Hydrological Budgets of Landfalling Tropical Cyclones

Floods are among the most destructive forces of nature, costing human lives and property. In order to study the causes and effects of flooding, characteristics of meteorology and hydrology must be combined. The purpose of this study is to examine tropical cyclone-induced flooding and simulate river characteristics to aid in flood forecasting. The Florida State University Global Spectral Model (FSUGSM) is used to generate precipitation forecasts for Hurricane Floyd of 1999, which caused major flooding along the eastern coast of the United States. A track forecast for Floyd is generated by the FSU hurricane superensemble, which provides real-time forecasts for the tracks and intensities of storms over the Atlantic Ocean. These forecasts, when compared to other Numerical Weather Prediction (NWP) models, have been shown to have reduced position errors out to three days. The FSU NWP superensemble (FSUSE) is used to produce precipitation forecasts for Tropical Storm Allison of 2001, a slow-moving storm that caused flooding along the southern Gulf Coast states. These forecasts are compared to the observed rainfall estimates from the Tropical Rainfall Measuring Mission (TRMM). Compared to its member models, the superensemble has been shown to improve precipitation forecasts out to three days. Because of the greater accuracy in storm track forecasts from the FSU hurricane superensemble, a method is proposed to adjust the forecast tracks of precipitation to the storm track positions for each event. Precipitation occurs on small scales ( 10 km) rainfall generated in NWP forecasts. A method for downscaling is proposed, which is then applied using a Land Analysis System (LAS) developed by the National Aeronautics and Space Administration (NASA). This is incorporated into the input variables for a land-surface hydrologic model. The hydrology model is used to predict streamflow for two river basins (from each tropical cyclone case) that suffered major flooding as a result of heavy precipitation. Using a digital elevation model at 1 km resolution, a watershed domain and stream network are delineated for each case. Selected vegetation, land, and soil parameters are also derived and used to drive the small-scale aspects of the hydrologic model. Simulated streamflow from the model is then compared to the observed streamflow for each case. The results are used to gain insight into the methodologies of flood forecasting. To improve simulated versus observed streamflow, accurate parameterizations within the hydrology model are required. Only daily observed data was available, so the averaging of hourly hydrology model output negatively affected the results. However, adjusting the FSUGSM precipitation forecasts for Hurricane Floyd towards the FSU hurricane superensemble track improved the location of heavy precipitation. These, in turn, slightly improved the simulated streamflow output from the hydrology model by producing a secondary peak near the time of Floyd's landfall. For the river basin affected by Tropical Storm Allison, the simulated streamflow related very closely to the observed streamflow, in regards to the timing and magnitude of peak streamflow. These results give confidence in the coupling of meteorology and hydrology data to forecast heavy precipitation, river conditions, and possible floods.