Peaks-over-Threshold Analysis for Watershed-Average Precipitation and Streamflow in the Baltimore Metropolitan Area

A changing climate disrupts many established interactions between the atmosphere and the water cycle. Warmer temperatures allow the air to hold more water vapor, enhancing moisture convergence and rain rates, thereby increasing the intensity of storm systems. Nonstationary probability analysis of radar precipitation data completed by Smith et. al, (in review) suggests that there is an increasing trend of extreme rainfall events over 22 years of record. The precipitation data combined reflectivity-based rainfall fields from 2000 - 2015 (HydroNEXRAD) and operational polarimetric rainfall fields from 2012 - 2021 (Digital Precipitation Rate product) and were bias-corrected using a network of rain gauges in the Baltimore metropolitan area.

For a set of gauged watersheds within the spatial domain described by Smith et al. (in review), we examined watershed-average versions of the same bias-corrected 15-minute radar rainfall time series for the period of April through September during the years 2000 to 2020. Nonstationary Peaks-Over-Threshold (POT) analysis was used to assess time trends in maximum rainfall accumulation for specified duration and exceedance probability. A similar nonstationary Peaks-Over-Threshold analysis was carried out with the four largest peak flows in each year for the U.S. Geological Survey stream gages at the outlets of those watersheds. Rainfall and streamflow results were compared to assess whether time trends in peak flow were similar to those observed for precipitation maxima over the same watershed. Our results suggest that trends in rainfall and runoff extremes over the sampled time period are similar and appear to be trending upwards in most of the watersheds studied.