CariCOOS: Improving high-resolution numerical weather prediction for the northeast Caribbean region

The northeast Caribbean as other insular regions lack reliable high-resolution weather forecast data of 10meter winds to serve as the engine of numerical ocean models targeting high resolution waves and currents forecasts for nearshore areas. This issue is exacerbated due to the complex orographic features of the Antilles, which govern topographic shadowing and incoming solar radiation fields responsible for the diurnally forced convection typical of tropical island weather. A solution to this problem is currently under evaluation for the CariCOOS region. This solution employs both dynamical numerical solvers of the Weather Research and Forecasting (WRF) Model, which are referred to as the ARW (Advanced Research WRF), and the NMM (Non-hydrostatic Mesoscale Model) cores. CariCOOS research and development efforts are executed in collaboration with the National Weather Service Weather Forecast Office San Juan (NWS WFO SJU). CariCOOS current operational WRF model setups are based on the NMM core at resolutions of 6-km, 2-km, and 1-km. These models target short and medium range forecast; the latest experimental WRF model setup is based on the ARW with resolutions of up to 500-m. The WRF-ARW setup is currently under evaluation to improve very-short-term weather forecast, in support of maritime operations of high-traffic ports and harbors (Bay of San Juan, PR, and Port of Yabucoa, PR). Noteworthy improvement have been achieved as evidenced by model skill assessments of forecasted wind speed, and wind direction of these WRF model setups when compared to in-situ observations of CariCOOS assets (land base weather stations and coastal buoys). The numerical weather prediction forecasting improvements realized via the implementation of both WRF dynamical cores are primarily driven by the increase in horizontal resolution. The most prominent improvements in weather forecasts is were achieved for the leeward side of Puerto Rico and U.S. Virgin Islands and harbor regions when simulated at very fine horizontal grid spacing resolution (less than 1-km). Considering forecast lead time requirements of the NWS WFO SJU, CariCOOS researchers constantly strive to optimize WRF model setups to its limit. Details of the various CariCOOS WRF model setups and implementations will be presented in this paper along with validation statistics.