Classification of Lake Michigan snow days for estimation of the lake‐effect contribution to the downward trend in November snowfall

Abstract

AbstractThe substantial impact of Lake‐effect snow in the Laurentian Great Lakes has led to interest in the impact of climate change on snowfall in the region. A recent assessment of Lake Michigan snowfall revealed a marked decrease in November snowfall since the 1950s, associated with a warming‐induced reduction in the fraction of precipitation days occurring as snowfall. Herein, in order to identify the trend contribution from Lake‐effect snowfall, snow days from the November 1950–2012 study period are classified as primarily System or Lake‐effect, with additional options of Insignificant, Both, Remnant, and Unclear. The classification is based on the snowfall distribution and visual map inspection for synoptic‐scale forcing, and results are compared with an objective classification based on clustering of daily snowfall. The regional snowfall patterns of Lake‐effect and System snow days are markedly different, with Lake‐effect snow days exhibiting a clear Lake‐effect signature downwind of Lake Michigan. The larger‐scale environments also differ, with much colder conditions in the Great Lakes and higher sea‐level pressure in the Great Plains on Lake‐effect days. With snowfall projected onto the classifications, the decrease of snowfall east of the lake is attributable to both reductions in Lake‐effect and System snowfall, while System snowfall changes are dominant west of the lake. The trends are consistent with the sensitivity to regional temperature, as well as an increasing prevalence of rain reports (within other sub‐regions) during snow days. Results based on the objective classification are largely congruent.