Representation of melt ponds for Global Circulation Models

• Main Authors: Sterlin, Jean, Fichefet, Thierry, Massonnet, François, Lecomte, Olivier, Vancoppenolle, Martin
• Format: Conference Object
• Language: English
• Published: Zenodo 2020
• Subjects: Melt ponds; Atmospheric reanalysis and forcing; Sea Ice Model; Ocean General Circulation Model,; Arctic climate; Albedo; Arctic; albedo; Sea ice
• Online Access: https://dx.doi.org/10.5281/zenodo.4545278; https://zenodo.org/record/4545278

During the Arctic summer, a fraction of the surface melt water and liquid precipitation collect on the ice surface in pools known as melt ponds. The albedo of the ponds is lower than the surrounding snow and ice surfaces. Consequently, the melt ponds are an important factor for the ice-albedo feedback. The feedback is a mechanism whereby a decrease in albedo results in greater absorption of solar radiation, further ice melt, and lower albedo. Several numerical schemes have been proposed for Global Circulation Models. They can be classified into two groups. The first makes use of an explicit relation to define the geometrical aspect of the ponds. The scheme of Holland et al. (2012) falls in this category. The second relies on theoretical considerations to infer the ponds characteristics. The scheme of Flocco et al. (2012) uses the ice thickness distribution to distribute the melt water among the ice categories. Despite their current complexity, the melt pond schemes fail to agree on the future evolution of the ponds in the next decades. The difference of trends casts doubts on the definition of the physical processes governing the melt pond evolution. It also raises questions about the sensitivity of the schemes to the atmospheric surface forcing method. In this study, we aim at identifying 1) the conceptual difference of the aspect ratio definition in melt pond schemes; 2) the role of the refreezing of the ponds; 3) the impact of the uncertainties in the atmospheric reanalyses on the simulations. To address these points, we have run the Louvain-la-Neuve Ice Model (LIM), part of the Nucleus for European Modelling of the Ocean (NEMO) version 3.6 along with two different atmospheric reanalyses as forcing sets: JRA-55 and DFS5.2. We implemented Holland et al. (2012) and Flocco et al. (2012) melt pond schemes in the model. We selected Holland et al. (2012) pond refreezing formulation for both schemes and tried two different threshold temperatures for refreezing. From the experiments, we state the importance of melt ponds for climate models. We describe the impact on the sea ice in the Arctic. We attempt at disentangling the separate effects of the atmospheric surface forcing method, the type of melt pond schemes, and the refreezing mechanism. From our results, we formulate a recommendation on the use of melt ponds in climate models.