Changes in glacier runoff in a warming Arctic: simulations from the Bayelva catchment, Svalbard

The Svalbard archipelago (76-81N) is undergoing increased warming compared to the global mean, which has major implications for freshwater runoff into the oceans from seasonal snow and glaciers. Quantifying changes of freshwater runoff  requires close integration of observations and process-based models.

Here, we use land-surface and ice-flow modelling in combination with satellite and in-situ observations, to simulate runoff from the Bayelva catchment, Svalbard (~30 km2, ~54% glacier cover), for the period 1991–2100. Runoff from seasonal snow and glaciers is simulated using the land surface model CRYOGRID, which includes a coupled energy balance-snow/firn model. Historical simulations (1991–2024) are forced by downscaled CARRA reanalysis data and evaluated against in situ measurements and geodetic mass balance observations. Future simulations (2024–2100) are driven by temperature and precipitation trends derived from CORDEX projections under the RCP4.5 and RCP8.5 scenarios.

To account for feedbacks between surface mass balance and glacier geometry, the runoff simulations are coupled to the 3D glacier evolution model IGM. Sentinel-1 surface velocity observations are used to constrain glacier sliding, while observed surface elevation changes are used to evaluate simulated thickness changes and in situ ice-thickness measurements to evaluate the initial glacier geometry.

For continued warming, glacier melt will intensify, thus increasing runoff, but at a later stage, the reduction of glacier area due to retreat will offset this effect, giving rise to a peak in glacier runoff. The simulations indicate that runoff from the Bayelva catchment is likely to peak within the next two decades. Under the RCP8.5 scenario, both glaciers within the Bayelva catchment, Austre and Vestre Brøggerbreen, are projected to largely disappear by 2100, resulting in a transition from glacier-dominated to snow-dominated runoff.