Sensitivity of Extreme Hurricane Precipitation to WRF Microphysics and Grid Spacing: Hurricane Melissa (2025) Landfall over Jamaica

Extreme precipitation associated with landfalling tropical cyclones poses major forecasting challenges, particularly over complex terrain. This study investigates the sensitivity of simulated hurricane rainfall to microphysics parameterization and horizontal resolution using the Weather Research and Forecasting (WRF) model for Hurricane Melissa, a Category 5 storm that made historic landfall over Jamaica in October 2025 and produced rainfall exceeding 1,000 mm in mountainous regions. Four WRF simulations were conducted using two commonly applied microphysics schemes, WSM6 (single-moment) and Morrison (double-moment), across two domain configurations: a single 9 km grid covering the Caribbean basin and a nested configuration with a 3 km convection-permitting inner domain centered over Jamaica. Model outputs were evaluated against satellite-based precipitation estimates from IMERG and CHIRPS. Results suggest that horizontal resolution strongly controls the spatial pattern of simulated precipitation. The 3 km nested simulations capture sharper gradients, localized maxima, and more physically consistent rainfall structures compared to the smoother and more diffuse patterns produced at 9 km resolution. Differences between microphysics schemes are secondary to resolution but remain evident, with the Morrison scheme producing more coherent and structured precipitation fields, while WSM6 generates more fragmented and spatially patchy rainfall. All simulations accurately reproduce the timing of peak precipitation during landfall, indicating weak sensitivity of storm evolution to microphysics choice. However, total rainfall amounts vary substantially across configurations, with convection-permitting simulations producing significantly higher accumulations. These totals exceed CHIRPS estimates, likely due to the underestimation tendency of extreme precipitation in complex terrain by CHIRPS, while agreement with IMERG varies by location and intensity. These findings highlight that accurate representation of extreme tropical cyclone precipitation requires convection-permitting resolution, while rainfall intensity remains sensitive to both microphysics selection and observational reference datasets.