Flash Flood Statistical Models for Predicting the Stage Heights on the Rocky Branch Creek in Downtown Columbia, SC

The Rocky Branch Creek watershed in downtown Columbia, South Carolina is prone to flash flooding on short time scales in convective heavy rain events. Rainfall runoff is enhanced due to both a significant elevation gradient in the basin and impervious urban land cover. The creek flows through a highly urbanized and commercial district known as Five Points and the southern portion of the University of South Carolina campus before it enters the Congaree River. Using upstream observed rainfall and response time as predictors, multiple linear regression models were developed to predict the stage heights at two locations on the creek: Main Street at Whaley Street and Pickens Street. The purpose of the models is to aid National Weather Service forecasters in flash flood warning decision-making by providing timely stage height predictions in near real time. A computer application was developed to ingest the predictor data in a timely manner and display forecast crest stages on the Rocky Branch Creek to the forecasters. It is hoped that this research will lead to an increase in flash flood warning lead times for the city of Columbia.