Agrometeorological drought early warning as a climate service: SPI projections using SARIMA models for seasonal risk management

Extreme weather events, such as frequent droughts, pose a significant threat to agriculture and livelihoods in countries such as Pakistan, where agriculture, which employs 62 % of the workforce, is heavily dependent on rainfall. In the current study, a climate service has been developed to develop early warnings for agrometeorological drought by applying Seasonal Autoregressive Integrated Moving Average (SARIMA) models to forecast the Standardized Precipitation Index (SPI) across 6- and 12-month intervals. This innovative approach aims to enhance the capacity for anticipating drought conditions, facilitate more effective agricultural management and decision-making in response to potential water scarcity. By using monthly precipitation data collected from 20 sites between 1991 and 2024, a comprehensive assessment of historical drought occurrences and projected seasonal conditions for the agricultural period from 2025 to 2030 and long-term 2–25 to 2050 are conducted. The best-fit SARIMA models demonstrated high accuracy (validation R2 values: 0.86–0.94; RMSE values: 0.31–0.49) across meteorological stations. From 2010 to 2024, the Quetta region experienced 17 months of extreme drought (SPI ≤ − 2.0), indicating that severe droughts were a recurrent phenomenon. Projections for 2025–2030 and 2025–2050, based on historical trends, predict prolonged mild drought conditions (SPI: −1.3 to − 1.7) during the Rabi season in Punjab and Sindh. Balochistan is expected to face severe arid conditions, with the SPI reaching − 2.1 by 2028. The SARIMA model showed high forecasting ability, with Nash-Sutcliffe Efficiency values > 0.81 across all stations, offering useful insights for irrigation planning and crop management. Our research will enable policymakers to forecast yield reductions of 25 %–35 % in drought-prone agrometeorological zones and prioritize resource allocation, providing a vital tool for seasonal risk assessment and serving as an early warning system to help plan climate-smart management practices, promote drought-tolerant crop varieties, and implement high-efficiency irrigation systems, thereby improving the resilience of rain-fed agricultural systems.