Results from ITMIX2 - the second phase of Ice Thickness Models Intercomparison eXperiment

Abstract:

Information about the ice thickness distribution of glaciers and icecaps is of fundamental importance for a range of applications. Yet, this information is limited by the scarcity of direct measurements and numerical models are used to overcome the gap. Here, we present results from the Second Phase of the Ice Thickness Models Intercomparison eXperiment (ITMIX2), which aimed at characterizing the capability of such models to extract information from limited subsets of thickness observations. The experiment was joined by 13 models run by a consortium of 21 scientists representing 20 different institutions, included 23 test cases comprising both real-world and synthetic glaciers, and comprised a set of 16 different experiments per test case mimicking various scenarios of data availability.

For locations without measured ice thickness, the results show high variability across models, and reveal typical deviations from local values in the order of 16% of the mean ice thickness. Still, limited sets of ice thickness observations are found to be effective in constraining the glacier total volume, which is of critical relevance for a number of questions. Whilst we find that the spatial distribution of the available observations has only a weak influence on the predicted ice thickness distribution, surveys restricted to the lowest glacier elevations typically result in an underestimation of the glacier’s total volume. Surveys that preferentially sample the thickest glacier parts, instead, found to be effective in preventing such biases. Although the response to various scenarios of data availability is specific to every model, no single best approach emerges from the experiment. This indicates that, wherever possible, a combination of models should be preferred.