Investigation of impact of climate change on sediment yield and streamflow in a subbasin of euphrates river using statistical downscaling and genetic programming

Studying the hydrological impacts of climate change is a critical area of research, as it directly affects regional water resources, ecosystem health, and human societies. Global Circulation Models (GCMs), while indispensable for understanding broad climatic patterns, exhibit limitations in representing local-scale climate features due to their coarse spatial resolution. Statistical downscaling techniques are employed to evaluate climate change's hydrological consequences at finer scales effectively. These approaches enhance GCM outputs, generating high-resolution climate projections that facilitate a more nuanced understanding and anticipation of climate change's specific implications for hydrological systems, thereby guiding the development of adaptation and mitigation strategies. This study focuses on predicting monthly discharge and sediment yield for a sub-basin of the Euphrates River. The analysis leverages scenarios derived from the Canadian Second Generation Earth System Model (CanESM2), coupled with a statistical downscaling method and artificial intelligence genetic programming models. Watershed delineation was conducted using the Watershed Modeling System (WMS), followed by the application of Gene Expression Programming (GEP) and Linear Genetic Programming (LGP) for the statistical downscaling process. The results were compared with linear regression outcomes. The machine-coded LGP model yielded the best correlation (R), lowest Mean Squared Error (MSE) and Mean Absolute Error (MAE) values for both discharge and sediment yield during training and testing periods. Regarding future projections under RCP 2.6, 4.5, and 8.5 scenarios, while both models yielded optimal results under the RCP 8.5 scenario, the GEP model demonstrated superior performance when comparing predictions to the observed period of 2006 – 2020. Overall, the models indicate a mild decreasing trend for both discharge (Q) and sediment yield (Qs) from 2021 to 2100, providing valuable insights for future hydrological management and planning.