Satellite derived Great Lakes sea surface temperatures (SST's) are compared to buoy observations during the fall and early winter season. Results of the comparison indicate that, at times, significant differences occur between buoy observed SST's that are directly measured by data buoys and those remotely sensed by satellites. Preliminary results suggest that overall, the satellite derived SST's showed a warm bias during this time frame. The warm bias is most likely due to the problems associated with remotely sensed SST retrievals when the Great Lakes are masked by thick cloud cover.

Currently, operational mesoscale NWP models are initialized off the satellite derived SST data set. It is hypothesized that an SST data set that is not updated due to extended periods of cloud cover could produce significant differences in model forecast snowfall for lake effect snow events. Sensitivity test were run using the workstation Eeta model to determine what impact warming or cooling the lakes by a couple of degrees would have on forecast snowfall. The results showed that snowfall totals varied widely based on slight changes in SST fields over the entire Great Lakes region.

This presentation compares satellite derived vs. buoy observed SST's over a 5-year period each fall and early winter. Next, select lake effect snow events are re-run with the workstation Eta model, first with the SST's warmed by C, then cooled by C. The resulting differences in lake effect snowfall span a range from insignificant accumulations to those that produce warning criteria snow amounts. Hopefully, this study will point out to the operational forecaster the impact that poorly initialized SST data due to persistent cloud cover may have in the model forecasts of lake effect snow. It may also lead to further research in developing other methods to improve the analysis of Great lakes SST's. As a result we should also see a similar improvement in model forecasts of lake effect snow.