Comparing the Impact of Differing Glacier Contribution Assessments on Modeled Discharge and River Temperature Across Southeast Alaska, USA

Southeast Alaska represents a highly variable hydrologic environment. Headwaters are dominated by alpine glaciers and larger river basins envelop significant glacier area. Daily estimates of streamflow and river temperature provide useful insight into regions undergoing rapid change. Here we present a river temperature model framework that adapts results from two different streamflow models applied in Southeast Alaska. We compare streamflow inputs from two land surface models: 1) a coupled simulation of RASM-CTSM at 4 km and 2) an offline run of RASM-WRF Hydro Glacier at 1 km. The land surface models primarily differ in glacier treatment, RASM-CTSM uses a traditional five-layer snow model to account for changes to glacier processes, while the RASM-WRF Hydro Glacier model applies the dynamic CROCUS snow model for improved glacier simulations in river basins with glacier presence. The RASM-WRF Hydro Glacier employs the three-layer snow model from the NoahMP land surface component to account for non-glaciated basins. Runoff from both models is routed to streamflow and the River Basin Model (RBM) is used to produce daily river temperature data. The models were run for a historical (1991-2020) period, including a parameter sensitivity test for river temperature, and then applied to a mid-century (2035-2064) climate scenario. We validated the river discharge and temperature estimates against USGS observations at seven basins. We assess model performance in terms of the Kling-Gupta Efficiency metric and its three components, correlation, volume bias, and variance bias. We also assess an overall percent bias between modeled and observed data.  In historical simulations, the RASM-WRF Hydro Glacier performs better than the RASM-CTSM model across all metrics in glaciated basins, while producing identical results in non-glaciated basins, indicating that the RASM-WRF Hydro Glacier model produces more accurate results and should be employed for future simulations. Estimates of future changes in discharge and stream temperature, together with glacier contributions to runoff, provide a holistic assessment of hydrologic system change in Southeast Alaska, useful for land management. The results provide the baseline for future investigations and modeling studies of the impacts of projected changes on fish habitats and hydrological trends in this glacier-dominated landscape.