Geochemical Indicators and Seasonality Controls to Arsenic Mobility in Groundwater in Langley, British Columbia
In this study, geochemical indicators are used to determine risk of arsenic mobilization in groundwater for an aquifer system in a temperate climate region. The study area is the township of Langley, BC, where arsenic in some well waters has been reported to be above the Canadian drinking water guideline (0.010 mg/l). Arsenic typically becomes mobile in reducing groundwaters and alkaline oxidizing groundwaters. Due to its mobility behavior, arsenic concentrations have been observed to correlate with anions and oxianions species in oxidizing waters, and to correlate with redox sensitive species in reducing waters. Historical groundwater geochemistry datasets available for the study area have variable data completeness and quality. Species observed to correlate with arsenic concentrations are queried in these datasets and used as indicators to produce an arsenic mobility risk map. A multi-proxy approach is used where a region is only considered to have high arsenic mobility potential if more than one indicator is observed. Preliminary data analyses show a positive correlation between arsenic and pH, with the highest arsenic concentrations occurring at a pH above 8. Geochemical modeling using PHREEQC demonstrates saturation with respect to hematite and goethite and undersaturation with respect to pyrite. No correlation is observed between arsenic and redox sensitive species, such as Fe, Mn, SO4 and NH4, while weak correlations were found between arsenic and pH sensitive species, as HCO3, Mo, B and F. These observations suggest that the groundwaters are oxidizing and that arsenic mobility is likely being controlled by desorption from solids as a result of groundwater alkaline conditions. Temporal variations in arsenic concentrations and associated species are examined with groundwater level variations to address seasonality controls to arsenic mobility and to identify potential long term trends. The results presented and the geochemical indicators used will be tested through extensive well water sampling and geochemical analyses starting in the summer of 2009.
Integrated characterisation of aquifer heterogeneity and landfill leachate plume migration
The understanding of groundwater flow and contaminant migration is based on our ability to characterize aquifers and represent these processes with numerical simulators. This understanding is required to efficiently remediate contaminated sites since the failure of remediation actions are often related to an insufficient understanding of aquifer heterogeneity. During the last decades, continuous development of numerical simulators allowed models to better represent complex flow systems. However, conventional hydrogeological characterization methods do not provide the data required to define aquifer heterogeneity. An original hydrogeological characterization approach was used to define aquifer heterogeneity and delineate landfill leachate plumes through the use and integration of varied techniques. The objective of the study is to develop a methodology to integrate hydrogeological, geophysical and geochemical data using geostatistical tools. The characterization program aims to better characterize the aquifer, delineate leachate plumes emitted by a former landfill, and guide a study of the natural attenuation of the plumes. The initial phase of the integrated multidisciplinary aquifer characterization program was carried out in a 12 km2 area of the sub-watershed surrounding the landfill of St-Lambert-de-Lauzon, Québec. In the study area, a 10-m thick sandy unconfined aquifer overlies clayey silt and till layers. In this relatively flat area, natural streams as well as agricultural and forestry drainage networks control groundwater flow. The first phase of the project focused on a regional hydrogeological and geochemical characterization where 5 field methods were combined: 1) surface geophysics (ground penetrating radar and electrical tomography) (GPR); 2) direct-push methods including a) cone penetration tests (CPT), b) soil sampling and c) installation of full- screened observation wells; 3) multilevel measurement of geochemical parameters and groundwater sampling with packers; 4) borehole geophysical logging; 5) high resolution hydraulic tests. The different types of data were integrated with multivariate geostatistical analysis and the results showed complex aquifer conditions. The aquifer base exhibits large topographic variations and semi-confined conditions seem to be present in certain locations. These conditions have a significant influence on groundwater flow and leachate migration. The geostatistical interpretation of multilevel geochemical parameters, combined with CPT data, provides a definition of groundwater geochemical spatial variability and indicates the likely extent of landfill leachate. This detailed knowledge of the aquifer serves as a base for the initial development of a numerical model considering heterogeneity and guides further characterization of the aquifer and plume. Keywords: characterization, heterogeneity, direct push, surface geophysics, numerical model, landfill leachate, natural attenuation.
Statistical Analysis of Road Salt Impact on Water Quality in New England Watersheds
Elevated concentrations of chemical components from winter de-icers in surface waters had been observed in many watersheds in New England since the 1970's. Chemical water quality data (unpublished and NAWQA- USGS) in combination with real time specific conductance monitoring records were statistically analyzed to establish correlations of the impact of de-icers by multivariate statistical analysis with other dissolved chemical components. Principal component analysis (PCA), factor analysis (FA), and chloride proxy time records (past five years) from selected watersheds illustrate a pattern characteristic across the entire southern New England region. Eight New England NAWQA watersheds in the southern segment show that the chemical data variations in these watersheds are constrained by only 4 independent variables from both PCA and FA analysis (80% of variance). Two of the most important statistical factors can be identified as (1) standard base flow mixed with NaCl de-icer in a uniform proportion, and (2) direct runoff with Na and Cl as the only chemical components. A similar trend is also observed on the multiyear Cl records derived from the specific conductance (SPC) records by calibrating SPC with chemically analyzed water samples. The records show winter peaks associated with local de-icer applications during snowstorms superimposed on an elevated uniform chloride trend throughout the entire year. We interpret this trend as base flow discharging from a uniformly contaminated reservoir of ground water mixed with the percolating denser chloride solutions derived during snow melting periods. This trend stays constant even during the summer precipitation events that are seen as temporarily diluting the base flow and then, as the precipitation event passes, returning to the more concentrated levels.
Inferring Watershed Characteristics Using Records of Multi-decade Stream Chemistry Response to Road Salt Application
There are many observations of increasing chloride (Cl-) concentrations in streams with road salt applied in their watersheds. However, multi-decade stream Cl- data sets of seven watersheds in the northern U.S. show that some streams with road salt application have little increase in stream Cl-. Relative to streams with increasing Cl-, streams with no Cl- increase have more extensive glacial sand and gravel deposits in their watersheds. We speculate that these sand and gravel deposits provide a large dilution volume that creates a lag between solute inputs and outputs over a comparable time scale as the multi-decade Cl- monitoring record. We also show that systems with little increase in Cl- concentrations have relatively high baseflow, as deduced from a comparison of flow-duration curves. This work provides a link among solute transport behavior, physical watershed characteristics, and flow-duration curves, perhaps providing insights to improve predictions of solute retention and transport when no direct stream chemistry data are available.
Distributed Denitrification in a Northeastern Agricultural Landscape
Denitrification may be an important sink of anthropogenic nitrogen (N) in eastern US watersheds. Denitrification occurs primarily under anaerobic conditions by heterotrophic microbes, and is therefore expected to be vigorous in wet soils containing high amounts of organic carbon. Actual rates of denitrification, however, have been difficult to quantify, and remain one of the critical unresolved N processes at the landscape scale. We measured denitrification rates in situ along hydrologic flow paths and across gradients of hydroperiodicities, i.e., frequencies and durations of saturated conditions, at Cornell University's Teaching and Research Center in Harford, NY (an active dairy farm). Denitrification rates were measured monthly using the ¬15N push-pull method from 14 mini-piezometers arrayed along a gradient of hydroperiodicity as indicated by a soil topographic index (STI). Measured rates of denitrification were spatially variable across sites and ranged from undetectable to over 200 µg N/kg soil/day with a mean of 55.9 ± 16.4 µg N/kg soil/day. Mean rates of denitrification increased with STI, which ranged from 10 to 23. This relationship was used to estimate distributed denitrification rates across the landscape and resolve a missing piece of the N budget for the farm. We found that 16% of the farm fell into areas of STI greater than 10. Using the distributed denitrification rates, this area accounts for 15-27% of the missing N balance for the farm (9.7-17.8 Mg N/yr). Improved understanding of the distribution and magnitudes of denitrification in agricultural landscapes has good potential to facilitate new, novel, and better management practices for controlling nitrogen loading to streams and rivers. Indeed, the very areas that appear to have a propensity to harbor denitrification, i.e., areas prone to be wet, are often artificially drained as part of standard agricultural practices which effectively increase N loading to rivers and contributes to downstream eutrophication. This practice not only reduces the frequency that these areas are likely to be anaerobic, but it constitutes a rapid N transport pathway between the landscape and streams.
Predicting nitrate concentrations of groundwater from land uses in Prince Edward Island, Canada
Elevated nitrate levels in groundwater are of growing concern for drinking water protection in Prince Edward Island (PEI). Elevated nitrate levels were identified being associated with the intensity of potato production. This study examined the relationship between land uses and nitrate levels in groundwater in PEI. A land use analysis was performed for each watershed (totaling 52) on PEI. Land uses in each watershed were categorized as A) agriculture and B) non-agriculture, and acreage for each category was calculated based on data of 1990. Category A was further subdivided into A1) land in potato production rotation and A2) pasture/grass based on data of period 1996-2000. Rotation lengths for Category A1 were also computed using GIS. LEACHM-N simulations, tile drain measurements, field N budgets and groundwater modeling were employed to define annual nitrate leaching to groundwater for each land use category, which was used to develop acreage-weighted nitrate leaching concentration of a watershed. With an assumption that the acreage- weighted nitrate concentration equal to mean nitrate concentration of well water, percentage of well water samples with nitrate level > 10 mg N/l was predicted based on a normal distribution derived from a statistics of well water samples. Mean nitrate concentration of well water in each watershed was calculated based on regular sample submissions to PEI Department of Environment during 2004-2008 (9512 samples) and compared against the acreage-weighted nitrate concentration based on current land use practice (assumed as the same of the year 20000). Acreage-weighted nitrate concentrations for a scenario of potato land to be adjusted to 3-yr rotation with a reduction of N rate on potato crop were used to predict the potential effects on nitrate reductions. The study showed mean nitrate level of well water appeared to be positively correlated with percentage of land used for potato production rotation in a watershed. Predicted acreage-weighted nitrate concentrations based on current land use were close to the statistics of well water samples. The study also predicted continuing current land use practices would lead to 23% (12/52) of watersheds in PEI having average nitrate of well water above 5 mg/l (corresponding to 5% of wells with nitrate above 10 mg N/l); the combination of adjusting current rotation length (~2.5 yrs) to a 3-yr rotation and reducing fertilizer N on potato crops from ~200 kg N/ha to 150 kg N/ha would only reduce the number of watersheds with average nitrate of well water above 5 mg N/l to 15% (8/52), suggesting further efforts on land use changes in some watersheds be required to restore their nitrate levels below 5 mg N/l.
Constraining Nitrogen, Phosphorus, and Carbon Exports in a Midwestern Agricultural Watershed
Export of nitrogen (N), phosphorus (P) and carbon (C) from agro-ecosystems of the U.S. Midwest significantly contributes to non-point source pollution in the Mississippi River drainage basin and to the development of the hypoxic zone in the Gulf of Mexico. Most NPC exports to streams occur during storm events and NPC concentrations in streams affect ecosystem productivity and biogeochemical processes at the watershed scale. The proposed research examines the flow pathways regulating nitrate, ammonium, dissolved and particulate organic N, P and C, and soluble reactive phosphorus in artificially drained landscapes of the US Midwest. These pathways include overland flow, matrix flow and macropore flow. Research will take place in Leary Weber Ditch (7.6 km2), IN, and water samples will be collected during storms at a high temporal resolution in overland flow, tile flow and the outlet of the watershed. NPC concentrations and standard hydrograph separations techniques using oxygen-18 isotopes of water measured throughout the watershed in precipitation, groundwater, tile flow, overland flow and the stream will inform us on the relative importance of major transport processes regulating NPC exports at the watershed scale during storms. A total of 12 storms corresponding to varying degrees of rainfall intensity and duration will be sampled from Spring 2009 through Fall 2010. It is hypothesized that although subsurface tile drainage may represent the majority of flow into the stream, overland flow might significantly affect water quality with respect to particulate N, P and C. Conversely, it is expected that stream water chemistry at the watershed outlet will likely reflect subsurface drain water chemistry with respect to nitrate and soluble reactive phosphorus. This work will be the first study to comprehensively address the relative importance of major hydrological export pathways at regulating stream water chemistry with respect to NPC in artificially drained landscapes of the U.S. Midwest and will help solute export model development for this part of the country and other artificially drained landscapes around the world.
Hydrobiogeochemical Influences on N2O Dynamics in an Agriculturally Impacted Riparian Wetland
Riparian zones (RZ) are known to act as buffers, reducing the transfer of potentially harmful nutrients from agricultural fields to surface water bodies. However, many of the same processes in the subsurface that help to reduce this nutrient loading, may also be leading to greenhouse gas (GHG) production and emissions from these areas. Agricultural riparian zones in Southern Ontario are often characterized by a sloped topography, with the highest topographic position being closest to the field edge, decreasing towards an adjacent stream or other surface water body. This topographic variability, combined with lateral chemical inputs from both upland areas and the stream, is expected to cause variable hydrochemical environments throughout the RZ, which may therefore lead to variable N2O dynamics between upland, mid-riparian and lowland areas. The objectives of this study were to examine these spatial trends in N2O production and resulting emissions, as related to the hydrochemical environment in these three distinct zones. Objectives were achieved by instrumenting 6 sites across two transects running perpendicular from the agricultural field edge, towards the stream edge, analyzing for subsurface N2O, moisture and temperature, groundwater NO3, NH4, dissolved organic carbon (DOC), dissolved oxygen, and surface fluxes of N2O. Subsurface N2O concentrations and ground water nutrient concentrations displayed distinct spatial trends in the three positions across the RZ, however N2O fluxes across the soil-atmosphere interface did not display consistent spatial trends. There is a disconnect between the subsurface and the fluxes at the surface, in that N2O fluxes do not reflect the N2O concentrations produced in the shallow soil profile, nor were the fluxes significantly related to the geochemical environment at each position. The lack of visible spatial trends in N2O fluxes may be due to an "oxic blanket" effect which may divide the surface from the subsurface soil profile. As N2O fluxes in this study were within the range observed at other, similar study sites, the oxic blanket doesn't appear to impede the movement of N2O across the soil-atmosphere interface. This may suggest that the N2O being released as a flux is being produced in the very shallow soil profile (0 - 6 cm). Subsurface concentrations of N2O were fairly high at times, which was not reflected in the fluxes. This may be a result of nitrifier denitrification reducing N2O to N2 before it reaches the surface. Aerobic denitrification may also be the cause of reduced N2O at the surface. Both of these can occur in oxygenated environments, but neither are well understood. The missing connection between subsurface N2O concentrations, ground water nutrients, and the surface fluxes was not a hypothesized result, and requires further research and analysis for a better understanding of the production and consequent movement of N2O.
Hydrochemical Characteristics of Wadi Systems in Arid and Semi-arid Regions in Southwest Saudi Arabia
Groundwater in the alluvial systems of Wadi Baysh and Wadi Habawnah has been the subject of geochemical and water quality investigations of this study. Water samples were collected from 62 and 20 hand dug wells in Wadi Habawnah and Wadi Baysh respectively. The groundwater sample concentrations of the major ions (Ca2+, Mg2+, Na+, K+, CO32-, HCO3-, SO42- and Cl-) and some trace elements (Fe2+, Mn2+, Ni2+, Pb2+ and Al3+) were determined. Graphical methods for interpreting the chemical data from the well fields in both wadis were adopted such as X_Y plots and the Piper diagram. The relationship between (Ca2++Mg2+) and (HCO3-+SO42-) shows a strong correlation (r = 0.9) in both wadis. This indicates that simple dissolution and mixing are the dominant processes. The Piper trilinear diagrams for Wadi Baysh and Wadi Habawnah show that the chemical compositions of waters from these two wadis are different. Generally, waters from wadi Baysh are characterized by a Ca-Na-SO4-Cl type, and waters from Wadi Habawnah are Ca-Mg-SO4-Cl. Cluster analysis was also used to further investigate the waters from both wadis and the results showed that groundwater mineralization of both wadis tend to increase in a down slope direction, which shows the influence of the aquifer lithology and the evaporation effects. All of the TDS analyses were above the EEC limit in the lower parts of both wadis. The sodium hazard is low to medium in both wadis The SAR of Wadi Baysh waters range between less than 1 and 7 and the SAR of Wadi Habawnah waters range between 2 and 6. Medium salinity water can be used for irrigation in most instances without special practices for salinity control.