Risk analysis of probable maximum precipitation estimates

Abstract

Flooding, caused by extreme precipitation, including probable maximum precipitation (PMP), causes considerable loss almost yearly in many parts of the U.S., such as Texas and Louisiana. A common method for estimating PMP is the statistical method but it involves sample mean, standard deviation, and frequency factor each of which can be considered as a random variable. As the uncertainty in the PMP values is due to the uncertainty in sample mean, sample standard deviation, and frequency factor, the relative contribution of each random variable to the total uncertainty was determined. The best-fit probability distribution was found and the hazard rate was calculated for different probability distributions. Using the best-fit probability distribution, design risk estimates along with probability bounds on the PMP values were determined. Considering the damage (in U.S. dollars) a PMP event can cause, risk analysis of extreme precipitation was done. The damage due to a single PMP event of 12-hour duration can be as high as $2 billion in Harris County, Texas.