A51P-0303
Intraseasonal Tropical Storm Activity Prediction with the NCEP CFSv2 45-day Forecasts

Friday, 18 December 2015
Poster Hall (Moscone South)
Jae-Kyung E Schemm and Lindsey N Long, Climate Prediction Center College Park, College Park, MD, United States
Abstract:
Global predictability of intraseasonal tropical storm (TS) activity is assessed using the 1999-2012 CFSv2 45-day hindcast suite and real-time predictions for 2014-2015. Weekly TS activities in the CFSv2 45-day forecasts were determined using the TS detection and tracking method devised by Carmago and Zebiak (2002). The forecast periods are divided into weekly intervals for Week 1 through Week 4. The TS activities in those intervals are compared to the observed activities based on the NHC HURDAT and JTWC Best Track datasets.

The CFSv2 45-day hindcast suite is made of forecast runs initialized at 00, 06, 12 and 18Z every day during the 1999 – 2012 period. For predictability evaluation, forecast TS activities are analyzed based on 20-member ensemble forecasts comprised of 45-day runs made during the most recent 5 days prior to the verification period. The forecast TS activities are evaluated in terms of the number of storms, genesis locations and storm tracks during the weekly periods. The CFSv2 forecasts are shown to have a fair level of skill in predicting the anomalous number of storms over most of the seven ocean basins during the 1999-2012 active seasons. The average temporal correlation score for Week 1 forecasts is between 0.50-0.52 in the Eastern and Western North Pacific, the South Indian and South Pacific basins, while correlations drop to around 0.20 for Week 4 forecasts. The forecast track is also examined using density distribution maps and false alarm statistics compiled using the hindcast analyses with Heidke Skill Scores peaking at around 0.35.

Real-time weekly TS activity predictions began in December 2013 using the climatological TS forecast statistics to make the model bias corrections in terms of the storm counts, track distribution and removal of false alarm storms. This operational implementation provides an objective tool for the CPC’s Global Tropical Hazards Outlooks. Verification and evaluation of the 2014 and 2015 seasons will be discussed.