Dissolved Organic Carbon in Groundwater Overlain by Irrigated Sugarcane by Thabo Thayalakumaran1,2 , Matthew J. Lenahan1,3 , and Keith L. Bristow1,4
Abstract Elevated dissolved organic carbon (DOC) has been detected in groundwater beneath irrigated sugarcane on the Burdekin coastal plain of tropical northeast Australia. The maximum value of 82 mg/L is to our knowledge the highest DOC reported for groundwater beneath irrigated cropping systems. More than half of the groundwater sampled in January 2004 (n = 46) exhibited DOC concentrations greater than 30 mg/L. DOC was progressively lower in October 2004 and January 2005, with a total decrease greater than 90% indicating varying load(s) to the aquifer. It was hypothesized that the elevated DOC found in this groundwater system is sourced at or near the soil surface and supplied to the aquifer via vertical recharge following above average rainfall. Possible sources of DOC include organic-rich sugar mill by-products applied as fertilizer and/or sugarcane sap released during harvest. CFC-12 vertical flow rates supported the hypothesis that elevated DOC (>40 mg/L) in the groundwater results from recharge events in which annual precipitation exceeds 1500 mm/year (average = 960 mm/year). Occurrence of elevated DOC concentrations, absence of electron acceptors (O2 and NO3 – ) and both Fe2+ and Mn2+ greater than 1 mg/L in shallow groundwater suggest that the DOC compounds are chemically labile. The consequence of high concentrations of labile DOC may be positive (e.g., denitrification) or negative (e.g., enhanced metal mobility and biofouling), and highlights the need to account for a wider range of water quality parameters when considering the impacts of land use on the ecology of receiving waters and/or suitability of groundwater for irrigated agriculture.
Introduction The chemical composition of groundwater is a measure of its quality and hence suitability as a source of water for human consumption, irrigation, and industry. Groundwater quality also influences ecosystem health and function, so it is important to be aware of and avoid the degradation of groundwater resources. In agricultural regions, considerable research has been carried out on groundwater quality problems related to salinity, nutrients, herbicides, and pesticides (Bohlke 2002; Di and Cameron 2002; Scanlon et al. 2007; Arias-Estevez et al. 2008; Lazorka-Connon and Achari 2009). Far less attention has been focused on the effects of agricultural practices on 1 Cooperative Research Centre for Irrigation Futures and CSIRO Water for a Healthy Country National Research Flagship, CSIRO Land and Water, PMB Aitkenvale, Townsville, Queensland 4814, Australia; [email protected] 2 Currently with Department of Environment and Primary Industries—Agriculture Research, 32 Lincoln Square Nth, Carlton, Vic 3053, Australia; [email protected] 3 Currently with AECOM, PO Box 5423, Townsville, Queensland 4810, Australia; [email protected] 4 Corresponding author: Cooperative Research Centre for Irrigation Futures and CSIRO Water for a Healthy Country National Research Flagship, CSIRO Land and Water, PMB Aitkenvale, Townsville, Queensland 4814, Australia. [email protected] Received February 2014, accepted July 2014. © 2014, Commonwealth of Australia. Groundwater © 2014, National Ground Water Association doi: 10.1111/gwat.12258
NGWA.org
dissolved organic carbon (DOC) leaching from soils and subsequent transport to the water table (Herbauts 1980; Ronen et al. 1987). DOC in groundwater can lead to several microbial and geochemical processes that affect the redox state of the aquifer system(s). Some of these processes, such as nitrate reduction, may be perceived as beneficial to the environment whereas those that affect oxygen availability and the mobility of metals and organic contaminants (Aravena and Wassenaar 1993; Dudal and Gerard 2004; Chapelle et al. 2012) are considered detrimental. Availability of surface DOC in connected surface water-groundwater systems may impact downstream ecosystems by reducing the availability of oxygen (Moore 2010) or facilitating the delivery of toxicants including mercury (Lamborg et al. 2004; Bone et al. 2007). Reported DOC concentrations in uncontaminated groundwater beneath cropped agricultural systems are less than 1 to 5 mg/L (Starr and Gillham 1993; Chomycia et al. 2008; Rivett et al. 2008), but specific agricultural and animal management practices can result in higher levels of DOC (up to 55 mg/L; Chomycia et al. 2008). Environmental and hydrologic conditions including intensive irrigation, high rainfall events, and high permeability soils enhance the transport of surface available DOC to shallow aquifer systems via various pathways (Kalbitz et al. 2000; Chapelle et al. 2013). In groundwater systems beneath an intensively irrigated sugarcane area, Thayalakumaran et al. (2008)
Vol. 53, No. 4–Groundwater–July-August 2015 (pages 525–530)
525
reported unusually high groundwater DOC concentrations and its widespread geographical occurrence. Thayalakumaran et al. (2008), however, found no specific spatial trend in the groundwater DOC concentrations, and the source of the DOC and delivery mechanisms to the groundwater was unknown. As agricultural management and hydraulic conditions follow seasonal patterns, it is likely that DOC loading to shallow groundwater will follow similar patterns if the DOC is sourced at or near the soil surface. We hypothesize that the DOC found in this groundwater system was sourced from the soil surface, transported to the groundwater via vertical recharge following years with particularly high levels of rainfall and that the DOC has a significant influence on the geochemistry of the shallow portions of the aquifer. The hypothesis was tested by (1) collecting groundwater samples across the lower Burdekin at three different times for analysis of DOC and various other constituents including DO, NO3 - , Fe2+ , and Mn2+ , and (2) using CFC-12 vertical flow rates and rainfall data to analyse the relationship between groundwater recharge and DOC concentrations in the groundwater.
Methods Study area The lower Burdekin (Figure 1) is one of Australia’s premier irrigation areas with a reputation for producing some of the highest yields and highest quality sugarcane in Australia (Bristow 2004). A detailed description of the area has been presented by Thayalakumaran et al. (2008), and catchment characteristics relevant to this study are summarized here. The lower Burdekin has a tropical climate characterized by hot wet summers (December to March) and dry winters (June to September). Mean annual rainfall for the region is 960 mm. Water for irrigation is sourced from the floodplain aquifer and/or Burdekin River. The aquifer typically provides in excess of 250,000 ML/annum for sugarcane irrigation. Sugarcane is harvested between June and December every year. The floodplain aquifer system comprises unconsolidated, alluvial, and marine sedimentary deposits that overlie a granitic basement. Numerous coarse-grained palaeochannels dissect the current floodplain (Alexander and Fielding 2006) and have been identified as important groundwater recharge areas (Lenahan and Bristow 2010). The presence of sandy to loamy soils and the discontinuous nature of the underlying clay layers means that the groundwater system is generally unconfined (Arunakumaren et al. 2000). The main recharge processes are infiltration of rainfall and irrigation water through an unsaturated zone of less than 15 m in depth, lateral discharge of river water during high flow periods, channel seepage, percolation through artificial recharge pits, and overbank flooding. Water loss from the aquifer is through groundwater discharge to the sea, wetlands, and local tributaries as well as extraction for irrigation and evapotranspiration. 526
Figure 1. Geographic setting of the lower Burdekin in tropical north Queensland, Australia. Shaded areas depict current sugarcane growing areas. Circles represent sampling locations where elevated DOC (>10 mg/L) was detected in groundwater in January 2004. Circle size is based on DOC concentration (labeled). Dotted lines highlight the locations of coarse-grained palaeochannel deposits in the floodplain.
Sampling and Analysis Water samples were collected from groundwater monitoring bores across the lower Burdekin floodplain at three different times: January 2004, October 2004, and January 2005. In January 2004, 22 single monitoring bore sites and 5 nested sites were sampled. Fifteen of these monitoring bores were sampled again in October 2004 and January 2005 and it is the data from these 15 bores that are reported in this paper. Monitoring bore depths ranged from 5.4 to 88.5 m and are typically located in high-yielding sandy units. Water samples were withdrawn using a stainless steel submersible pump with a Teflon discharge line connected to a flow cell where electrical conductivity (EC), pH, redox potential (Eh), and dissolved oxygen (DO) were continuously monitored. Bores were purged until either three bore volumes had been extracted or until the EC, pH, Eh, and DO had stabilized. DO measurements were made using a DO probe CellOx 325 (WTW GmbH & Co. KG, Weilheim, Germany) combined with a WTW Multiline P4 Universal meter (WTW GmbH & Co. KG, Weilheim, Germany). All the meters were calibrated in the field before making measurements. Subsamples were collected from each bore for Fe2+ , Mn2+ , and NO3 − analysis after filtering through a 0.45-μm filter. For Fe2+
T. Thayalakumaran et al. Groundwater 53, no. 4: 525–530
NGWA.org
and Mn2+ analysis, water samples were collected in vials containing 4.0 mL of 1 M sodium acetate and 2.5 mL of 0.1% phenanthroline. Subsamples for DOC analysis were collected in glass bottles with no head space after filtering with 0.45-mm glass-microfiber filters and acidified. All the water samples were immediately chilled to 4 ◦ C and transported to the laboratory. The analysis was performed within 48 h of collection. Nitrate, Fe2+ , and Mn2+ were determined via colourimetry and DOC was analyzed using the Shimadzu Automatic carbon analyser (Shimadzu Corporation, Tokyo, Japan). Statistical analysis was performed using Genstat 14th edition.
(a)
Results and Discussion The DOC concentrations were highest in January 2004 with more than half of the 46 samples having concentrations greater than 30 mg/L (Figure 1). The maximum concentration recorded (82 mg/L) is, to the best of our knowledge, the highest measured DOC value recorded in groundwater beneath irrigated cropping systems. The DOC concentrations declined through the two additional sampling periods with 14 of the 15 locations showing greater than 90% decline (Figure 2a). Although most of the temporal trends in DOC depicted in Figure 2a represent groundwater samples from less than 20-m depth, similar declining trends were also observed at depths of 35 and 52 m (data not shown). Possible sources of DOC are related to agricultural practices such as application of sugarcane by-products to amend soils and/or sugarcane burning and harvesting and consequent release of sugarcane sap to the surface soil. Sugarcane by-products known as filter mud and boiler ash are applied at rates of 100 to 200 wet tons/ha and contain 8 to 10% (w/v) organic carbon (Barry et al. 2001). Extremely high levels of DOC, up to 260 mg/L in surface runoff following sugarcane harvest have been reported in this area (Bohl et al. 2002). This suggests that the surface applied organic matter may account for, or at least contribute to, the elevated DOC in shallow groundwater. In order to better understand the mechanisms and timeframes of DOC delivery from the surface to the aquifer, CFC-12 data presented by Thayalakumaran et al. (2008) were reconsidered. CFC-12 modeled ages for groundwater sampled from less than 20 m below the water table increase linearly with depth and indicate a vertical flow velocity of 0.28 m/year (Thayalakumaran et al. 2008). For this study, we adopted the same assumptions of piston flow transport and that the CFC-12 concentration has not been affected by chemical processes following recharge. Fielding et al. (2006) reported that the upper 30 m of large parts of the aquifer is comprised of medium to coarse-grained alluvial and deltaic deposits and can be treated as largely unconfined. Assuming an average porosity of 35%, the vertical velocity corresponds to a recharge rate of 0.1 m/year. With this information the date of groundwater recharge, or recharge year, can be estimated from the sampling depth below the water table NGWA.org
(b)
Figure 2. Temporal variation in (a) DOC and (b) Fe2+ in groundwater from 15 monitoring bore locations.
for groundwater having elevated DOC (as measured in January 2004). Figure 3 displays the relationship between historical rainfall data and the estimated recharge year. The estimate of the recharge year for groundwater containing elevated DOC in January 2004 is rudimentary, but the data does suggest that elevated DOC concentrations correspond to periods of high annual rainfall, or immediately thereafter (Figure 3). Conversely, low DOC concentrations correspond to periods of low annual rainfall. These observations support the hypothesis that the DOC is sourced at or near the surface and supplied to the aquifer via vertical recharge following above average rainfall. This finding is consistent with the reported positive correlations between annual rainfall and DOC delivery to shallow groundwater systems throughout the United States (Chapelle et al. 2012). Furthermore, higher DOC concentrations have been observed below the root zone of soils having greater water fluxes (Kalbitz et al. 2000; Kaiser and Guggenberger 2005; Mertens et al. 2007; Don and Schulze 2008). The spatial variability observed in Figure 1 could be
T. Thayalakumaran et al. Groundwater 53, no. 4: 525–530
527
Figure 3. Relationship between DOC concentrations in shallow groundwater (20 mg/L) have been measured in shallow groundwater (Thorburn et al. 2003). However, during January 2004, DO and NO3 − were low (
Abstract Elevated dissolved organic carbon (DOC) has been detected in groundwater beneath irrigated sugarcane on the Burdekin coastal plain of tropical northeast Australia. The maximum value of 82 mg/L is to our knowledge the highest DOC reported for groundwater beneath irrigated cropping systems. More than half of the groundwater sampled in January 2004 (n = 46) exhibited DOC concentrations greater than 30 mg/L. DOC was progressively lower in October 2004 and January 2005, with a total decrease greater than 90% indicating varying load(s) to the aquifer. It was hypothesized that the elevated DOC found in this groundwater system is sourced at or near the soil surface and supplied to the aquifer via vertical recharge following above average rainfall. Possible sources of DOC include organic-rich sugar mill by-products applied as fertilizer and/or sugarcane sap released during harvest. CFC-12 vertical flow rates supported the hypothesis that elevated DOC (>40 mg/L) in the groundwater results from recharge events in which annual precipitation exceeds 1500 mm/year (average = 960 mm/year). Occurrence of elevated DOC concentrations, absence of electron acceptors (O2 and NO3 – ) and both Fe2+ and Mn2+ greater than 1 mg/L in shallow groundwater suggest that the DOC compounds are chemically labile. The consequence of high concentrations of labile DOC may be positive (e.g., denitrification) or negative (e.g., enhanced metal mobility and biofouling), and highlights the need to account for a wider range of water quality parameters when considering the impacts of land use on the ecology of receiving waters and/or suitability of groundwater for irrigated agriculture.
Introduction The chemical composition of groundwater is a measure of its quality and hence suitability as a source of water for human consumption, irrigation, and industry. Groundwater quality also influences ecosystem health and function, so it is important to be aware of and avoid the degradation of groundwater resources. In agricultural regions, considerable research has been carried out on groundwater quality problems related to salinity, nutrients, herbicides, and pesticides (Bohlke 2002; Di and Cameron 2002; Scanlon et al. 2007; Arias-Estevez et al. 2008; Lazorka-Connon and Achari 2009). Far less attention has been focused on the effects of agricultural practices on 1 Cooperative Research Centre for Irrigation Futures and CSIRO Water for a Healthy Country National Research Flagship, CSIRO Land and Water, PMB Aitkenvale, Townsville, Queensland 4814, Australia; [email protected] 2 Currently with Department of Environment and Primary Industries—Agriculture Research, 32 Lincoln Square Nth, Carlton, Vic 3053, Australia; [email protected] 3 Currently with AECOM, PO Box 5423, Townsville, Queensland 4810, Australia; [email protected] 4 Corresponding author: Cooperative Research Centre for Irrigation Futures and CSIRO Water for a Healthy Country National Research Flagship, CSIRO Land and Water, PMB Aitkenvale, Townsville, Queensland 4814, Australia. [email protected] Received February 2014, accepted July 2014. © 2014, Commonwealth of Australia. Groundwater © 2014, National Ground Water Association doi: 10.1111/gwat.12258
NGWA.org
dissolved organic carbon (DOC) leaching from soils and subsequent transport to the water table (Herbauts 1980; Ronen et al. 1987). DOC in groundwater can lead to several microbial and geochemical processes that affect the redox state of the aquifer system(s). Some of these processes, such as nitrate reduction, may be perceived as beneficial to the environment whereas those that affect oxygen availability and the mobility of metals and organic contaminants (Aravena and Wassenaar 1993; Dudal and Gerard 2004; Chapelle et al. 2012) are considered detrimental. Availability of surface DOC in connected surface water-groundwater systems may impact downstream ecosystems by reducing the availability of oxygen (Moore 2010) or facilitating the delivery of toxicants including mercury (Lamborg et al. 2004; Bone et al. 2007). Reported DOC concentrations in uncontaminated groundwater beneath cropped agricultural systems are less than 1 to 5 mg/L (Starr and Gillham 1993; Chomycia et al. 2008; Rivett et al. 2008), but specific agricultural and animal management practices can result in higher levels of DOC (up to 55 mg/L; Chomycia et al. 2008). Environmental and hydrologic conditions including intensive irrigation, high rainfall events, and high permeability soils enhance the transport of surface available DOC to shallow aquifer systems via various pathways (Kalbitz et al. 2000; Chapelle et al. 2013). In groundwater systems beneath an intensively irrigated sugarcane area, Thayalakumaran et al. (2008)
Vol. 53, No. 4–Groundwater–July-August 2015 (pages 525–530)
525
reported unusually high groundwater DOC concentrations and its widespread geographical occurrence. Thayalakumaran et al. (2008), however, found no specific spatial trend in the groundwater DOC concentrations, and the source of the DOC and delivery mechanisms to the groundwater was unknown. As agricultural management and hydraulic conditions follow seasonal patterns, it is likely that DOC loading to shallow groundwater will follow similar patterns if the DOC is sourced at or near the soil surface. We hypothesize that the DOC found in this groundwater system was sourced from the soil surface, transported to the groundwater via vertical recharge following years with particularly high levels of rainfall and that the DOC has a significant influence on the geochemistry of the shallow portions of the aquifer. The hypothesis was tested by (1) collecting groundwater samples across the lower Burdekin at three different times for analysis of DOC and various other constituents including DO, NO3 - , Fe2+ , and Mn2+ , and (2) using CFC-12 vertical flow rates and rainfall data to analyse the relationship between groundwater recharge and DOC concentrations in the groundwater.
Methods Study area The lower Burdekin (Figure 1) is one of Australia’s premier irrigation areas with a reputation for producing some of the highest yields and highest quality sugarcane in Australia (Bristow 2004). A detailed description of the area has been presented by Thayalakumaran et al. (2008), and catchment characteristics relevant to this study are summarized here. The lower Burdekin has a tropical climate characterized by hot wet summers (December to March) and dry winters (June to September). Mean annual rainfall for the region is 960 mm. Water for irrigation is sourced from the floodplain aquifer and/or Burdekin River. The aquifer typically provides in excess of 250,000 ML/annum for sugarcane irrigation. Sugarcane is harvested between June and December every year. The floodplain aquifer system comprises unconsolidated, alluvial, and marine sedimentary deposits that overlie a granitic basement. Numerous coarse-grained palaeochannels dissect the current floodplain (Alexander and Fielding 2006) and have been identified as important groundwater recharge areas (Lenahan and Bristow 2010). The presence of sandy to loamy soils and the discontinuous nature of the underlying clay layers means that the groundwater system is generally unconfined (Arunakumaren et al. 2000). The main recharge processes are infiltration of rainfall and irrigation water through an unsaturated zone of less than 15 m in depth, lateral discharge of river water during high flow periods, channel seepage, percolation through artificial recharge pits, and overbank flooding. Water loss from the aquifer is through groundwater discharge to the sea, wetlands, and local tributaries as well as extraction for irrigation and evapotranspiration. 526
Figure 1. Geographic setting of the lower Burdekin in tropical north Queensland, Australia. Shaded areas depict current sugarcane growing areas. Circles represent sampling locations where elevated DOC (>10 mg/L) was detected in groundwater in January 2004. Circle size is based on DOC concentration (labeled). Dotted lines highlight the locations of coarse-grained palaeochannel deposits in the floodplain.
Sampling and Analysis Water samples were collected from groundwater monitoring bores across the lower Burdekin floodplain at three different times: January 2004, October 2004, and January 2005. In January 2004, 22 single monitoring bore sites and 5 nested sites were sampled. Fifteen of these monitoring bores were sampled again in October 2004 and January 2005 and it is the data from these 15 bores that are reported in this paper. Monitoring bore depths ranged from 5.4 to 88.5 m and are typically located in high-yielding sandy units. Water samples were withdrawn using a stainless steel submersible pump with a Teflon discharge line connected to a flow cell where electrical conductivity (EC), pH, redox potential (Eh), and dissolved oxygen (DO) were continuously monitored. Bores were purged until either three bore volumes had been extracted or until the EC, pH, Eh, and DO had stabilized. DO measurements were made using a DO probe CellOx 325 (WTW GmbH & Co. KG, Weilheim, Germany) combined with a WTW Multiline P4 Universal meter (WTW GmbH & Co. KG, Weilheim, Germany). All the meters were calibrated in the field before making measurements. Subsamples were collected from each bore for Fe2+ , Mn2+ , and NO3 − analysis after filtering through a 0.45-μm filter. For Fe2+
T. Thayalakumaran et al. Groundwater 53, no. 4: 525–530
NGWA.org
and Mn2+ analysis, water samples were collected in vials containing 4.0 mL of 1 M sodium acetate and 2.5 mL of 0.1% phenanthroline. Subsamples for DOC analysis were collected in glass bottles with no head space after filtering with 0.45-mm glass-microfiber filters and acidified. All the water samples were immediately chilled to 4 ◦ C and transported to the laboratory. The analysis was performed within 48 h of collection. Nitrate, Fe2+ , and Mn2+ were determined via colourimetry and DOC was analyzed using the Shimadzu Automatic carbon analyser (Shimadzu Corporation, Tokyo, Japan). Statistical analysis was performed using Genstat 14th edition.
(a)
Results and Discussion The DOC concentrations were highest in January 2004 with more than half of the 46 samples having concentrations greater than 30 mg/L (Figure 1). The maximum concentration recorded (82 mg/L) is, to the best of our knowledge, the highest measured DOC value recorded in groundwater beneath irrigated cropping systems. The DOC concentrations declined through the two additional sampling periods with 14 of the 15 locations showing greater than 90% decline (Figure 2a). Although most of the temporal trends in DOC depicted in Figure 2a represent groundwater samples from less than 20-m depth, similar declining trends were also observed at depths of 35 and 52 m (data not shown). Possible sources of DOC are related to agricultural practices such as application of sugarcane by-products to amend soils and/or sugarcane burning and harvesting and consequent release of sugarcane sap to the surface soil. Sugarcane by-products known as filter mud and boiler ash are applied at rates of 100 to 200 wet tons/ha and contain 8 to 10% (w/v) organic carbon (Barry et al. 2001). Extremely high levels of DOC, up to 260 mg/L in surface runoff following sugarcane harvest have been reported in this area (Bohl et al. 2002). This suggests that the surface applied organic matter may account for, or at least contribute to, the elevated DOC in shallow groundwater. In order to better understand the mechanisms and timeframes of DOC delivery from the surface to the aquifer, CFC-12 data presented by Thayalakumaran et al. (2008) were reconsidered. CFC-12 modeled ages for groundwater sampled from less than 20 m below the water table increase linearly with depth and indicate a vertical flow velocity of 0.28 m/year (Thayalakumaran et al. 2008). For this study, we adopted the same assumptions of piston flow transport and that the CFC-12 concentration has not been affected by chemical processes following recharge. Fielding et al. (2006) reported that the upper 30 m of large parts of the aquifer is comprised of medium to coarse-grained alluvial and deltaic deposits and can be treated as largely unconfined. Assuming an average porosity of 35%, the vertical velocity corresponds to a recharge rate of 0.1 m/year. With this information the date of groundwater recharge, or recharge year, can be estimated from the sampling depth below the water table NGWA.org
(b)
Figure 2. Temporal variation in (a) DOC and (b) Fe2+ in groundwater from 15 monitoring bore locations.
for groundwater having elevated DOC (as measured in January 2004). Figure 3 displays the relationship between historical rainfall data and the estimated recharge year. The estimate of the recharge year for groundwater containing elevated DOC in January 2004 is rudimentary, but the data does suggest that elevated DOC concentrations correspond to periods of high annual rainfall, or immediately thereafter (Figure 3). Conversely, low DOC concentrations correspond to periods of low annual rainfall. These observations support the hypothesis that the DOC is sourced at or near the surface and supplied to the aquifer via vertical recharge following above average rainfall. This finding is consistent with the reported positive correlations between annual rainfall and DOC delivery to shallow groundwater systems throughout the United States (Chapelle et al. 2012). Furthermore, higher DOC concentrations have been observed below the root zone of soils having greater water fluxes (Kalbitz et al. 2000; Kaiser and Guggenberger 2005; Mertens et al. 2007; Don and Schulze 2008). The spatial variability observed in Figure 1 could be
T. Thayalakumaran et al. Groundwater 53, no. 4: 525–530
527
Figure 3. Relationship between DOC concentrations in shallow groundwater (20 mg/L) have been measured in shallow groundwater (Thorburn et al. 2003). However, during January 2004, DO and NO3 − were low (
