John Cangialosi and Shuyi S. Chen

The location and complex terrain of the Caribbean islands are known to affect tropical cyclones (TC’s). More specifically, they can have some major impacts on TC structure and rainfall distribution at landfall. The interaction of a TC with mesoscale topographic features is not well understood. Significant variations in wind, pressure and precipitation distribution have been observed over mountainous regions. The objective of this study is to further investigate this subject matter with the high-resolution (1.67 km) multi-nested non-hydrostatic 5th generation PSU-NCAR Mesoscale Model (MM5). Hurricane Georges tracked directly over the terrain of Puerto Rico and Hispaniola; therefore, it is a perfect candidate for such a study. The model is initialized with the initial and boundary conditions from the Navy Operational Global Atmospheric Prediction System (NOGAPS) on September 21, 1998 at 0000 UTC and integrated for 96 hours. Hurricane Georges tracked directly over the terrain of Puerto Rico and Hispaniola; therefore, it is a perfect candidate for such a study. While Hurricane Georges interacted with the topography of these islands heavy rainfall, up to 1000 mm, occurred on the windward side of the mountains with significantly less precipitation on the leeward side of the mountains. The simulated storm displays good agreement with the observed storm in terms of track and intensity with landfalls occurring on both Puerto Rico and Hispaniola. The model simulated precipitation compares very well with the observations from the WP-3D aircraft radars before the landfall and the WSR-88D over land. Both model and observations show a very strong asymmetry with dominance in the rear quadrants of the storm and little convection in the northwest quadrant while interacting with Puerto Rico. In addition, from the simulation, orographic effects are clearly indicated over both islands, with tremendous rainfall totals observed on the upslope side of the mountains. Over Hispaniola, significant drying and heating occurred on the downslope side of the mountains, weakening the modeled hurricane 23 mb in 24 hours. We are in the process of understanding the physical mechanisms affecting the storm-land interaction by analyzing model results in comparison with the airborne Doppler radar derived 3-D wind fields. This will shed light on the evolution and structural change of a powerful hurricane when interacting with complex topography.