Modeling the history of Lake of the Woods since 11,000 cal yr B.P. using GIS

Abstract

Because of differential isostatic rebound, many lakes in Canada have continued to change their extent and depth since retreat of the Laurentide Ice Sheet. Using GIS techniques, the changing configuration and bathymetry of Lake of the Woods in Ontario, Manitoba, and Minnesota were reconstructed for 12 points in time, beginning at 11,000 cal yr B.P. (∼9.6 14C ka B.P.), and were also projected 500 years into the future, based on the assumption that Lake of the Woods continued to have a positive hydrological budget throughout the Holocene. This modeling was done by first compiling a bathymetric database and merging that with subaerial data from the Shuttle Radar Topography Mission (SRTM). This DEM file was then adjusted by: (1) isobase data derived from Lake Agassiz beaches prior to 9000 cal yr B.P. (∼8.1 14C ka B.P.) and (2) modeled isostatic rebound trend analysis after 9000 cal yr B.P. Just after the end of the Lake Agassiz phase of Lake of the Woods, only the northernmost part of the basin contained water. Differential rebound has resulted in increasing water depth. In the first 3000 years of independence from Lake Agassiz, the lake transgressed >50 km to the south, expanding its area from 858 to 2857 km2, and more than doubling in volume. Continued differential rebound after 6000 cal yr B.P. (∼5.2 14C ka B.P.) has further expanded the lake, although today it is deepening by only a few cm per century at the southern end. In addition, climate change in the Holocene probably played a role in lake level fluctuations. Based on our calculation of a modern hydrological budget for Lake of the Woods, reducing runoff and precipitation by 65% and increasing evaporation from the lake by 40% would end overflow and cause the level of the lake to fall below the outlets at Kenora. Because this climate change is comparable to that recorded during the mid-Holocene warming across the region, it is likely that the area covered by the lake at this time would have been less than that determined from differential isostatic rebound alone.