Sensitivity of lake ice regimes to climate change in the Nordic region

A one-dimensional process-based multi-year lake ice model, MyLake, was used to simulate lake ice phenology and annual maximum lake ice thickness for the Nordic region comprising Fennoscandia and the Baltic countries. The model was first tested and validated using observational meteorological forcing on a candidate lake (Lake Atnsjoen) and using downscaled ERA-40 reanalysis data set. To simulate ice conditions for the contemporary period of 1961-2000, the model was driven by gridded meteorological forcings from ERA-40 global reanalysis data downscaled to a 25 km resolution using the Rossby Centre Regional Climate Model (RCA). The model was then forced with two future climate scenarios from the RCA driven by two different general circulation models (GCMs) based on the Special Report on Emissions Scenarios (SRES) A1B. The two climate scenarios correspond to two future time periods namely the 2050s (2041-2070) and the 2080s (2071-2100). To take into account the influence of lake morphometry, simulations were carried out for four different hypothetical lake depths (5 m, 10 m, 20 m, 40 m) placed at each of the 3708 grid cells. Based on a comparison of the mean predictions in the future 30-year periods with the control (1961-1990) period, ice cover durations in the region will be shortened by 1 to 11 weeks in 2041-2070, and 3 to 14 weeks in 2071-2100. Annual maximum lake ice thickness, on the other hand, will be reduced by a margin of up to 60 cm by 2041-2070 and up to 70 cm by 2071-2100. The simulated changes in lake ice characteristics revealed that the changes are less dependent on lake depths though there are slight differences. The results of this study provide a regional perspective of anticipated changes in lake ice regimes due to climate warming across the study area by the middle and end of this century.