Some of the success of the systematic approach to tropical cyclone track forecasting is attributed to the ability to recognize likely errors associated with specific dynamical forecast models. These model error traits were identified after assessing thhe skill associated with an exhaustive set of dynamical forecast cases. Extension of track forecasts to 120 h would benefit from better identification of the location and timing of tropical cyclone formation. This study thus complements the systematic approach to forecasts of tropical cyclone tracks via identification of dynamical model error characteristics associated with forecasts of tropical cyclone formation.

An algorithmn has been developed to detect and track in dynamical model fields those circulations that have become, or are forecast to become, tropical cyclones. Additionally, specific environmental conditions in the model analysis and forecast fields are attached to the tracked circulation. Examples of environmental parameters include low-level vorticity, sea-level pressure, convective and total precipitation amounts, vertical wind shear, and mid-tropospheric moisture. Threshold valuues of these environmental parameters associated with the failed and the successful model predictions of tropical cyclone formation are identified at each forecast interval to 120 h.

Based on the circulation tracks from algorithmn, an assessment of successful and failed 12-h through 120-h forecasts of tropical cyclone formation have been made based on forecasts from 2002 and 2003 tropical cyclone seasons over the western North Pacific and North Atlantic Oceans. Failed forecasts may be actual formations that were not forecast or forecast formations that never occurred. Classification of environmental parameters associated with successful and failed forecasts will be used to calculate threshold measures that define a measure of confidence that a forecast circulation will indeed develop. Such a confidence measure could be used by operational forecasters to maximize the probability of detection of a tropical cyclone formation or minimize a false alarm.