Stochastic Recharge Model for Edwards Aquifer in Central Texas

Abstract

A condensed disaggregation procedure has been employed to generate synthetic recharge sequences for the Edwards aquifer in the San Antonio region. Two stochastic processes ae used in the procedure: a multivariate annual flow generator, and an annual-to-monthly disaggregation model. The processes are multivariate, first-order autoregressive, implemented in SPIGOT, and fit to a 58-yr historical record based on U.S. Geological Survey (USGS) recharge estimates for the nine river basins of the Edwards aquifer recharge zone. Transformations of basin recharge are employed in order for the autoregressive models to reproduce the marginal distributions of recharge. Validation of the synthetic recharge generation procedure shows some discrepancies between the marginal distributions, correlations, and drought characteristics of the historical and synthetic flows.