Assessment of monthly hydroclimatic patterns and rainfall-runoff modeling for hydrometric forecasting in the Upper Inaouene Basin, Northern Morocco

Climate change and consecutive droughts have exacerbated the water resource crisis in many regions, especially in semi-arid Mediterranean areas like Morocco. This study aims to conduct a monthly spatial assessment of the hydroclimatic regime and compare the performance of two different monthly models for runoff forecasting at the hydrological stations of the Upper Inaouene watershed. The models evaluated were the Hydrological Model of Rural Engineering with two monthly parameters (GR2M) and the Artificial Neural Network (ANN) model. Hydrometeorological data were obtained using the Thiessen and averaging methods. The models' effectiveness was assessed using the Nash-Sutcliffe criterion (NS), mean absolute error (MAE), root mean square error (RMSE), correlation coefficient (r), and coefficient of determination (R2). The results indicate that the basin's climate is Mediterranean semi-arid, characterized by significant spatial and temporal variability and two distinct annual periods. The runoff deficit is influenced by thermal factors, with evapotranspiration systematically exceeding runoff. Flow variability is attributed to infiltration from karstic aquifers and contributions from domestic sewage. The GR2M model demonstrated efficiencies of 85.45% for the correlation coefficient (r), 71.31% for the coefficient of determination (R2), and 71.08% for the Nash-Sutcliffe criterion (NS). In comparison, the ANN model achieved 89.4% for r and 80% for both R2 and NS, outperforming GR2M in accuracy. These models, particularly the ANN, provide essential hydrological data for water resource management, addressing the lack of flow rate data.