RESUMO
The phenomenon of non-significant trends in ice-off date under a warming climate was quantitatively explained by three efforts: exploring possible driving factors where possible and defining new factors to represent snow conditions, identifying the contributing factors through correlation and trend tests, and evaluating relative contributions through partial Mann-Kendall method. Why the ice-off became only slightly earlier over 62 years at Lake of Bays has been satisfactorily assessed: the increased winter temperature, increased total rain and decreased days of snow on ground acted as three promoting drivers to earlier ice-off date, but their promoting functions were effectively offset by adverse changes in four other factors (snowfall slope, precipitation slope, snowpack slope, and last day of snow). The ice-off date at Lake Nipissing did not have a significant trend over 58 years, although there were five factors contributing to the ice-off decline without sufficient offsetting, suggesting that the ice-off of this lake may not be sensitive, or basically elastic, to the climatic variation stressor. Relative contributions of drivers as calculated helped explain how much they contributed to ice-off trends or how much they offset the influences.