Uncertainty Analysis of Recharge to the Edwards Aquifer using Bayesian Model Averaging of Numerical Models.

The Edwards Aquifer is one of the most prolific artesian aquifers in the world catering to the water demands of urban and suburban regions of San Antonio, Texas. In recent decades, water demands in the region have far exceeded the capacity of this aquifer. Appropriate estimation of recharge will prove beneficial in planning Best Management Practices (BMPs) in the region. But a lot of uncertainties are involved in quantifying the exact amount of recharge entering the aquifer. These uncertainties in recharge estimates arise due to underlying uncertainties inherent to conceptual hydrologic model structures, model parameters, and input/forcing data. Our study explores the possibility of using a suite of hydrologic models, instead of a single model, to estimate recharge and evaluate the associated uncertainties in the Edwards aquifer region. A multi-model combination of the numerical models using Bayesian Model Averaging (BMA) scheme is expected to improve the predictions by better addressing the model structural uncertainty.