Probabilistic Intense Rainfall Prediction from Landfalling Tropical Cyclones using Convective-scale Ensemble Data Assimilation System

The NOAA’s National Severe Storms Laboratory is developing a regional, on-demand, sub-hourly, cycled ensemble data assimilation and prediction system, known as the Warn-on-Forecast System (WoFS). The WoFS provides continuous flow of 0–6 h probabilistic forecasts of evolving weather threats from individual convective storms (i.e. tornadoes, intense rainfall, flash flood, extreme local winds and damaging hail). The goal is to provide NWS forecasters with numerical model guidance that will enable earlier and more accurate communication of hazardous thunderstorm threats for better preparedness.

The WoFS uses the WRF-ARW model and GSI-EnKF data assimilation system and assimilates MRMS reflectivity, WSR-88D radial velocity, GOES-16 cloud water path and NWS conventional observations. These observations are assimilated continuously every 15-min and 0–6 h ensemble forecasts are generated every hour. This presentation will show the performance of WoFS for several recent record-breaking landfalling tropical cyclone (i.e., Harvey 2017, Florence 2018 and Michael 2018) events. Visual inspections and probabilistic verification metrics demonstrate promise of WoFS to highlight areas where intense rainfall from LTC can result in flash flooding. Areas requiring further development and improvements will also be discussed.