CS38B - Concurrent Session 38B: Probable Maximum Precipitation (PMP): From Storm Based Estimates to Numerical Weather Modeling and Probabilistic Estimates

Defined at the time as “the depth of precipitation which can be reached but not exceeded under known meteorological circumstances”, Maximum Possible Precipitation (MPP) was developed starting in the 1940s by the U.S. Weather Bureau (HMR 23, 1947). The term Probable Maximum Precipitation (PMP) replaced MPP in the 1950’s (HMR 33, 1956). PMP was based on data from historic extreme rainfall events together with procedures to increase that rainfall to its plausible maximum. Often referred to as the storm-based method, this approach has incorporated new data (e.g. radar precipitation, longer records for both precipitation and dew points) along with improvements in analysis procedures but has otherwise remained essentially unchanged. Recently, high-resolution numerical models have become increasingly capable of reproducing observed rainfall patterns. Model-based PMP approaches have been under development for several years which essentially follow the storm-based logic but use the model equations, which represent our current understanding of atmospheric physics, to replace some of the assumptions required by earlier methods and explicitly determine rainfall amounts. Building on these developments and to address known shortcomings (limited data and understanding of meteorology at the time the methods were developed, meaning assumptions had to be made which may cause inconsistencies among studies) of the current definition of PMP, a June 2024 report from the National Academies of Science (NAS) on “Modernizing PMP” recommends a probabilistic definition of PMP using extreme precipitation from large ensemble atmospheric model simulations. This paper reviews storm-based methods for PMP determination, emerging alternatives using numerical modeling and summarizes the recommendations from the recent NAS report.