The HFIP High Resolution Hurricane Forecast Test: Beyond the traditional verification metrics

Tropical cyclones are a serious concern for the nation, causing significant risk to life, property and economic vitality. The National Oceanic and Atmospheric Administration (NOAA) National Weather Service has a mission of issuing tropical cyclone forecasts and warnings, aimed at protecting life and property and enhancing the national economy. In the last 10 years, the errors in hurricane track forecasts have been reduced by about 50% through improved model guidance, enhanced observations, and forecaster expertise. However, little progress has been made during this period toward reducing forecasted intensity errors. To address this shortcoming, NOAA established the Hurricane Forecast Improvement Project (HFIP) in 2007. HFIP is a 10-year plan to improve one to five day tropical cyclone forecasts, with a focus on rapid intensity change.

Recent research suggests that prediction models with grid spacing less than 1 km in the inner core of the hurricane may provide a substantial improvement in intensity forecasts. The 2008-09 staging of the High Resolution Hurricane (HRH) Test focused on quantifying the impact of increased horizontal resolution in numerical models on hurricane intensity forecasts. The primary goal of this test was an evaluation of the effect of increasing horizontal resolution within a given model across a variety of storms with different intensity, location and structure. The test focused on 69 retrospectives cases from the 2005 and 2007 hurricane seasons. Six modeling groups participated in the HRH test. The models used for HRH included three configurations of the Weather Research and Forecasting (WRF) model, the operational GFDL model, the Navy's tropical cyclone model, and a model developed at the University of Wisconsin-Madison (UWM). The Development Testbed Center (DTC) was tasked with providing objective verification statistics for a variety of metrics. Bernardet et al provides an overview of the HRH Test and a summary of the standard verification results. This presentation will discuss the results obtained by applying new verification tools developed at the DTC that assess changes in forecast skill for Rapid Intensification (RI) and Rapid Weakening (RW) events, as well as changes in forecast consistency resulting from changes in horizontal grid spacing within a given model configuration.