Understanding and Improving Greenland Ice Albedo in Climate Models

Meltwater production from the Greenland Ice Sheet (GrIS) is increasing rapidly, accelerating global sea level rise.However, the uncertainty of the projected sea level rise prevents proper planning of mitigations against the effects of sea level rise.This uncertainty is in large part due to a lack of understanding of the physical processes that control surface melting.Specifically, we lack understanding of the physical processes controlling ice albedo, a major driver of surface melting.Therefore, climate models often use a constant value or an overly simplified equation to model ice albedo.Consequently, ice albedo is currently not properly represented in climate models (Figure 1), with implications for surface melting and sea level rise predictions (Antwerpen et al., 2022).Here, we show how we use the available, but under-used, information from climate model output and satellite imagery to improve the representation of ice albedo.We use physics-informed machine learning (ML) tools to explore the relations between 1) the modeled atmospheric and glaciological parameters of bare ice and 2) the observed bare ice albedo.