Global Sensitivity Analysis for a coupled Hydro-economic model under a groundwater management policy in Kern County, California.

The adaptation of the coupled agricultural-water system to surface water variability and groundwater management policies has high uncertainty, particularly during long dry periods. Hydro-economic models using optimization methods have been widely used to forecast the impact of endogenous and exogenous changes in the system (e.g. policies, change in prices, climate change, drought), however the impact of the agricultural production system on groundwater depth requires study of the two systems in tandem and their feedback interactions. To assess groundwater management policy and develop an informed forecast tool, we must consider the uncertainty of the variables and parameters used in the model. A global sensitivity analysis was performed for a dynamic hydro-economic model that couples an artificial neural network, which forecasts groundwater depth change, with a calibrated production model. The results show the sensitivity of irrigated land, cropping patterns, and groundwater pumping to uncertainties in economic and physical variables, as well as calibration parameters. Furthermore, the results give insight into potential changes in groundwater depth to fulfill groundwater sustainability requirements.