Ice thickness inversion assessment: A comparison study for Waldemarbreen and Irenebreen glaciers, Svalbard

Abstract

Accurate ice thickness estimates are crucial for solving both glaciological and hydrological problems, and ice thickness inversion models are generally used to obtain these estimates. Several studies have assessed the accuracy of model inversions, however, usually the evaluation is done against a sparse set of field measurements. In this study, we employ the open-source models GlabTop2 and Open Global Glacier Model (OGGM) to estimate ice thickness and volume for the Waldemarbreen and Irenebreen glaciers (Svalbard archipelago), evaluating their accuracy against ground-penetrating radar (GPR) measurements. The results show that the ice thickness root-mean-square errors for GlabTop2 and OGGM are 28.5 and 13.9 m and the relative volume errors are 17.4 and 39.9%, respectively. Calibration of glacier shape factor slightly improves GlabTop2 results, however, this model tends to provide good average estimates but falls short on accurate spatial thickness distribution. On the contrary, the OGGM model is highly dependent on the chosen dataset for mass balance calibration and tends to provide better spatial distribution while performing worse regarding average thickness and overall glacier volume. Both inversion models present severe thickness errors among studied glaciers, implying the need for improvements in models and underlying theoretical assumptions that require continuous and extensive field measurements.