Search

The Greenland ice sheet mass loss is one of the main sources of contemporary sea-level rise. The mass loss is primarily caused by surface melt and the resulting runoff. During the melt season, the ice sheet’s surface receives energy from sunlight absorption and sensible heating, which subsequently heats the subsurface snow and ice. The energy from the previous melt season can also enhance melting in the following summer as less heating is required to bring the snow and ice to the melting point. Subsurface temperatures are therefore both a result and a driver of the timing and magnitude of surface melt on the ice sheet. We present a dataset of more than 3900 measurements of ice, snow and firn temperature at 10 m depth across the Greenland ice sheet spanning the years from 1912 to 2022. We construct an artificial neural network (ANN) model that takes as input the ERA5 reanalysis monthly near-surface air temperature and snowfall for the 1954-2022 period and train it on our compilation of observed 10-meter temperature. We use our dataset and the ANN to evaluate three broadly used regional climate models (RACMO, MAR and HIRHAM). Our ANN model provides an unprecedented and observation-based description of the recent warming of the ice sheet’s near-surface and our evaluation of the three climate models highlights future development for the models. Overall, these findings improve our understanding of the ice sheet’s response to recent atmospheric warming and will help reduce uncertainties of ice sheet surface mass balance estimates.

The quantification of snow accumulation and the temporal evolution of the snow pack is essential when investigating the mass balance conditions of mountain glaciers. In particular, accumulation regions become smaller due to the gradual increase of the equilibrium line, thus reducing mass input into the glacier system. This will have severe consequences on ice flux and thus the mass balance conditions across many mountain regions worldwide. The mass redistribution within the accumulation regions is considerably influenced by migration of melt water in the snow and firn pack and the induced mass and density changes. Here, we study snow and firn processes at a high mountain accumulation plateau on 3470 m asl at Vernagtferner, Austria. Vernagtferner is a major glacier in the drainage basin of Rofenache, with an area of about 6.9 km², covering altitudes between 2900 m and 3550 m. A snow monitoring station, including an upward-looking ground penetrating radar (upGPR) was installed at the highest accumulation basin in 2018. This station allows the continuous determination of the snow pack stratigraphy and of the snow water equivalent (SWE) (Heilig et al. (2009, 2010), Schmid et al. 2014, Heilig et al., 2015). We compare numerical simulations of the 1-dimensional snow cover model SNOWPACK (Bartelt and Lehning, 2002), driven by automatic weather station data, with continuous observations of the installed upGPR system and bi-annual in-situ data. The analysed upGPR data enable continuous evaluation of the SNOWPACK simulations over several melt and accumulation seasons. The upGPR data show that even at high elevations frequent melt-freeze crusts develop during the accumulation period. Even though the crusts are several centimetres, melt water rapidly percolates trough these layers, once the snow pack reaches isothermal conditions in late spring. The simulation results demonstrate, that SNOWPACK is able to reproduce this fast advance of the melt front accurately, while the up-GPR measurements provide an independent proof of the model performance. These measurements also show that firn layers (previous summer surfaces) block water infiltration into depth only for a very short period, indicating that SWE measurements of glacier accumulation only provide realistic values, if carried out before or just at the onset of spring melt. This feasibility study provides important indication on how to extend such studies to larger glacier systems, also in less monitored regions, where in-situ data might be sparse.

Recent studies indicate that a declining surface mass balance will dominate the Greenland Ice Sheet’s (GrIS) contribution to 21st century sea level rise. It is therefore crucial to understand the liquid water balance of the ice sheet and its response to increasing temperatures and surface melt if we want to accurately predict future sea level rise. The ice sheet firn layer covers ~90% of the GrIS and provides pore space for storage and refreezing of meltwater. Because of this, the firn layer can retain up to ~45% of the surface meltwater and thus act as an efficient buffer to ice sheet mass loss. However, in a warming climate this buffer capacity of the firn layer is expected to decrease, amplifying meltwater runoff and sea-level rise. Dedicated firn models are used to understand how firn layers evolve and affect runoff. Additionally, firn models are used to estimate the changing thickness of the firn layer, which is necessary in altimetry to convert surface height change into ice sheet mass loss.Here, we present the latest version of our firn model IMAU-FDM. With respect to the previous version, changes have been made to the handling of the freshly fallen snow, the densification rate of the firn and the conduction of heat. These changes lead to an improved representation of firn density and temperature. The results have been thoroughly validated using an extensive dataset of density and temperature measurements that we have compiled covering 126 different locations on the GrIS. Meltwater behaviour in the model is validated with upward-looking GPR measurements at Dye-2. Lastly, we present an in-depth look at the evolution firn characteristics at some typical locations in Greenland.Dedicated, stand-alone firn models offer various benefits to using a regional climate model with an embedded firn model. Firstly, the vertical resolution for buried snow and ice layers can be larger, improving accuracy. Secondly, a stand-alone firn model allows for spinning up the model to a more accurate equilibrium state. And thirdly, a stand-alone model is more cost- and time-effective to use. Firn models are increasingly capable of simulating the firn layer, but areas with large amounts of melt still pose the greatest challenge.

Perennial snow, or firn, covers 80 % of the Greenland ice sheet and has the capacity to retain part of the surface meltwater, buffering the ice sheet’s contribution to sea level. Multi-layer firn models are traditionally used to simulate the firn processes and estimate meltwater retention. We present the output from nine firn models, forced by weather-station-derived mass and energy fluxes at four sites representative of the dry snow, percolation, ice slab and firn aquifer areas. We compare the model outputs and evaluate them against in situ observations. Models that explicitly account for deep meltwater percolation overestimate percolation depth and consequently firn temperature at the percolation and ice slab sites although they accurately simulate the recharge of the firn aquifer. Models using Darcy's law and a bucket scheme compare favourably to observations at the percolation site but only the Darcy models accurately simulate firn temperature and thus meltwater percolation at the ice slab site. We find that Eulerian models, that transfer firn through fixed layers, smooth sharp gradients in firn temperature and density over time. From the model spread, we find that simulated densities (respectively temperature) have an uncertainty envelope of ±60 kg m−3 (resp. ±14 °C) in the dry snow area and up to ±280 kg m−3 (resp. ±15–18 °C) at warmer sites.

Increasing melt over the Greenland Ice Sheet (GrIS) recorded over the past several years has resulted in significant changes of the percolation regime of the ice sheet. It remains unclear whether Greenland's percolation zone will act as a meltwater buffer in the near future through gradually filling all pore space or if near-surface refreezing causes the formation of impermeable layers, which provoke lateral runoff. Homogeneous ice layers within perennial firn, as well as near-surface ice layers of several meter thickness have been observed in firn cores. Because firn coring is a destructive method, deriving stratigraphic changes in firn and allocation of summer melt events is challenging. To overcome this deficit and provide continuous data for model evaluations on snow and firn density, temporal changes in liquid water content and depths of water infiltration, we installed an upward-looking radar system (upGPR) 3.4 m below the snow surface in May 2016 close to Camp Raven (66.4779∘ N, 46.2856∘ W) at 2120 m a.s.l. The radar is capable of quasi-continuously monitoring changes in snow and firn stratigraphy, which occur above the antennas. For summer 2016, we observed four major melt events, which routed liquid water into various depths beneath the surface. The last event in mid-August resulted in the deepest percolation down to about 2.3 m beneath the surface. Comparisons with simulations from the regional climate model MAR are in very good agreement in terms of seasonal changes in accumulation and timing of onset of melt. However, neither bulk density of near-surface layers nor the amounts of liquid water and percolation depths predicted by MAR correspond with upGPR data. Radar data and records of a nearby thermistor string, in contrast, matched very well for both timing and depth of temperature changes and observed water percolations. All four melt events transferred a cumulative mass of 56 kg m−2 into firn beneath the summer surface of 2015. We find that continuous observations of liquid water content, percolation depths and rates for the seasonal mass fluxes are sufficiently accurate to provide valuable information for validation of model approaches and help to develop a better understanding of liquid water retention and percolation in perennial firn.