Extreme Precipitation Events Observed by the U.S. Climate Reference Network and the NOAA Hourly Precipitation Dataset

Differences in temporal and spatial resolutions, gauge shielding, gauge type, and minimum reported amounts between precipitation datasets make them harder to compare, especially for studies of relatively rare events like extreme rainfall. The United States Climate Reference Network (USCRN) provides triple-sensor 5-minute precipitation records at 114 sites across the conterminous United States that are intended to serve as reference points when comparing measurements between networks. This study evaluates the ability of a high quality, measurement-dense, but sparsely distributed network such as the USCRN to capture precipitation extremes in comparison to another more densely distributed network, NOAA’s Hourly Precipitation Dataset (HPD), for the same period of record. USCRN precipitation totals were computed for rolling duration intervals and compared with thresholds based on gridded NOAA Atlas 14 precipitation frequency estimates. Exceedances of the NOAA Atlas 14 precipitation frequency for one to five-year return intervals over a full range of durations from 5 minutes to multiple days were determined for the beginning of the period of record through 2022. Throughout the NOAA Atlas 14 coverage area, a larger number of exceedances per year than predicted by the return interval were found for durations in the five-minute to one-day time frame. This methodology was applied to NOAA’s Hourly Precipitation Dataset (HPD) for the same period of record as the USCRN network, with one set of results comprised of all HPD stations and another set of results for only the closest station match within 100 km of a USCRN station. The analysis highlighted the influence of station sparsity on the USCRN’s ability to capture precipitation extremes for time intervals longer than one day, although the USCRN and HPD results were overall similar from 15-minute to one-day intervals. This suggests that USCRN can be put to further use as a reference for studies evaluating changes in the frequency of sub-daily extreme precipitation events. The capture of a larger than expected number of exceedances for both networks, especially for shorter duration events, highlights the need for updates to the Atlas 14 extreme precipitation event thresholds to account for climate non-stationarity. This will be remedied by the anticipated release of NOAA Atlas 15 by the National Weather Service.