An Analysis of Recent Dynamical Model Guidance for Eastern North Pacific Basin Tropical Cyclones Near the Coast of Mexico

Dynamical tropical cyclone (TC) track models have been shown to possess characteristic basin-wide error statistics (Goerss 2007). However, little attention has been devoted to the study of forecast errors for smaller subsets of tropical cyclones (e.g., those near land). In this paper, we examine forecast error statistics and skill of the regional and global dynamical models for tropical cyclones located close to the Mexican coast. Later, forecast error statistics for the same models are analyzed for eastern North Pacific TCs once watches and warnings have been issued for coastal locations in Mexico. Anecdotal evidence suggests that regional models such as the GFDL and HWRF exhibit a poleward bias for TCs near the Mexican mainland, while global models like the UKMET tend to have longer westerly tracks. Dynamical model and official NHC track prediction biases will be presented and compared, followed by a discussion of the validity of the anecdotal evidence (forecaster perception) with regard to track model guidance.