Subseasonal to Seasonal (S2S) Forecasting of Tropical Cyclones

Abstract

Tropical cyclones (TCs) can have major impacts on planning and executing Navy and USMC operations. For operational planning, it is important to forecast TCs at subseasonal to seasonal (S2S) lead times. The existing NPS long lead TC forecasting system produces skillful, ensemble-based, probabilistic forecasts of TC formations at leads out to 90 days. These forecasts have substantially higher resolutions (1 degree, daily to monthly) than other S2S TC forecasting systems. The NOAA Climate Prediction Center (CPC), JTWC, COMSUBPAC, and others use the NPS system forecasts in their product development and planning. The NPS system uses statistical models to post-process dynamical forecasts of the large-scale environment and generate TC forecasts. Newly emerging data sets and dynamical forecasting systems have the potential to improve significantly both the dynamical and statistical components of the system. We will investigate improvements based on using new: (a) data sets for the statistical component (especially the ERA5 and updated IBTrAC, JTWC, and NHC data sets); and (b) inputs to the dynamical components (especially from the Navy ESPC system). Our objective will be to assess the potential of new: (a) data sets to improve the resolution and accuracy of the NPS system statistical models; and (b) dynamical inputs to increase the forecast lead times and skill. Our main questions will be: How do upgrades in the data sets and dynamical inputs affect the: (a) efficiency and skill of the NPS system; and (b) the operational utility and value of the system's forecasts? We will also work with JTWC, FWCs, FNMOC, CPC, and staff oceanographers to assess the revised system's forecasts and their operational applications. Our main deliverable will be a final report on our data, methods, results, operational applications of the results, and recommendations. This topic addresses N2N6E FY21 RDTE priorities 3.a and 3.e.