The Potential of Multiple Discriminant Analysis for seasonal snowfall forecasting in the continental United States
Dr. Daria Kluver, Assistant Professor
Department of Earth and Atmospheric Sciences
Institute for Great Lakes Research
Central Michigan University
10:30AM, Lake Superior Hall
Great Lakes Environmental Research Lab
4840 South State Road
Ann Arbor, MI 48108
Multiple Discriminant Analysis (MDA) is used to develop statistically based seasonal forecast models for several snowfall regions across the contiguous United States. Forecast skill ranges from correct forecasts near 70% to greater than 90% of the time, for broad snowfall categories. For winter snowfall amounts and frequencies, previously documented relationships are corroborated (with the PNA, NAO, and ENSO), and new relationships are identified, such as with Eurasian snow cover extent. On a monthly time scale, the variables are less consistent, however, a few regions clearly show a progression of influential teleconnection patterns changing throughout the winter season. An example is the Western United States, which is more highly influenced by the PDO earlier in the snow season and the PNA later in the season. This research not only verifies previous studies connecting large-scale teleconnection patterns with snowfall, but also supports recent work positing a relationship between northern latitude surface cover (snow and sea ice) and its relationship to North American snow.
Or, reserve your Webinar seat now at:
https://www2.gotomeeting.com/register/778831122














