Abstract. With the aim of studying the recent Greenland ice sheet (GrIS) surface mass balance (SMB) decrease relative to the last century, we have forced the regional climate MAR (Modèle Atmosphérique Régional; version 3.5.2) model with the ERA-Interim (ECMWF Interim Re-Analysis; 1979–2015), ERA-40 (1958–2001), NCEP–NCARv1 (National Centers for Environmental Prediction–National Center for Atmospheric Research Reanalysis version 1; 1948–2015), NCEP–NCARv2 (1979–2015), JRA-55 (Japanese 55-year Reanalysis; 1958–2014), 20CRv2(c) (Twentieth Century Reanalysis version 2; 1900–2014) and ERA-20C (1900–2010) reanalyses. While all these forcing products are reanalyses that are assumed to represent the same climate, they produce significant differences in the MAR-simulated SMB over their common period. A temperature adjustment of +1 °C (respectively −1 °C) was, for example, needed at the MAR boundaries with ERA-20C (20CRv2) reanalysis, given that ERA-20C (20CRv2) is ∼ 1 °C colder (warmer) than ERA-Interim over Greenland during the period 1980–2010. Comparisons with daily PROMICE (Programme for Monitoring of the Greenland Ice Sheet) near-surface observations support these adjustments. Comparisons with SMB measurements, ice cores and satellite-derived melt extent reveal the most accurate forcing datasets for the simulation of the GrIS SMB to be ERA-Interim and NCEP–NCARv1. However, some biases remain in MAR, suggesting that some improvements are still needed in its cloudiness and radiative schemes as well as in the representation of the bare ice albedo. Results from all MAR simulations indicate that (i) the period 1961–1990, commonly chosen as a stable reference period for Greenland SMB and ice dynamics, is actually a period of anomalously positive SMB (∼ +40 Gt yr−1) compared to 1900–2010; (ii) SMB has decreased significantly after this reference period due to increasing and unprecedented melt reaching the highest rates in the 120-year common period; (iii) before 1960, both ERA-20C and 20CRv2-forced MAR simulations suggest a significant precipitation increase over 1900–1950, but this increase could be the result of an artefact in the reanalyses that are not well-enough constrained by observations during this period and (iv) since the 1980s, snowfall is quite stable after having reached a maximum in the 1970s. These MAR-based SMB and accumulation reconstructions are, however, quite similar to those from Box (2013) after 1930 and confirm that SMB was quite stable from the 1940s to the 1990s. Finally, only the ERA-20C-forced simulation suggests that SMB during the 1920–1930 warm period over Greenland was comparable to the SMB of the 2000s, due to both higher melt and lower precipitation than normal.

Abstract. The Antarctic ice sheet mass balance is a major component of the sea level budget and results from the difference of two fluxes of a similar magnitude: ice flow discharging in the ocean and net snow accumulation on the ice sheet surface, i.e. the surface mass balance (SMB). Separately modelling ice dynamics and SMB is the only way to project future trends. In addition, mass balance studies frequently use regional climate models (RCMs) outputs as an alternative to observed fields because SMB observations are particularly scarce on the ice sheet. Here we evaluate new simulations of the polar RCM MAR forced by three reanalyses, ERA-Interim, JRA-55, and MERRA-2, for the period 1979–2015, and we compare MAR results to the last outputs of the RCM RACMO2 forced by ERA-Interim. We show that MAR and RACMO2 perform similarly well in simulating coast-to-plateau SMB gradients, and we find no significant differences in their simulated SMB when integrated over the ice sheet or its major basins. More importantly, we outline and quantify missing or underestimated processes in both RCMs. Along stake transects, we show that both models accumulate too much snow on crests, and not enough snow in valleys, as a result of drifting snow transport fluxes not included in MAR and probably underestimated in RACMO2 by a factor of 3. Our results tend to confirm that drifting snow transport and sublimation fluxes are much larger than previous model-based estimates and need to be better resolved and constrained in climate models. Sublimation of precipitating particles in low-level atmospheric layers is responsible for the significantly lower snowfall rates in MAR than in RACMO2 in katabatic channels at the ice sheet margins. Atmospheric sublimation in MAR represents 363 Gt yr−1 over the grounded ice sheet for the year 2015, which is 16 % of the simulated snowfall loaded at the ground. This estimate is consistent with a recent study based on precipitation radar observations and is more than twice as much as simulated in RACMO2 because of different time residence of precipitating particles in the atmosphere. The remaining spatial differences in snowfall between MAR and RACMO2 are attributed to differences in advection of precipitation with snowfall particles being likely advected too far inland in MAR.

The Antarctic ice sheet (AIS) is classified as a polar desert where, similar to other deserts around the world, the annual precipitation is dependent on a few episodic precipitation events. Recent research has highlighted that certain regions of the AIS receive 40%-60% of their total annual precipitation from the largest 10% of daily precipitation events (Turner et al., 2019). There is a high coast-inland snowfall gradient, as most

Abstract. Observations and models agree that the Greenland Ice Sheet (GrIS) surface mass balance (SMB) has decreased since the end of the 1990s due to an increase in meltwater runoff and that this trend will accelerate in the future. However, large uncertainties remain, partly due to different approaches for modelling the GrIS SMB, which have to weigh physical complexity or low computing time, different spatial and temporal resolutions, different forcing fields, and different ice sheet topographies and extents, which collectively make an inter-comparison difficult. Our GrIS SMB model intercomparison project (GrSMBMIP) aims to refine these uncertainties by intercomparing 13 models of four types which were forced with the same ERA-Interim reanalysis forcing fields, except for two global models. We interpolate all modelled SMB fields onto a common ice sheet mask at 1 km horizontal resolution for the period 1980–2012 and score the outputs against (1) SMB estimates from a combination of gravimetric remote sensing data from GRACE and measured ice discharge; (2) ice cores, snow pits and in situ SMB observations; and (3) remotely sensed bare ice extent from MODerate-resolution Imaging Spectroradiometer (MODIS). Spatially, the largest spread among models can be found around the margins of the ice sheet, highlighting model deficiencies in an accurate representation of the GrIS ablation zone extent and processes related to surface melt and runoff. Overall, polar regional climate models (RCMs) perform the best compared to observations, in particular for simulating precipitation patterns. However, other simpler and faster models have biases of the same order as RCMs compared with observations and therefore remain useful tools for long-term simulations or coupling with ice sheet models. Finally, it is interesting to note that the ensemble mean of the 13 models produces the best estimate of the present-day SMB relative to observations, suggesting that biases are not systematic among models and that this ensemble estimate can be used as a reference for current climate when carrying out future model developments. However, a higher density of in situ SMB observations is required, especially in the south-east accumulation zone, where the model spread can reach 2 m w.e. yr−1 due to large discrepancies in modelled snowfall accumulation.

Abstract. We compare the performance of five different regional climate models (RCMs) (COSMO-CLM2, HIRHAM5, MAR3.10, MetUM, and RACMO2.3p2), forced by ERA-Interim reanalysis, in simulating the near-surface climate and surface mass balance (SMB) of Antarctica. All models simulate Antarctic climate well when compared with daily observed temperature and pressure, with nudged models matching daily observations slightly better than free-running models. The ensemble mean annual SMB over the Antarctic ice sheet (AIS) including ice shelves is 2329±94 Gt yr−1 over the common 1987–2015 period covered by all models. There is large interannual variability, consistent between models due to variability in the driving ERA-Interim reanalysis. Mean annual SMB is sensitive to the chosen period; over our 30-year climatological mean period (1980 to 2010), the ensemble mean is 2483 Gt yr−1. However, individual model estimates vary from 1961±70 to 2519±118 Gt yr−1. The largest spatial differences between model SMB estimates are in West Antarctica, the Antarctic Peninsula, and around the Transantarctic Mountains. We find no significant trend in Antarctic SMB over either period. Antarctic ice sheet (AIS) mass loss is currently equivalent to around 0.5 mm yr−1 of global mean sea level rise (Shepherd et al., 2020), but our results indicate some uncertainty in the SMB contribution based on RCMs. We compare modelled SMB with a large dataset of observations, which, though biased by undersampling, indicates that many of the biases in SMB are common between models. A drifting-snow scheme improves modelled SMB on ice sheet surface slopes with an elevation between 1000 and 2000 m, where strong katabatic winds form. Different ice masks have a substantial impact on the integrated total SMB and along with model resolution are factored into our analysis. Targeting undersampled regions with high precipitation for observational campaigns will be key to improving future estimates of SMB in Antarctica.

Abstract. The future surface mass balance (SMB) will influence the ice dynamics and the contribution of the Antarctic ice sheet (AIS) to the sea level rise. Most of recent Antarctic SMB projections were based on the fifth phase of the Coupled Model Intercomparison Project (CMIP5). However, new CMIP6 results have revealed a +1.3 ∘C higher mean Antarctic near-surface temperature than in CMIP5 at the end of the 21st century, enabling estimations of future SMB in warmer climates. Here, we investigate the AIS sensitivity to different warmings with an ensemble of four simulations performed with the polar regional climate model Modèle Atmosphérique Régional (MAR) forced by two CMIP5 and two CMIP6 models over 1981–2100. Statistical extrapolation enables us to expand our results to the whole CMIP5 and CMIP6 ensembles. Our results highlight a contrasting effect on the future grounded ice sheet and the ice shelves. The SMB over grounded ice is projected to increase as a response to stronger snowfall, only partly offset by enhanced meltwater run-off. This leads to a cumulated sea-level-rise mitigation (i.e. an increase in surface mass) of the grounded Antarctic surface by 5.1 ± 1.9 cm sea level equivalent (SLE) in CMIP5-RCP8.5 (Relative Concentration Pathway 8.5) and 6.3 ± 2.0 cm SLE in CMIP6-ssp585 (Shared Socioeconomic Pathways 585). Additionally, the CMIP6 low-emission ssp126 and intermediate-emission ssp245 scenarios project a stabilized surface mass gain, resulting in a lower mitigation to sea level rise than in ssp585. Over the ice shelves, the strong run-off increase associated with higher temperature is projected to decrease the SMB (more strongly in CMIP6-ssp585 compared to CMIP5-RCP8.5). Ice shelves are however predicted to have a close-to-present-equilibrium stable SMB under CMIP6 ssp126 and ssp245 scenarios. Future uncertainties are mainly due to the sensitivity to anthropogenic forcing and the timing of the projected warming. While ice shelves should remain at a close-to-equilibrium stable SMB under the Paris Agreement, MAR projects strong SMB decrease for an Antarctic near-surface warming above +2.5 ∘C compared to 1981–2010 mean temperature, limiting the warming range before potential irreversible damages on the ice shelves. Finally, our results reveal the existence of a potential threshold (+7.5 ∘C) that leads to a lower grounded-SMB increase. This however has to be confirmed in following studies using more extreme or longer future scenarios.

Abstract. The ERA5 reanalysis, recently made available by the European Centre for Medium-Range Weather Forecasts (ECMWF), is a new reanalysis product at a high resolution replacing ERA-Interim and is considered to provide the best climate reanalysis over Greenland to date. However, so far little is known about the performance of ERA5 over the Greenland Ice Sheet (GrIS). In this study, we compare the near-surface climate from the new ERA5 reanalysis to ERA-Interim, the Arctic System Reanalysis (ASR) as well as to a state-of-the-art polar regional climate model (MAR). The results show (1) that ERA5 does not outperform ERA-Interim significantly when compared with near-surface climate observations over GrIS, but ASR better models the near-surface temperature than both ERA reanalyses. (2) Polar regional climate models (e.g., MAR) are still a useful tool to downscale the GrIS climate compared to ERA5, as in particular the near-surface temperature in summer has a key role for representing snow and ice processes such as the surface melt. However, assimilating satellite data and using a more recent radiative scheme enable both ERA and ASR reanalyses to represent more satisfactorily than MAR the downward solar and infrared fluxes. (3) MAR near-surface climate is not affected when forced at its lateral boundaries by either ERA5 or ERA-Interim. Therefore, forcing polar regional climate models with ERA5 starting from 1950 will enable long and homogeneous surface mass balance reconstructions.

Abstract. Antarctic ice sheet mass loss is currently equivalent to around 1 mm year−1 of global mean sea level rise. Most mass is lost due to sub-ice shelf melting and calving of icebergs. Ice sheet models of the Antarctic ice sheet have thus largely concentrated on parameterising sub-shelf and calving processes. However, surface mass balance (SMB) is also of crucial importance in controlling the stability and evolution of the vast Antarctic ice sheet. In this paper we compare the performance of five different regional climate models (COSMO-CLM2, HIRHAM5, MAR3.10, MetUM and RACMO2.3p2) in simulating the near surface climate and SMB of Antarctica. Our results show that, when regional climate models (RCMs) are forced by the ERA-Interim reanalysis, the integrated Antarctic ice sheet ensemble mean annual SMB is 2329 ± 94 Gigatonnes (Gt) year−1 over the common 1987 to 2015 period. However, individual model estimates vary from 1961 ± 70 to 2519 ± 118 Gt year−1. The large differences are mostly explained by different SMB estimates in West Antarctica and the peninsula as well as around the Transantarctic mountains. The calculated annual average SMB is very sensitive to the period chosen but over the climatological mean period of 1980 to 2010 the ensemble mean is 2486 Gt year−1. The interannual variability in SMB is consistent between the models and dominated by variability in the driving ERA-Interim reanalysis. The declining trend in Antarctic SMB reported in other studies is also very sensitive to period chosen and models disagree on the sign and magnitude of the trend in Antarctic SMB over the ERA-Interim period. Evaluation of models shows that they simulate Antarctic climate well when compared with daily observed temperature (Pearson correlation of 0.85 and higher) and pressure (bias ranges from −0.39 hPa in HIRHAM5 to −6.01 hPa in MAR with a mean of −3.49 hPa over all models) and nudged models, constrained within the domain as well as at lateral boundaries, perform better than un-nudged models. We compare modelled surface mass balance with a large dataset of observations which, though biased by undersampling in some regions, indicates that many of the biases in modelled SMB are common between models. The inclusion of drifting snow schemes improves modelled SMB on ice sheet slopes between 1000 and 2000 m where strong katabatic winds form but other regions where precipitation rates are high lack observations needed for the evaluation of different SMB estimates. Different ice masks have a substantial impact on the integrated total SMB and along with model resolution is therefore factored into our analysis. The majority of the different values for continental SMB are due to differences in modelled precipitation at relatively few grid points in coastal areas. Our analysis suggests that targeting coastal areas for observational campaigns will be key to improving and refining estimates of the total surface mass balance of Antarctica.