Verification of National Hurricane Center forecasts of extratropical transition
John L. Beven II, NOAA/AOML/NHC/TPC, Miami, FL
The process of a tropical cyclone becoming an extratropical cyclone – extratropical transition (ET) – is an issue of increasing importance in tropical cyclone forecasting and impact assessment. However, no tropical cyclone warning center currently performs a rigorous verification of ET forecasts. This paper presents result from an experimental verification of ET forecasts at the National Hurricane Center from 1996-2007.
The verification has two parts. The first is a version of the 2 x 2 contingency diagram verification used by the National Weather Service for severe weather warning verification – was the event forecast versus did it actually occur. This includes the various skill statistics such as critical success index, false alarm ratio, and probability of detection. The second is an examination of the timing errors of forecasts of the cyclone being extratropical. To maintain a homogeneous data set, the second part includes a 5-day forecast verification for 2001-2007 and a 3-day verification from 1997-2007.
The results show that the NHC forecasts have skill according the contingency diagram metrics. Also, the NHC forecasts seem to have become more consistent in terms of event bias since the development of the Cyclone Phase Space. The paper will highlight some cyclones that have been very problematic in terms of large forecast timing errors, incorrect transition forecasts, and missed transitions.
Extended Abstract (120K)
Session 10C, Extratropical Transition II: Forecast Challenges
Wednesday, 30 April 2008, 10:15 AM-12:00 PM, Palms H
Previous paper Next paper
Browse or search entire meeting
AMS Home Page