Abstract
BACKGROUND: Leptospirosis remains a major public health concern in Sri Lanka, a country with a tropical climate conducive to transmission. Despite ongoing surveillance, there is limited evidence on the spatial and climatic determinants driving long-term disease dynamics. This study aimed to investigate the spatiotemporal distribution and climatic sensitivity of leptospirosis from 2009 to 2024 using advanced statistical modelling. METHODS: District-level monthly leptospirosis case data for the period 2009–2024 were obtained from the Epidemiology Unit of the Ministry of Health, Sri Lanka. Corresponding monthly district-level climatic variables, including total rainfall, mean temperature, minimum temperature, maximum temperature, and mean relative humidity, were retrieved from the NASA POWER satellite dataset. Associations between climatic variables and leptospirosis incidence, as well as spatial heterogeneity in case distribution, were assessed using a Generalized Additive Model for Location, Scale, and Shape with a Zero-Adjusted Gamma distribution, while spatial clustering and autocorrelation were examined using Moran’s I. RESULTS: A total of 81,629 confirmed cases of leptospirosis were recorded during the study period. The ZAGA-GAMLSS model identified several climatic, spatial, and temporal predictors that were significantly associated with district-level incidence. Relative humidity and maximum temperature showed immediate negative associations with incidence, while humidity, mean temperature, and rainfall demonstrated positive lag-dependent effects at 1–3 months. In contrast, maximum and minimum temperatures exhibited predominantly negative association at 3-month lag. Spatial heterogeneity was evident in both incidence rates and zero-inflation, and temporal dependence was detected at 1 and 12-month lags. Higher relative humidity three months earlier was linked to more stable leptospirosis incidence rates across districts. Spatial analysis further revealed significant clustering, with hotspots identified in the districts of Ratnapura, Galle, Matara, and Hambantota, while Colombo was identified as a spatial outlier. CONCLUSIONS: Leptospirosis in Sri Lanka exhibits distinct spatiotemporal patterns influenced by climatic variability, with elevated risk following monsoon rains in the south‑western part of the country. Climate‑sensitive modelling, as demonstrated in this study, supports the integration of meteorological surveillance into early warning and response systems to enhance leptospirosis control, particularly in identified hotspot regions. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12879-026-12533-1.