Assessing the nitrate vulnerability of shallow aquifers under Mediterranean climate conditions.

The EU Nitrates Directive calls for urgent integration of process-oriented indicators of nitrate fate with map overlaying approaches for assessing nitrate vulnerable zones (NVZs). In the region of Campania (southern Italy) groundwater contamination represents a serious concern because of the presence of intensive agricultural practices and livestock farming. A protocol was proposed to assess the probability distribution of the following three indicators of nitrate transport across the vadose zone (obtained by using a physically-based model): i) an annual cumulative nitrate flux entering the shallow aquifer, ii) annual cumulative root nitrate uptake, and iii) nitrate transit time across the vadose zone. This method involves numerical simulations of soil water flow and solute transport using Hydrus-1D for a representative 10-m-thick soil profile beneath an irrigated maize plot located in a study area within the Sele plain, Campania. Two scenarios are built by running a set of one hundred (20-yr-long) simulations at a daily time resolution: i) 195 kg N ha-1 of nitrate fertilizer (urea) is applied annually to estimate the probability distributions of annual cumulative nitrate leaching and root nitrate uptake (Scenario 1), ii) 195 kg N ha-1 of nitrate fertilizer (urea) is applied once to estimate the nitrate transit time distribution across the soil profile (Scenario 2). In each Scenario, the simulations consider two agricultural practices with either one or three annual nitrate fertilizer applications subject to randomly generated climate forcing using a Monte Carlo approach. Variations in soil and rainfall properties are described by the Miller-Miller geometric similitude and the Poisson parameterization, respectively. In Scenario 1, roots absorb on average 80.3 kg N ha-1 every year (corresponding to 40% of applied nitrate), and the median annual cumulative nitrate flux across the soil profile bottom is 74.9 kg N ha-1 (representing about 38% of applied nitrate), when urea is applied in a single treatment. In contrast, when fertilizer is applied in three treatments, 112.4 kg N ha-1 (corresponding to almost 60% of applied nitrate) of nitrate is removed by root water uptake, and the median annual cumulative nitrate leaching is 52.4 kg N ha-1 (corresponding to about 27% of applied nitrate). In Scenario 2, mean transit time values are 2741 days, 2707 days, and 2650 days when urea is applied on April 1st, June 1st, and August 1st, respectively. Our model simulations provide useful indicators of nitrate transport and can be integrated with map overlaying procedures for delineating nitrate vulnerable zones