Abstract
Both the intensities of individual extreme rainfall events and the frequency of such events are important for infrastructure planning. We develop a new statistical extreme value model, the PGEV model, which makes it possible to use high-quality annual maximum series data instead of less well-checked daily data to estimate trends in intensity and frequency separately. The method is applied to annual maximum data from Vol. 10 of NOAA Atlas 14, dating from approximately 1900 to 2014, showing that in the majority of 333 rain gauge stations in the northeastern United States the frequency of extreme rainfall events increases as mean temperature increases, but that there is little evidence of trends in the distribution of the intensities of individual extreme rainfall events. The median of the frequency trends corresponds to extreme rainfall becoming 83% more frequent for each 18C of temperature increase. Naturally, increasing trends in frequency also increase the yearly or decadal risks of very extreme rainfall events. Three other large areas in the contiguous United States, the Midwest, the Southeast, and Texas, are also studied, and show similar but weaker trends than those in the Northeast.