Mapping solute deep percolation fluxes at regional scale by integrating a process-based vadose zone model in a Monte Carlo approach

Abstract

Interpreting and predicting the evolution of non-point source (NPS) pollution of soil and surface and subsurface water from agricultural chemicals and pathogens, as well as overexploitation of groundwater resources at regional scale, are continuing challenges for natural scientists. Accordingly, in this study we present a regional-scale modeling approach for vadose zone solute leaching that is based on stochastic application of a deterministic vadose zone model describing the water flow and solute transport processes in the unsaturated zone using the Richards equation (RE) and the advective-dispersive equation (ADE), respectively. The stochastic framework (Monte Carlo technique) allows accounting for uncertainty in the vulnerability outputs. As the approach is built on physically-based equations, it may be extended to the predictions of water fluxes (i.e., groundwater recharge) in the vadose zone. The approach relies on available datasets coming from different sources (detailed pedological information, hydrological properties in different soil horizons, water table depth, spatially distributed climatic temporal series and land use) and offers quantitative answers to soil and groundwater vulnerability to NPS of chemicals at regional scale within a defined confidence interval. Interpolation of these local distributions by geostatistical tools provides areal distributions of the statistical moments of solute and water fluxes. A preliminary evaluation of methodology was carried out for quantifying contaminant transport and groundwater recharge profile in the Metaponto plain in Southern Basilicata, Italy. Results showed large differences in the magnitude of the different travel times and related uncertainties among different profiles. The lower or higher vulnerability was found to be mainly related to the average silt content of the soil profiles.