Parametric and numerical analysis of the estimation of groundwater recharge from water-table fluctuations in heterogeneous unconfined aquifers

Abstract

The groundwater recharge produced by discrete precipitation events in unconfined aquifers is often estimated from water-table fluctuations (WTFs) recorded in shallow wells. This recharge estimate is prone to uncertainties when recharge is not instantaneous, when there is groundwater drainage, and when there are other processes producing WTFs. A numerical analysis of these uncertainties is presented, which accounts for noninstantaneous recharge and the changes in the stage of a river connected to the unconfined aquifer. This analysis is performed for idealized synthetic unconfined aquifers with one-dimensional (1-D) and 2-D numerical flow models which account for the anisotropy and the spatial heterogeneity of the hydraulic conductivity (K). The logarithm of K is assumed to be a Gaussian random field with a spherical semivariogram. Groundwater recharge may be grossly underestimated with the WTF data when the recharge is not instantaneous. Estimation errors are especially important near the river. On the other hand, the recharge may be largely overestimated when the river stage rises simultaneously during the recharge episode. The errors increase with the variance of the Ln K value and depend on the main direction of anisotropy and the spatial connectivity of the most permeable areas near the river. The errors are large along the most permeable zones connected to the river in the main direction of anisotropy. The recharge estimation errors are largest when the main direction of anisotropy is perpendicular to the river and are smallest when the main direction of anisotropy is parallel to the river.