Using a coupled lake model with WRF for dynamical downscaling

The Weather Research and Forecasting (WRF) model is used to downscale a coarse reanalysis (National Centers for Environmental Prediction–Department of Energy Atmospheric Model Intercomparison Project reanalysis, hereafter R2) as a proxy for a global climate model (GCM) to examine the consequences of using different methods for setting lake temperatures and ice on predicted 2 m temperature and precipitation in the Great Lakes region. A control simulation is performed where lake surface temperatures and ice coverage are interpolated from the GCM proxy. Because the R2 represents the five Great Lakes with only three grid points, ice formation is poorly represented, with large, deep lakes freezing abruptly. Unrealistic temperature gradients appear in areas where the coarse‐scale fields have no inland water points nearby and lake temperatures on the finer grid are set using oceanic points from the GCM proxy. Using WRF coupled with the Freshwater Lake (FLake) model reduces errors in lake temperatures and significantly improves the timing and extent of ice coverage. Overall, WRF‐FLake increases the accuracy of 2 m temperature compared to the control simulation where lake variables are interpolated from R2. However, the decreased error in FLake‐simulated lake temperatures exacerbates an existing wet bias in monthly precipitation relative to the control run because the erroneously cool lake temperatures interpolated from R2 in the control run tend to suppress overactive precipitation. Key Points Unrealistic lake temperatures and ice result when interpolating from global data WRF coupled with the FLake model improves Great Lakes temperatures and ice cover Positive precipitation bias increases despite better representation of lakes