The climatic influence on a large cold/polythermal Arctic ice cap with substantial surface melt was investigated by calculating the surface energy balance (SEB)and subsurface properties. This method was applied on Austfonna (∼8100km2,NE Svalbard) during melt season 2004. Surface mass balance was calculated from April 23rd to September 26th on the grid of a 1000m resolution digital elevation model(DEM). Hourly in-situ meteorological measurements by an automatic weather station(AWS) located on the glacier were used to force the model. Precipitation during the model run was taken from ERA40-reanalysis data, while initial conditions such as, snow cover, firn extent, snow and ice temperatures and densities were based on in-situ and satellite observations. Mass balance during the 2003-04period was then obtained by including the snow cover used as model input as an estimate for the 2003-04 winter mass balance.
During the melt season of 2004, short wave radiation contributed 86 % of the total energy available for melting ice at the glacier surface, whereas the remaining 14 % was supplied by sensible heat. Long wave radiation and the ground flux were two major energy sinks, while some energy was lost through sublimation and the latent energy flux. Spatial variations in the available energy for melt werecontrolled by snow depth variations and the firn extent, through albedo and melt water retention.
The overall modeled specific SMB 2003/04 was -41 [cm w.eq] which equals -3.3 [km3 w.eq.]. All of Austfonna’s drainage basins had a negative mass balance this year, with an overall accumulation area ratio (AAR) of 15 %. The estimated equilibrium line altitude (ELA) ranged from 570m in ESE to 680m in NW.
The model performance was fairly good compared to SMB derived from 16 stakes, with a r=0.84 correlation. Uncertainties in the model performance are dominated by two components: Firstly, quality of model input data such as initial snow, firn, snow and ice temperatures and densities and the simple extrapolations of meteorological parameters measured at the AWS. Secondly, uncertainties connected with model parameterizations for albedo and runoff, and scaling parameters used for calculations of the turbulent fluxes.