All Lake Erie lake-effect days for a 10-year period prior to the 1976–77 snowfall season were utilized in the development of an operational lake-effect snowfall forecast model. Upper air and surface observations were combined with overtake data and analyzed, using stepwise multiple-discriminant analysis. A nine-predictor mesoscale forecast model resulted from this statistical test and its performance was evaluated during the 1976–77 and 1977–78 snowfall seasons. The results of this evaluation indicate that it is possible to predict six intensifies of the Lake Erie lake-effect snowstorm using this mesoscale model.