Probability Distribution for Water Fluxes in a Heterogeneous Unsaturated Zone Using an Ensemble of 1D Simulations

Abstract

Groundwater recharge plays an important role in water resource management. In arid and semi-arid areas, the groundwater recharge under agricultural fields constitutes a large portion of the total recharge. Estimating the agricultural recharge is important for groundwater management and assessing contamination risk. However, estimations of the recharge fluxes are difficult due to the soil heterogeneity and the complex nature of the flow processes in the soil. Here, we suggest a new method for estimating the recharge flux probability distribution that accounts for the soil heterogeneity. The method is demonstrated using a unique data set of long-term continuous measurements of water content under an agricultural field. The measurements are used to assign weights to a large ensemble of hydraulic parameter sets. The weights are then used to derive the probability distribution of flux under different agricultural practices. We found that the distribution of the recharge fluxes varies considerably between one annual crop and a cycle of three crops over 2 years. However, the distribution of the fraction of the input water that infiltrates does not vary much between the two practices. Under natural conditions, with no irrigation, the recharge flux is much smaller, and its distribution is narrower. Our approach provides the full probability distribution of fluxes rather than a single expected value of the flux or its mean.