Available lake ice data for some freshwater lakes in the United States, not including the Laurentian Great Lakes, were assembled. The principal data were: ice~in (freeze over) dates, ice~off dates, ice thickness and ice cover duration on lakes.. The data were analyzed and interpreted to look for trends over time and relationships with lake characteristics and geographical location (weather). A multiple variable linear regression was performed for the ice~in dates, ice-out dates and ice cover duration. Using the regression equations, projections under a 2xC02 climate scenario were made. A trend (change) over recent periods of up to 140 years duration was most apparent for ice-off dates. The trend was toward earlier ice-off dates. For ice-on dates the trend was positive for some lakes and negative for others. The data analysis showed that ice-off dates, ice~on dates, ice thicknesses and ice cover durations on lakes depend on geographic latitude and elevation as crude measures of the climate at a lake. Mean seasonal (3-month average) values of air temperature were also used as a more direct measure of local climate at a lake. Three of the ice cover parameters were shown to depend on these mean seasonal temperatures. Ice-off was related to the April through June mean air temperatures. Ice-on was related to the October through December mean air temperatures. In addition to weather, ice-on' dates also depend on the mean depth of a lake, as was shown in other previous studies. Ice-on is delayed by 1.3 to 1.4 days for every additional meter of mean lake depth. Ice cover duration was related to mean annual air temperature. For a 2xC02 climate scenario, the analytical relationships obtained from the data project a delay in ice-on dates by 15 to 27 days, an advance in ice-off dates by 18 to 23 days and a shortening of ice cover duration by 37 to 50 days.
|Original language||English (US)|
|State||Published - Aug 1997|