Abstract
Rainfall during boreal fall (September-November: SON) is crucial for water resources and flood risk in Sri Lanka (SL), yet it shows significant interannual variability driven by large-scale tropical climate modes, particularly the Indian Ocean Dipole (IOD) and the El Ni & ntilde;o-Southern Oscillation (ENSO). Despite their recognised importance, the distinct influences of IOD & ENSO on mean and extreme rainfall over SL remain insufficiently understood. This study examines the combined and independent influences of IOD and ENSO on mean rainfall (Rmean) and heavy rainfall days (R10mm) during SON for 1981-2023 using partial correlation and regression analyses to isolate their respective effects, together with Generalised Extreme Value (GEV) theory to assess the statistical representation of R10mm. The analyses reveal that IOD has dominant independent influences on Southwest SL rainfall, ENSO predominantly affects West SL, and co-occurring IOD-ENSO affects both regions. These relationships come from condition-dependent atmospheric mechanisms, with combined positive IOD-ENSO conditions associated with stronger circulation, moisture transport and convergence, and convective activity than the independent forcing of either mode. Both Rmean and R10mm show increasing tendencies during the study period, particularly in the southwest and west SL. The GEV analysis further shows that non-stationary models, including climate drivers, better represent R10mm in SL than stationary models. These findings improve understanding of SL regional rainfall variability, which is associated with climate drivers, and provide useful insights for seasonal prediction, flood risk assessment, and climate resilience planning in SL.