An 825-year long varve record from Lillooet Lake, British Columbia, and its potential as a flood proxy

Abstract

Lillooet River in southwest British Columbia has produced damaging floods many times during the past century. The floods are recorded in Lillooet Lake, into which the river flows, as anomalously thick clastic varves. In order to determine whether an 825-year long varve record obtained from 12 percussion and vibracores can be used as flood proxy, we compare river discharge records dating back to 1914 to the thickness of the varves deposited during the same time period. Correlations between varve thickness and a variety of historical discharge measures are low to moderate for the periods 1914–2004 (r2 = 0.37) and 1914–1945 (r2 = 0.40), but higher for the period 1946–2004 (r2 = 0.55). The best correlation (r2 = 0.55) is between maximum fall discharge and varve thickness during the most recent period (1946–2004). Varve thickness for the earlier period of hydrometric data (1914–1946), which is a time of rapid glacier retreat and warmer temperatures in British Columbia, is best explained with a discharge proxy combining nival runoff, glacier runoff, and maximum fall discharge. Landslides, glacier fluctuations, river dyking, artificial lowering of Lillooet Lake, as well as lag effects of storms are responsible for the considerable unexplained variance in the relation between discharge measures and varve thickness over the historic period. The cores contain many anomalously thick varves, some of which we attribute to previously dated prehistoric landslides in the watershed or to local landslides into the lake. We conclude that many historic and prehistoric floods are faithfully recorded as anomalously thick clastic varves, but that other processes operating in the watershed preclude using this record as a reliable paleo-flood proxy.