Characterization and Prediction of Water Stress Using Time Series and Artificial Intelligence Models

In agroecosystems, drought is a critical climatic phenomenon that affects evapotranspiration and induces water stress in plants. The objective this study was to characterize forecast the Hyderabad region of India using artificial intelligence models. monthly precipitation data for period 1982–2021 characterized by standardized index (SPI) modeled classical autoregressive integrated moving average (ARIMA) model (AI), i.e., neural network (ANN) support vector regression (SVR) model. results show on short-term SPI3 time scale studied experienced extreme deficit 1983, 1992, 1993, 2007, 2015, 2018, while mid-term SPI6 scale, 1991, 2011, 2016 were extremely dry. addition, prediction at both scales AI models outperformed ARIMA both, training validation sets. Among applied models, SVR performed better than other modeling predicting (confirmed root mean square error—RMSE), Diebold–Mariano test confirmed output significantly superior. A reduction error 48% 32% (vs. ARIMA), 21% 26% ANN) observed sets scales. These may be due ability account nonlinear complex patterns input against linear contribute more sustainable efficient management resources/stress cropping systems.