APPLYING DYNAMICALLY DOWNSCALED CLIMATE PROJECTIONS TO A MOUNTAINOUS WATERSHED IN WESTERN WASHINGTON TO ESTIMATE FUTURE PEAK FLOWS

Climate change is projected to increase river flooding in the Puget Sound region of Washington State by reducing snowpack and yielding more intense storm events. Hydrologic models paired with meteorological inputs (forcings) that are downscaled from global climate models (GCMs) are a robust tool for simulating watershed response to projected atmospheric conditions. Dynamical downscaling offers a way of increasing the temporal and spatial resolution of forcings in a physically-based manner. More specifically, dynamically downscaled forcings capture the intense rainfall events that drive flooding in Western Washington - an application for which statistically-based forcings are not well suited. Here we apply climate projections, dynamically downscaled using the Weather Research and Forecasting model (WRF), to the Stillaguamish watershed in northwest Washington State using the physically-based Distributed Hydrology Soil Vegetation Model (DHSVM). We simulate hourly streamflows for 12 high emissions scenarios (i.e., Representative Concentration Pathway 8.5) throughout the 1,700-square-kilometer basin from 1970 through 2099. Our projections indicate that as the climate warms, snowpack will recede to higher elevations and the basin will shift from being mixed rain-snow to rain-dominant, leading to an increase in average winter streamflows and a decline in spring and summer streamflows. To estimate future changes in peak flows over time, we use conventional flood frequency analysis techniques on simulated streamflows. Our results indicate that peak flows will increase by about 10-30% across multiple return intervals and flow durations by the 2080s. This shift in streamflow timing and magnitude has implications for flooding within the developed lowlands of the watershed, and for threatened salmonid populations which are culturally and economically critical to the region.