SPI-gamma random forest modelling for meteorological drought characterization and prediction in the Bengal Delta, Indo-Bangladesh region

This study explores the nature and severity of drought in the Bengal Delta based on monthly precipitation data (1980–2020) and its future projections. The standardized precipitation index (SPI) with gamma probability function was employed to characterize drought, while a Random Forest (RF) model was adopted for drought forecasting for 2030 and 2050. Results reveal that extremely dry conditions prevailed in the western part of West Bengal, with Purulia district experiencing 14 years of extremely dry conditions between 1980 and 2020, compared to 9 years for Bankura and 5 years for Paschim Medinipur. In Bangladesh, the highest negative SPI values were found in Shylet (-3.02) in 1981 and Cox’s Bazar (-2.94) in 2018. Seasonal and spatial characteristics indicate a significant above-normal trend in the West Bengal pre-monsoon season. Bangladesh shows a noteworthy and consistent above-normal trend in the pre-monsoon season for 6, 12, and 24-month timescales. Future predictions indicate an increasing nature of droughts in the Indo-Bangladesh region. This comprehensive assessment of drought patterns in the Bengal Delta region fills a critical knowledge gap and provides valuable insights for developing sustainable drought management strategies. The study’s findings are crucial for enhancing community resilience and supporting informed decision-making in the face of increasing drought risks in this agriculturally important and densely populated region. A novel drought management framework is proposed for integrated drought mitigation and adaptation.