Assessment of the impact of drought on water resources in the Gharb plain: Approach using the Support Vector Regression (SVR)

Abstract

The Gharb plain is one of Morocco’s most fertile agricultural regions, as it is located in the lower part of the Sebou basin at a low altitude. This study combines the Standardized Precipitation–Evapotranspiration Index (SPEI) for identifying drought events, correlation analysis to assess their relationship with flow, and the support vector regression (SVR) model for hydrological modeling and prediction. The results reveal a strong relationship between drought and river flow. The SPEI-12 index shows the best performance, with a coefficient of determination (R 2 = 0.91) and a root mean square error (RMSE = 62.39), while SPEI-6 also exhibits a significant relationship (R 2 ≈ 0.79). All models demonstrate very high statistical significance (P < 10 –18 ), confirming the ability of the SPEI index to explain flow variability in the study area.