One of the most profound ways the Laurentian Great Lakes impact the region's climate is by augmenting snowfall in downwind locations during autumn and winter months. Intense lake-effect snows can have large impacts on communities surrounding the lakes. Understanding long-term trends of the lakes' contribution to snowfall can help maximize economic gains, mitigate property loss and deaths, and increase awareness of regional impacts of climate change. The consensus among recent studies is that lake-effect snowfall increased by varying degrees during the 20th Century. The goals of the present study are to better understand and update the previous studies using the most recent quality-controlled observations available.

Daily snowfall observations from nine National Weather Service Cooperative Observer sites near Lake Michigan, found by a previous study to be suitable for long-term trend analysis, were chosen for this study. Sites were grouped into three cross-lake transects such that each transect had sites located west and east of the climatologic impact region of Lake Michigan's influence, i.e. �snowbelt�, and one site inside the lake's snowbelt. Months with 10% (3 days) or less missing daily snow observations or failed quality control checks, as well as only common data among three sites within each transect, are included in this study.

To describe climatic-scale surplus to local snowfall downwind of lakes, we introduce the terminology lake-contribution snowfall (LCS) which accounts for both lake-effect and lake-enhanced events. We demonstrate that considerable variability exists in 5-year periods of LCS east and south of Lake Michigan from 1920 to 2005. We found a general increase in LCS from the early 1920s to the early 1960s (southern areas) and late 1970s (central areas). LCS then decreased from these maxima through the early 2000s, indicating a distinct reversal of trends previously not reported by earlier studies. Potential reasons for this reversal will be discussed.