.
[2008] and Weisbrod and Dragila [2006]. In their work, they proposed that air convection within the void of a vertical open fracture was thought to start when the
Rayleigh number (Ra), which compares buoyant and viscous forces, exceeded the critical value of 40 [Nachshon et al., 2008; Nield, 1982]:
52 Supporting information – Paper 4
∆TαgkL Ra = (1) νκ
where ∆T (°C) is the temperature difference between the crack air at the top and the bottom of the crack over the length scale L (m), g is the gravitational constant (9.80 m s-2), k is the crack permeability (m2) [k = (2b)2/12, where 2b is the crack aperture
[Shemin, 1997]], ν is the kinematic viscosity (1.51 x 10-5 m2 s-1) and κ is the thermal diffusivity (2.0 x 10-5 m2 s-1). Within the range of 0 to 80° C, air density can be assumed to be a linear function of temperature and is expressed by the thermal expansion coefficient [α = 0.00367(1/°C)]. Eq. (1) considers temperature-dependent volume changes of gas to be negligible, such that there is no latent heating due to gas compressibility, which is justifiable for the situation considered [Weisbrod and
Dragila, 2006].
Once convection is initiated, the entry of cool dry fingers of atmospheric air and
the venting of the crack were thought to continue until atmospheric warming reversed
the thermal gradient in the crack. Based on the suggestion of Weisbrod and Dragila
[2006], in which the velocity of the entering cool and dry atmospheric air fingers is
strongly influenced by the crack's aperture, the average air venting speed (uave) was
approximated by Poiseuille flow (assuming laminar conditions persist throughout the process):
gα∆Tk u = (2) ave ν
The rate by which vapor is lost from the crack was computed using Weisbrod and
Dragila's [2006] equation:
QC = V (∆C)N (3) where V is the volume of fracture air (m3), ∆C is the difference in water vapor concentration (kg m-3) between the fracture air and the atmospheric air (∆C =
53 Supporting information – Paper 4
(∆RH)R(T) ρair), ∆RH is the difference in relative humidity between the fracture air and the atmospheric air, R(T) = 0.0145 at 20° C is the saturated mixing ratio of water
-3 vapor in the air, and ρair = 1.2 (kg m ) is the density of moist air, N is the number of
venting cycles per nighttime (uave/L), and L is the depth of the convection cell (m).
References
Nachshon, U., N. Weisbrod, and M. I. Dragila (2008), Quantifying air convection
through surface-exposed fractures: A laboratory study, Vadose Zone J., 7,
948–956, doi : 10.2136/vzj.2007.0165.
Nield, D. A. (1982), Onset of Convection in a Porous Layer Saturated by an Ideal-
Gas, Int. J. Heat Mass Transf., 25, 1605–1606, doi : 10.1016/0017-
9310(82)90039-4.
Shemin, G. (1997), A governing equation for fluid flow in rough fractures, Water
Resour. Res., 33, 53–61, doi : 10.1029/96WR02588.
Weisbrod, N., and M. I. Dragila (2006), Potential impact of convective fracture
venting on salt-crust buildup and ground-water salinization in arid
environments, J. Arid. Environ., 65, 386–399, doi :
10.1016/j.jaridenv.2005.07.01.
54 Discussion and Summary
3. Discussion and Summary
3.1 Key observations and findings
This study evaluated the impact of unlined dairy waste lagoons constructed in expansive (smectite-dominated) clay on groundwater degradation, based on water flow, solute migration and biogeochemical interactions in the unsaturated zone. The first stage of the research focused on characterization of the desiccation-crack networks formed in the clay soil at the study site (see Paper 1 – Fig. 3). Repeated measurements of the same desiccation cracks during two years (2008/2009 and 2009/2010) showed that the cracks were never fully closed, even during the rainy seasons. In most cases, swelling of the sediment was not uniform along a crack line, leaving closed and unclosed sections. Detailed inspections of places where the cracks were fully closed on the land surface showed that the cracks remained open in their deeper subsurface sections (below 20 cm) (see Paper 1 – Fig. 5). Core sampling and tracking of the propagation of the traced solution (brilliant blue and bromide solution) were also used in the attempts to quantify the crack distribution and dimensions in the subsurface. However, these methods were found to be ineffective since (a) the tracers were adsorbed to the sediment and didn’t move with the water, and (b) the subsurface structure was destroyed during the coring, excavation and relaxation of the sediment. Additional factor that prevented quantitative characterization of the crack dimensions in the subsurface was the small aperture (down to micron scale) of the cracks in the deep (<1 m) unsaturated sections (as suggested by Chertkov and Ravina 1998).
The water percolation dynamics through the cracked clayey vadose zone underlying a dairy waste lagoon and its margins was investigated using three independent vadose-zone monitoring systems (VMSs) that enabled in-situ sampling of the sediment pore water and continuous measurements of the temporal variation in vadose zone water-content profiles, from land surface to 30 m below the surface (see Paper 1 – Fig. 1). The field observations indicated temporal variations in measured water contents, which were associated with infiltration of local runoff during significant rain events (see Paper 1 – Fig. 4) and with infiltration of wastewater during fluctuations in the lagoon level (see Paper 2 – Fig. 3). Infiltrating water propagated through the entire unsaturated clay cross section within 2 to 15 h of most of the
55 Discussion and Summary
significant rain/flooding events, and indicated significant deep (13.3 m) and fast (within 36 to 48 h) recharge into the underlying calcareous sandstone (see Paper 1 – Fig. 4a and Fig. 6). The observed infiltration-propagation pattern indicated domination of a preferential flow mechanism, even at high water contents (0.50–0.60 m3 m-3), rather than gradual slow (mm d-1) propagation in the porous clayey matrix. It was further suggested that rapid infiltration of water to deep sections of the vadose zone, via the desiccation-crack networks, jeopardizes groundwater quality, since it bypasses the sediment’s most biogeochemically active parts. These findings, along with field observations of natural hormones (testosterone and estrogen) in the deep vadose zone and in the groundwater underlying the waste lagoon at the site (Arnon et al., 2008), demonstrate the potential for groundwater pollution underlying expansive clays, and the nontrivial relations between clay water content and the dimensions of the desiccation-crack network. The research clearly shows that the perception of using/regarding expansive clay soils as a "hydrological barrier" in arid and semi-arid environments might be wrong and warrants a second thought. Overall the results of this work could not provide quantitative estimates of the ratio between the water that infiltrates through the desiccation cracks (preferential flow) and the water infiltrating through the matrix (matrix flow). Nonetheless, the ratio between the infiltration via preferential flow paths and the infiltration through the matrix can be estimated based on the following mass balance:
AxClperf+ BxClmatrix= (A+B)xClGW_init - where, ClGW_init is the average Cl concentration in the groundwater prior to the initial -1 cultivation (200 mg L ; data available at the Beer Tuvia Arcive), Clperf is the average Cl- concentration in the local runoff (75 mg L-1 ;based on field measurements with the - CWS, see paper 5 section 2.2.2), Clmatrix is the average Cl concentration in the pore water of the vadose zone – 4000 mg L-1 (based on data presented in the work of Kurtzman and Scanlon 2011) and A and B are the recharge fluxes through the preferential flow paths and the matrix flow, respectively. Solving of the above equation, with the presented concentrations, shows that the preferential flow flux is an order of magnitude higher than the matrix flux (i.e. A~30B). This simple mass balance supports the field observations of rapid and deep (>12 m) infiltration, highlighting the dominance of the preferential infiltration mechanism.
56 Discussion and Summary
The effects of the desiccation-crack network on subsurface aeration were studied, based on the fate of nitrogen species in the subsurface. The subsurface under the lagoon and its margins was divided into two main zones: (i) that under the permanently flooded area and (ii) that under the banks; this division was based on the infiltration mechanism and the sediment water content, and their impact on N transformations. The subsurface under the permanently flooded area of the waste lagoon was unsaturated (70% saturation) due to formation of an organic seal at the lagoon's bottom and continuous slow (mm d-1) infiltration through it (Paper 2 – Fig.2 and Fig.7). High resemblance between the water content values at saturation in sediment samples with high (>10 g g-1) and low organic matter content, indicated limited contribution of the organic matter to the saturation value (see Paper 2 Fig. 2). Accordingly, the subsurface under the banks was closer to saturation (90% saturation) due to rapid (m h-1) and deep (>8 m) preferential infiltration and redistribution of wastewater through the desiccation-crack network to the sediment matrix, which occurs regularly during lagoon-level fluctuations and rainstorms (Paper 2 – Fig.7). In + both zones, organic-N and NH4 were completely oxidized in the upper 0.5 m of the + - unsaturated sediment (Paper 2 – Fig.4). NH4 oxidation was coupled with NO3 reduction, and indicated that natural microbial activity removes 90 – 100% of the leaching N mass from the subsurface, with higher removal under regions with higher water contents. By correlating the nitrogen removal to the water content, this work sheds a new light on previous works on lagoon leakage including the work of Raveh et al. (1972), which reported 99% removal of nitrogen under percolating pits (waste ponds).
Molecular data indicated that both archeal and microbial NH3 oxidizing groups + are present in the subsurface under the lagoon. The similarities between the NH4 concentration patterns (decreasing with depth down to 0.5 m BLS; Paper 3 – Fig. 1), the abundance of the AOB 16S rRNA gene copies (Paper 3 – Fig. 3) and the nitrification potentials (Paper 3 – Fig. 2) suggest that AOB play a major role in the N transformations under the lagoon. Nonetheless, an unchanged, high abundance of 7 -1 AOA amoA gene copies (10 gene copies g dry sediment) down throughout the profile + imply that archeal NH4 oxidation might play a significant role in the removal of N + from the subsurface, especially in niches with low oxygen and NH4 availability. The
57 Discussion and Summary
- lack of NO2 accumulation in the subsurface under the lagoon may stem from either - the efficient activity of the nitrite-oxidizing bacteria (NOB), NO2 consumption by the anammox bacteria and heterotrophic denitrifiers, or a result of their combined - 15 - 18 activity. The isotopic composition of NO3 in the subsurface (δ N-NO3 and δ O- - NO3 ) indicated vast denitrification activity (Paper 2 – Fig. 6), with peak denitrification activities between the bottom of the lagoon and depth of 2 m BLS.
+ - The findings of the research suggest that neither NH4 nor NO3 can alone serve as indicators of lagoon leakage and that CND (concurrent microbial process), rather than the cation exchange capacity (CEC) of the sediment, regulates the fate of
N in the vadose zone under permanently flooded unlined waste lagoons. The microbial abundance (AOA, AOB, NOB and AMX) and the complete oxidation of the - infiltrating N-species indicated that NO3 accumulation in the vadose zone probably - resulted from complete aerobic nitrification (NH3 oxidation to NO3 ) and as a byproduct of anammox activity. The abundance of AMX, AOB (16S rRNA gene - copies) and AOA (amoA gene copies) and the NO3 accumulation also indicated that desiccation cracks in the clay sediment enhanced the aeration of the subsurface and thus microbial abundance even at near saturation conditions (90% saturation).
The effects of the desiccation-crack network on subsurface evaporation and - salinization processes were studied, based on Cl concentrations and isotopic values 2 2 18 (δ H-H O and δ O-H2O) in the vadose zone based on sediment and water samples. The isotopic values in the sediment samples were measured in aluminum bags where equilibrium was formed between the isotopic composition of the pore water and the vapor in the bag. Two main processes controlled the isotopic values of water in the clayey vadose zone: (a) infiltration of rainwater (δ18O –4.6±0.7 and δ2H –19.1±4.8 ‰) at the margins of the lagoon that is exposed to evaporation down to ~3 m below the surface, and (b) infiltration of wastewater (δ18O – 2.5±0.8 and δ2H – 20.1±2.7 ‰) from the permanently flooded lagoon that is mixed with rainwater infiltrating from the margins along the vadose zone (Paper 4 – Fig.4 and Fig. 5). Chloride (Cl-) concentrations in the vadose zone pore water substantially increased with depth (up to a 5.5 fold increase), and supported the isotopic signature of deep subsurface evaporation (Paper 4 – Fig. 6). The absence of plant roots and soluble salts in the subsurface (>0.5 m) indicated that these factors were not the main cause for the
58 Discussion and Summary
subsurface increase in the salinization. A desiccation-crack-induced salinization (DCIS) conceptual model was proposed in which salinization of the sediment in the deep section of the vadose zone is induced by subsurface evaporation due to convective air flow in the desiccation crack’s void (Paper 4 – Fig.8).
Sediment profiles from 36 well logs in the region indicated that there isn’t a continuous clay layer in the subsurface, and that the sediment profile under the lagoon is representative for most of the region. Accordingly, all the observations from the monitoring of the vadose zone under the dairy waste lagoon and its surroundings and from the regional groundwater surveys were integrated to assess the regional impact of dairy waste lagoons on groundwater quality. The integrated values accounted for both the preferential infiltration and the slow infiltration through the matrix. Mass balances were based on GIS interpolations and on the assumption that the region acts as a closed basin (due to the Coastal Drainage System) where all the pumped water are used locally for irrigation. Results showed that from the initiation of the dairy - - industry in the 1960s up until 2010, 14.7 and 1.9 kiloton of Cl and NO3 , respectively, reached the groundwater from dairy lagoons leachates (Appendix A – Table 3). This - - mass accounts for 5.6 % and 14.3 % of the Cl and NO3 , respectively, of the mass added to the groundwater in the study site. Close inspection of the relative - - contribution of leachates from the waste lagoons to the Cl and NO3 mass in the groundwater under all the settlements in the region (the supposed source of - contamination), indicated greater contributions of up to 19 % for Cl and up to 42% - - for NO3 . Despite the apparent contribution of dairy farms to groundwater Cl and - 2 NO3 concentrations, no statistically meaningful correlations (R values) were found - - between the Cl and NO3 concentration in the groundwater and the distance to dairy farms.
The mass balances for the vadose zone under the agricultural area were initially based on the average concentrations from sediment extractions. This method - led to an underestimation of 71.9 kiloton Cl and an overestimation of 75.2 kiloton - NO3 (Appendix A. Table 3), and highlighted the inaccuracy in using sediment extraction for mass balances. These estimates were corrected based on calculations of detailed data collected from the vadose zone under the dairy lagoon and based on groundwater concentrations. The corrected masses indicated the agricultural fields in
59 Discussion and Summary
the region may be the main cause for the groundwater salinization (estimated Cl- -1 concentration of ~1300 mg L in the propagating pore water). They also indicated - - substantial flushing of NO3 below the root zone in the fields (estimated NO3 -1 concentration of ~60 mg L in the propagating pore water). The chemical composition of the groundwater in the observation well under the agricultural field monitored by us (Paper 5, RFW2) along with the results from the regional survey, indicated that the ongoing pumping in the region mixes the recharging solutions with the ambient water, and generates a mixed saltier layer in the upper 40 m of the groundwater. The dissolved oxygen levels in the groundwater under the dairy waste lagoon and under the agricultural field were below saturation (2.9 – 6 mg L-1), suggesting that denitrification activity may occur in the groundwater.
I believe based on current literature review, that the findings of this research regarding the nitrogen cycle and salinization under the dairy waste lagoons are representative to the whole Coastal Aquifer area in Israel. None the less, the salinization intensity would differ based on the clay content and the presence of desiccation cracks and subsurface evaporation.
Kanfi et al. (1983) suggested, based on the work of Reinhorn and Avnimelech (1974), that the main source for nitrate to the Israeli Coastal aquifer, came from mineralization and mobilization of organic matter stored in uncultivated soils and swamps, due to initial cultivation and drying up. Based on the work of Kanfi et al. (1983), Ronen et al. (1983) estimated that the nitrate concentration in the groundwater of the Israeli Coastal aquifer would reach a steady state or may even decrease with time, due to pumpage of water with high nitrate content and dilution by natural replenishment. The results of the regional mass balances, along with nitrate being the st main reason for closure of drinking-water wells in Israel in the beginning of the 21 century (Elhanany, 2009) are in disagreement to the work of Ronen et al. (1983). The latter study may have reduced the interest in studying nitrate behavior in this aquifer compared with the intense work performed in the 1970s on this topic (e.g. Avnimelech and Raveh, 1976; Kahanovich and Blank, 1974; Mercado, 1976; Raveh et al., 1972; Ronen, 1972; Saliternik et al., 1972a; 1972b). Nonetheless, new studies on the impact of agricultural activity on groundwater quality in the Israeli Coastal
60 Discussion and Summary
aquifer support and strengthen the findings of this work (Dahan et al., under review; Kurtzman et al., 2013).
3.2 Conclusions The main conclusions of this dissertation are: In arid and semi-arid environments, desiccation cracks exist year-round in unsaturated expansive clay sediments. Desiccation cracks serve as water conduits that cross the entire clay layer, enabling preferential transport of water (recharge) and pollutants into the deep, less biogeochemically active, part of the vadose zone. Significant recharge occurs under expansive clay soils; accordingly, the common perception of using/regarding expansive clay soils as a "hydrological barrier" in arid and semi-arid environments might be wrong and warrants a second thought. Wastewater from unlined waste lagoons constructed in expansive clay soils may preferentially infiltrate via desiccation cracks formed at the banks into the deep vadose zone during lagoon-level fluctuations, jeopardizing groundwater quality. Air flow via desiccation cracks induces the aeration of the vadose zone, even at + near saturation water contents, and enables microbial NH4 oxidation, even under anaerobic water sources with a very high organic load. -1 - AOB, AOA and AMX coexist in saline (>1500 mg L Cl ) environments with high organic and nitrogen loads.
The relative contributions of the bacterial and the archeal NH3 oxidizing groups to
the NH3 oxidation in the studied system could not be determined. Yet, archeal amoA gene abundance suggests that AOA may be more active than AOB at low levels of ammonia and oxygen. Coupled nitrification-denitrification and anammox activities naturally remove from the vadose zone more than 90% of the infiltrating N-species masses. Thermally driven convective air flow in the desiccation crack’s void leads to subsurface evaporation and salinization down to depth of at least 3 m below the land surface. Spatial interpolations of detailed chemical and physical data from the vadose zone underlying a pollution point source can be used to estimate the relative
61 Discussion and Summary
contribution of a point source to the regional groundwater contamination processes.
3.3 Future perspectives As with most scientific investigations, this work led to a number of open questions. There remain several scientifically interesting issues for future research regarding the role of desiccation cracks and dairy lagoon effluents on the environment Development of a computer model that would enable simulations and calibration of the preferential and matrix air and water flow and the associated transport and subsurface evaporation and salinization. Establishment of statistical parameters that would enable the determination of the relations between the water content of the clay sediment, the surface aperture, the subsurface aperture and the level of connectivity between the cracks. Evaluate the maximal depth to which water losses due to thermally driven evaporation inside natural desiccation-crack voids can occur. Examine the effects of air convection cells that form inside major desiccation cracks on the aeration of the minor and horizontal desiccation cracks connected to them. Evaluate the relationships between bacterial and archaeal nitrogen transformations and the relative saturation and aeration of expansive clay sediment. Evaluating the spatial contribution of a common farming practice on the groundwater degradation using geochemical indicators.
62 References
4. Additional References (not included in published papers)
Almasri, M.N. and Kaluarachchi, J.J., 2004a. Assessment and management of long-
term nitrate pollution of ground water in agriculture-dominated watersheds.
Journal of Hydrology, 295: 225-245.
Almasri, M.N. and Kaluarachchi, J.J., 2004b. Implications of on-ground nitrogen
loading and soil transformations on ground water quality management. Journal
of the American Water Resources Association, 40: 165-186.
Avnimelech, Y., Raveh, Y., 1976. Nitrate leakage from soils differing in texture and
nitrogen load. Journal of Environmental Quality, 5: 79-82.
Bernstein, A., Adar, E., Nejidat, A. and Ronen, Z., 2011. Isolation and
characterization of RDX-degrading Rhodococcus species from a contaminated
aquifer. Biodegradation, 22: 997-1005.
Bobier, M.W., Frank, K.D. and Spalding, R.F., 1993. Nitrate-N movement in a fine-
textured vadose zone. Journal of Soil and Water Conservation, 48: 350-354.
Brown, C.D. et al., 1999. Investigation into the effect of tillage on solute movement to
drains through a heavy clay soil. II. Interpretation using a radio-scanning
technique, dye-tracing and modelling. Soil Use and Management, 15: 94-100.
Brown, V. and Associates, 2003. Review of Animal Waste Management Regulations,
Task 2 Report: Evaluate Title 27 Effectiveness to Protect Groundwater
Quality. Central Valley Regional Water Quality Control Board.,
http://www.waterboards.ca.gov/centralvalley/water_issues/dairies/historical_d
airy_program_info/index.shtml (viewed October 2012).
Burkart, M.R. and Stoner, J.D., 2007. Nitrate in aquifers beneath agricultural systems.
Water Science and Technology, 56: 59-69.
63 References
Burkholder, J. et al., 2007. Impacts of waste from concentrated animal feeding
operations on water quality. Environmental Health Perspectives, 115: 308-
312.
Dahan, O., Babad, A., Lazarovitch, N., Eliani, E., Kurtzman, D., under review. Nitrate
leaching from intensive organic farming to groundwater. Environmental
Science & Technology.
DeSutter, T.M. and Pierzynski, G.M., 2005. Evaluation of Soils for Use as Liner
Materials: A Soil Chemistry Approach. Journal of Environmental Quality. 34:
951–962.
Elhanany, S., 2009. Challenges in preserving the quality of water resources in Israel,
The Jerusalem Conference for Environment and Nature, Jerusalem, Israel (in
Hebrew).
Graham, C.B., 2004. Evaporation from Bare Soil and Soil Cracks. Ms.C, Thesis,
Oregon State University, Corvallis,Oregon, United States of America, Dep. of
Geol. and Geophys, 52 pp.
Kahanovich, Y., Blank, D., 1974 Nitrate groundwater pollution in Israel-data
processing. Tahal Report No. 01 74 47, p. 25 (in Hebrew).
Kass, A., Gavrieli, I., Yechieli, Y., Vengosh, A. and Starinsky, A., 2005. The impact
of freshwater and wastewater irrigation on the chemistry of shallow
groundwater: a case study from the Israeli Coastal Aquifer. Journal of
Hydrology, 300: 314-331.
Kelly, B.P. and Pomes, M.L., 1998. Preferential flow and transport of nitrate and
bromide in claypan soil. Ground Water, 36: 484-494.
64 References
Kurtzman, D., 2011. The soil, the vadose zone and groundwater salinization in the
south-eastern part the Coastal Aquifer. A paper of the Agricultural Research
Organization, Israel (in Hebrew).
Melloul, A.J. and Wollman, S.H., 2003. Qualitative hydrological and land-use
planning tool for the Israel Coastal aquifer. The Science of The Total
Environment, 309: 1-17.
Melo, A. et al., 2012. Impact of intensive horticulture practices on groundwater
content of nitrates, sodium, potassium, and pesticides. Environmental
Monitoring and Assessment, 184: 4539-4551.
Mercado, A., 1976. Nitrate and chloride pollution of aquifers: a regional study with
the aid of a single-cell model. Water Resources Research. 731-747.
Mercado, A., 1985. The use of hydrogeochemicl patterns in carbonate sand and
sandstone to identify the flushing of saline water. Ground Water, 23: 635-646.
Mermut, A.R., Dasog, G.S. and Dowuona, G.N., 1996. Soil morphology. In: N.
Ahmed and A. Mermut (Editors), Vertisols and technologies for their
management. Elsevier Science, Amsterdam, pp. 89-110.
Oostindie, K. and Bronswijk, J.J.B., 1995. Consequences of preferential flow in
cracking clay soils for contamination-risk of shallow aquifers. Journal of
Environmental Management, 43: 359-373.
Raveh, Y., Avnimelech, Y., Saliternik, C., 1972. Nitrates concentrations in the soil
layer above the Coastal Plain aquifer. Tahal Report No. HR. 72 073, p. 149 (in
Hebrew)
Reinhorn, T., Avnimelech, Y., 1974. Nitrogen release associated with the decrease in
soil organic matter in newly cultivated soils. Journal of Environmental Quality.
3: 118 - 121.
65 References
Ronen, D., 1972. Nitrate contamination of groundwater in the Rishon Lezion-Rehovot
district. Tahal Report No. HR. 072 046, p. 100 (in Hebrew)
Ronen, D., Kanfi, Y., Magaritz, M., 1983. Sources of Nitrates in Groundwater of the
Coastal-Plain of Israel - Evolution of Ideas. Water Research. 17, 1499-1503.
Saliternik, C., Kahanovich, Y., Raveh, Y., Shevach, Y., Ronen, D., Shwartz, Y.,
Blank, D., 1972a. Groundwater contamination in Israel. Collection and
prosessing of data. Tahal Report No. HR. 72 043 p. 57 (in Hebrew)
Saliternik, C., Kahanovich, Y., Shevach, Y., 1972b. Groundwater nitrate pollution
sources Tahal Report No. HR. 072 042, p. 68 (in Hebrew)
Scanlon, B.R., 1992. Moisture and Solute Flux Along Preferred Pathways
Characterized by Fissured Sediments in Desert Soils. Journal of Contaminant
Hydrology, 10: 19-46.
Shavit, U. and Furman, A., 2001. The location of deep salinity sources in the Israeli
Coastal aquifer. Journal of Hydrology. 250: 63-77.
Sliekers, A.O. et al., 2002. Completely autotrophic nitrogen removal over nitrite in
one single reactor. Water Research. 36: 2475-2482.
Tolmach, Y., 1977. Hydrogeological atlas of Israel - geological sections in the Coastal
Plain area, 1-6 (in Hebrew). Israel Hydrological Service, Jerusalem.
Wilkison, D.H. and Blevins, D.W., 1999. Observations on preferential flow and
horizontal transport of nitrogen fertilizer in the unsaturated zone. Journal of
Environmental Quality. 28: 1568-1580.
Zektser, I.S. and Everett, L.G. (Editors), 2004. Groundwater resources of the world
and their use. IHP-VI, Series on groundwater No. 6. United Nations
Educational, Scientific and Cultural Organization, Paris, 342 pp.
66 Appendix
5. Appendix A
67 Appendix A: Regional impact
Assessing the impact of dairy waste lagoons on groundwater quality using a spatial analysis of vadose zone and groundwater information: example from the Beer-Tuvia region Israel
Baram, Sa., Kurtzman, Db., Ronen, Za., Peeters, Ac., Dahan, Oa.
aDepartment of Environmental Hydrology & Microbiology, Zuckerberg Institute for Water Research, Blaustein Institutes for Desert Research, Ben-Gurion University of the Negev, Sede Boqer Campus 84990, Israel bInstitute of Soil, Water and Environmental Sciences, Agricultural Research Organization, The Volcani Center, Bet Dagan 50250, Israel cUnit of Desert Architecture and Urban Planning, The Swiss Institute for Dryland Environmental and Energy Research, J. Blaustein Institutes for Desert Research, Ben- Gurion University of the Negev, Sede Boqer Campus 84990, Israel
Email addresses: [email protected] (S. Baram), [email protected] (D. Kurtzman), [email protected] (Z. Ronen), [email protected] (A. Peeters), [email protected] (O. Dahan).
Corresponding author: Shahar Baram, [email protected], 972-8-6563513
Abstract Dairy waste lagoons are considered to be a major source of groundwater contamination. The objective of this work is to introduce a methodology to assess past and future impacts of dairy waste lagoons on regional groundwater. The method is - - based on a spatial statistical analysis of chloride (Cl ) and nitrate (NO3 ) concentration distributions in the saturated and unsaturated zones, in order to quantify the relation between the locations of dairy farms and the spatial variability in contaminant
68 Appendix A: Regional impact
concentrations in groundwater. The method was applied to the Beer-Tuvia region, Israel, where intensive dairy farming has been practiced for over 50 years above the local phreatic aquifer. Mass balance calculations accounted for the various groundwater recharge and abstraction sources in the entire region. The mass balances showed that although the groundwater in the area suffers from continuous degradation, leachates from dairy waste lagoons have only contributed to the aquifer - - 5.6 % and 14 % of the total mass of Cl and NO3 , respectively. The chemical composition of the aquifer and vadose zone water suggested that irrigated agricultural - - activity in the region is the main contributor of Cl and NO3 to the groundwater. A - - low spatial correlation between the Cl and NO3 concentrations in the groundwater and the on-land location of the dairy farms strengthened this assumption, despite the dairy waste lagoon being a point source for groundwater contamination by Cl- and - NO3 . Mass balance calculations, for the vadose zone of the entire region, indicated that drying of the lagoons would have a relatively minor effect on the regional groundwater salinization process (5.7 % of the total Cl- load is stored under lagoons). - Nevertheless, a more considerable effect on the groundwater contamination by NO3 - is expected (14.1 % of the total NO3 load is stored under lagoons). Results demonstrate that analyzing vadose zone and groundwater data by spatial statistical analysis methods can significantly contribute to the understanding of the relations between groundwater contaminating sources, and to assessing appropriate remediation steps.
Key words: dairy farm; waste lagoon; groundwater; contamination; GIS; spatial statistics
1. Introduction
One of the most complex and intense agricultural sources of groundwater pollution is concentrated animal feeding operations (CAFOs), such as dairy farms (Burkart and Stoner, 2007; UNESCO, 2006). The direct and unavoidable result of such facilities is that large amounts of solid and liquid waste that contains high concentrations of nitrogen (N), phosphorus (P), salts, organic compounds and other possible contaminants are stored in waste lagoons in the dairy farm area. For economic reasons, in many cases, the liquid waste is stored in earthen (soil-lined or unlined)
69 Appendix A: Regional impact
lagoons rather than in plastic-lined lagoons, steel tanks or concrete tanks (Ham and DeSutter, 2000). The long-term impact of such dairy farm waste management has been frequently shown to affect the quality of the underlying groundwater, due to direct leakage from the dairy farm area and waste lagoons (Baram et al., 2012a; DeSutter et al., 2005; Gooddy et al., 2001; Ham and Baum, 2009; Harter et al., 2002; Parker et al., 1999c). However, estimation and quantification of the direct contribution of lagoon seepage to the groundwater degradation, especially on a regional scale, remain a challenge.
Numerous studies have focused on detecting or characterizing the impact of - dairy lagoons on groundwater salinization and contamination by nitrate (NO3 ). DeSutter et al. (2005), Parker et al. (1995; 1999a; 1999b) and Ham (2002) found, using sediment extractions, substantial loads (i.e., the total mass) of different + - nitrogen-species (N-species; i.e., organic, ammonium (NH4 ) and NO3 ) and evidence for Cl- leaching to the subsurface under cattle-feedlot runoff lagoons and dairy waste lagoons. They observed spatial variability in the concentrations under the lagoons. Even though their methods provided estimations of the infiltration fluxes, they could not provide accurate estimates of the cumulative loads in the entire vadose zone, or the yearly contaminant loads to the underlying groundwater. All of the abovementioned studies relied on data obtained through sediment sampling. This method, however, provides only a snapshot in time and space of the pollutants' distribution in the vadose zone with no insight into the dynamic processes of flow, transport and microbial transformations in the subsurface. Moreover, sediment extracts provide information on the content of total pollutants in a sediment sample and not on the mobility properties of a specific pollutant in an undisturbed sediment profile (Amiaz et al., 2011; Rimon et al., 2011). Korom and Jeppson (1994) overcame some of these limitations by using suction cups to sample the concentrations in the - propagating pore water. They reported high NO3 concentrations in the recharging - solutions, yet their study was designed to prove NO3 leaching from lagoons constructed in coarse sediment and not to try and estimate the yearly recharging loads into the underlying groundwater. Gooddy et al. (1998) and Withers et al. (1998) observed elevated N-species and Cl- concentrations in pore water extracted from sediment samples taken from nearly the entire vadose zone (>40 m) under a dairy storage lagoon, and associated them with groundwater contamination. However, high
70 Appendix A: Regional impact
variability with time in the groundwater concentrations prevented them from providing estimates of the yearly recharge loads. Gooddy et al. (2001) concluded, based on pore-water extractions from sediment samples and on travel time estimates in the vadose zone and groundwater, that proper characterization of unsaturated zone processes is the key factor for determining the impact of agricultural pollution on groundwater quality in the UK. Baily et al. (2011) showed, using dual isotopic composition of nitrate (18O and 15N) in the groundwater and travel time estimates in - the vadose zone, that the major source for groundwater NO3 contamination was cattle farm effluents and long-term irrigation with diluted slurry (soiled water). However, they could not determine if reduction in the N-loads occurred due to denitrification in the vadose zone or in the groundwater. Harter et al. (2002) found, using a - groundwater monitoring network, that the electrical conductivity (EC) and the NO3 concentrations in the groundwater under dairy lagoons were 2.4 and 2.6 times higher, respectively, than up-gradient values. They estimated the infiltration water flux from the lagoons to be 0.8 m y-1. Van der Schans et al. (2009) estimated average nitrogen leaching from lagoons to shallow groundwater (1 – 3 m below land surface (BLS)) to be 807 kg ha-1 y-1. Their estimates are based on a calibration of a sub-regional scale and a farm-scale three-dimensional groundwater flow and transport model to observations from a groundwater monitoring network. They estimated the N concentrations in the recharging solutions based on a known recharge rate and on mass balances with the concentrations in the groundwater. Van der Schans et al. (2009) concluded that independent measurements of lagoon leaching are needed to better assess the loading rates from these manure management units. Their conclusions highlight the inaccuracies associated with estimations of the recharging loads and the loads stored in the vadose zone based on groundwater data. Such inaccuracies may be derived from: (a) the chemical composition of groundwater, as obtained from observation wells, which reflects a dynamic mixture between the chemical properties of the vertical flow components of the percolating pore water in the vadose zone and the chemical properties of horizontal groundwater flow components in the aquifer (especially when local recharge makes a significant contribution to groundwater flow to wells), and (b) the fact that when samples are taken from pumping wells, the concentrations in the pumped water reflect the mixing of the deep and shallow water in the well, and the mixing of the recharge from
71 Appendix A: Regional impact
different land management units, which may obscure the properties of the recharging solutions. Based on the above, it is difficult to accurately assess the past and the future impacts of lagoon leachates on the underlying groundwater, especially on the regional scale. Baram et al. (2012a, 2012b, under review) overcame some of these limitations through continuous measurements of the temporal variation in the vadose zone water content and frequent in-situ sampling of the chemical composition of the vadose zone pore water and groundwater underlying a dairy waste lagoon and its near surroundings. Their monitoring method provided new insights on nitrogen transformations and salinization processes in the vadose zone and enabled direct quantification of the groundwater-recharge and pollution processes under the waste lagoon and its surroundings. The aim of this study is to introduce a methodology by which spatial analysis of detailed data from the vadose zone and groundwater data can be used to assess the past and future impacts of dairy waste lagoons on groundwater degradation. In this study, the methodology is used for evaluating the relative contribution of leachates from dairy waste lagoons in the past 50 years to the - - Cl and NO3 loads in the vadose zone and groundwater in the Beer-Tuvia region in Israel. The methodology is based on: (i) the temporal and spatial distribution of pollutant concentrations and the sediment water content along the entire cross-section of a thick vadose zone (> 40 m) underlying dairy waste lagoons (data previously published by Baram et al. (2012a, 2012b, under review)), (ii) the spatial distribution of pollutants in the region's groundwater, (iii) the lithological structure of the region's subsurface, (iv) the history of dairy farming in the region, and (v) the spatial - - correlations between Cl and NO3 concentrations in the groundwater and the proximity to dairy farms.
2. Materials and methods
2.1 Study area
The study area is located in the Beer Tuvia region (40 km2), above the southern part of the Coastal Aquifer in Israel. The phreatic aquifer in the area is composed of calcareous sandstone with localized lenses of clay of varying thickness (1 to >10 m)
72 Appendix A: Regional impact
(Kurkar Group) overlaid by Pleistocene age clays (Issar, 1968; Weinberger, 2007). The land surface area is covered by a clay layer, dominated by 2:1 smectite minerals, with thicknesses that vary from several centimeters to up to 20 m over a distance of tens of meters. The climate is Mediterranean with average summer and winter temperatures of 24.3°C and 14.2°C, respectively. The average annual precipitation is ~450 mm, occurring during the winter season (November to March), mostly during five to eight rainy episodes. The main recharge to the aquifer is related to the percolation of seasonal rainwater and agricultural return flow from irrigated and rain- fed fields that have been intensively cultivated in the area for over 60 years. Since the mid 1970s, the natural flow of fresh groundwater from the Coastal Aquifer in the region into the sea was stopped by production wells (Coastal Drainage wells), and since then, the region has acted as a closed basin.
The Beer Tuvia region has been highly cultivated from the early years of the 1950s up to the present day. The agricultural activity in the region was originally based on local groundwater and fertilizers (organic and inorganic). However, groundwater salinization over the years [Cl- concentration in the local groundwater has increased from ~200 mg Cl- L-1 in the 1950s to 600-900 mg Cl- L-1 in 2010 (Vengosh and Ben-Zvi, 1994; Weinberger, 2007)] led to the shutting down of many production wells, and the agricultural activity in the region was then based on imported water from the Israeli National Water Carrier, or on mixtures between the two. Dairy farming in the region began in the 1950s (a small number of cows per farm) and intensified with time, up until the late 1980s, when the number of lactating cows in the region had stabilized at ~12000. Currently, the region hosts approximately 12,500 lactating cows in 140 dairy farms, which are located within the villages (the dairy farming history of each village is presented in the supporting information 1 (SI- 1, Table 1-1)). All the dairy farms in the region use a free stall barn with a center feed lane, and effluents from the barns are either stored in the farms in unlined earthen lagoons or spread on nearby agricultural fields as a readily available fertilizer (without regulation). Israel's Ministry of Environmental Protection reformed the dairy farming industry at the end of the 1990s. The reform led to the closure of small family-owned farms due to mergers with other farmers. The merging of farms has decreased the number of potential contamination point sources, yet increased the dairy effluent loads in the merged farms. As of today, all the farms in the region have upgraded their
73 Appendix A: Regional impact
facilities to include concrete tanks to store their effluents. Nevertheless, a centralized end-solution, such as a wastewater treatment plant or biogas generation plant, has not yet been completed, and the major portion of the dairy effluents is still stored in unlined lagoons, or spread on nearby agricultural fields. Even though many farms were closed during the reform, due to the mergers, the pollutants stored in the vadose zone beneath them may still affect the groundwater quality for many years to come.
2.2 Monitoring approach and methods used
2.2.1 Groundwater, local runoff, and sediment sampling
Groundwater and sediments from the vadose zone were sampled from a number of boreholes in the study area, including groundwater production wells, groundwater observation wells and shallow boreholes for sediment sampling. A dry-drilling method with a 6" bucket auger was used for sediment sampling from the entire vadose zone cross-section (>40 m) and the upper 10 m of the saturated zone under a dairy waste lagoon (RFW1), under a rain-fed agricultural field 800 m up gradient from the dairy farm (RFW2), and under the sediment adjacent to the dairy farm waste channel 200 m from the lagoon (RFW3). The three wells were cased and screened, with the screen interval from the water table at 42 - 46 m below surface (bls) to the bottom of the well (52 – 56 m bls, respectively). Sediment samples were collected from four 3" shallow (<10 m) boreholes drilled around the dairy farm area, using the same dry- drilling method. Three of these boreholes were drilled 0.5 to 5 m from the waste channel, and two were drilled 100 m from the waste channel in an undisturbed plot (representing the background). Two additional 2" boreholes were drilled manually to a depth of 1 m, 2 m away from the waste channel, using a hand Edelman auger (Eijkelkamp, Giesbeek, Netherlands). Local runoff generated during high-intensity rain events (>15 mm day-1) was collected from inside the desiccation cracks formed in the upper clay layer, by channeling the flowing water from the side walls of the desiccation cracks into a water collecting unit, using custom-made crack-water samplers (CWSs) (see SI-2 for more details). Though it is considered here as local runoff, it is also regarded as infiltrating water as it was sampled in the fractures below land surface. Therefore, it may be considered as infiltrating storm water. The infiltrating runoff water samples
74 Appendix A: Regional impact
were collected during the winter of 2008-2009, at 15 different locations around the dairy farm. A regional groundwater survey was conducted in 2007, 2008, and 2011, and water was sampled either directly from the production wells or from the observation wells, using a submersible pump (model MP1, Grundfos, Denmark) after purging three well-volumes. Groundwater in the observation well under the agricultural field (RFW2) and under the dairy farm waste lagoon (RFW1) was sampled at 1 and 6 m below the water table, using packers to separate perforated intervals into close-to- water table and deeper sub-intervals. All water and sediment samples were stored in polypropylene bottles and in plastic bags, respectively, and kept on ice until they + reached the laboratory (<12 h). Sediment samples were extracted for NH4 using 2 M - KCl (Maynard et al., 2008) and using double-distilled water for major ion (Cl and - NO3 ) analyses. Both the water samples and the extracted solutions were filtered (45- µm glass fiber filter) and kept at 4°C until analysis (<2 weeks). Chemical analyses to - - + determine the concentrations of N species (NO3 , NO2 , NH4 and total N) were done - - - within 48 h after sampling. NO3 , NO2 , and Cl concentrations were measured using ion chromatography (Dionex-4500i, Sunnyvale, CA). The phenate method was used + for NH4 (APHA 4500-NH3 F). Total N (TN) was determined using the persulfate digestion method (APHA 4500-N C) (APHA, 1998). All concentrations are presented as averages with standard deviations and with the number of samples, n, in brackets.
2.2.2 Mass balances
The total loads of Cl- and TN that were discharged into the environment by dairy farms from 1960 to 2011 were compared to the loads stored in the vadose zone and groundwater in the study site. The masses of Cl- and TN that were excreted from dairy farms in the study area were based on results reported by Baram et al. (2012a, 2012b, under review) and Sher et al. (2012). Since the waste lagoons at the site stored only the cows’ secretions, unlike in other locations where the stored wastewater includes collected rainfall and added irrigation water (Pettygrove et al., 2010; Viers et al., 2012), the masses discharged by each cow were calculated by dividing the loads in the monitored waste lagoon by the number of cows in the farm. The regional mass balances for each decade from 1960 to 2011 were calculated by multiplying the Cl- and TN mass discharged yearly by an individual cow with the number of cows at each
75 Appendix A: Regional impact
village during that decade (SI-1, Table 1-2). All the mass balance calculations are based on the assumption that no mass was lost from the region, and therefore, the mass of Cl- and TN that was discharged from dairies between 1960 and 2011 is stored either on the land surface in waste lagoons and in dry manure piles or in the vadose zone and groundwater. Loss of TN mass due to microbial transformations was accounted for in the vadose zone data.
Mass balance calculations for the vadose zone, underneath the dairy farms in the study area, were based on the assumption that the concentrations in the vadose zone under the different zones that were monitored in the studied dairy farm are representative for the Beer-Tuvia region, and that all dairy farms in the area had similar waste disposal systems, including a waste lagoon and a drainage channel (personal communication with Mr. Bashari Meshulam, Head of the Agricultural Department of the Beer-Tuvia local municipality). The calculations were based on results from infiltration tests (Baram et al., 2012a), vadose zone monitoring (sediment water content profiles and pore-water chemistry) (Baram et al., 2012a, 2012b, under review), extraction from sediment samples and on spatial interpolations of the lithological structure (i.e., clay or sand layers). In all the calculations, it was assumed that the percolating pore water is the sole source of Cl- and TN.
Yearly inputs of water, TN and Cl- into the vadose zone underlying the wastewater sources and their margins were estimated based on multiplication of their surface area (A) by the infiltration water flux (J) and the concentration in the infiltrating water (C) (Table 1). The TN and Cl- mass in the vadose zone at a given
time were calculated using the following equation: M vz nARZsand sand C nARZclayclayC where n is the number of dairy farms in a specific village at a given time, A (m2) is the total land surface area of the waste disposal and storage system on each farm (i.e., a summation of the surface area of the waste lagoon, its drainage channels and their margins) (Table 2), Z (m) is the thickness of the clay and sand layers (SI-1, Table 1- 4), θ is the average water content (m3 m-3) and C (g m-3) is the average concentration in the percolating pore water (Table 2). The average dairy farm size and the average lagoon size in each village in the region have changed with time, and differ from the monitored farm (hosting 60 dairy cows). Accordingly, the loads infiltrating into the vadose zone in each village were up-scaled and down-scaled using a factor, R, which
76 Appendix A: Regional impact
represents the ratio between the number of cows per dairy farm in the village at a given time and the number of cows in the monitored dairy farm (60) (SI-1, Table 1-3). The masses in the vadose zone underneath the areas outside the villages (background) were calculated using the similar equation with n,R = 1. The water, TN and Cl- masses that have reached the groundwater (recharge) and the mass stored in the vadose zone underneath the dairy farms were estimated based on an estimated travel time of 20 y through the 40 m deep vadose zone [based on an average pore-water velocity of 2 m y-1; assuming yearly flux of 0.2 m y-1 (Baram et al., under review) and volumetric water content of 0.1 m3 m-3 (Table 2)]. In all the calculations, the sole input sources of water and Cl- to the vadose zone in the dairy farm area were the infiltrating wastewater and the local runoff.
Table 1. Yearly input of water, TN and Cl- into the vadose zone underlying the permanently flooded regions and their margins. Aa J b Load Lagoon and channel combined Cl- TN m2 m y-1 Ton y-1 Ton y-1 Flooded regions 500 0.88c 0.704d 3.685d Margins of the flooded regions 1840e 0.45f 0.062g 0.170g
a Surface area; b water flux; c average measured infiltration flux from the lagoon and channel bed (Baram et al., 2012a); d average concentration in the wastewater (1600 mg-Cl- L-1 and 8376 mg-TN L-1 (Baram et al., under review)), multiplied by A and J; e equals the length of the flooded zone banks times four meters (Baram et al., under review).f average yearly precipitation in the study site, assuming no runoff; g average measured concentration in the preferentially infiltrating local runoff (75 mg-Cl- L-1 and 205 mg-TN L-1measured in the field using the CWSs).
77 Appendix A: Regional impact
- - Table 2. Chloride (Cl ) and NO3 concentrations in the vadose zone pore water, the volumetric water content of the sediment and the surface area of each location around the dairy farm that were used for the spatial interpolations Vadose zone location Concentration in the pore Water content Area No. of -1 water (mg L ) (m3 m-3) (m2) samples (n) - - Cl NO3 Clay Sand Background 630±96a 330±210a 31±10b 6±2c 8 Waste lagoon 4080±585d 740±820d 41±1e 10±2e 200 95 Waste lagoon margins 4080±482 d 780±636d 55±7e 10f 220 27 Waste channel 6400±1470d 90±160d 56±4e 10±3e 300 50 Waste channel margins 8300±1470d 1270±1200d 33±4e 9±4e 1600 30
a Background concentrations were calculated from eight sediment extractions
- - assuming that pore water is the sole source of Cl and NO3 . Concentrations were -1 calculated based on the equation Cpw = CextVtotVpw , where Cpw is the concentration in
the pore water, Cext is the concentration in the sediment extraction, Vpw is the water
volume in the extracted sediment sample and Vtot is the volume of water that was used b in the extraction plus Vpw; estimated in the lab from 46 sediment samples taken from the subsurface (down to 10 m) at four locations around the dairy farm; c estimated in the lab based on sediment samples taken from the subsurface (down to 40 m) under a non-irrigated agricultural field in the study area; d average of n pore-water samples from the vadose zone (data presented in Baram et al.(2012a; under review)); e average of three years of in-situ monitoring of the water content using FTDR probes; f not measured directly, estimated based on the high resemblance between the water content values in the sand under the waste lagoon and the waste channel and its margins.
The lithological profile of the Beer-Tuvia region was estimated based on 36 well logs (data stored in the Israel Water Authority archive). The entire vadose zone was divided into 5 m deep sections. Based on the work of Tulmatz (1977) that demonstrated that the clay in the subsurface has a localized lens-like structure, the ratio between the sand and the clay in each 5 m increment was based on their presence in the well logs. The volume of water in the saturated zone was estimated based on a
78 Appendix A: Regional impact
similar clay to sand ratio, assuming the water content at saturation for the clay and sand to be 0.6 and 0.4 m3 m-3, respectively (Baram et al., 2012a; Shavit and Furman, 2001). Superposition of the wells’ screen intervals, the groundwater table and the Cl- concentrations in the wells showed that the chloride concentrations in the deep wells (elevation deeper than -30 m below sea level) have hardly changed in the past 40 years (remained at ~200 mg L-1). At the same time, the concentrations in the shallow wells (elevation shallower than -30 m below sea level) have substantially increased in the past 40 years (Fig .1). Accordingly, all mass balances were done for a saturated zone with a thickness of 40 m (elevation of 10 to -30 m below sea level; Fig. 1).
- - Mass balances between the masses of Cl and NO3 stored in the groundwater today and the masses that have reached and were pumped out from the groundwater underlying each one of the villages in the region were calculated. In our calculations, we estimated that the boundaries of the villages in the Beer-Tuvia region have not changed since the 1960s, and therefore, the area of each village represents the area affected by leachates from dairy waste lagoons--assuming even distribution of the farms in each village (Fig. 2). Additionally, due to the lack of data, we estimated that all the groundwater production wells in the study area pumped similar annual amounts of water.
79 Appendix A: Regional impact
Cl- (g m-3) 0 200 400 600 800 1000 20
0 )
-20
-40 relative to sea level (
1960s Elevation of well screens (m) -60 1990 - 2010
Figure 1. Chloride (Cl-) concentrations and elevation of the screen intervals at different wells (of monitoring and pumping usage) in the Beer-Tuvia region. The groundwater table in the wells is located at 5 – 12 m above sea level.
2.2.3 Spatial interpolations using a geographic information system (GIS)
To understand the relation between the location of a dairy farm and the spatial variability in groundwater contamination, a GIS-based spatial analysis was applied, namely, spatial interpolation. Spatial interpolation is a set of techniques that are based on estimating the values of a surface phenomenon (e.g., groundwater concentration) for the entire study area from a set of known sample points (wells). The modeled surface represents the spatial distribution of the phenomenon. (Lloyd 2010). In addition, spatial statistical methods were used to examine and quantify the impact and correlation between dairy farms as surface pollution point sources and groundwater - - contamination. A set of sampled data, consisting of Cl and NO3 concentrations in the groundwater of the Beer-Tuvia region from the last five decades (1960-2010), was used for: (a) recognizing patterns in the groundwater concentrations; (b) testing their statistical significance and (c) evaluating quantitatively to what extent dairy farms
80 Appendix A: Regional impact
impact the contamination of groundwater. The analysis was performed using the ArcGIS® Desktop software package (ESRI, 2011).
- Chloride and NO3 concentrations in the groundwater across the entire region, for the years 1960, 1970, 1980, 1990, 2000 and 2010, were estimated based on an interpolation of a limited set of groundwater samples collected from wells in the - - regions (see SI-1, Tables 1-5 and 1-6, for a detailed description of the Cl and NO3 concentrations used). Due to the nature of the sampled data (highly heterogeneous, unevenly distributed and sparse), several interpolation methods, namely the Inverse Distance Weighted (IDW) method, the Local Polynomial Interpolation (LPI) method and different Spline methods, were compared and evaluated for their accuracy, using a cross-validation method. Cross-validation operates by removing one data sample at a time, predicting its value using the rest of the data and comparing between the predicted value and the observed value at the removed point. The cross-validation outputs’ measurements of accuracy, such as a root mean square error and a root mean square standardized error, are used for comparing between different interpolations. A Spline method was selected as it resulted in the most accurate surfaces. In addition, the Spline method can produce relatively accurate surfaces with only a few sample points. The IDW interpolations that were tested were less accurate than Spline, probably due to the sparse and uneven distribution of the available data and the short correlation distances. The Spline method in ArcGIS is a Radial Basis Function (RBF), and surfaces are created similarly to fitting a rubber sheet through sample points. Different Spline methods have been developed that differ in the smoothness of the surface (Kurtzman and Kadmon, 1999). A Spline with a tension method was selected since it performed best in accurately modeling the phenomenon.
A zonal statistical analysis was applied in order to compare between the mean concentrations within the villages and the mean concentrations outside of the village boundaries. The analysis was based on the assumptions that (a) the areas of the villages did not change within the period of time from 1960 to 2010 and that (b) the village boundaries represented the impact area of the dairy farms. The process of zonal statistics calculates statistics, such as mean, median and standard deviation for - - cells, in a raster and, in this case, on the interpolated surfaces of Cl and NO3 , based on defined zones, i.e., on the polygons that represent the areas of the villages.
81 Appendix A: Regional impact
Figure 2. Map of the Beer-Tuvia region, and the boundaries of each of the villages in the region that were used for our mass balance estimates. Circles represent the pumping and monitoring wells that were used in the spatial interpolations. Base image from Bing™ Maps Aerial, ESRI World Map Background, Copyright © 1995 - 2012 ESRI.
82 Appendix A: Regional impact
- - 2.2.4 Spatial correlations between Cl and NO3 concentrations in the groundwater and the proximity to dairy farms
To answer the question, “to what extent do dairy farms influence groundwater contamination?” a Geographically Weighted Regression (GWR) was applied. GWR (Fotheringham et al., 2002), as opposed to an Ordinary Linear Regression, provides a local model that fits a regression equation to every data point and outputs a set of regression coefficients for each data point (Lloyd, 2010), providing a truly spatial regression method. This method provides a powerful statistical analysis for modeling linear spatial relations that vary in space. In addition, GWR not only analyzes the correlation but also demonstrates the degree (strength) of correlation for each independent variable at each location and how this correlation changes in space. GWR was applied to a set of 100 random points generated for the Beer Tuvia region - - (SI-3 Fig. 3-1). Cl and NO3 concentration values were extracted from the interpolated surfaces at each of the random sample locations for the years 1960-2010. To examine the effect of dairy farm location, for each random point, a straight line distance (Euclidean distance) to the nearest dairy farm was computed--assuming that the locations of dairy farms in the Beer-Tuvia village (52, active and non-active) have hardly changed in the past 50 years (data available in the Beer-Tuvia village archive). The distance to the farm was considered as the independent (explanatory) variable and the concentration value as the dependent variable.
3. Results and discussion
All the interpolated groundwater concentration maps indicate a clear and systematic pattern of a salt plume, for both the Cl- and the NO3-, within the Beer Tuvia region, that agrees with other groundwater concentration maps for the region (Vengosh and Ben-Zvi, 1994; Weinberger, 2007) (Fig. 3). It should be noted, though, that the data set used for the interpolations of the regional groundwater Cl- and NO3- concentrations consisted of a different number of sample points (wells), and therefore, the extent of the interpolation varies between the decades (SI-1, Table 1-5 and Table 1-6).
83 Appendix A: Regional impact
- - Figure 3. Spline interpolations of groundwater Cl (a-c) and NO3 (d-f) concentrations for the years 1960, 1990 and 2010. Base image from Bing™ Maps Aerial, ESRI World Map Background, Copyright © 1995 - 2012 ESRI.
3.1 Spatial correlations between the on-land location of the dairy - - farm and the Cl and NO3 concentrations in the groundwater
Application of GWR to a set of 100 random points could not indicate a statistically 2 - - meaningful correlation (R values) between the Cl and NO3 concentrations in the groundwater and the on-land location of the 52 dairy farms in the Beer-Tuvia village.
84 Appendix A: Regional impact
The results of the GWR (SI-3 Fig. 3-2 and Fig. 3-3) indicate that there is much variation in the nature and in the strength of the relationship between the distance - - from the dairy farms and the concentrations of Cl and NO3 in the groundwater at the 2 - - random points. R values for both Cl and NO3 ranged from 0.0003 to 0.95. The spatial location of the R2 values of Cl- indicated that most of the low R2 is concentrated at points that are close to the farms, suggesting a weak relation between the points that are near farms and the concentrations at the points. High R2 values were observed in the north-western part of the region. We believe this strong correlation does not necessarily imply that the increase in distance results in a decrease in concentration, since this section is in proximity to artificial recharge, where the recharging water (surplus from the national water carrier) has a relatively low concentration of Cl- (~250 g m-3) (e.g., note the north-western part in Fig. 3c). 2 - The spatial location of the R values of NO3 indicated a similarly weak relation between the points that are close to the farms and the concentrations at the points. High R2 values were observed in the southern part of the region and not in the north- western part of the region, as with Cl-. The high R2 values in the southern part suggest - that NO3 concentrations decrease with distance from the dairy farms. It is very likely that this trend was only observed in the southern part, and not in the other parts, since the random points at the north-eastern part are located near other dairy farms in Orot village that were not accounted for (SI-3, Fig. 3-3). It is also possible that in the north- - western part, the dairy-farm contribution is obscured by the NO3 concentration in the artificially recharging water (~40 g m-3), which is close to the regional concentration.
Several processes in the region may have obscured the pollution signature of the dairy waste lagoons. It is reasonable to assume that due to the intensive groundwater pumping, which has been carried out in the region since the 1960s, the pollution plume has shifted all over the region and away from the pollution sources (dairy farms). For instance, from 1990 to 2010, 24 million m3 has been pumped from seven wells around the Beer-Tuvia village (personal communication with the water manager at Beer-Tuvia). The volume of the pumped water is ~150 times higher than the volume of the estimated recharge from dairy lagoons leachates (0.16 million m3), meaning that the recharging solutions are constantly pumped and mixed with deeper less-polluted groundwater; this is similar to the conclusions made by Gooddy et al. (2001) and to observations made by Harter et al. (2002). Another process that may
85 Appendix A: Regional impact
- mask the NO3 contribution of lagoon leachates is denitrification within the saturated zone, as observed by Singleton et al. (2007). Lastly, the chloride and nitrate concentrations in the upper (>1 m) groundwater of the monitoring well, installed by us under an agricultural field typical to the region (RFW2 – SI1 Table 1-5 and Table 1-6), were 1.5 and 1.3 times higher than the deeper groundwater (<6 m) (Fig. 4), indicating a substantial contribution of salts to the groundwater under the irrigated agricultural fields. Such concentrated recharging solutions are the outcome of the common agricultural practice in the region of salt flushing and unregulated dairy manure application to the fields, which is known to affect the underlying groundwater (Baker and Hawke, 2007; Chang and Entz, 1996; Harter et al., 2002; van der Schans et al., 2009). - -1 Cl- (mg L-1) NO3 (mg L ) 30 60 90 600 800 1000 1200 0
1
6
- NO3 -
Depth below the water table (m) Cl 7
- - Figure 4. Nitrate (NO3 ) and chloride (Cl ) concentrations at 1 and 6 m below the groundwater table under an agricultural field typical to the Beer-Tuvia region. The boundaries of the box closest to and farthest from zero indicate the 25th and 75th percentile, respectively, the solid and the dashed lines within the box mark the median and the mean, respectively, and the error bars indicate the 90th and 10th percentiles. Outlier measurements are represented by circles.
86 Appendix A: Regional impact
3.2 Regional mass balances
From 1960 to 2011, the number of dairy cows in the study area has nearly tripled from 4750 to 12625, respectively (SI, Table 1-1). Based on the concentrations in the wastewater, each cow was found to discharge yearly 41 kg of Cl- and 214 kg of total + nitrogen, from which 51.4 kg was as NH4 -N [assuming a daily secretion of 70 L of - + feces (ASAE, 2005), and average Cl , TN and NH4 -N concentrations in the wastewater are 1600, 8375 and 2012 mg L-1, respectively (n=14)]. These masses are in agreement with other works on cow excretions (Pettygrove et al., 2010; Van Horn et al., 1991). The following mass balances are based on the Cl- and nitrogen loads secreted by cows, the loads that have recharged the groundwater and that are still stored in the vadose zone (assuming a 40 m deep vadose zone and an average pore- water velocity of 2 m y-1), and the loads pumped out from the region in the production wells.
An examination of the potential contribution of dairy farming, from its initiation in the 1960s up until 2010, indicated that within that time frame, 18.8 kilotons of Cl- and 98.1 kilotons of TN were discharged into dairy waste lagoons in the study area (SI, Table 1-2). Based on calculated travel time of 20 y in the vadose zone (Baram et al., under review), leachates from the waste lagoons have contributed salts to the underlying groundwater (recharge) from 1980 onwards. Based on the volume of water that was pumped out from the regional groundwater from 1980 to - - 2010, and the estimated Cl and NO3 concentrations in the pumped water and in the - - aquifer, we estimate that 262.7 kilotons of Cl and 13.3 kilotons of NO3 (the dominant N-specie in the groundwater) were added to the groundwater at the study - - site (Table 3). During that time period, the masses of Cl and NO3 that had reached the groundwater from dairy lagoon leachates, based on data from the vadose zone (Table 2), were 14.7 and 1.9 kilotons, respectively (Table 3). The estimated yearly groundwater recharging loads from lagoons leachates were 0.42 ton-N ha-1 y-1 and 13.4 ton-Cl- ha-1 y-1 (weighted mean for observation from the entire vadose zone underlying the lagoon and channel area; Table 2). The estimated recharging Cl- loads are very close (96 %) to the calculated loads leaching from the lagoon bed (14.0 ton ha-1 y-1; based on data presented in Baram et al. (under review)). On the other hand, the estimated recharging N loads account for only a small portion (2.5 %) of the
87 Appendix A: Regional impact
calculated loads leaching through the lagoon bed (17.5 ton ha-1 y-1; based on data presented in Baram et al. (2012a; under review)). The substantial differences between the fate of the Cl- (conservative) and the N (bio-geo-reactive) in the vadose zone represent the profound impact that microbial transformations have on the overall groundwater loading rates from such manure management units. Although our calculations indicate that 97.5 % of the infiltrating N-loads are naturally removed from the vadose zone underlying lagoons, the estimated recharging N-loads in this study are an order of magnitude higher than the loads estimated by van der Schans et al. (2009) (0.01 – 0.091 ton-N ha-1 y-1 for the lagoons and the corral area). We believe that these large differences result from both the lower concentrations in the stored wastewater (the concentrations observed by Baram et al. (2012a) are almost an order of magnitude higher than the ones observed van der Schans et al. (2009), due to dilution with runoff), and from the spatial variability in the N-species concentrations observed under lagoons (DeSutter et al., 2005; Ham, 2002; Parker et al., 1995; 1999b). Baram et al. (2012a) emphasized the spatial variability in the N-species and indicated the close relation between the water content of the sediment and the fate of the N-species in the vadose zones-- based on four years of in-situ monitoring of the vadose zone underlying the area from the lagoon banks towards its center by 23 independent probes at different spatial locations and depths. Furthermore, groundwater sampling in this study indicated high variability over a short distance - (200 m down-gradient) between the NO3 concentrations in the groundwater directly under the lagoon bank and the concentrations in groundwater adjacent to the lagoon drainage channel (RFW1 vs. RFW3 in SI Table 1-6), even though both locations are - subjected to high NO3 concentrations in the recharging solutions (Table 2).
The masses added to the groundwater from artificial recharge in the study area - - (mainly around Azrikam) were 11.5 kilotons of Cl and 0.34 kilotons of NO3 . Subtraction of these masses and the masses added from lagoon leachates, from the - - total addition to the groundwater, indicated that 236.6 and 11 kilotons of Cl and NO3 were added from agricultural activity, in the area surrounding the villages (Fig. 2). Based on our average concentrations, which were measured in extractions of sediment - - samples, we estimated that 164.6 kilotons of Cl and 86.2 kilotons of NO3 entered the groundwater under the agricultural fields in the area. We believe that the differences - - between the estimated additions and the actual additions of Cl and NO3 mass to the
88 Appendix A: Regional impact
groundwater (underestimation of 71.9 kilotons Cl- and overestimation of 75.2 kilotons - NO3 ; Table 3) stemmed from the fact that the average concentrations under the agricultural field measured by us did not represent the average concentrations in the propagating solutions under the agricultural fields of the region. The small portion (1.3 %) of the land surface area occupied by waste lagoons in the region, and the similarities between the dairy lagoons in the region and the detailed monitoring of their subsurface, support this assumption. In order to complete the mass balances, the - - Cl and NO3 concentrations in the propagating water under the agricultural fields of the region should have been 1357 and 63 g m-3, respectively. These values are within - - the range of the Cl and NO3 concentrations observed by us in the vadose zone under an agricultural field typical to the region (365 – 10200 g m-3 and 1 – 297 g m-3, - - respectively) (Fig. 5). The high variability in the Cl and NO3 concentrations throughout the vadose zone under the field could have resulted from the differences between two agricultural practices: (i) irrigation of fields with locally pumped water (high Cl- concentrations) and (ii) rain-fed irrigation (low Cl- concentrations) (personal communication with the land owner), and from other factors, such as the irrigation regime, flushing of pre-cultivation immobile salty vadose zone pore water, the fertilizer input, the type of crop grown, etc. (Burow et al., 2010; Kurtzman and Scanlon, 2011; Melo et al., 2012). Observations in the vadose zone under land subjected to different agricultural practices (orchards, open fields and greenhouses) in the nearby region, using a vadose zone monitoring technique similar to the one on which this research is based, showed analogous high variability between the Cl- and - NO3 concentrations in the pore water percolating down in the vadose zone (up to differences of three orders of magnitude; unpublished field observations). Assuming that the concentrations in the recharging solutions under the fields are 1357 g-Cl- m-3 - -3 and 63 g-NO3 m means that the yearly recharging loads from the fields are 0.018 ton-N ha-1 y-1 and 0.089 ton-Cl- ha-1 y-1. These values are much lower than the values estimated by the UCCE (2005) and by van der Schans et al. (2009) (0.43 and 0.49 ton-N ha-1 y-1, respectively). We believe these differences reflect the difference in the lagoon usage. The work of the UCCE (2005) and the work of van der Schans et al. (2009) represent scenarios in which wastewater (a mixture of dairy effluents and collected runoff) from the lagoons is used mainly for irrigating the nearby fields, i.e., high quantities of wastewater with high N loads per area that result in high leaching
89 Appendix A: Regional impact
loads. On the other hand, the results of this study represent scenarios in which wastewater (effluents without local runoff) from the lagoons is used mainly as a soil amendment (readily available fertilizer), usually prior to the seeding of the fields, i.e., small quantities of wastewater with low N loads per area that result in lower leaching loads.
- - Table 3. Chloride (Cl ) and NO3 mass balances - - Cl NO3 kiloton Addition to groundwater 139.7 7.5 Pumped from ground water 123.1 5.8 Recharge from lagoon leachates 14.7 1.9 Recharge from agricultural activity 164.6 86.2 Artificial recharge a 11.5 0.34 Delta 71.9 -75.2
a - - Volumes of the artificial recharge and the Cl and NO3 concentrations in them were based on data from Vengosh and Ben-Zvi (1994) and from Israel's Water Authority database.
- -3 - -3 Cl (g m pore-water) NO3 (g m pore-water)
0 2000 4000 6000 8000 10000 12000 0 50 100 150 200 250 300 350 0
10
20
Depth (m) 30
40 a b
- - Figure 5. Chloride (Cl ) (a) and nitrate (NO3 ) (b) profiles under an agricultural field typical to the region, 15 years after the farmer stopped irrigating the field with locally pumped water.
The spatial interpolations of the Cl- concentrations in the regional ground water, from 1960 to 2010, clearly indicate that the concentrations in the groundwater increased with time. The activation of a saline source from the underlying Saqiye
90 Appendix A: Regional impact
Group and the contribution of salts from the eastern margins of the aquifer have been suggested by different studies as the sources for the ongoing salinization (Avisar et al., 2004; Gvirtzman, 2002; Rosenthal et al., 1992; Shavit and Furman, 2001; Vengosh and Ben-Zvi, 1994). The chemical data collected by us during the groundwater surveys, along with data stored in the database of Israel's Water Authority, show high variability in the Cl- concentrations over short periods of time (for instance: well Beer-Tuvia 2, June 2002 vs. July 2003, 517 vs. 969 mg L-1, respectively; July 2008 vs. August 2009, 641 vs. 853 mg L-1, respectively). These observations, along with new deep observation wells that reached the base of the aquifer and found fresh water rather than saline (the suggested underlying saline source) (Negev et al., 2011), strengthen our assumption that the groundwater salinization process results mainly from agricultural return flow (average 1357 g Cl- m-3, as previously discussed) and from the circulation of salts in the groundwater. It is - - likely that upon initiation of the agricultural activity in the region, Cl and NO3 that had accumulated in the subsurface over time were washed down (Kanfi et al., 1983; Kurtzman and Scanlon, 2011), a mechanism that intensified the salinization process.
3.3 Local impacts under each village
The regional groundwater-quality survey conducted in this study, along with the data - provided by Israel's Water Authority, indicated that the NO3 concentrations in most of the study area have remained below the Israeli standard for drinking water (70 mg L-1). Nevertheless, the spatial interpolations indicated that from the 1960s to today, - the NO3 concentrations in the groundwater under the Beer-Tuvia village have remained higher than in the surrounding groundwater (Fig. 3). Four years of frequent sampling of the upper groundwater, under and near the dairy waste lagoon, indicated that the concentrations dramatically decreased from ~300 mg L-1 under the lagoon to ~80 mg L-1 250 m down-gradient from the lagoon (RFW1 and RFW3, respectively in - SI-1, Table 1-6). Since we did not know the isotopic composition of the NO3 , we could not determine if the decrease in concentrations over such a short distance resulted from denitrification within the saturated zone (as observed by Singleton et al. (2007)) or from dilution with the ambient groundwater. It is also possible that due to - the irregular groundwater pumping regime in the region, the NO3 contamination
91 Appendix A: Regional impact
plume was shifted by a local hydraulic gradient away from our assumed down- gradient observation well. We assume that denitrification processes within the upper groundwater layer (close to the capillary fringe), along with dilution, are the main - - causes for the observed decrease in the NO3 concentrations, since high NO3 concentrations, as observed under the waste lagoon, were not observed in any other well around the region, although some of the wells were located tens to hundreds of meters away from active dairy farms. Based on the above, and on the small temporal - changes that were observed in the groundwater NO3 concentrations, we ignored the data from the monitoring well under the dairy farm (RFW1) in our mass balances, and used mean values that are based on interpolations of observations from the year 2000.
Close examination of the ratio between the masses that have reached the groundwater from dairy waste lagoon leachates and the mass added to the groundwater only under the villages in the Beer-Tuvia region, from the initiation of the dairy industry in the 1960s up until the year 2010, indicated contributions of up to - - 25 % Cl and 60 % NO3 . On average, leachates from dairy farms have contributed 19 - - % and 42 % of the masses of Cl and NO3 that were leached to groundwater under the villages in the Beer-Tuvia region (Table 4). These values indicate that even though the waste lagoons contribute to the groundwater salinization and contamination, they are not the main contributors. A chloride profile from under an agricultural field typical to the region showed high concentrations (7000 – 10000 mg L-1) in the deep section (25 – 40 m bls) and much lower concentrations (400 – 3400 mg L-1) in the shallower section (>25 m bls) (Fig. 5). A personal conversation with the field owner indicated that the profile represents the agricultural history of the field: high concentrations for the time period up to ~1990 when the field was irrigated with locally pumped water, and low concentrations for the time period from ~1990 to nowadays when the field was only rain-fed. These observations suggest that if the agricultural activity in the region would shift to being only rain-fed agriculture or low salinity irrigation water, the salinization process would stop within the next 20 – 30 years. It is more reasonable to assume that the fields in the region would continue to be irrigated. Nonetheless, due to the high Cl- concentration in the groundwater in the region, the groundwater can no longer be used for domestic or agricultural use, and farmers are shifting to irrigation with alternative water sources, such as treated domestic wastewater. We believe that due to the growing usage of desalinated
92 Appendix A: Regional impact
seawater in Israel, the Cl- concentrations in the treated wastewater will decrease and, accordingly, the salinization will decrease.
- - Table 4. Ratios between the Cl and NO3 masses that have reached the groundwater from dairy waste lagoon leachates, and the total mass addition to the groundwater under the village boundaries (Figure 2), from the initiation of the dairy industry in the 1960s up until the year 2010. - - Cl NO3 Avigdor 0.19 0.56 Azrikam 0.19 0.40 Beer-Tuvia 0.13 0.26 Kefar-Varburg 0.25 0.60 Orot 0.17 0.27
3.4 Remediation implications
Unlined dairy waste lagoons are still being used to store dairy effluents in the Beer- Tuvia region, since no other economically justified end solution exists. The results of - - our mass balances indicate that 21.5 kilotons of Cl and 2.7 kilotons of NO3 are still percolating in the vadose zone underlying dairy waste lagoons. These loads (16.1 ton- N ha-1) are 1.3 – 2.5 times higher than the loads reported by DeSutter et al. (2005). The fact that the average pore-water velocity in the vadose zone is 2 m y-1 means that even if the lagoons are dried out, in the twenty years following the drying, ~1075 tons - of Cl and ~135 tons of NO3 that leached from the lagoons would leach into the underlying aquifer every year. Based on the work of Baram et al. (2012a) that indicated the complete nitrification of the leaching nitrogen species within the upper 0.5 m of the vadose zone under the lagoons in the region, we do not anticipate that the - drying up of the lagoons would lead to a substantial addition of NO3 to the vadose zone, unlike in other locations (DeSutter et al., 2005; Ham, 2002; Harter et al., 2002). - - Nonetheless, today, 5.7% of the Cl mass and 14.1% of the NO3 mass that are stored in the vadose zone of the region are located under dairy lagoons (assuming the average concentrations in the vadose zone under the agricultural fields are 1357 g Cl- -3 - -3 m and 63 g NO3 m ). Accordingly, drying of the lagoons would have a relatively
93 Appendix A: Regional impact
minor effect on the regional groundwater contamination and salinization process. It is very likely that due to the unregulated application of manure slurry to agricultural fields in the region, the actual contribution of the dairy effluents to the pollution process is much higher. This assumption is strengthened by the observation of a high nitrate concentration in the vadose zone under land subjected to such an agricultural practice (unpublished field observations) and by many other studies from around the world (Baker and Hawke, 2007; Chang and Entz, 1996; van der Schans et al., 2009). We highly recommend that strategies for managing nutrients from manure application, such as those suggested by Van Horn et al. (1991) and others, would be adopted to avoid environmental pollution from agricultural practices. Additional thought should be given to the high risk for accelerated contamination that is associated with the preferential transport of manure and water through desiccation cracks that form in the clayey soils of the region (Baram et al., 2012a; 2012b), especially since these cracks may continue to exist below the plow zone (Mermut et al., 1996).
4. Summary
The study shows a method based on spatial statistical analysis of vadose zone and groundwater data for assessing the regional impact of dairy waste lagoons on - - groundwater quality. Regional Cl and NO3 mass balances were based on spatial interpolations of previously published data from detailed monitoring of the subsurface underlying dairy waste lagoons in the Beer-Tuvia region, Israel, on sediment samples and on a groundwater quality survey. From the 1960s to 2010, leachates from dairy - - lagoons contributed 5.6 % and 14 % of the Cl and NO3 masses that were added to the groundwater in the region. The mass balances for the lagoons and the underlying vadose zone indicated that 97.5 % of yearly infiltrating N-loads are naturally removed from the vadose zone due to microbial transformations. In the groundwater under the - - villages in the region, dairy farms contributed 19 % and 42 % of the Cl and NO3 masses, respectively, indicating that the whole area of the village, rather than just the - - area underlying the lagoons, should be addressed as a point source for Cl and NO3 contamination when conducting a regional analysis. The mass balances suggested that leachates from irrigated agricultural activity in the region are the main contributors of
94 Appendix A: Regional impact
- - Cl and NO3 to the groundwater. Sediment extractions from the vadose zone under an - - agricultural field typical to the region, along with elevated Cl and NO3 concentrations in the upper groundwater (>1 m below the water table), supported the
mass balances. Finally, the developed method can be applied to other case studies, as well as used to provide quantitative evaluation of the spatial impact of pollution point sources on the extent of groundwater contamination.
5. Acknowledgments
We thank Michael Kogel for his extensive help in the field, the dairy farm owner for
allowing us to conduct this research on his farm, Udi Galili for helping us in the
regional groundwater surveys and Dr. Irena Pankratov for her lab work. We wish to
express our gratitude to Mrs. Sara Elhanany from Israel's Water Authority (IWA) for
the close collaboration and the fruitful discussions. The project was funded by Israel's
Water Authority.
6. References
Amiaz, Y., Sorek, S., Enzel, Y., Dahan, O., 2011. Solute transport in the vadose zone and groundwater during flash floods. Water Resour. Res. 47, W10513. APHA (American public health association), 1998. Standard methods for the examination of water and wastewater, 20th ed. United Book Press, Baltimore, MD. ASAE, 2005. Manure production and characteristics (D384.2 standards), American Society of Agricultural Engineers, St. Joseph, MI, p. 19. Avisar, D., Rosenthal, E., Shulman, H., Flexer, A., Ben Havraham, Z., Kronfeld, J., 2004. The pliocene Yafo formation in Israel: hydrogeologically inert or active? Hydrogeol. J. 12, 291–304. Baily, A., Rock, L., Watson, C.J., Fenton, O., 2011. Spatial and temporal variations in groundwater nitrate at an intensive dairy farm in south-east Ireland: Insights from stable isotope data. Agr. Ecosyst. Environ. 144, 308-318. Baker, T.M., Hawke, R.M., 2007. The effects of land application of farm dairy effluent on groundwater quality - West Coast 2001. J. Hydrol. 46, 105-115.
95 Appendix A: Regional impact
Baram, S., Arnon, S., Ronen, Z., Kurtzman, D., Dahan, O., 2012a. Infiltration mechanism controls nitrification and denitrification processes under dairy waste lagoons J. Environ. Qual. 41, 1623-1632. Baram, S., Kurtzman, D., Dahan, O., 2012b. Water percolation through a clayey vadose zone. J. Hydrol. 424-425, 165-171. Baram, S., Kurtzman, D., Külls, C., Dahan, O., under review. Desiccation-crack- induced salinization in deep clay sediment. Hydrol. Earth Syst. Sc. Burkart, M.R., Stoner, J.D., 2007. Nitrate in aquifers beneath agricultural systems. Water Sci. Technol. 56, 59-69. Burow, K.R., Nolan, B.T., Rupert, M.G., Dubrovsky, N.M., 2010. Nitrate in groundwater of the United States, 1991-2003. Environ. Sci. Technol. 44, 4988- 4997. Chang, C., Entz, T., 1996. Nitrate leaching losses under repeated cattle feedlot manure applications in southern Alberta. J. Environ. Qual. 25, 145-153. DeSutter, T.M., Pierzynski, G.M., Ham, J.M., 2005. Movement of lagoon-liquor constituents below four animal-waste lagoons. J. Environ. Qual. 34, 1234-1242. ESRI, 2011. ArcGIS Desktop 10.0. Environmental Systems Research Institute, Inc., Redlands. Fotheringham, A.S., Brunsdon, C., Charlton, M., 2002. Geographically Weighted Regression: The Analysis of Spatially Varying Relationships. John Wiley & Sons, Chichester. Gooddy, D.C., Clay, J.W., Bottrell, S.H., 2002. Redox-driven changes in porewater chemistry in the unsaturated zone of the chalk aquifer beneath unlined cattle slurry lagoons. Appl. Geochem. 17, 903-921. Gooddy, D.C., Hughes, A.G., Williams, A.T., Armstrong, A.C., Nicholson, R.J., Williams, J.R., 2001. Field and modelling studies to assess the risk to UK groundwater from earth-based stores for livestock manure. Soil Use Manage. 17, 128-137. Gooddy, D.C., Withers, P.J.A., McDonald, H.G., Chilton, P.J., 1998. Behaviour and impact of cow slurry beneath a storage lagoon: II. Chemical composition of chalk porewater after 18 years. Water Air Soil Poll. 107, 51-72 Gvirtzman, H., 2002. Israel Water Resources, Chapters in Hydrology and Environmental Sciences, Yad Ben-Zvi Press, Jerusalem (in Hebrew).
96 Appendix A: Regional impact
Ham, J.M., 2002. Seepage losses from animal waste lagoons: A summary of a four- year investigation in Kansas. Trans. ASAE. 45, 983-992. Ham, J.M., Baum, K.A., 2009. Measuring seepage from waste lagoons and earthen basins with an overnight water balance test. Trans. ASABE. 52, 835-844. Ham, J.M., DeSutter, T.M., 2000. Toward site-specific design standards for animal- waste lagoons: Protecting ground water quality. J. Environ. Qual. 29, 1721- 1732. Harter, T., Davis, H., Mathews, M.C., Meyer, R.D., 2002. Shallow groundwater quality on dairy farms with irrigated forage crops. J. Contam. Hydrol. 55, 287- 315. Issar, A., 1968. Geology of central coastal plain of Israel. Israel J. Earth Sci. 17, 16- 29. Kanfi, Y., Ronen, D., Magaritz, M., 1983. Nitrate trends in the coastal-plain aquifer of Israel. J. Hydrol. 66, 331-341. Korom, S.F., Jeppson, R.W., 1994. Nitrate Contamination from Dairy Lagoons Constructed in Coarse Alluvial Deposits. J. Environ. Qual. 23, 973-976. Kurtzman, D., Kadmon, R., 1999. Mapping of temperature variables in Israel: a comparison of different interpolation methods. Clim. Res. 13, 33-43. Kurtzman, D., Scanlon, B.R., 2011. Groundwater recharge through vertisols: irrigated cropland vs. natural land, Israel. Vadose Zone J. 10, 662-674. Lloyd, C.D., 2010. Spatial Data Analysis. Oxford University Press, Oxford. Maynard, D.G., Kalra, Y.P., Crumbaugh, J.A., 2008. Nitrate and exchangeable ammonium nitrogen, in: Carter, M.R., Gregorich, E.G. (Eds.), Soil Sampling and Methods of Analysis (2nd ed.). CRC Press, Taylor and Francis Group, Boca Raton, pp. 71-80. Melo, A., Pinto, E., Aguiar, A., Mansilha, C., Pinho, O., Ferreira, I., 2012. Impact of intensive horticulture practices on groundwater content of nitrates, sodium, potassium, and pesticides. Environ. Monit. Assess. 184, 4539-4551. Mermut, A.R., Dasog, G.S., Dowuona, G.N., 1996. Soil morphology, in: Ahmed, N., Mermut, A. (Eds.), Vertisols and Technologies for Their management. Elsevier Science, Amsterdam, pp. 89-110.
97 Appendix A: Regional impact
Negev, I., Rozenthal, A., Lidgi, I., Guttman, J., 2011. Salinization mechanism in the Coastal Aquifer near Beer-Tuvia, Israel: reassessments based on updated data from new and existing wells, EGU general assembly, Vienna, Austria. Nordstedt, R.A., Baldwin, L.B., Hortenstine, C.C., 1971. Multistage lagoon systems for treatment of dairy farm waste, Livestock Waste Management and Pollution Abatement Proc. ASAE Publication, pp. 77-80. Parker, D., Nienaber, J., Eisenhauer, D., Schulte, D., 1995. Unsaturated seepage from a feedlot runoff storage pond. American Society of Agricultural Engineers Special Meetings and Conferences Papers. Parker, D.B., Eisenhauer, D.E., Schulte, D.D., Martin, D.L., 1999a. Modeling seepage from an unlined beef cattle feedlot runoff storage pond. Trans. ASAE 42, 1437-1445. Parker, D.B., Eisenhauer, D.E., Schulte, D.D., Nienaber, J.A., 1999b. Seepage characteristics and hydraulic properties of a feedlot runoff storage pond. Trans. ASAE 42, 369-380. Parker, D.B., Schulte, D.D., Eisenhauer, D.E., 1999c. Seepage from earthen animal waste ponds and lagoons - An overview of research results and state regulations. Trans. ASAE 42, 485-493. Parker, D.B., D.E. Eisenhauer, D.D. Schulte, and J.A. Nienaber. 1999b. Seepage characteristics and hydraulic properties of a feedlot runoff storage pond. Trans. ASAE 42, 369-380. Pettygrove, G.S., Heinrich, A.L., Eagle, A.J., 2010. Dairy manure nutrient content and forms. University of California Cooperative Extension, Manure Technical Guide Series, 10. Rimon, Y., Nativ, R., Dahan, O., 2011. Physical and chemical evidence for pore-scale dual-domain flow in the vadose zone. Vadose Zone J. 10, 322-331. Rosenthal, E., Vinokurov, A., Ronen, D., Magaritz, M., Moshkovitz, S., 1992. Anthropogenically induced salinization of groundwater: A case study from the Coastal Plain aquifer of Israel. J. Contam. Hydrol. 11, 149-171. Sewell, J.I., 1978. Dairy lagoon effects on groundwater quality. Trans. ASABE 21, 948-0952. Shavit, U., Furman, A., 2001. The location of deep salinity sources in the Israeli Coastal aquifer. J. Hydrol. 250, 63-77.
98 Appendix A: Regional impact
Sher, Y., Baram, S., Dahan, O., Ronen, Z., Nejidat, A., 2012. Ammonia transformations and abundance of ammonia-oxidizers in a clay soil underlying a manure pond. FEMS Microbiol. Ecol. 81, 145-155. Singleton, M.J., Esser, B.K., Moran, J.E., Hudson, G.B., McNab, W.W., Harter, T., 2007. Saturated zone denitrification: Potential for natural attenuation of nitrate contamination in shallow groundwater under dairy operations. Environ. Sci. Technol. 41, 759-765. Tulmatz, Y., 1977. Hydrogeological atlas of Israel – the Coastal Basin. The state of Israel, Department of Agriculture, Israel Water Authority, the Hydrological Service, Jerusalem. UCCE, 2005. University of California Committee of Experts on Dairy Manure Management. Managing dairy manure in the Central Valley of California. University of California, Revised version of June 2005. UNESCO, 2006. Water a shared vision. The United Nations Water Development Report 2. UNESCO-WWAP, Paris, France. van der Schans, M.L., Harter, T., Leijnse, A., Mathews, M.C., Meyer, R.D., 2009. Characterizing sources of nitrate leaching from an irrigated dairy farm in Merced County, California. J. Contam. Hydrol. 110, 9-21. Van Horn, H.H., Newton, G.L., Nordstedt, R.A., French, E.C., Kidder, G., Graetz, D.A., Chambliss, C.F., 1991. Dairy manure management: strategies for recycling nutrients to recover fertilizer value and avoid environmental pollution. Gainesville, Fla.: Florida Cooperative Extension Service, Institute of Food and Agricultural Sciences, University of Florida. Vengosh, A., Ben-Zvi, A., 1994. Formation of a salt plume in the Coastal-Plain Aquifer of Israel - the Beer-Toviyya region. J. Hydrol. 160, 21-52. Viers, J.H., Liptzin, D., Rosenstock, T.S., Jensen, V.B., Hollander, A.D., McNally, A., King, A.M., Kourakos, G., Lopez, E.M., De La Mora, N., Fryjoff-Hung, A., Dzurella, K.N., Canada, H., Laybourne, S., McKenney, C., Darby, J., Quinn, J.F., Harter, T., 2012. Nitrogen sources and loading to groundwater, Technical Report 2: Assessing Nitrate in California’s Drinking Water With a Focus on Tulare Lake Basin and Salinas Valley Groundwater, Report for the State Water Resources Control Board Report to the Legislature, p. 323.
99 Appendix A: Regional impact
Weinberger, G., 2007. The developement and use of water resources in Israel until autumn 2006. Ministry of National Infrastructure, Jerusalem, p. 405 (in Hebrew). Withers, P.J.A., McDonald, H.G., Smith, K.A., Chumbley, C.G., 1998. Behaviour and impact of cow slurry beneath a storage lagoon: 1. Groundwater contamination 1975-1982. Water Air Soil Poll. 107, 35-49.
100 Appendix A: Supporting information – Paper under review, Journal of Environmental Management
Supporting information to "Assessing the impact of dairy waste lagoons on groundwater quality using a spatial analysis of vadose zone and groundwater information: example from the Beer-Tuvia region Israel"
Shahar Baram1, Zeev Ronen1, Daniel Kurtzman2, Aviva Peeters3 and Ofer
Dahan1
1Department of Environmental Hydrology & Microbiology, Zuckerberg Institute for
Water Research, Blaustein Institutes for Desert Research, Ben-Gurion University of the Negev, Sede Boqer Campus, Israel 84990.
2Institute of Soil, Water and Environmental Sciences, Agricultural Research
Organization, The Volcani Center, Bet Dagan 50250, Israel.
3Unit of Desert Architecture and Urban Planning, The Swiss Institute for Dryland
Environmental and Energy Research, J. Blaustein Institutes for Desert Research, Ben-Gurion
University of the Negev, Sede Boqer Campus 84990, Israel
101 Appendix A: Supporting information – Paper under review, Journal of Environmental Management
Supporting information 1
Table 1-1. Number of dairy cows (calves, heifers and lactating) and farms in the villages in the studied region between the years 1960 – 2010
Village 1960 1970 1980 1990 – 2010 Cows DF Cows DF Cows DF Cows DF Avigdor 450a 60a 1000b 60c 1500d 60e 2470f 19f Orot 400g 8c 400b 8c 570d 8e 405f 5f Beer-Tuvia 1440h 90h 2200b 60c 4300d,i 61i 5400f 49f Kefar-Varburg 1260h 80h 1500b 60c 1500d 60e 2350f 43f Azrikam 1200j 70j 1600j 40j 1280j 32j 2000f 23f Total 4750 308 5800 2289150 221 12625 139
DF - Dairy farm; a (Atzmon, 1959); b (Ben-Zvi, 1971); c calculated based on the assumption that the ratio between the cows and the dairy farms in the Azrikam village represents the whole region; d (Zelinger, 1983; Zelinger, 1985); eno data was found, therefore remains unchanged; f data from the archive of the Agricultural Department in the Beer-Tuvia local municipality; g personal communication with Mr. Zvi Marks, one of the founders of the Orot village; h (Atzmon, 1964; Shmargad, 1962); i(Zoler, 1980); j personal communication with Mr. Eli Z'orno, senior dairy farmer in Azrikam village.
102 Appendix A: Supporting information – Paper under review, Journal of Environmental Management
- + Table 1-2. Chloride (Cl ), total-nitrogen (TN) and ammonium-nitrogen (NH4 -N) discharged from dairy farms into the environment at the research site between the years 1960 – 2010. -a a + a Village Cl TN NH4 -N Ton Ton Ton Avigdor 3235 16885 4056 Orot 894 4665 1121 Beer-Tuvia 7683 40104 9634 Kefar-Varburg 3674 19174 4606 Azrikam 3313 17291 4154 Total 18799 98119 23571 a Based on the assumption that each cow discharges: 41 kg y-1 Cl-; 214 kg y-1 TN, -1 + from which 51.4 kg y as NH4 -N.
Table 1-3. The ratio between the number of cows per dairy farm in the village at a given time and the number of cows in the monitored dairy farm.
Village 1960a 1970a 1980a 1990-2010a Orot 0.8 0.8 1.2 1.4 Beer Tuvia 0.3 0.6 1.2 1.8 Kefar Varburg 0.3 0.4 0.4 0.9 Azrikam 0.3 0.7 0.7 1.4 Avigdor 0.1 0.3 0.4 2.2 a Calculated by dividing the average dairy size during each decade (number of cows divided by number of dairy farms; data presented in SI1-Table 1-1) by the size of the monitored dairy farm (60 dairy cows).
103 Appendix A: Supporting information – Paper under review, Journal of Environmental Management
Table 1-4. Thickness of the clay and sand units in the saturated and unsaturated zones used for mass balance calculations.
Village Vadose zonea Saturated zonea Clay Sand Clay Sand
(m) (m) (m) (m) Orot 8.75 36.25 0 40 Beer Tuvia 7 38 0 40 Kefar Varburg 15.25 29.75 0 40 Azrikam 3.1 36.9 6 34 Avigdor 8.3 36.7 3.3 36.7 Rest 10.55 34.45 1.8 38.2 a Thicknesses represent the sum of the ratios between the sand and the clay in 5 m deep increments of 36 well logs from the region (data stored in the Israel Water Authority archive).
104 Appendix A: Supporting information – Paper under review, Journal of Environmental Management
Table 1-5. Chloride concentrations in the wells spanning the upper 40 m of the groundwater that were used for the spatial interpolations. Well ID 1960 1970 1980 1990 2000 2010 Cl- (mg L-1) Avigdor 12/7 12812201 230a 712b Avigdor A 12412501 190 a 216 a 300 a 384 a 648 b Avigdor B 12412401 260 a 260 a Avigdor C 12312503 240 a 240 a 300 a Avigdor D 12412502 200 a 259 c 336 c 520 b Avigdor E 12312502 180 a 77 b Avigdor F 12312401 180 a 311 a Azrikam 3 12712001 190 a 242 b Azrikam 4 12612001 182 a Beer-Tuvia 1 12712201 191 a 196 a 250 a 295 a Beer-Tuvia 14/5 12812401 164.5 a Beer-Tuvia 2 12612201 343 a 464 a 517 c 480 c 752 b Beer-Tuvia 3 12612403 400 a 485 a 657 a 730 c 1060 a Beer-Tuvia 4 12712302 352 a 491 a 563 a 677 c 790 a 866 b Beer-Tuvia 5 12612402 250.5 293 a 419 a 533 c 547 a 794 b Bitania 2 12812205 180 d 200 d 242 d 263 d 314 d 294 d Hazor 13/4 12912201 279 a Kefar Ahim 14/6 12812501 245 a Kefar Ahim 14/7 12812701 194 a Kefar Ahim 4 12712601 211.5 a 200 a 243 a 480 c Kefar Ahim 6 12712602 220 a Kefar Varburg 12/6 12512301 190 a 646 b Kefar Varburg 6 12512202 243 a Kefar Varburg A 12512401 250 a 280 a 390 a 471 a 727 a Kefar Varburg B 12512303 248 a 310 a Kefar Varburg D 12512302 220 a 336 a 515 a 545 a 682 a Magen Avraham Eliezer A 12612502 180 a 190 a 222 a 250 a Magen Avraham Eliezer B 12612501 190 a 210 a 233 a 422 a 675 a 898 b Orot A 12712401 200 a 258 a 475 a 353.5 a RFW1 (Dairy farm) 921 e RFW2 (Ag. field) 641 e RFW3 645 b a Concentrations are taken from the Israel Water Authority’s database; b average concentration in water sampled in this project (2007 to 2011); c concentrations are taken from Vengosh and Ben-Zvi (1994); d average concentrations for each decade based on Mekorot – Israel National Water Company’s database; e average of four years of monthly sampling.
105 Appendix A: Supporting information – Paper under review, Journal of Environmental Management
Table 1-6. Nitrate concentrations in the wells spanning the upper 40 m of the groundwater that were used for the spatial interpolations.
Well ID 1960 1970 1980 1990 2000 2010 - -1 NO3 (mg L ) Avigdor 12/7 12812201 19a 21 b Avigdor A 12412501 48 a 45 c 36 b Avigdor B 12412401 31 a 31 a 35 b Avigdor C 12312503 32 c Avigdor D 12412502 41 a Avigdor F 12312401 36.9 a Azrikam 3 12712001 47 a 30 b Azrikam 4 12612001 39.4 a Beer-Tuvia 1 12712201 37 a 52.2 a 29 a 26 a Beer-Tuvia 14/5 12812401 1.4 a Beer-Tuvia 2 12612201 37.7 a 48.6 a 52 b Beer-Tuvia 3 12612403 50.5 a 55.8 a 55 a 62 a 103 a Beer-Tuvia 4 12712302 60 a 63 a 51 a 63 a 95 a 82 b Beer-Tuvia 5 12612402 50.5 a 60.4 a 53 a 48 a 51 a 54 b Bitania 2 12812205 2 d 23 d 42 d 33 d 46 d 36d Hazor 13/4 12912201 0 a Kefar Ahim 14/6 12812501 2 a Kefar Ahim 14/7 12812701 2 a Kefar Ahim 4 12712601 34.7 a 28 a 29 a Kefar Varburg 12/6 12512301 53 b Kefar Varburg 6 12512202 17.8 a Kefar Varburg A 12512401 51.3 a 42 a 42 a 46 a Kefar Varburg B 12512303 35.8 a Kefar Varburg D 12512302 19.8 a 18 a 26 a 32 a Magen Avraham Eliezer A 12612502 49.4 a 43 a 32 a Magen Avraham Eliezer B 12612501 54.9 a 46 a 43 a 52 a 47 b Orot A 12712401 74.7 a 66 a 33.9 a RFW1 (Dairy farm) 309 e RFW2 (Ag. field) 67 e RFW3 80 b a Concentrations are taken from the Israel Water Authority’s database; b average concentration in water sampled in this project (2007 to 2011); c concentrations are taken from Vengosh and Ben-Zvi (1994); d average concentrations for each decade based on Mekorot – Israel National Water Company’s database; e average of four years of monthly sampling.
106 Appendix A: Supporting information – Paper under review, Journal of Environmental Management
Supporting information 2
Crack water samplers (CWSs) were built from an 8 x 20 cm PVC pipe cut in half with rubber stoppers at both ends and a 50-mL cupped polypropylene water-collecting unit at its center with a small-diameter pipe in it (Fig. 2-1a). The CWSs were installed inside desiccation cracks at 15 different locations around the dairy farm, and collected the water flowing on the side walls of the desiccation cracks (Fig. 2-1b). Water samples were sucked out from the water-collecting unit through the small-diameter pipe using a syringe.
Fig. 2-1. Crack water sampler (a), and the way it operates inside a desiccation crack (b).
107 Appendix A: Supporting information – Paper under review, Journal of Environmental Management
Supporting information 3
Figure. 3-1. Set of 100 random points generated for the Beer Tuvia region, from - - which Cl and NO3 concentration values were extracted from the interpolated surfaces for the years 1960 – 2010, and the dairy farms in the village. Base image from Bing™ Maps Aerial, ESRI World Map Background, Copyright © 1995 - 2012 ESRI.
108 Appendix A: Supporting information – Paper under review, Journal of Environmental Management
Figure 3-2. Geographically weighted regression (GWR) for the set of 100 random points generated for the Beer Tuvia region and the groundwater Cl- concentrations in the year 2010. R2 values indicate the strength of the relation between the distance from dairy farms and between concentration values at random points. Base image from Bing™ Maps Aerial, ESRI World Map Background, Copyright © 1995 - 2012
ESRI.
109 Appendix A: Supporting information – Paper under review, Journal of Environmental Management
Figure 3-3. Geographically weighted regression (GWR) for the set of 100 random - points generated for the Beer Tuvia region and the groundwater NO3 concentrations in the year 2010. R2 values indicate the strength of the relation between the distance from dairy farms and between concentration values at random points. Base image from Bing™ Maps Aerial, ESRI World Map Background, Copyright © 1995 - 2012 ESRI.
110 Appendix A: Supporting information – Paper under review, Journal of Environmental Management
References
Atzmon, Y., 1959. Cattle show in Avigdor village. Dairy Farming Journal, 31: 14-15 (in Hebrew). Atzmon, Y., 1964. Meeting again with Beer-Tuvia. Dairy Farming Journal, 60: 3-13 (in Hebrew). Ben-Zvi, A., 1971. Summary of milk yield 1969/70. Dairy Farming Journal, 111: 2-15 (in Hebrew). Shmargad, A., 1962. Summary of milk and fat yield 1960/61. Dairy Farming Journal, 43: 4-12 (in Hebrew). Vengosh, A. and Ben-Zvi, A., 1994. Formation of a Salt Plume in the Coastal-Plain Aquifer of Israel - the Beer-Toviyya Region. Journal of Hydrology, 160: 21- 52. Zelinger, Y., 1983. Herd book summaries 1981. Dairy Farming Journal, 183: 25-28 (in Hebrew). Zelinger, Y., 1985. Summary of complete lactations 1983/84. Dairy Farming Journal, 194: 9-16 (in Hebrew). Zoler, Y., 1980. On Dairy farming in Beer-Tuvia. Dairy Farming Journal, 165: 53-54 (in Hebrew).
111 תקציר
- - תשטיפים מבריכות שפכי רפתות תרמו % 5.6 ו- % 14 מכלל מסת ה- Cl וה- NO3 שנוספה ושנשאבה ממי התהום באזור יחדיו. בחינה מקרוב של היחס בין המסות שהגיעו אל מי התהום מתשטיפי בריכות שפכי רפתות לבין המסות שנוספו אל מי התהום מתחת לכל אחד מיישובי אזור - - באר טוביה מ- 1960 ועד ל- 2010 הצביעה על תרומה ממוצעת של % Cl 19 ו- % NO3 42. ריכוזים גבוהים אלו הצביעו על כך שכתוצאה מהשאיבה הממושכת של מי התהום באזור, יש - - להתייחס אל כלל שטח היישוב כאל מוקד זיהום ב- Cl וב- NO3 ולא רק לשטח שמתחת
- - 2 לבריכות. הנחה זו גובתה בערכי מתאם נמוכים ( R) בין ריכוז ה - Cl וה - NO3 במי התהום לבין המרחק מהרפתות על פני השטח; על אף שרפתות נמצאו כמקור זיהום של מי התהום ב - -Cl וב-
- - - NO3. חישובים אף הצביעו על כך ש % 5.7 ו- % 16.4 ממסות ה- Cl וה- NO3, בהתאמה הנמצאות בתווך הלא רווי באזור נמצאות בתווך הלא רווי מתחת לבריכות שפכי הרפתות. בהתאם לכך, ייבוש הבריכות ישפיע באופן מועט על תהליך ההמלחה האזורי ובאופן די משמעותי - על תהליך הזיהום ב NO3. אם זאת, ייבוש בריכות השפכים יעצור את תהליך הזיהום וההמלחה הנקודתי של מי התהום.
מילות מפתח: סדימנט חרסיתי, סדקי כיווץ, זרימה במסלולי זרימה מועדפים, המלחה בתת הקרקע, אידוי מתת הקרקע, בריכות שפכי רפתות, זיהום מי תהום, רפתות.
ד תקציר
- ככל הנראה מניטריפיקציה (nitrification) מלאה בתנאים אירוביים (חמצון NH3 ל - NO3) וכתוצר לוואי של פעילות אנאמוקס. כמו כן, תצפיות אלו הצביעו על כך שהיווצרות סדקי כיווץ בסדימנט החרסיתי הגביר את אוורור תת - הקרקע אף בתנאי רטיבות הקרובים לרוויה. תהליכי - ניטריפיקציה ודה-ניטריפיקציה [חיזור NO3 לחנקן גזי (N2)] מצומדים נמצאו כגורם העיקרי - המווסת את גורל ה- NO3 בתווך הלא רווי מתחת לבריכות עפר. בתכולות רטיבות נמוכות (45% רוויה במחקר זה) תהליכי ניטריפיקציה ודה-ניטריפיקציה מצומדים הובילו להפחתה מינורית - במסת ה- NO3 הנעה בתווך הלא רווי, בעוד שבתכולות רטיבות בינוניות (70% רוויה במחקר זה) וגבוהות (90% רוויה במחקר זה) תהליכים מצומדים אלו הובילו להסרה של 90% וכמעט 100%, - בהתאמה ממסת ה- NO3 הנעה בתווך הלא רווי. על אף שתהליכי ניטריפיקציה ודה-ניטריפיקציה מצומדים הובילו להסרה של מעל 90% ממסת החנקן המחלחלת לתווך הלא רווי מאזור בריכת השפכים, ריכוז החנקן בתמיסות המעשירות את מי התהום בסביבה זו נותר מאוד גבוהה (>400 - 1 - - 1 מ"ג ל ). ריכוז ה- NO3 במי התהום מתחת לבריכת השפכים (>300 מ"ג ל ) היה גבוהה מהותית (פי 5 ויותר) מאשר במי התהום באזור המחקר, והצביע על כך שבריכת השפכים מהווה מקור - זיהום נקודתי למי התהום ב- NO3.
ריכוז הכלוריד (-Cl) במי הנקבים מתחת לבריכת השפכים ולסביבתה עלה עד פי 5.5 לאורך התווך הלא רווי החרסיתי. ערכי איזוטופים יציבים של חמצן ומימן (δ18O ו - δ2H, בהתאמה) במי הנקבים ובדוגמאות סדימנט מהתווך הלא רווי הצביעו על העשרה איזוטופית (הכבדה) עם העומק, והצביעו על תהליכי אידוי בתת הקרקע עד לעומק של 3~ מטר מתחת לפני השטח. תנודות יומיות בגרדיינט בין טמפרטורת האוויר על פני השטח וטמפרטורת האוויר בתוך מפתח הסדק, נמצאו כמספיקות בכדי ליצור תנועת אוויר קונבקטיבית במפתח הסדק, ולהוביל לאידוי מים מתת הקרקע. מודל קונספטואלי בשם: המלחה מואצת בסדקי כיווץ [-desiccation (crack-induced salinization (DCIS] הוצע בכדי להסביר את תצפיות השדה. המודל הקונספטואלי תומך במודלים קודמים על המלחה של התווך הלא רווי ומי התהום בסלע סדוק בסביבות יבשות (arid) ומרחיב את תקיפותם לקרקעות חרסיתיות בסביבות חצי יבשות (-semi arid). חלחול תמיסות בעלות מליחות גבוהה למדי (1600 מ"ג כלוריד ל 1-) מסביבת בריכת השפכים, ותכולת הרטיבות הגבוהה בסדימנט תחתיה גורמים להגברת האידוי מתת הקרקע ומאפשרים הצטברות של מלח בתת-הקרקע. בהתאם לכך, בריכת השפכים מהווה מקור נקודתי הממליח את מי התהום.
תוצאות ניטור התווך הלא רווי מתחת לבריכת השפכים וסביבתה והרכבם הכימי של שמונת המטרים העליונים של מי התהום מתחת לבריכה, צורפו לסקר אזורי של הרכב מי התהום בכדי להעריך את תרומתם האזורית של בריכות שפכי רפתות על איכות מי התהום. בחינת - - תרומת בריכות השפכים נעשתה באמצעות מאזני מסה אזוריים של Cl ו - NO3 תוך שימוש באינטרפולציות מרחביות באמצעות תוכנת GIS. תוצאות מאזני המסה הראו כי מראשית התפתחות ענף הרפתות באזור באר טוביה בשנות השישים של המאה הקודמת ועד לשנת 2010
ג תקציר
תקציר
בריכות עפר (בור חפור בקרקע או בור עם דיפון של חרסית) משמשות לאגירה של שפכים ממוקדי חקלאות בעלי חיים אינטנסיביים כגון רפתות. לצמצום החלחול מבריכות אלו תחתית וגדות הבריכה לרוב מדופנות בסדימנט חרסיתי כבד, אשר נחשב כבעל חדירות נמוכה עקב המוליכות ההידראולית הנמוכה שלו בתנאי רוויה. לרוע המזל, לעיתים תכופות נמצא כי פרקטיקה זו אינה מונעת חלחול נרחב של מזהמים לתת הקרקע ומובילה לזיהום של מי התהום. אחד החששות העיקריים מחלחול מבריכות האוגרות שפכי רפתות הינו הריכוז הגבוה של צורני חנקן שונים [ + חנקן אורגאני, אמוניה (NH3), ואמוניום ( NH4)] בתמיסות המחלחלות. מחקרים על גורל צורני החנקן בתווך הלא רווי ובמי התהום מתחת לבריכות אגירה לשפכי רפתות הראו כי גורלם יכול להשתנות במרחב ובזמן הן בין ברכות שונות והן מתחת לברכה אחת. במחקר זה, מנגנוני זרימת המים והסעת מזהמים וכן גורל צורני החנקן בתווך הלא רווי ובמי התהום מתחת לבריכת עפר האוגרת שפכי רפתות, נלמדו תוך שימוש במדידות שדה (in-situ) לאורך טווח זמן ארוך ובניסויי מעבדה תומכים. המחקר בחן בתחילה את המאפיינים הפיזיקאליים של מערכות סדקי הכיווץ הנוצרות באופן טבעי בסדימנט החרסיתי בסביבת בריכת השפכים, ואת השפעתם על חלחול המים ועל התקדמותם בתווך הלא רווי. מדידות רציפות של שינויים עיתיים בתכולת הרטיבות בחתך הסדימנט הלא רווי בשדה, הצביעו על חלחול מהיר והתקדמות של חזית ההרטבה לעומק (>12 מ) החתך תוך שעות ספורות מאירועי גשם גדולים ובעת שינויים במפלס השפכים בבריכה. דפוס התקדמות המים בתווך הלא רווי הצביע על כך שכמות מים נכבדה חוצה את שכבת החרסית ומעשירה את שכבת הכורכר תחתיה. על בסיס התצפיות הוצע כי סדקי הכיווץ חוצים את כל שכבת החרסית ונותרים פתוחים על פני השטח ובתת הקרקע לאורך כל השנה, אף במהלך החורף ובתנאי רטיבות הקרובים לרוויה. בנוסף לאירועי חלחול מהיר במסלולי זרימה מועדפים, נמצא כי קיים חלחול קבוע ואיטי (מ"מ יום) מתחת לחלקי הבריכה המוצפים באופן תמידי. חלחול השפכים מאזור זה נשלט על ידי המוליכות ההידראולית של שכבת חומר אורגאני דק אשר שקע על קרקעית הבריכה והפחית את חדירותו של מטריקס הסדימנט החרסיתי. הפחתת המוליכות ההידראולית באזור המוצף, הובילה להיווצרותם של תנאי אי רוויה בסדימנט לאורך כל השנה.
דיגום תכוף של מי הנקבים מהתווך הלא רווי באמצעות מערכות ניטור בשדה יחד עם גלעיני סדימנט, הצביעו על כך שכלל צורוני החנקן המחלחלים מסביבת הבריכה (חנקן אורגני, + NH3ו- NH4 ) חומצנו ב 0.5 המטר העליונים של התווך הלא רווי. תהליך החמצון לווה - בהצטברות של NO3 בתווך הלא רווי. נפיצותן של בקטריות מחמצנות - אמוניה ( -ammonia oxidizing bacteria- AOB) ובקטריות - אנאמוקס (anammox bacteria – AMX) נמצאה גבוהה ביותר בשכבת הקרקע העליונה (<0.2 מטר; 107 העתקי גנים ג 1- סדימנט יבש) ופחתה באופן ניכר עם העומק (עד לעומק של 0.5 מטר). לעומת זאת, נפיצותן של הארכיאות מחמצנות-האמוניה (ammonia-oxidizing archaea- AOA), כמעט ולא השתנתה לאורך שכבת הקרקע העליונה (<0.5 מטר) ונותרה על 107 העתקי גנים ג 1- סדימנט יבש. נפיצות הבקטריות יחד עם החמצון - המלא של צורוני החנקן המחלחלים הצביעה על כך שהצטברות ה- NO3 בתווך הלא רווי נבעה
ב עבודה זו נעשתה בהדרכתם של "ד ר עפר דהן ופרופ' זאב רונן המחלקה להידרולוגיה ומיקרוביולוגיה של הסביבה, מכון צוקרברג לחקר המים, המכונים לחקר המדבר על שם יעקוב בלאושטיין, אוניברסיטת בן-גוריון בנגב . . "ד ר דניאל קורצמן המכון למדעי הקרקע, המים והסביבה, מינהל המחקר החקלאי, מרכז וולקני, בית דגן . .
א
שפכי רפתות וזיהום מי תהום – תובנות מהתווך הלא רווי
מחקר לשם מילוי חלקי של הדרישות לקבלת תואר "דוקטור לפילוסופיה " "
מאת
שחר ברעם
הוגש לסינאט אוניברסיטת בן גוריון בנגב
אישור המנחה ______
אישור דיקן בית הספר ללימודי מחקר מתקדמים ע"ש קרייטמן ______
ל ניסן התשע ג" 10 אפריל 2013
באר שבע
שפכי רפתות וזיהום מי תהום – תובנות מהתווך הלא רווי
מחקר לשם מילוי חלקי של הדרישות לקבלת תואר "דוקטור לפילוסופיה " "
מאת
שחר ברעם
הוגש לסינאט אוניברסיטת בן גוריון בנגב
ל ניסן התשע ג" 10 אפריל 2013
באר שבע