Climatology and persistence are often good predictors of tropical cyclone motion; they are routinely used as a benchmark to assess forecast skill. Numerical models developed in the last decade have led to significant reduction in track forecast errors. However, with a few exceptions, this rapid advance in skill of numerical modeling has resulted in a decline in development and improvement of statistical prediction methods (Bessafi et al. 2002). Such statistical methods still have important forecast applications because they provide quick, simple predictions.

The primary objective of this research has been to develop a concise graphical forecasting tool that displays several components of climatological data in an easy-to-interpret summary format. Tropical cyclone forecasters have many demands on their time. Often, a “quick look” presentation of climatology data will aid the forecaster in the decision-making process. Instead of giving a point forecast, this graphical tool shows the spread of previous tropical cyclone motion bearings and assigns each thirty-degree motion sector a specific probability. Each motion sector is colorized by speed or intensity, increasing the utility to the forecaster. Furthermore, the 33-year data set can be filtered by time-of-year and intensity to augment the first-order climatology. Finally, filtering by previous motion bearing – a second-order approach – adds even more value to the forecasting graphic.

Beyond its application as a graphical forecast tool, the climatological data can be correlated with NWP model output. Knowing a particular dynamical model’s error-to-climatology correlation allows forecasters to place appropriate weight when considering its predictions. The initial correlation between 2001 and 2002 COAMPS and NOGAPS errors and the 33-year Atlantic basin climatology was surprisingly low. However, by recalculating the correlation after filtering the database by time-of-year, the results are more promising.

To continue exploring possible relationships between model errors and climatology, specific cases from the active 2003 Atlantic season, including Hurricane Isabel, were examined. Isabel’s model errors have stronger links to climatology because the hurricane followed a traditional track. As a result, point climatology values are higher than those used in the more general study. Results will be presented in detail at the 2004 AMS tropical conference.