This paper presents a numerical modeling experiment of Tropical Storm Allison using a very high resolution model (1 km) and all available data. Tropical Storm Allison developed about 200 km south of Galveston, Texas on June 5, 2001 and moved inland, subjecting Houston and the surrounding areas to record rainfall totals and catastrophic flooding. It stalled and remained organized over Southeast Texas as a depression for 4 days before drifting back over the Gulf waters on June 10. Although Allison exhibited some subtropical characteristics during the first 5 days of its existence, the National Hurricane Center classified it as a tropical storm or tropical depression until June 10 when it developed more evident baroclinic instability and became classified as a subtropical depression. Allison reintensified into a subtropical storm and moved east-northeast causing extensive flooding in southeastern Louisiana. The storm reached its minimum pressure of 1000 mb in southern Mississippi and weakened into a subtropical depression as it continued its track to North Carolina and then along the East coast and out to sea.

Our intention was to model Allison’s erratic track, highly variable rainfall pattern, and intensity changes. Since the storm's environmental flow was weak in its initial days, small convective processes were the predominant steering influence and were responsible for Allison's erratic track. Rainfall amounts between places only a few kilometers apart varied considerably because intense rainbands would develop and remain stationary over the same area for many hours. For instance, during a 10 hour period between 2300 UTC June 8 and 0900 UTC June 9, the Green's Bayou station in Houston recorded 26.93 inches (680 mm) of rain while a station just 7 km away recorded 9.57 inches (240 mm).

The NCEP operational models generally did a poor job of predicting Allison, especially in the first days of its existence. We chose to use the ARPS (Advanced Regional Prediction System) available from CAPS (Center for Analysis and Prediction of Storms) to model Allison. The very high resolution of ARPS (1 km) and the availability of post-analysis data helped us obtain a clearer picture of Allison than can be expected of the NCEP operational models.

The latest information regarding this research can be found at http://weather.ou.edu/~bobwein/Allison.html