Impacts of Initial and Boundary Conditions on Hurricane Katrina Track Forecasts

It has been recognized recently that skillful hurricane intensity and structure forecasts require using high-resolution (<5 km grid space) numerical models in order to resolve inner core structures. While previous studies showed improved capability of high resolution models capturing intensity and structure changes of individual storms, the track forecast errors ranged from less than 100 km to above 400 km over the storm life-cycle for 48-72 h forecasts. It remains an important task to continue to improve hurricane track forecasts at high resolution so that meaningful comparisons can be drawn between observations and model forecasts for detailed wind and precipitation distribution and evolution. In this study, a series of COAMPS experiments were designed to investigate the impacts on COAMPS track forecasts of initial and boundary conditions from two operational global model analyses: Navy's Operational Global Atmospheric Prediction System (NOGAPS) and the Global Forecast System (GFS) of the National Center for Environmental Prediction. It is found that not only the initial synoptic flow and but also the vortex structure in the global model analysis play an important role in the high resolution COAMPS track forecasts. Sensitivity tests using different combinations of initial condition and boundary conditions also reveal that COAMPS track forecasts are influenced by the boundary conditions, although to a lesser extent. Detailed examinations of these impacts will be provided at the conference.