Verification and expansion of probabilistic tropical cyclogenesis forecasts by Dvorak analyses

The lack of a standardized data set of pre-genesis tropical disturbances was one of the greatest inhibiting factors toward the development of guidance for tropical cyclone genesis. To address this issue, Dvorak intensity estimates, which are operationally performed on tropical features of interest, were collected and have allowed a basic investigation into the climatological likelihood of tropical cyclogenesis. This work continues previous research (Cossuth et al. 2010), which gathered Dvorak data from 2001-2009 by the Tropical Analysis and Forecast Branch (TAFB) at the National Hurricane Center (NHC) and the Central Pacific Hurricane Center (CPHC), and extends it through 2011 as well as adds Dvorak analyses from 2005-2011 by the NESDIS Satellite Analysis Branch (SAB). Data from multiple agencies are compared to gain further predictive power and confidence in cyclogenesis forecasts. Additionally, verification of Dvorak genesis probabilities is performed by out of sample testing and is shown to perform well as a first approximation for genesis forecasting.

Results from the analysis show trends in tropical disturbance locations and Dvorak intensity estimates that foster TC genesis predictability. The Atlantic main development region shows the highest incidence of Dvorak estimates on tropical disturbances, with smaller maxima in the western Caribbean and off the southeast US coast. There is also spatial variability in average Dvorak intensity and genesis rate through the basin. The eastern/central Pacific is also characterized by large maximum of in the location of disturbances, but with genesis likelihood generally decreasing westward through the basin. Further, genesis probabilities verification show predictive skill, especially in the middle percentages, and may be used to inform NHC probabilities.