Winter Ice-Wave Modeling With WAVEWATCH III in Lake Erie

Operational wave forecasts for the Great Lakes originate from the NOAA Great Lakes Waves Unstructured version 2 system. The model uses a simple ice blocking (IC0) parameterization for ice-wave damping, so ice-covered portions of the lakes are treated as land in the modeling system. Although simple and effective, the simple block can impede forecasting by eliminating wave forecast guidance from areas with thin or partial ice cover. We evaluate 12 ice-wave damping parameterizations within WAVEWATCH III (WW3, version 6.07.1) for Lake Erie, by comparing model results against wave observations made at several locations using moored acoustic wave and current profilers during the winters of 2010-2011 and 2012-2013. The comparisons show that the IC4M4 module performs the best among 12 ice modules with a root mean square error (RMSE) of 0.32-0.39 m and a root bias of -0.06 to -0.11 m, outperforming the existing IC0 parameterization (RMSE: 0.46-0.59 m; bias: -0.23 to -0.34 m) during the 2010-2011 analysis year. WW3 ice modules are mostly derived from measurements and studies of the Arctic and Antarctic Ocean. The dominant wave frequency is about 0.05 similar to 0.10 Hz in the Arctic Ocean compared to 0.15 similar to 0.2 Hz in the lake. Thus formulas built on frequency based on the studies from deep oceans may not be suitable for the shallow lakes because they cause too much damping. Although the IC4M4 ice module is from the study of the Antarctic Ocean , the wave attenuation formula based on incoming wave height is also suitable for Lake Erie.