Limnol. Oceanogr., 44(3), 1999, 530–540

In the Great Lakes region, the observational evidence for climatic change has been primarily limited to changes in lake-ice conditions, with no long-term trends identified in water temperatures. Seven nearshore water intake sites (Bay City, Michigan; Green Bay, Wisconsin; Sault Ste. Marie, Michigan; St. Joseph, Michigan; Sandusky Bay, Ohio; Put-In-Bay, Ohio; and Erie, Pennsylvania) in the Great Lakes were chosen, and their data were examined for any climatic trends. Regression results on the annual mean temperatures showed varying support in favor of a warming trend at five of the seven sites. A new approach facilitated determination of the interannual variability in the timing of the 48C temperature of maximum density. Two of the three sites with data records extending back to the early part of this century (Sault Ste. Marie and Put-In-Bay, respectively) showed a 4and a 6-h yr21 rate of increase in the maximum potential duration of summer stratification (DSS). Over the time span of these two data sets, this equates to a 14and 18-d increase in the potential DSS, respectively. The rate of increase in the duration data was skewed, with most of the increase due to an earlier transition to springlike conditions. Finally, the data do not extend far enough back in time to know if these climatic trends are part of an unresolvable natural cycle or forced by anthropogenic activity. General circulation models (GCM) representing physical processes from both the atmosphere and ocean have seen widespread application in attempting to forecast the longterm implications of increasing atmospheric CO2. Observational evidence for global warming centers on long-term temperature records (Hansen and Lebedeff 1988; Wigley et al. 1989) and proxy climate variables, which infer historical global temperatures based on ice core analyses (Barnola et al. 1987), tree ring data (Feng and Epstein 1996), bore hole temperatures, and the retreat of mountain glaciers (Ott 1997). At the same time, new insights suggest that much of the observed warming may be partially explained by a reduction in the diurnal temperature range for most parts of the world (Easterling et al. 1997). Both modeling and data approaches toward understanding global processes are hampered by uncertainty. Grotch (1988) documents a 2–38C departure from global median temperatures with four different GCMs when they were used to simulate current conditions. Observational data also suffer from possible bias due to changes in land use patterns and inadequate spatial coverage. The limitations of GCMs on global