Comparison of seasonal snowfall among the Great Lakes snowbelts

Seasonal snowfall varies among the Great Lakes snowbelts due to differences in latitude, cyclone paths, interaction between Lake water and overlying air, spatial orientation of each Lake, surrounding terrain, and perhaps different impacts of seasonally persistent long-wave tropospheric patterns. This work examined seasonal (October-May) snowfall of each Great Lake snowbelt (averaging data of several stations within each snowbelt) for the thirty-eight year period 1952-1990.

Time trends of seasonal snowfall varied among the Great Lakes snowbelts for the period 1952-1990, in contrast to past studies that found similar time trends among the snowbelts. Linear regression analysis of seasonal snowfall showed different time trends between northern and southern snowbelts; polynomial regression analysis showed a more definitive northwest-to-southeast geographic regionalization of seasonal snowfall time trends among the snowbelts. The six snowbelts were divided into three groups based on polynomial regression analysis of seasonal snowfall: (1) the Northwestern Snowbelts (Lake Superior and northern Lake Michigan) each showed an increase in seasonal snowfall for the period (linear trend, slope>0); (2) the Central Snowbelts (southern Lake Michigan and Lake Huron) each showed an increase in seasonal snowfall during the first half of the period and a decrease during the second half of the period (quadratic trend); (3) the Southeastern Snowbelts (Lake Erie and Lake Ontario) each showed no trend toward increase or decrease in seasonal snowfall for the period (linear trend, slope=0). Strength of correlation of seasonal snowfall among the snowbelts also reflected the northwest-to-southeast geographic regionalization.

As a final measure, winter-to-winter variability of seasonal snowfall was compared among stations within each snowbelt. Single-linkage classification and correlation analysis were employed to cluster together stations that showed similar variability. Geographic boundaries between clusters of stations were mapped, depicting subregions of a snowbelt. Four of the snowbelts showed division into subregions; two of the snowbelts did not show subdivision. The division into subregions of the Lake Superior, northern Lake Michigan, southern Lake Michigan, and Lake Erie snowbelts showed that a single station should not be used to represent the seasonal snowfall variability of the entire snowbelt (1952-1990); whereas, nondivision of the Lake Ontario and Lake Huron snowbelts showed that a single station could be used to generally represent the seasonal snowfall variability of the entire snowbelt (1952-1990).

Differences in seasonal snowfall among the Great Lakes snowbelts for the period 1952-1990 indicated that climate changes may not have been homogeneous among the Great Lakes snowbelts, or that seasonal snowfall response varied among the snowbelts. Spatial and temporal variability of seasonal snowfall was exemplified further in the division into subregions of four of the Great Lakes snowbelts.