| Summary: | This dissertation describes the research results from applying regional climate models (RCMs) to simulate the near-past and future evolution of Vatnajökull, the largest ice cap in Iceland, and investigate the sensitivity of ablation to albedo and spring snow cover. The output of the RCM HIRHAM5 is first validated over Vatnajökull using available automatic weather station and mass balance observations, and the albedo parameterisation is improved by adding an ice albedo map based on MODIS observations. Although the ice albedo map improves the simulations, there is still an average overestimation of the albedo simulations over the ablation season due to an overestimation in the amount of snowfall and the lack of simulations of the effect of dust storms. As a result, the mean ablation over the melt season is generally underestimated compared to observations. However, the specific mass balance is well simulated by the model, which captures the trends in mass loss since the 1991-92 glaciological year. The overestimation of the accumulation in HIRHAM5 was suspected to be partly due to the hydrostatic approximation used in the atmospheric model, and experiments forcing the well-validated snow pack model from HIRHAM5 with meteorological parameters from other climate models were therefore conducted. The snow pack model simulates the surface energy balance, the total melt, and subsurface processes like melt percolation and refreezing. The meteorological forcing from the non-hydrostatic (i.e. solves the vertical momentum equation) numerical weather prediction model (NWP) HARMONIE-AROME is shown to provide improved simulation over Vatnajökull, in particular due to improvements in the simulation of accumulation. The HARMONIE-AROME forced model is used to simulate the runoff from 1980-2015 and estimate the sensitivity of runoff to variations in spring snow thickness. It is found that the runoff is much more sensitive to changes in the summer weather overall, although the north-facing outlet glaciers Brúarjökull and ...
|
|---|