The use of a SHIPS ensemble for probabilistic tropical cyclone intensity forecasts

Several variables factor into the intensification of a tropical cyclone. Some of these variables are a result of large scale features. Sea surface temperature, wind shear, and storm track all influence the intensification of a tropical cyclone.

A 48-member ensemble of SHIPS forecasts is used to produce probabilistic forecasts for tropical cyclone intensity. While the ensemble mean provides a deterministic forecast, the spread in the ensemble members provides an estimate of forecast certainty.

SHIPS ensemble members are created by varying the tracks (using tracks from GFS, CMC, UKMO, and NOGAPS), varying the initial intensity (+/- 5 knots from the NHC value), varying the atmospheric parameters - such as wind shear (using the aforementioned models), and varying the sea surface temperature (by using different analyses).

Ensemble forecasts for the 2002-2007 seasons are compared to the actual intensities, for storms in the Atlantic Basin. In addition to assessing forecast skill, we assess which variables contribute most to ensemble spread.