A numerical study of the distribution of precipitation in Hurricane Bonnie (1998)

As the damage that accompanied the floods from Hurricanes Georges and Mitch of 1998 and Floyd and Irene of 1999 dramatically highlighted, rainfall is one of the most significant impacts that accompanies a landfalling tropical cyclone. However, quantitative precipitation forecasting (QPF) is one of the most difficult problems in tropical cyclone forecasting. Rainfall measured at a particular location during the landfall of a tropical cyclone depends on many factors, such as the location with respect to the storm's track, the intensity and distribution of rainfall around the storm, the translational speed of the storm, and local effects such as topography and orientation of the coast. Understanding how these and other factors influence the precipitation distribution of landfalling hurricanes is an important step in improving hurricane QPFs.

In this study the distribution of precipitation from Hurricane Bonnie (1998) is explored by performing a numerical simulation using the primitive-equation, nonhydrostatic mesoscale model MM5. A six-day simulation is performed, and precipitation fields from the innermost domain (5-km grid length) are compared against observations. The wealth of observational information that accompanied the third phase of the NASA Convection and Moisture Experiment (CAMEX-3) field program, including the NASA TRMM satellite and various airborne observations from the NOAA P-3's and G-IV and the NASA DC-8 and ER-2, are used to validate the simulation. Statistics related to precipitation distribution, such as radial profiles of rainfall, radial-height profiles of hydrometeors, and probability distributions of various rainfall thresholds, are presented for the simulation and the observations from the TRMM satellite. Validation of these parameters will lend confidence in the skill of the model in capturing the precipitation distribution and enable an investigation of the physical processes governing the distribution to be performed in a later study.