Diagnosing Hurricane Track and Intensity Change as Predicted by the Geophysical Fluid Dynamics Laboratory Hurricane Model

The Geophysical Fluid Dynamics Laboratory (GFDL) hurricane model is presently one of the best track and intensity forecast models available. Nevertheless, there are situations where the GFDL hurricane model makes substantial forecast errors. In this study we intend to diagnose the reasons for good and bad forecasts from the GFDL hurricane model using the technique of potential vorticity (PV) inversion. The results of our PV inversions will be used to improve forecasts of tropical cyclone motion and better predict intensity change events.

Observations have indicated that divergence can be large in real hurricanes (Montgomery and Franklin 1998). Therefore we use the asymmetric balance (AB) theory of Shapiro and Montgomery (1993) which is, in contrast to the balance equations, capable of accommodating large divergence. Idealized primitive equation simulations of a moving hurricane have been successfully diagnosed using the AB theory (Möller and Jones 1998). The diagnostic framework has provided valuable insight into the influence of friction and heating on the PV distribution and corresponding velocity fields in tropical cyclones. The objective of the present study is to apply this technique to more realistic model data.

A data set from the GFDL model (Hurricane Opal of 1995) is presently being used for development and testing of the PV inversion technique. This data set will be replaced by a higher resolution model run when it becomes available. In all cases the data are processed by J. Persing (Persing, Montgomery and Tuleya 1999).

We believe that a careful diagnosis of the model results should give some insight into the dynamics of the GFDL forecasts, and will help to improve this and other forecast models. In this regard, the diagnosis of vertical motion in terms of the AB version of "Q"-vectors may be helpful. Results from the developmental analysis will be presented as available.