Spectral snow albedo in a regional climate model: application to the Greenland and Antarctic ice sheets

Abstract

An important part of the climate system of Earth are the ice sheets of Greenland and Antarctica. Crucial for the development of an ice sheet is the reflectivity of sunlight, or albedo, at the surface. To better understand a process like albedo and the climate system of the past, present and future, we use climate models. More specifically, we use a regional climate model to study the climate of ice sheets. There is still room for improvements in such models, however, as the description of some processes like snow albedo is rather limited. The goal of this research is therefore to implement a new snow albedo scheme in the regional climate model RACMO2, and analyze its impact on the development of the major ice sheets. In this thesis, we present the steps taken, discuss the implemented new physical processes and evaluate the results. To improve the representation of snow albedo in RACMO2, we implement the two-stream radiative transfer in snow model TARTES. TARTES will allow for more physical processes to be included in RACMO2, introducing an explicit wavelength-dependent albedo and subsurface energy absorption. As TARTES is a spectral model, it performs radiation calculations for sets of single wavelengths. RACMO2, in contrast, uses fourteen wavelength bands, the so-called narrowbands, for its atmospheric radiation calculations. Therefore, we developed the spectral-to-narrowband albedo coupling module SNOWBAL. With SNOWBAL, TARTES has been successfully implemented into RACMO2. The modeled albedo of this new RACMO2 version matches well with in situ and remote sensing observations of the Greenland ice sheet, with a negligible domain-averaged bias for the interior. Some discrepancies in the albedo are, however, observed around the ice sheet margins. The surface mass balance and energy budget are in good agreement with observations. An important improvement is that absorption of sunlight beneath the surface leads to higher snow temperatures, which is also in agreement with observations. Compared to the previous RACMO2 version, the albedo is higher in the bare ice zone during the ablation season. Furthermore, melt and refreezing have changed more substantially. Using the new RACMO2 version, we performed several sensitivity experiments for the Antarctic ice sheet. By changing one parameter at a time, we investigated the impact of the fresh snow and refreezing grain size, snow metamorphism, and internal heating on temperature and melt. We show that temperature and melt are exceptionally sensitive to changes in the snowpack, and that modeling a correct snowpack evolution is vital for a correct estimation of melt. When fully tuned, however, the new RACMO2 version compares well with observations of the surface mass and energy balance, melt, (sub)surface temperature, grain size and albedo. To summarize, we have developed new methods to successfully introduce a new snow albedo scheme that includes more physical processes in the regional climate model RACMO2. It has led to many new insights and improved model simulations. With this research, we hope to contribute to better constrained future climate projections of the Greenland and Antarctic ice sheets.