Glaciers and ice sheets cover roughly 10% of Earth’s land surface and store approximately 69% of Earth’s freshwater (Meredith et al., 2019), thus serving as a critical natural resource. Moreover, glaciers and ice sheets are important components of Earth’s climate system—responding to global temperature changes, and are a major driver of sea level change (Meier et al., 2007). There has been increased surface melting of glaciers and ice sheets in recent decades (Nghiem et al., 2012; Lenaerts et al., 2013) resulting from both land surface air temperature increases as well as reductions in glacier surface albedo (Box et al., 2012; Ji et al., 2014; Cook et al., 2020; Tedstone et al., 2020)—a change to which the biodiversity of glacier and glacier adjacent habitats is sensitive (Anesio et al., 2017; Cauvy-Fraunié and Dangles, 2019; Stibal et al., 2020). Pigmented algae inhabiting glacier and ice sheet surfaces are part of the light absorbing particulates present on snow and ice surfaces. Such particulates have been shown to decrease surface albedo and increase surface melt rates of both snow and bare ice surfaces (Benning et al., 2014; Dumont et al., 2014; Lutz et al., 2016; Cook et al., 2020; Williams et al., 2020). Among these particulates, pigmented snow and glacier ice algal blooms play a major role in changing the albedo during the late spring to late summer melt season. In the dark zone on the western margin of the Greenland Ice Sheet, glacier ice algae contribute between 13% and 26% of the total annual surface melt (Cook et al., 2020) by lowering the albedo. As the main phototrophic primary producers on glacier surfaces, algae are also important for carbon and nutrient cycling (Anesio et al., 2017) as they produce dissolved organic carbon (DOC; Hood et al., 2009, 2015; Singer et al., 2012) that acts as a substrate for heterotrophic organisms (Stibal et al., 2012) and fertilizes downstream ecosystems when exported in meltwater (Lawson et al., 2014; Musilova et al., 2017; Irvine-Fynn et al., 2021).

Algae are found in numerous different habitats on glaciers, including cryoconite holes (Hamilton et al., 2013; Lutz et al., 2019), snow (Lutz et al., 2015a; Havig and Hamilton, 2019), bare ice (Yallop et al., 2012; Lutz et al., 2016, 2018; Williamson et al., 2018), biofilms (Lutz et al., 2017), and glacial streams (Smith et al., 2017). Such pigmented algae are often visible in distinct red or deep purple blooms on the snow and glacial ice surfaces (Lutz et al., 2016, 2018; Cook et al., 2020; Stibal et al., 2020; Williamson et al., 2020). The most commonly reported green and red snow algae are those of the phyla Chlorophyta and Ochrophyta, and the genera Chloromonas, Chlamydomonas, and Raphidonema (Lutz et al., 2015a), while Hydrurus has been reported as the dominant genus in yellow snow (Remias et al., 2013). In contrast, algal communities on bare ice surfaces are typically dominated by the phylum Streptophyta (Zygnematophyceae) consisting primarily of Mesotaenium (recently reclassified as Ancylonema alaskana; Procházková et al., 2021) and Ancylonema nordenskioeldii (Remias et al., 2009; Yallop et al., 2012; Lutz et al., 2018). Bacteria, archaea, fungi, and viruses are also found in glacial surface habitats (Larose et al., 2010; Maccario et al., 2014; Bellas et al., 2015; Lutz et al., 2015b; Perini et al., 2019). Most of the bacteria and fungi inhabiting glacier surfaces are heterotrophs that can feed on algal exudates (Lutz et al., 2015b), and they drive important biogeochemical processes including the fixation of atmospheric nitrogen (Telling et al., 2011; Larose et al., 2013).

Studies of microbial processes in supraglacial ecosystem have been limited to the spring and summer melt seasons, and an evaluation of variation in summer vs. winter dynamics is lacking. Challenges associated with sampling glacial and high-latitude environments during winter conditions include inaccessibility of sampling sites, adverse weather and light conditions, and low rates of biological activity (Carpenter et al., 2000). Furthermore, deep snow tends to cover glacier surface during winter, making the ice surface less accessible for sampling. Bacterial communities in snowpacks have been well characterized, even in early spring prior to melt (Larose et al., 2010; Maccario et al., 2014), however knowledge of the seasonal dynamics of bacterial, archaeal, fungal, viral, or algal communities is poor. Microbial communities have been analyzed seasonally in other cryospheric environments, including the active layers of permafrost (Schostag et al., 2015), thermokarst lake water columns (Vigneron et al., 2019), and alpine soils (Lazzaro et al., 2015), yet these systems differ significantly from glacial ice and snow and therefore offer little to no comparison of biological, chemical and physical processes.

Here, we assessed the microbial community composition and nutrient variations of glacier surface ice and snow habitats on various Icelandic glaciers, with sampling carried out during two consecutive summers and one winter time period. We analyzed the structure of glacial microbial communities using bacterial and eukaryotic marker genes and measured physico-chemical parameters including the concentrations of DOC, major ions, pH, and conductivity. For eukaryotic communities, we used two marker genes (18S rRNA and ITS) as previous studies (Lutz et al., 2018; Engstrom et al., 2020) showed these to give a more complete overview of eukaryotic diversity. Our data revealed distinct differences between summer and winter supraglacial microbial communities and we discuss their links to variable geochemical signatures of the surface snow and ice environments.

Our data reveal clear seasonal shifts in the microbial community with a link to the geochemical characteristics of the glacier surface environment.

The three dominant eukaryotic phyla found in summer snow and ice samples (Chromista, Fungi, and Plantae) have been found previously on Icelandic glaciers (Lutz et al., 2015a) and elsewhere including snow from the Rocky Mountains (United States; Havig and Hamilton, 2019) and cryoconite holes in Queen Maud Land, Antarctica (Lutz et al., 2019), however in lower relative abundance than we detected here. The glacier ice algae Ancylonema, Mesotaenium, and Cylindrocystis (Figure 4) are also typical of glacier ice samples, as determined by conventional microscopic analysis (Takeuchi, 2001; Takeuchi and Kohshima, 2004; Uetake et al., 2010; Remias et al., 2012; Williamson et al., 2018). However, so far, only two studies have documented a large dominance of Ancylonema and Mesotaenium in glacier surface ice samples using molecular techniques (Lutz et al., 2018; McCutcheon et al., 2021). The clade Phragmoplastophyta, which was present in our 18S rRNA data but not in our ITS2 data (pointing toward a potential PCR primer bias) has previously been shown to comprise the majority of algae inhabiting glacier ice (Takeuchi, 2001; Takeuchi and Kohshima, 2004; Uetake et al., 2010; Remias et al., 2012; Lutz et al., 2018; Williamson et al., 2018; McCutcheon et al., 2021).

The summer snow algal community in our samples largely reflects findings from other similar (i.e., glacier surface snow) environments. In particular, Raphidonema and Hydrurus sp. have been shown to be common on Icelandic glaciers (Lutz et al., 2015a). Interestingly, the red-colored summer snow–ice interface sample from Snæfellsjökull (IS19-12) was dominated by Chlainomonas and not by the previously reported Chloromanas and Chlamydomonadaceae (Lutz et al., 2015a). Chlainomonas has rarely been documented among snow algal communities (Novis, 2002b), and has previously been reported in few locations (Hoham, 1975; Novis, 2002a, b; Remias et al., 2016; Engstrom et al., 2020). The abundance and diversity of algae found in our summer samples was not mirrored in our winter samples; they were fully absent from genomic data from our winter samples. Eukaryotic ice and snow algae require liquid water to bloom (Lutz et al., 2014; Anesio et al., 2017), and the scarcity of water during winter likely prohibits or slows algal metabolism. Moreover, in several of our winter sampling locations (i.e., Snæfellsjökull and Langjökull) glacier ice algae may have been present in deeper ice layers below the snow cover that we could not reach in our sampling efforts. The presence of such layers representing algal communities “frozen-in” at the end of the previous melt season is not well established and should be the target of future studies. Also, it remains unclear why glacial ice algae Ancylonema and Mesotaenium are so under-represented on Icelandic glaciers in comparison to the Greenland Ice Sheet or other Arctic settings where they are often found to dominate the glacier ice algae community (Lutz et al., 2017, 2018).

Bacteria were detected in all samples, but notably, the community composition differed significantly between summer and winter. The winter bacterial community was similar to air samples from Antarctica (Archer et al., 2019) and the Austrian Alps (Els et al., 2020), and also to fresh early spring snow from Svalbard, Norway (Larose et al., 2010; Maccario et al., 2014), based on previous studies that analyzed the same marker gene (16S rRNA) or the entire metagenome. The bacterial communities present in our summer samples shared similarities with those communities characterized in previous studies of summer-time glacial ecology (Hell et al., 2013; Lutz et al., 2015b; Hamilton and Havig, 2017; Havig and Hamilton, 2019).

Heterotrophic groups such as Bacteriodetes, Alphaproteobacteria, and Gammaproteobacteria dominated in all summer samples and belong to known groups that have been identified in marine systems to degrade algae-derived polysaccharides (Ruff et al., 2014; Teeling et al., 2016). These heterotrophs express carbohydrate-active enzymes known as CAZymes (Lombard et al., 2014), such as glycoside hydrolases (GH) that are most likely involved in the degradation of algae exudates (e.g., polysaccharides; Teeling et al., 2016). Cyanobacteria, while previously shown to inhabit cryoconite holes on glaciers and ice sheets (Stibal et al., 2008; Edwards et al., 2011), ice surfaces (Yallop et al., 2012), and some snow (Lutz et al., 2015a, 2016), often classify as Chloroplast in 16S rRNA gene sequencing. However, in this case the majority of sequences classified as Chloroplast are likely derived from the high abundance of glacier ice algae in our algal rich summer samples.

The variability in microbial communities between winter and summer was mirrored in the geochemical signals measured in the snow and ice samples. High concentrations of DOC were associated with summer algal blooms. During summer, glacial meltwater exports significant quantities of DOC (Lawson et al., 2014; Musilova et al., 2017), while during winter, freezing conditions on the glacier surface likely prevent supraglacial DOC export (Hood et al., 2009; Singer et al., 2012). It has recently been shown that primary production by supraglacial algae and other microorganisms over a summer melt season leads to an increase in the bioavailability of the meltwater-exported DOC (Kellerman et al., 2020). The high DOC concentrations measured in our summer 2018 snow samples (between 4 and 20 ppm, Table 2) are at the upper-most range of what has previously been measured for glacial ice (Hood et al., 2015), exceeding typical glacial snowpack and glacial outflow concentrations (Singer et al., 2012; Stubbins et al., 2012; Yan et al., 2016), and are largely in line with measured DOC concentrations in green snowpacks (i.e., snow containing abundant Chlorophyta) in Svalbard (Lutz et al., 2015b), and in ice samples from Antarctica (Barker et al., 2006). We suggest that the relatively high DOC concentrations in our samples compared to others (Singer et al., 2012; Stubbins et al., 2012; Hood et al., 2015; Yan et al., 2016) could arise from (i) our sampling at the end of the summer season (i.e., ~mid-September 2018) compared to other studies sampling between May and August (Singer et al., 2012; Stubbins et al., 2012; Wright et al., 2013; Lutz et al., 2015a,b; Yan et al., 2016), and (ii) the relatively high nutrient loading (and therefore high primary productivity) experienced by Icelandic glaciers due to volcanic input (Bagnato et al., 2013; Galeczka et al., 2017).

The combined organic δ¹³C signature and the C/N ratios for the summer samples provide an indication of the origin of the organic material (Lamb et al., 2006). Summer DOC (Table 2) is likely strongly influenced by microbially derived organic matter, evidenced by low C/N ratios (<13) and low δ¹³C-values (← ~ −25‰), typical of freshwater values (Lamb et al., 2006; Spijkerman et al., 2012). Compared to other glaciers in Svalbard and Sweden, the C/N and δ¹³C characteristics of our summer samples were similar to previously reported green snow and dirty ice samples (Lutz et al., 2017).

We found seasonal differences in the major macro- and micronutrients that are relevant for catabolic and anabolic cell processes (Goldman et al., 2012), for co-factors for enzyme activity, and that are used in anti-freeze protection mechanisms (Garnham et al., 2011). The concentrations of major dissolved ions in all our snow and ice samples (Table 2), fell within the range of concentrations reported in previous studies for Iceland snow and ice samples (Lutz et al., 2015a; Galeczka et al., 2017). In most cases, the ion concentrations were higher during winter than summer regardless of habitat (i.e., bare ice or snow; Table 2). Low summer nutrient and trace elements concentrations (e.g., Na, K, Mg, Ca, SO₄²⁻, NO³⁻, Mn, Mo, and Fe) correspond to the timing of algal blooms (i.e., summer). The availability of nitrogen and phosphorous are known to limit biological productivity of glacier surface ecosystems (Wolff, 2013; Williams et al., 2016; McCutcheon et al., 2021).

The high concentrations of NO₃²⁻ detected in Snæfellsjökull, Skaftfellajökull, and Sólheimajökull winter snowpacks may arise from direct dry deposition of reactive inorganic nitrogen in snow (as observed in Svalbard; Kühnel et al., 2012, 2013; Björkman et al., 2013), biological nitrogen fixation (for example, by Cyano- and Proteobacteria; Telling et al., 2011), by cycling of organic nitrogen (Holland et al., 2019) or microbial necromass from glacial or pro-glacial environments (Bradley et al., 2016). While some chemical measurements point toward nitrification (Telling et al., 2014; Wadham et al., 2016), only one study directly identified putatively involved microorganisms (Segawa et al., 2014). Ammonia-oxidizing microorganisms such as Nitrosomonadaceae (Prosser et al., 2014) were present in our ice samples but were rarely found in our snow samples. Sulfate, in both our summer and winter samples, was likely derived from atmospheric deposition as sea-spray (e.g., Lutz et al., 2015a, 2016), or aerosol sulfate.

Our data reveal seasonal characteristics of glacier surface ecosystems. We show a clear shift in the microbial community structure and corresponding geochemical environmental characteristics from the summer to the winter season.

Summer periods were characterized by a higher availability of microbially derived DOC and low nutrient concentrations, whereas winter was characterized by lower DOC and high nutrient concentrations.

The summer eukaryotic algae inhabiting the glacier surface ice and snow were distinct and were closely associated with a diverse community of heterotrophic bacteria likely relying on algal-delivered DOC. Winter microbial communities contained bacteria resembling communities typically found in air and snow and lacked eukaryotic algae.

The low abundance of glacial ice algae Ancylonema and Mesotaenium during summer in Iceland in comparison to the dominance in Greenlandic glaciers (Lutz et al., 2018; McCutcheon et al., 2021) is still unclear and should be further explored. Additionally, future work should identify the seasonal dynamics of glacier ice algae—specially to clarify whether their apparent absence in winter samples (as suggested in this study) is a consequence of sampling (i.e., algae may be present in deeper snow/ice interface layers) or is due to other ecological factors that drive algal distribution and abundance seasonally. In addition, summer and winter are punctuated by spring thaw and autumn freezing periods, which must also be studied to fully close the remaining gaps in understanding year-round characteristics of glacial ecosystems.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.

MW and LGB designed the study. MW, CBT, and RM carried out laboratory work. MW and CBT performed bioinformatic analysis. MW wrote the first draft of the manuscript. All authors contributed to the sample collection in Iceland and contributed guidance and edits to the final manuscript preparation. All authors contributed to the article and approved the submitted version.

MW, RM, CBT, and LGB were funded through a Helmholtz Recruiting Initiative grant no. I-044-16-0. During the writing of this manuscript, LGB was also supported through an ERC Synergy grant (“DeepPurple” grant # 856416) from the European Research Council (ERC) funded under the European Union’s Horizon 2020 research and innovation program. JAB was part funded by an Alexander von Humboldt Foundation postdoctoral fellowship and a UK NERC grant (NE/T010967/1). The open access publication of this paper was supported within the funding program “Open Access Publikationskosten” Deutsche Forschungsgemeinschaft (DFG, German ResearchFoundation)—project number 491075472.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.