Federally regulated nuclear power plants and dams are designed to withstand extreme storm rainfall events. Some Federal agencies design their structures to the Probable Maximum Precipitation (PMP), the theoretical upper limit of precipitation for a specified area and duration. Other agencies use a risk-based approach where the probability of occurrence of precipitation events (e.g. 1 in 1,000-year storm) is determined using statistical methods, and the structure is designed according to the appropriate risk. The risk-based approach still considers PMP in their calculations; PMP is the operational upper limit of the precipitation frequency curve.

For the United States, the most current methodology used to determine PMP is provided in Hydrometeorology Reports (HMRs) published by the National Weather Service. For example, HMR 51 provides guidance to determine design rainfall estimates for the eastern United States. However, HMR 51 was published in 1978 from a storm database that was last updated in the early 1970s.

There is a need to update the storm database and to revisit PMP. Federal agencies continue to utilize the outdated database and HMR reports. Observed isohyetal patterns from storms are overlaid on the drainage basin of Federal dams to examine rainfall-runoff relationships. Temporal patterns of observed rainfall are input into hydrologic models to accurately represent the time distribution of precipitation. Documented storms occurring at several frequencies assist in more completely filling in the precipitation frequency information for the risk-based framework. Most importantly, examination by the Bureau of Reclamation of extreme precipitation events occurring recently suggests that several events have the potential to meet or exceed current PMP estimates (e.g. Hurricane Floyd in 1999; Hurricane Fran in 1996; storm event in northern Idaho in November 2006, storm event in central/northern California in October 2009).

It is problematic if storms are producing precipitation in excess of PMP, especially since Federal nuclear power plants and dams consider PMP to be the absolute upper threshold. If the spillway of a dam is designed and constructed to withstand the flood caused by the PMP storm, but the PMP is too low, then the actual PMP storm would overtop the dam and possibly result in dam failure. In the risk-based approach, if the precipitation frequency curve is calculated using an outdated storm dataset, then there is potential for the curve and subsequent risk estimates to be biased too low. Based on recent observations of extreme precipitation events, there is a definite need to update the extreme storm database and for PMP to be revisited.