11.1 Evaluation of Lake Process Simulations within the Weather Research and Forecasting Model

Jin, Jiming and Hongping Gu, Utah State University, Michael B. Ek, Environmental Modeling Center, National Centers for Environmental Prediction

A one-dimensional physically-based lake model was coupled with the Weather Research and Forecasting (WRF) model to improve lake-effect precipitation simulations for the Great Lakes region. Initial simulations with the WRF-Lake model show that the seasonal cycle of lake surface temperature (LST) was greatly exaggerated especially for the deep lakes such as Lake Superior. The exaggerated LST seasonal cycle results from insufficiently simulated turbulent mixing in the lake. A series of sensitivity tests with WRF-Lake were performed to optimize the eddy diffusivity that controls water mixing in the lake scheme and is a function of surface wind and roughness length. The coupled model is able to realistically reproduce the LST seasonal cycle with the optimized eddy diffusivity. In addition, we performed multi-year simulations at 10 km resolution for the period of 2003-2008 forced with 32 km resolution North American Regional Reanalysis data to validate the coupled model.  The results reveal that the simulated LSTs are in very good agreement with surface buoy observations and Moderate Resolution Imaging Spectroradiometer satellite data. The realistic LST simulations also generate more accurate lake-effect precipitation when compared with that produced by the release version (3.2) of WRF without a lake scheme.