Ocean-glacier interaction on the large regional scale

Abstract

Glaciers are important regulators of water availability in many regions of the world and their retreat can lead to increased geohazards. Glacier melt has contributed significantly to sea-level rise in the past and has become the biggest single source of observed sea-level rise since 1900, even if the ice mass stored in glaciers is small compared to the Greenland and Antarctic ice sheets (<1%). Glacier melt has and will continue to be a major source of sea-level rise in the 21st century. Therefore, it is a pressing task to improve the knowledge of how glaciers change when subjected to climate change, both natural and anthropogenic. About 30% of the glaciers on earth terminate in the ocean and frontal ablation (mass loss by calving and frontal melting) is a major component of the mass budget of tidewater glaciers, strongly affecting their dynamics.

Most global scale ice volume estimates to date still suffer from considerable uncertainties related to i) the implemented frontal ablation parameterization or ii) not accounting for frontal ablation at all in the glacier model. To improve estimates of the ice thickness distribution of tidewater glaciers, it is thus important to identify and test low-cost and robust parameterizations of this process. By implementing such parameterization into the ice-thickness estimation module of the Open Global Glacier Model (OGGM v1.1.2), this thesis conducts a first assessment of the impact of accounting for frontal ablation on the estimate of ice stored in glaciers located in Alaska and Greenland.

OGGM is the first globally applicable, open source, community-driven model for consistently simulating past and future global scale glacier change. It's ice thickness inversion scheme relies on a mass-conservation approach, this thesis found that if frontal ablation is neglected from the mass balance budget, the model systematically underestimated the mass turnover, and therefore the thickness and volume of tidewater glaciers. This underestimation can amount to up to 19% on a regional scale in Alaska, and up to 14% in Greenland's Peripheral glaciers (PG's). For individual glaciers volume underestimation can be up to 30% (e.g Columbia Glacier in Alaska). The effect is independent of the size of the glacier.

Additionally, this study performs different sensitivity experiments to study the influence of i) a constant of proportionality (k) used in the frontal ablation parameterization, ii) Glen's temperature-dependent creep parameter (A) and iii) a sliding velocity parameter, on the regional dynamics of Alaska tidewater glaciers. OGGM is able to reproduce previous regional frontal ablation estimates by applying a number of combinations of values for k, Glen's A and sliding parameter. The sensitivity studies also show that differences in thickness between accounting for and not accounting for frontal ablation occur mainly at the lower parts of the glacier, both above and below sea level. This indicates that not accounting for frontal ablation will have an impact on the estimate of the glaciers' potential contribution to sea-level rise.

In Greenland, there are no regional observations or estimates of frontal ablation to constrain model parameters. Forcing the study to develop two independent methods to calibrate the calving parameterization implemented in OGGM. The first method constrains the calving constant of proportionality k, with surface velocity fields derived from the MEaSUREs Multi-year Greenland Ice Sheet Velocity Mosaic. Whereas the second method constrains the k parameter using frontal ablation fluxes, derived from Surface Mass Balance (SMB) means over an equilibrium reference period (1961-1990), obtained from the monthly output of the Polar Regional Climate Model RACMO, statistically downscaled to 1 km resolution. The second calibration method is based on the strong assumption that most PG's during that time have a balanced budget (i.e. did not experience any mass loss or gain). Considering an equilibrium between what the glacier gained and calved might not reflect real frontal ablation fluxes, but such estimates, serve as a base to asses the dynamic mass loss of glaciers when combined with frontal ablation estimates constrained from velocity observations. By comparing the model output after applying both calibration methods, this thesis finds that the model is not able to predict individual tidewater glacier dynamics, if it relies only on SMB estimates and the assumption of a closed budget to constrain k values. Velocity observations are essential to constrain model parameters and estimate the dynamic mass loss of PG's.