Impact of Climate Change on River Discharge and Groundwater Level Using Copula Functions (Case Study: Siminehrood River Basin, Iran)

Abstract

The aim of this research is to simulate and predict the groundwater level in the Siminehrood River Basin, which is situated south of Lake Urmia, Iran. This simulation was conducted using copula functions while accounting for changes in river discharge influenced by climate change. A total of 26 large-scale CMIP6 models were utilized in this study. Precipitation data were downscaled and simulated using the LARS WG 7.0 model. Subsequently, precipitation data for both the baseline period (1988-2018) and the future period (2031-2050) were predicted for three scenarios: SSP1-2.6, SSP2-4.5, and SSP5-8.5 through a weighted average method. Following the simulation and prediction of precipitation in the Siminehrood River Basin, copula functions were employed to simulate and predict both river discharge and groundwater levels. Prior to fitting the copula function, correlations between pair of parameters precipitation-river discharge and river discharge-groundwater level were examined using Kendall's tau coefficient; correlation values obtained were 0.43 for precipitation-river discharge and 0.44 for river discharge-groundwater level. After selecting marginal distributions and examining these correlations, ten different copula functions were fitted to each pair of parameters in order to identify the most suitable model among them. The results from predicting precipitation related to climate change indicated that annual precipitation under all three scenarios would decrease compared to the measured precipitation. Annual precipitation reductions were projected to be 5.1 mm, 31.5 mm, and 34.8 mm under the scenarios SSP1-2.6, SSP2-4.5, and SSP5-8.5, respectively. Analysis through copula functions revealed that the Clayton copula provided optimal performance when creating a joint distribution for these pair of parameters during simulation phases concerning river discharge as well as groundwater levels and its accuracy was validated based on evaluation criteria including NSE (Nash-Sutcliffe Efficiency), RMSE (root mean square error), and R² (coefficient of determination). Furthermore, it was concluded that reductions in annual precipitation would lead to decreases in annual river discharge ranging from 2.9 m³/s to 6.6 m³/s alongside an annual drop in groundwater levels estimated between 0.3 m and 1.5 m.