Control of snow albedo decay on Greenland surface melt using the land surface model ORCHIDEE

The Greenland ice sheet (GrIS) constitutes a major contributor to sea level rise, with significant mass losses dominated by surface melt and runoff over recent decades. The enhanced surface melt is concomitant with an observed decrease in albedo and a retreat of the snow line to higher elevations, amplified by melt-albedo feedbacks. A realistic representation of snow and ice albedo over the GrIS is therefore essential in climate models to correctly represent the surface energy balance, the meltwater production and the ice sheet surface mass balance.

In ORCHIDEE, the land surface component of the IPSL-CM Global Climate Model, snow albedo is currently represented with a decay parameterization based on snow age. Such simplified parametrizations are common in land surface schemes of CMIP earth system models. Our objective is to improve the representation of snow albedo over the GrIS using observations from PROMICE stations and MODIS retrievals. Our approach relies on the use of the History Matching calibration tool, based on Gaussian processes emulators, to identify the snow albedo parameters best adapted to Greenland conditions.

We first perform an offline calibration over Greenland to assess the ability of the ORCHIDEE model to reproduce the snow albedo decay  and the bare ice extent in summer. We show that a single year calibration allows for a good representation of snow albedo decay over the 2000–2019 period. We aim to revisit this approach by using the atmospheric component of IPSL-CM, LMDZ, coupled with ORCHIDEE, in a regional configuration. This will allow us to assess whether atmosphere-surface feedback needs to be taken into account during the calibration procedure. Finally, we assess if the improved representation of albedo decay enables the IPSL model to produce a more realistic decrease of the GrIS surface mass balance over recent decades.