Identifying nitrate-vulnerable zones for surface water pollution in Flanders based on the depth of the redoxcline and aquifer thickness

As in other areas with intensive agriculture in Europe, Flanders struggles with bringing surface water quality in line with the EU Nitrates Directive. At about 25% of the 874 surface water measurement locations in Flanders monitored by the Flanders Environment Agency (VMM), the 90^th percentile from monthly measurements conducted during 2018-2023 exceeded a NO₃^- concentration of 50 mg/L.

A large fraction of the nitrate that leaches out of the root zone is denitrified in the groundwater. However the denitrification rate varies spatially, so mitigation measures are best targeted to zones from where NO₃^- is transported to the surface water without undergoing significant denitrification in the aquifer. We used a novel methodology to predict NO₃^- concentrations in the outlets of small catchments that explicitly considers hydrochemical variation within aquifers and is based on the thickness of oxidized and reduced zones in an aquifer.

The depth of the redoxcline, the boundary between the oxidized and reduced zone, was determined from the phreatic groundwater monitoring network of VMM consisting of 2089 multilevel groundwater wells in Flanders. Analysis of the time series of hydro-chemical data (redox potential and dissolved NO₃^-, O₂, Fe, Mn) allowed us to classify the filters as being in oxidized or reduced zone. For each well, the depth of the first ‘reduced’ filter was taken as the depth of the redoxcline.

Assuming a spatially uniform nitrate concentration in the groundwater recharge, the nitrate concentration of the water reaching the catchment outlet can be estimated as:

NO₃^- at catchment outlet = (thickness of oxidized zone / equivalent aquifer thickness ) × NO₃^-   in recharge water        (1)

This simple approach assumes that all nitrate is denitrified once the groundwater flowline crosses the redoxcline. Instead of the true aquifer depth, we used the aquifer's equivalent thickness, utilizing the Hooghoudt equation based on the average distance between watercourses. The nitrate input in the recharge water for this calculation was taken from VMM’s NEMO model, which provides the nitrate leachate from agricultural fields to groundwater.

Our investigation covered 68 small agricultural catchments (0.4-20.4 km²). The ratio of oxidized to entire aquifer thickness varied from 0.03 to 1, averaging 0.33. Therefore, on average, 33% of the agricultural areas would show high nitrate vulnerabilities because flowlines originating from there do not cross the redoxcline. However, the ratio of the average observed NO₃^- concentration in surface water to that in recharge water is 0.46, indicating an overestimation of denitrification. The estimated nitrate concentrations in surface water, calculated using Equation (1), showed a reasonable agreement with observed values (R² = 0.32). A much lower R² (0.08) is observed when replacing the ratio of thicknesses in Equation (1) with the average thickness ratio of 0.33. This suggests that the variation in nitrate concentration at the catchment outlet is predominantly governed by the relative thickness of the oxidized zone.

This method identifies nitrate-vulnerable areas along water courses in a catchment that can be used to better target mitigation measures across Flanders.