The impact of deep glacial water diversions from a hydroelectric reservoir in the thermal dynamics of a sub-arctic lake

Abstract

Interbasin water diversions associated with hydroelectric power operations can influence the physics and water quality of downstream receptor lakes. Little is known about the impact of such diversions in sub-arctic and arctic lakes, which are characterized by weak summer stratification and a high relative contribution of cold and highly turbid tributaries of glacial origin. From 2003 to 2007, Lake Lagarfljót (53 km2, 65°N) experienced a series of changes in its natural hydrological conditions as part of the 690 MW Kárahnjúkar hydroelectric project. The most significant change was the damming and diversion of a second glacial river into the lake, reducing its hydraulic residence time by a factor of three, and increasing its background turbidity levels by one order of magnitude. Here we developed a three-dimensional hydrodynamic modeling approach to assess the impacts of this project on the Lake’s physics. We accounted for the uncertainty in pre-dam suspended-solids data and conducted a stepwise assessment of each alteration in the hydrological forcing and lake background conditions. The results revealed abrupt changes in the lake's summer thermal dynamics. Advective heat fluxes now have a similar magnitude to atmospheric heat exchanges or even dominate the lake’s heat balance during the second half of the thermal stratification period. The diversion of deep water from Hálslón reservoir led to a net cooling of both the surface (median ≈1°C) and deep (median ≈0.6 °C) layers of the lake during the thermal stratification period, to a ≈ 45 % decrease in the average summer water column Schmidt stability and a two-week decrease in the duration of the thermal stratification period. The lower stability of the water column and deep river-plume intrusions now facilitate nearly complete hypolimnetic water renewal during the thermal stratification period. The proposed modeling approach, which deconstructs the lake’s conditions from post- to pre-operational stages, is suitable for lakes with limited pre-operational in-situ data.