『Abstract
Groundwater contamination by nitrate was investigated in an agricultural
area in southern Quebec, Canada, where a municipal well is the
local source of drinking water. A network of 38 piezometers was
installed within the capture zone of the municipal well to monitor
water table levels and nitrate concentrations in the aquifer.
Nitrate concentrations were also measured in the municipal well.
A Water flow and Nitrate transport Global Model (WNGM) was developed
to simulate the impact of agricultural activities on nitrate concentrations
in both the aquifer and municipal well. The WNGM first uses the
Agriflux model to simulate vertical water and nitrate fluxes below
the root zone for each of the seventy agricultural fields located
within the capture zone of the municipal well. The WNGM then uses
the HydroGeoSphere model to simulate three-dimensional variably-saturated
groundwater flow and nitrate transport in the aquifer using water
and nitrate fluxes computed with the Agriflux model as the top
boundary conditions. The WNGM model was calibrated by reproducing
water levels measured from 2005 to 2007 in the network of piezometers
and nitrate concentrations measured in the municipal well from
1997 to 2997. The nitrate concentrations measured in the network
of piezometers, however, showed greater variability than in the
municipal well and could not be reproduced by the calibrated model.
After calibration, the model was validated by successfully reproducing
the decrease of nitrate concentrations observed in the municipal
well in 2006 and 2007. Although it cannot predict nitrate concentrations
in individual piezometers, the calibrated and validated WNGM can
be used to assess the impact of changes in agricultural practices
on global nitrate concentrations in the aquifer and in the municipal
well.
Keywords: Agricultural pollution; Modeling; nitrate contamination;
Groundwater; Capture zone; Municipal well』
1. Introduction
2. Site description
3. Equipment and material
4. Modeling
4.1. Agriflux
4.1.1. General structure
4.1.2. Variables and parameters for Agriflux simulations
4.2. HydroGeoSphere
4.2.1. General structure
4.2.2. Variables and parameters of HydroGeoSphere
4.2.2.1. Hydraulic conductivities and porosity
4.2.2.2. Nitrate and water fluxes
4.3. Run for simulations
5. Model calibration
5.1. Sensitivity analysis (Steps 1 and 4)
5.1.1. Sensitivity analysis for hydraulic calibration (Step
1)
5.1.2. Sensitivity analysis for nitrate calibration (Step 4)
5.2. History matching (Steps 2 and 5)
5.2.1. hydraulic history matching (Step 2)
5.2.2. Nitrate transport history matching (Step 5)
5.3. Model validation (Steps 3 and 6)
5.3.1. Flow validation (Step 3)
5.3.2. Nitrate transport validation (Step 6)
6. Results and discussion
6.1. Results of the sensitivity analysis (Steps 1 and 4)
6.1.1. Sensitivity analysis for hydraulic head (Step 1)
6.1.2. Sensitivity analysis for nitrate transport (Step 4)
6.2. Results of history matching (Steps 2 ad 5)
6.2.1. History matching for hydraulic head (Step 2)
6.2.2. History matching for nitrate transport (Step 5)
6.3. Results of model validation
6.3.1. Validation for hydraulic head (Step 3)
6.3.2. Validation for nitrate transport (Step 6)
6.4. Global (capture zone) versus local (piezometer) model predictions
7. Conclusion
Acknowledgments
References