High-altitude glacier archives lost due to climate change-related melting

Global warming has caused widespread surface lowering of mountain glaciers. By comparing two firn cores collected in 2018 and 2020 from Corbassière glacier in Switzerland, we demonstrate how vulnerable these precious archives of past environmental conditions have become. Within two years, the soluble impurity records were destroyed by melting. The glacier is now irrevocably lost as an archive for reconstructing major atmospheric aerosol components.

chosen as a new location complementing the existing ice-core sites in the European Alps (Col du Dôme on Mont Blanc and Colle Gnifetti on Monte Rosa).On this glacier, both winter and summer snow is preserved, leading to a higher snow accumulation rate of 0.86-0.98m water equivalent per year (w.e.yr −1 ) than at Colle Gnifetti (0.45 m w.e.yr −1 ) 16 , where most of the winter snow is eroded by wind.At the same time, this is a lower annual snow accumulation than at Col du Dôme (0.5-2.4 m w.e.yr −1 ) 17,18 , so it was expected that this archive reaches further back in time.
In this Brief Communication we compare the major ion concentration and stable oxygen isotope ratio in water (δ 18 O) records from two firn cores drilled on Corbassière glacier in 2018 (14.1-m length, GC18) and 2020 (18.1 m, GC20), respectively.Firn is the intermediate stage in the transformation of snow to glacier ice and is porous, whereas glacier ice is nearly impermeable to liquid water.The δ 18 O signals of both firn cores are in excellent agreement for the overlapping period 2011-2018 (Fig. 2a) and generally follow the seasonal temperature trend (Extended Data Fig. 1).In the common, upper part (2016-2018), similar major ion concentrations are observed for ammonium ( ).In the part of GC20 below 8-m depth (year 2016), ion concentrations are generally substantially smoother, and the annual mean values are lower (Fig. 2b-d; Extended Data Fig. 2, same data on a log scale; Extended Data Fig. 3).The annual means began to be significantly lower at the 95% confidence level in 2016 (NO  ) back in time (Fig. 2f-h).The observed discrepancy cannot be explained by the spatial variability of the ion profiles, which is typically much smaller

Brief Communication
https://doi.org/10.1038/s41561-023-01366-1cumulative deposition fluxes in the GC20 core suggest that ions were either relocated with meltwater to even deeper layers or removed entirely from the archive with the percolating meltwater (runoff).In contrast, δ 18 O as the matrix component was not yet affected by the at high-altitude glaciers.This is illustrated in the example of the Colle Gnifetti glacier, because no such data exist for the summit plateau of the Corbassière glacier (Extended Data Fig. 4).Instead, we attribute the observed deviation to melt-induced elution processes.The lower

Brief Communication
https://doi.org/10.1038/s41561-023-01366-1amount of melt (Fig. 2e).Among the ions, SO 4 2− shows the strongest depletion (Fig. 2h), as expected based on previous observations [19][20][21] .Those concluded that NH 4 + , and to some degree also NO 3 − , are preferentially incorporated inside the ice crystal structure, whereas SO 4 2− is enriched at the firn grain surfaces 20,21 , thus being especially sensitive to relocation (or removal) by percolating meltwater as observed here.Further evidence for melt processes is the presence of refrozen infiltration ice layers from the top in both cores (Extended Data Fig. 5), producing positive outliers in the density profiles.Overall, the ice layer records of both cores show similar features, although thicker ice layers appear already in the top part of GC20, in contrast to the GC18 core.Wet firn conditions were encountered below 14-m depth during drilling in 2020, most probably related to the presence of a nearly 1-m-thick ice layer at 17-m depth, preventing further downward percolation of meltwater.We assume that the meltwater ran off along this impermeable layer, because ions were not enriched at this depth.
Daily air temperatures increased continuously over the period from 2007 to 2020, and the trend is significant (Mann-Kendall test: slope = 0.000,27, P < 0.000,07; Fig. 1c).In addition, after 2015 there is a shift to a significantly higher sum of air temperatures above the melting point per day (positive degree-day sum (pdd), Fig. 1d; average 2015-2020, 0.24; significance of shift, t-test, P < 0.000002) and a lower fraction of the day with negative temperatures (ndd, Fig. 1e; average 2015-2020, 0.89; t-test, P < 0.00000005) compared to the years before (2007-2014; pdd average, 0.15; ndd average, 0.93).This regime shift implies that there was more energy available for melting and fewer days with temperatures below 0 °C in the years 2015 to 2020 than before.
We hypothesize that alterations in the preservation of chemical impurity signals observed between 2018 and 2020 in the Grand Combin firn cores were caused by a combination of the following factors: • The active layer describes the uppermost firn layer, in which thermal heat transfer from the seasonally varying surface air temperatures causes corresponding seasonal temperature changes (decreasing in amplitude with depth) 22 .The active layer depth at the Grand Combin drilling site is unknown.On Colle Gnifetti (312 m higher, 45 km east of Grand Combin), borehole temperatures suggest an active layer thickness of ~10-20 m (ref.23).On Col du Dôme (112 m higher, 35 km southwest of Grand Combin), borehole temperatures below a depth of 20 m are consistent 24 .We therefore assume an active layer depth at the Grand Combin drilling site of ~10-20 m.Meltwater formed at the surface in summer percolated through this part without causing a detectable alteration in the signal, because cold from the previous winter was stored in this layer.The quantity of meltwater formed at the surface and the corresponding available thermal energy was not sufficient to raise temperatures to 0 °C in this layer, therefore not causing additional melt in the firn and no detectable alteration of the recorded proxy signals (Fig. 2b-d, back to 2016 at 8-m depth in GC20).• In the porous firn of the active layer, preferential percolation in channels might have occurred, thus leaving a vast area of the firn (including the area of the GC20 ice core) unaffected.Below, preferential percolation is less likely because of increasing density and lower porosity.• Below the active layer, summer meltwater refroze, as indicated by the thick ice layer, thereby releasing latent heat.At this depth, where the firn is isolated from the cold winter conditions by the active layer above, this additional influx of heat led to a temperature well above the average air temperature, leading to meltwater accumulation 25 .A high water content at this depth was confirmed by the observation of very wet firn during drilling.The high water

Brief Communication
https://doi.org/10.1038/s41561-023-01366-1content enabled percolation of water to even deeper parts, or, considering porosity, eventual horizontal runoff along the firn-ice transition.The water-soluble chemical impurities were thereby continuously removed.• With the recent air temperature increase, surface meltwater has formed every summer.Firn layers older than one year have therefore experienced more than one percolation event, leading to a cumulative effect (a basic principle of an increasing number of chromatography elution cycles with time).This process explains why, even though the temperature shift had already occurred in 2015, no effect was apparent in the 2018 core.
Our findings underline the delicate condition even glaciers at the highest altitudes in the Alps have reached due to climate warming.Within a short period of only two years, the summit plateau of the Corbassière glacier has reached a tipping point, leaving this natural archive unsuitable for reconstructing major aerosol components at least for the recent period covered by the firn part.

Online content
Any methods, additional references, Nature Portfolio reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at https://doi.org/10.1038/s41561-023-01366-1.

Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons. org/licenses/by/4.0/.© The Author(s) 2024 https://doi.org/10.1038/s41561-023-01366-1

Methods
Altogether, four firn cores were collected on the summit plateau of the Corbassière Glacier, in the Grand Combin massif, one on 5 October 2018 (GC18; 7.29269° E, 45.93814° N; 4,138 m a.s.l., to a depth of 14.1 m) and three from 15 to 19 September 2020 (7.29265° E, 45.9378° N; 4,123 m a.s.l., to depths of 17.6 m, 18.1 m and 25.3 m) as part of the ICE MEMORY initiative 15 .The firn core GC18 and blank ice were cut in Venice at the Institute of Polar Sciences.The 18.1-m firn core (GC20) and blank ice were cut in the cold room (−20 °C) at the Paul Scherrer Institut (PSI).All the samples were analysed at PSI.
Concentrations of the major ions (for example, NH 4 + , NO 3 − and SO 4 2− , discussed in this study) were determined with ion chromatography (IC, 850 Professional, Metrohm) 20 .Aliquots of the samples were used to analyse the stable isotopes in water (δD and δ 18 O) by cavity ring-down spectroscopy (WS-CR DS, L2130-I Analyser, Picarro) 20 .The firn cores were dated by annual layer counting, considering seasonal variations in δ 18 O and NH 4 + concentration.To estimate the temperatures at the Grand Combin drill site, the instrumental temperature record from Capanna Margherita 27 (Monte Rosa Massif, 4,560 m.a.s.l., 45 km from Grand Combin) was used, because the altitude difference is the smallest of all surrounding stations.Temperatures were corrected with a lapse rate of 0.599 °C per 100 m for a vertical elevation difference of 436 m.The lapse rate was obtained by linear extrapolation of the instrumental temperatures from nearby high-altitude stations in the Swiss Alps (Piz Corvatsch, Gornergrat, Col du Grand St-Bernhard, Jungfraujoch; IDAWEB, Mete-oSwiss) over the period 2007-2020 within a radius of 200 km, including those from Capanna Margherita.The obtained rate agrees well with the environmental lapse rate (0.65 °C per 100 m), which has been shown to vary regionally 28 .Any gaps in the derived Grand Combin temperature series were filled using quantile mapping 29,30 of the hourly data from seven stations (additionally Stockhorn, Mont Blanc Colle Major and Mont Blanc Observatoire Vallot).Missing values in the initial series were replaced by the corresponding arithmetic means of the composite series.
Using the derived Grand Combin temperature record, we calculated the positive degree-day sum (pdd) 31 and the fraction of the day with negative temperature (ndd) 31 .Using pdd and ndd is a simple approach for a melt model relying solely on temperature, as there are no radiation data available for the drill site. .The Colle Gnifetti is a cold glacier site, not influenced by meltwater disturbance, located ca. 45 km east from the Corbassière glacier.This site is prone to wind erosion mainly removing dry winter snow and thus, a high spatial variability is expected.
Accordingly, the major ion concentration records of the two Colle Gnifetti cores in the overlapping period 2003-2015 do partially show interannual differences, but the long-term trends and the cumulative deposition fluxes perfectly agree.The δ 18 O signal as a temperature proxy is more sensitive to the amount of winter snow removal than the major ions, which are transported as aerosol particles by convection in summer.This is reflected in the slightly diverging cumulative δ 18 O records with depth (time).Nevertheless, the example of the Colle Gnifetti site illustrates that even if there is some spatial variability of major ion concentrations, cumulative deposition fluxes between close-by sites commonly agree.
Extended Data Table 1 |

Fig. 1 |
Fig. 1 | Location of the drill site and temperature data.a,b, The Corbassière glacier drill site, shown on a map of western Europe 26 (a) and a picture of the Grand Combin massif (courtesy of Riccardo Selvatico; b).c-e, Air temperature (c), positive degree-day sum (pdd; d) and the fraction of the day with negative temperature (ndd; e), presented for the period 2007-2021, comparing daily and yearly averages.A different scaling was used for daily/yearly average temperatures in c.