The increasing demand for water, especially in semiarid regions, poses new challenges for water resources managers. Hence, they must understand the interactions between surface water and groundwater systems, especially under uncertainty. In groundwater systems, uncertainty arises from boundary conditions and aquifer parameters, while in surface water systems, uncertainty stems from the natural variability of the climatic inputs such as precipitation, air temperature, and river flows and, to a lesser degree, from surface model parameters. The main objective of this paper is to present a simulation model for examining the interaction(s) between surface water and groundwater under geologic uncertainty. This model was specifically developed for the study of the Najafabad semiarid plain in Iran. The plain aquifer under study is recharged by lateral groundwater flow from boundaries, irrigation percolation, canals and river seepage, and precipitation directly on the plain. Groundwater flow and groundwater/surface-water interactions were simulated for steady and transient flow conditions using MODFLOW-2000 for average monthly stress conditions in the system during 2001-2004. Calibration of the model was performed by comparing simulated and observed heads, and observation measurements made on a monthly basis during 2002-2004 in 49 observation wells. The Monte Carlo method was used for uncertainty analysis and the investigation of how different model structure alternatives would affect the estimated parameter values. This procedure has a fairly general applicability including systems with scare measurements. The stochastic modeling and analysis for hydraulic conductivity and river conductance was performed using normal distributions under 100 realizations. The results show that return flows from irrigated lands and seepage from the Zayandehrood River form significant components of the water budget, while boundary conditions play a minor role in the total mass balance.

Note: Bounding box set based on full extent of Iran since exact location of model is unknown.