Multi-Scale Spatiotemporal Drought Assessment in Sri Lanka Using the Standardized Precipitation Index (SPI) and Probabilistic Modeling

Drought is a pre-existing threat to agriculture, water, and socio-economic stability of Sri Lanka. The research evaluates the spatiotemporal variability of droughts with reference to five geographically distinct areas characterized by different climates, such as China Bay, Colombo, Nuwara Eliya, Puttalam, and Kurunegala using the Standardized Precipitation Index (SPI) at several time scales (1 to 24 months) between the years 2010 and 2023. The monthly precipitation data were modeled using Gamma, Exponential, and Normal distributions. The most appropriate model for each region was identified based on the Akaike Information Criterion (AIC). Findings indicate that the regions in the Dry Zone especially China Bay and Puttalam are most affected and high by frequent severe short-term drought whereas Kurunegala has the most vulnerability to long-term drought. The paper also discovers that the choice of probability distribution is important for deriving SPI at various time scales. These results are an indication of the regional differences in the drought risk and the necessity of the customized early warning systems, agricultural plans, and strategies of water management. This multiscale, probabilistic method can provide a very solid design of enhancing drought resilience in Sri Lanka.