Investigation of Irma & Maria Track Forecasts Using an Ensemble Approach within Basin-Scale HWRF

Although deterministic models continue to provide valuable guidance on tropical cyclone (TC) forecasts, they only provide a singular solution and, therefore, fail to convey the forecast uncertainty therein. As a result, ensemble prediction systems have gained more attention recently as tools to better understand the range of realistic solutions for a given TC forecast and the probability that each forecast will verify. Under the support of the Hurricane Forecast Improvement Project (HFIP), the NOAA/AOML/Hurricane Research Division (HRD) led a collaborative effort to develop the “basin-scale” Hurricane Weather Research and Forecasting Model (HWRF-B), an advanced, experimental version of the community HWRF. HWRF-B, configured with a large outermost domain that covers ~¼ of the globe and multiple sets of movable nests to produce high-resolution forecasts for multiple TCs, is an ideal tool to study TC interactions with the environment and other TCs at high resolution. In this latest development, HWRF-B was configured as a 20-member ensemble prediction system, initialized from the Global Ensemble Forecast System (GEFS), to investigate track forecasts for Hurricane Irma (2017) and Hurricane Maria (2017). As Irma and Maria moved near land, nearby synoptic-scale systems induced uncertainty to track forecasts, measured here by the spread of ensemble members. In particular, deviations in the near-TC environment amongst ensemble members were evaluated to better understand track forecast spread for these high-profile hurricanes.