Accuracy of tropical cyclone intensity forecasts in the North Pacific and Atlantic

Consensus methods require that the techniques have no bias and have skill. The accuracy of six statistical and dynamical model tropical cyclone intensity guidance techniques was examined for western North Pacific, eastern North Pacific, and North Atlantic tropical cyclones during the 2003-2004 seasons using the climatology and persistence techniques called ST5D or SHF5 as measures of skill. A framework of three phases: (i) initial intensification, (ii) maximum intensity with possible decay/reintensification cycles; and (iii) decay was used to examine the skill.

From an initial study for the 2003-2004 western North Pacific seasons, only about 60% of the 24-36 h forecasts during both the formation and intensification stages were within +/- 10 kt, and the predominant tendency was to under-forecast the intensity. None of the guidance techniques predicted rapid intensification well. All of the techniques tended to under-forecast maximum intensity and miss decay/reintensification cycles. Whereas about 60-70% of the 12-h to 72-h forecasts by the various techniques during the decay phase were within +/- 10 kt, the strong bias was to not decay the cyclone rapidly enough. In general the techniques predict too narrow of a range of intensity changes for both intensification and decay.

From an initial study for the 2003-2004 eastern North Pacific and Atlantic seasons, the Decay Statistical Hurricane Intensity Prediction (DSHIPS) technique was the best technique in both basins during the formation phase. When the forecast errors during formation exceed +/- 10 kt, the statistical techniques tend to over-forecast and the dynamical models tend to under-forecast. Whereas DSHIPS was also the best technique in the Atlantic during the early intensification stage, the Geophysical Fluid Dynamics Laboratory model was the best in the eastern North Pacific. All techniques under-forecast periods of rapid intensification and the peak intensity, and have an overall poor performance during decay/reintensification cycles in both basins. Whereas the DSHIPS was the best technique in the Atlantic during decay, none of the techniques excelled during the decay phase in the eastern North Pacific. All techniques tend to decay the tropical cyclones in both basins too slowly, except that the DSHIPS performed well (13 of 15) during rapid decay events in the Atlantic.