Richard N. Palmer
This research uses projected changes in future precipitation to calculate the changes in the magnitude, frequency, and timing of streamflow, particularly peak annual flows, in the U.S. Northeast through 20 representative watersheds. Temperature and precipitation data on a 3-hourly time step from five climate projection from the North American Regional Climate Change Assessment Program (NARCCAP) are downscaled and bias- corrected using the Spatial Downscaling and Bias Correction (SDBC) method. These data are used to force a regional hydrological model (WRF-Hydro) to estimate daily future streamflow. The changes in magnitude at various return intervals of peak flow events are determined through the comparison between peak annual flow values during the historical period (1968-1999) and during the future period (2038-2070). The frequencies of high daily streamflow in each month are evaluated using a peak-over-threshold (POT) analysis of both high precipitation days and high streamflow days to understand the correlation (if any) between the two in this particular region. The results indicate an overall average increase of 10%, 15%, and 18 % in the 2-year, 50-year, and 100-year return interval magnitude of peak floods in the U.S. Northeast region, respectively. The POT analysis reveals increases in the number of extreme precipitation days during the winter months (DJF) which is expected to result in higher peaks in streamflow. This correlation is less apparent during the summer (JJA) months, suggesting a significant role of soil moisture and snowpack. The degree of climate change impacts vary by season, lending to differing flow regimes. Shifts in the seasonality of future peak flow events are observed in the results and further explain the changes in flood magnitudes and frequencies. They suggest similar trends in the inundation processes that directly influence soil moisture; consequently exacerbating flood and drought events that require new adaptation and mitigation strategies in the region.