Numerous studies suggest the importance of lake and river ice break-up as an index of climate variability and change, especially as related to temperature and snow cover (Palecki and Barry, 1986; Schindler et al., 1990; Robertson et al., 1992; Assel and Robertson, 1995; Anderson et al., 1996; Wynne et al., 1998; Magnuson et al., 2000). Records of lake and river ice can be used to independently evaluate changes of temperature and, to some extent, snow cover. Like other proxy measurements they have limitations, and are subject to their own time-dependent biases such as changes in observers and protocols related to the identification of "ice on" and "ice off" conditions. Larger lakes often have the best records, but are often located near human settlements which can affect the homogeneity of the record, e.g., associated cooling water discharges and urban heat islands, so care is needed to select suitable lakes.
A recent analysis has been made of trends in 39 extensive Northern Hemisphere lake and river ice records over the 150-year period from 1846 to 1995. Ice break-up dates now occur on average about nine days earlier in the spring than at the beginning of the record, and autumn freeze-up occurs on average about ten days later (Magnuson et al., 2000). Only one of the 39 records, in Japan, showed changes that indicate a slight cooling.
Other reports in this collection