The accurate and timely prediction of lake-effect snowstorms to the lee of the Great Lakes is of critical importance to commerce and industry in New York State. Large-scale operational numerical models, such as the Nested Grid Model (NGM), are not capable of predicting these mesoscale events. However, data from the large-scale models can be used to initialize mesoscale models developed to simulate such storms.
The mesoscale model used in this project was developed by R. Ballentine at SUNY Oswego and tested by faculty and students at SUNY Oswego and SUNY Brockport. The grid has a 10-km horizontal resolution and covers the Lake Ontario region with a mesh of 45x31 grid points. For these experiments, the model uses 9 prediction layers. In one mode of operation, mandatory and significant level sounding data and standard-level wind data from Buffalo are used to initialize the model and to provide inflow boundary conditions. In the second mode of operation, initial and forecast data from the 11 lowest layers of the NGM (interpolated to Buffalo or Syracuse) are used to construct initial and boundary data. This mode is much simpler for forecasters to use; minimal effort is required to carry out a 24-hour simulation.
Preliminary tests using data from two lake-effect storms indicated that the Oswego model predicted the approximate location and time of formation of the snowbands, as well as the correct movement and changes in intensity. A set of case studies will be assembled by the forecasters at Buffalo to test the accuracy of the Oswego model in predicting the onset time, location, movement, and precipitation rate of lake-effect snowbands. Once a technique has been chosen, a protocol will be established for using the Oswego mesoscale model in an operational setting.