The focus of our study on the Hydrology, Stream Temperature and Sediment Impacts of Climate Change in the Skagit River Basin is to improve our understanding of the Skagit River earth processes systems using a coupled glacio-hydrology model and future climate projections. Our modeling work focused on the Sauk sub-watershed, but also include results for naturalized streamflow at Skagit River Hydroelectric Project reservoir locations (Ross, Diablo, Gorge) and at sixteen tributaries using future climate change scenarios. This project utilized data products from the Integrated Scenarios of the Future Northwest Environment project, which identified a core set of 10 global climate models (GCMs) of the Coupled Model Intercomparison Project Phase 5 (CMIP5; Mote et al., 2015) as the best performing models based on comparisons of observed 20th century climate of the Pacific Northwest. To simulate streamflow, we used the Distributed Hydrology Soil Vegetation Model (DHSVM) – a coupled glacio-hydrology model, and advanced the model code to include RBM stream temperature model, and new advances in sediment transport modeling. The model domain included the entire Skagit River basin at 150m digital elevation model (DEM) resolution, with nested models of 50m resolution of selected subbasins (Thunder Creek and Cascade Creek) that have the major glacier ice cover at their high elevations. The modeling steps included: (a) hydrometeorology bias correction, (b) spin up of the glacier model to develop realistic glacier cover in the glaciated uplands prior to watershed hydrology and streamflow predictions; (c) calibration of DHSVM using select model parameters and climate forcing bias correction in select subbasins; (c) model validation using historical streamflow observations; (d) projections of streamflow into the future using CMIP5 models; (e) bias-corrections of modeled streamflow to match observations based on monthly mean, (f) streamflow temperature modeling, and (g) suspended sediment modeling. Results include modeled streamflow for 90% exceedance probability flows in summer months. Validation and corrections to the glacio-hydrology model were conducted using empirical data (collected by North Cascades National Park), naturalized flows at reservoir locations (three reservoirs), and observed stream gauges (where and when available at 16 Skagit River tributaries). Future projections were calculated using GCMs for multiple overlapping fifty year periods starting from 2010 to 2099.

In glaciated high elevation basins, the current conditions of approximately 100 km2 of glacier ice are projected to decrease to less than 50 km2 by 2050. If global emissions stop increasing by the 2050s, it is likely that the highest elevation glaciers will continue to store pockets of ice and provide some glacier melt in the summer months. By the 2050s, the low flows will show a wide range of change conditioned on elevation. Low summer flows (10 year low flows; 7Q10) are projected to decrease 10-20% in low-elevation rain-dominated tributaries, (e.g., Rinker, Clear Creek), and ~30% in mid-elevation mixed rain and snow tributaries (e.g., South Fork Sauk). High flows and peak event are projected to slightly increase by 2050, with statistically significant peak events (100 year flood) expected to increase 20-30% for a conservative climate change scenario (RCP 4.5 Ensemble), and +40% for other estimates (RCP 8.5 Ensemble). Historic Maximum Daily Maximum Temperature during summers were 17.8 oC, 19.2 oC and 17.4 oC at the Sauk R. above Suiattle, the Sauk River near Darrington, and the White Chuck River, respectively, and projected to increase by 2-3 oC for all scenarios and sites; the Sauk River near Darrington shows the largest increase in MDMT and the White Chuck River the least. Historical suspended sediment load is ~ 40% lower than the 5-year mean annual historic yield. In two future scenarios, using data from 10 future models, suspended sediment load increases over time due to changes in hydrology, with moderate climate models giving up to threefold increases in suspended load.