Response of Lake Superior to mesoscale wind forcing: A comparison between currents driven by QuikSCAT and buoy winds

The satellite scatterometer QuikSCAT wind field has been available every 12 hours on Lake Superior since 19 July 1999. The wind data cover most of the interior area of the lake with a spatial resolution of about 25 km. Driving the three-dimensional Lake Superior circulation model by the QuikSCAT winds, we resimulated the 1999 seasonal variability of currents in Lake Superior. A comparison was made with our previous simulation results, which relied on the wind field interpolated from moored buoys and land-based meteorological stations. The model driven by QuikSCAT winds improved the simulation of the spatial coverage area of the cold band during upwelling favorable wind events and the current jet during downwelling favorable wind events observed along the Keweenaw coast in July-October 1999. A statistical analysis shows that these improvements were mainly reflected in the low-frequency variation of the long-shore current, even though the overall deviation between computed and observed surface temperature and currents was measurably reduced. This study suggests that the wind field constructed from either moored buoys plus land-based meteorological stations or QuikSCAT is not sufficient to provide a reliable and accurate simulation of coastal currents and stratification in Lake Superior. A mesoscale meteorological model assimilated with observed winds on all the available weather measurement sites or stations or QuikSCAT is needed to provide an accurate meteorological forcing for the Lake Superior physical model. Copyright 2004 by the American Geophysical Union