Validation of gridded precipitation datasets for flood-typing in select conterminous U.S. basins

Gridded precipitation datasets are required for flood-typing historical annual peak streamflow events in basins across the Conterminous United States. Selected gridded precipitation datasets were validated over the period 1981–2013 through comparisons with gage data from the NOAA Global Historical Climatology Network daily (GHCNd). The ability of each gridded dataset to capture the spatiotemporal characteristics of daily precipitation, including multi-day extremes over six selected regions, was assessed using the Kling-Gupta Efficiency metric and its component statistics. Overall, the Parameter-elevation Regression on Independent Slopes Model and Livneh-unsplit were found to best match the spatiotemporal variability of the GHCNd precipitation data, including extremes. The Analysis of Record for Calibration was found to be the third best-performing dataset in most regions except in the western U.S. The performance of reanalysis datasets evaluated appears to be poor compared to gage-based datasets. The reanalysis datasets might not be able to skillfully capture precipitation amounts at the correct location and time. Gage- and radar-based datasets were found to have relatively small biases (within +/-10% on an annual basis), while reanalysis datasets were found to have larger positive apparent biases, especially in winter and spring in most regions. It is possible that the apparent overestimation of winter and spring precipitation in the reanalysis datasets might reflect snow undercatch at gages especially in the central U.S. An overall deterioration of performance for correlation and/or variability was also observed for the summer season compared to other seasons in the reanalysis datasets. Various precipitation datasets might need to be used for flood-typing during different periods from the late 19th century to present. Datasets from different sources have different biases and errors and might have to be homogenized using downscaling and bias-adjustment methods. Alternatively, precipitation thresholds used in some flood-typing schemes might have to be adjusted as a function of time.