High-Impact Weather Forecasts Using a Prototype Warn-on-Forecast System

The Warn-on-Forecast (WoF) is a multi-year research and development program carried out by NOAA’s National Severe Storms Laboratory (NSSL), with a goal of increasing the forecast accuracy and warning lead time for tornadoes, flash floods, large hail, and damaging winds. Increased warning lead time of these high-impact severe weather events will reduce the loss of life and property damage caused by these phenomena. To accomplish this goal, NSSL is actively working on developing a prototype WoF system, which is a frequently updated convection-allowing ensemble data assimilation and forecast system using the WRF-ARW model to provide probabilistic guidance on severe weather hazards. Several past studies show that the correct representation of land surface and atmospheric exchange processes in land surface models (LSM) coupled to numerical weather prediction models is critical for forecasting deep moist convection. This study analyzes the impact of two operationally used LSM schemes on the forecast accuracy of a prototype WoF system. Two sets of WoF experiments are conducted using either the RUC LSM or the NOAH LSM schemes for several severe weather events during the past year. Accuracy of and differences between the two sets of experiments are analyzed for the cases using composite reflectivity, precipitation, 2-meter temperature and dew point, 10-meter wind speed, soil moisture, and soil temperature. Results of the comparison show that overall, the RUC LSM has a drier soil surface than the NOAH LSM, and this has a noticeable impact on forecast accuracy for the variables studied.