SPEI Drought Forecasting in Mexico

This study analyzes and compares three Standardized Precipitation and Evapotran-spiration Index (SPEI) prediction models at different time scales: (1) Kalman Filter with Exogenous Variables (DKF-ARX-Pt, FK), (2) Closed Recurrent Units (GRU) and (3) Autoregressive Neural Networks with External Input (NARX). The evaluation was performed on observed data from meteorological stations in the State of Mexico and Mexico City, considering several performance metrics, such as mean absolute error (MAE), mean square error (MSE), root mean square error (RMSE), coefficient of de-termination (R²), Nash-Sutcliffe efficiency coefficient (NSE) and Kling-Gupta efficien-cy (KGE). The results indicate that the FK model with exogenous variables is the most accurate model for SPEI prediction at different time scales, standing out in terms of stability and low variance in prediction error. GRU networks showed acceptable per-formance on long time scales (SPEI12 and SPEI24), but with lower stability on short scales. In contrast, NARX networks presented the worst performance, with high errors and negative efficiency coefficients at several time scales. It is concluded that models based on Kalman filters can be key tools to improve drought mitigation strategies in vulnerable regions.