LONG‐TERM RECONSTRUCTION OF WATER LEVEL CHANGES FOR LAKE ATHABASCA BY ANALYSIS OF TREE RINGS1

Publisher Website
Google Scholar
Abstract
The fact that dendrochronology can be a useful tool in water resources management is demonstrated by a study of past changes in lake levels. The study was necessitated by the recent (1967) closure of the gates on the W.A.C. Bennett Dam on the Peace River in British Columbia, causing a drop in water levels of Lake Athabasca, the lake levels control water level undulations in a series of smaller lakes which lie in the 1.5 million acre delta area along its western edge in Alberta. The W.A.C. Bennett Dam has by its regulation of the Peace River affected the river's historical role of creating a hydraulic dam during early summer flood stages thereby preventing outflow from Lake Athabasca and inducing annual inundation of the delta area. Because the ecololgy of the lake and adjacent delta region has depended on the now attenuated snow‐melt flooding from the Upper Peace River Basin, it became necessary to consider some means of artificially inducing this annual inundation. It was not known, however, what the long‐term water level changes were around which the present ecology had developed. Continuous historical lake level records exist only for the period 1935–1967. Was this period one of anomalously high or low water levels? Relatively old white spruce trees growing along natural levees of the channels in the delta region were found to contain tree‐ring records that reflected the water stages in the channels. Because the water levels in the channels could be correlated with lake levels, it was possible to use the three‐ring series to extend the known 33‐year record of lake level changes to 158 years. By using canonical analysis and 10‐day mean lake levels for three different subperiods in the 33‐year period of calibration, along with tree‐ring series from appropriately chosen stands of white spruce, reconstructions were made of the long‐term record for late May, early July, and late September. The reconstructed record shows that the May 21–30 lake levels have been three times as variable in the past as in the period of historical record (1935–1967), the July 11–20 levels twice as variable, and the September 21–30 levels 10% less variable. However, the mean water level for each of the three subperiods for the long‐term record is very close to the means for the period of historical record. The reconstructed record shows that, before the dam gates were closed, there was only one three‐year period (1866–1868) in which the lake levels were as low as they have been since closure of the gates (that is, 1967–1970). (Levels were nearly as low, however, during the period 1942–1945.