Deposition and postdeposition mechanisms as possible drivers of microbial population variability in glacier ice

Authors


  • Editor: Ian Head

Correspondence: Shu-Rong Xiang, Institute of the Tibetan Plateau Research, Chinese Academy of Science (CAS), Beijing 100085, China. Tel.: +86 10 62849309; fax: +86 10 62849886; e-mail: srxiang@ns.lzb.ac.cn

Abstract

Glaciers accumulate airborne microorganisms year by year and thus are good archives of microbial communities and their relationship to climatic and environmental changes. Hypotheses have focused on two possible drivers of microbial community composition in glacier systems. One is aeolian deposition, in which the microbial load by aerosol, dust, and precipitation events directly determines the amount and composition of microbial species in glacier ice. The other is postdepositional selection, in which the metabolic activity in surface snow causes microbial community shifts in glacier ice. An additional possibility is that both processes occur simultaneously. Aeolian deposition initially establishes a microbial community in the ice, whereas postdeposition selection strengthens the deposition patterns of microorganisms with the development of tolerant species in surface snow, resulting in varying structures of microbial communities with depth. In this minireview, we examine these postulations through an analysis of physical–chemical and biological parameters from the Malan and Vostok ice cores, and the Kuytun 51 Glacial surface and deep snow. We discuss these and other recent results in the context of the hypothesized mechanisms driving microbial community succession in glaciers. We explore our current gaps in knowledge and point out future directions for research on microorganisms in glacial ecosystems.

Introduction

Dust and aerosol deposition (dry deposits) and the association with snow fall (wet deposit) cycles have significant consequences for the glacial ecosystem (Jones, 1999). They clearly affect the frequency and duration of ‘wet/dry’ cycles on a glacier. They directly control the processes in the glacial system through dry/wet deposition cycles, sustain the mass balance of the glacier through addition of snow (Hodson, 2006), and alter the flux of microorganisms into the glacier by microbial transportation from outside environments (Gloster et al., 1982; Shuval et al., 1989; Catranis & Starmer, 1991; Abyzov, 1993; Ma et al., 2000; Prospero et al., 2005; Price et al., 2008). However, the relationship of microorganisms in the glacier to global climatic and environmental changes is not well known. The ephemeral nature of the effect of dry/wet deposition cycles on the microbial and nutrient fluxes into the glacier and the response of glacial microorganisms to seasonal changes in temperature and availability of water contribute to the difficulty of obtaining reliable information on the effects of dust and snow events on the glacial ecosystem. Moreover, the factors controlling the microbial community structure in the glacier remain uncertain. Field investigations have indicated that factors such as geographical location (Yoshimura et al., 1997; Mueller & Pollard, 2004; Takeuchi & Kohshima, 2004), prevalent wind direction, light intensity, and hydrological conditions affect the distribution of microorganisms in a glacier (Christner et al., 2000; Mueller & Pollard, 2004; Takeuchi & Kohshima, 2004). It is essential to understand better how dust and snow events interact with intrinsic spatial and temporal heterogeneities of nutrient sources and temperature on the microbial community in the glacier.

How microorganisms respond to the climatic and environmental changes as soon as they are deposited onto a glacier and are buried in the deep glacier ice during the processes of snow compaction (e.g. firnification) is still an open question, although it has been widely accepted that microorganisms are blown and deposited (aeolian deposition) with snow onto a glacier (Cameron et al., 1972; Baylor et al., 1977; Sun et al., 1978; Castello et al., 1999; Segawa et al., 2005; Yao et al., 2006, 2008; Zhang et al., 2007; Price et al., 2008). On the surface of a glacier, microorganisms are often exposed to an environment with light, nutrients, water, and temperatures as high as 0 °C, which could support microbial activity in snow. Previous investigations suggest the growth of microorganisms in the surface snow (Vincent et al., 1993; Pinckney & Paerl, 1996; Fritsen & Priscu, 1998; Paerl & Priscu, 1998; Säwström et al., 2002; Campen et al., 2003; Stibal et al., 2007). Laboratory experiments (Christner, 2002; Rohde & Price, 2007), modeling (Price, 2000; Price & Sowers, 2004; Rohde & Price, 2007), and conceptual work (Price, 2007; Priscu et al., 2008b) have indicated in situ metabolic activity of microorganisms in glaciers. However, in the deep firn and ice, where the temperature drops to −8 to −56 °C (Salamatin et al., 1998; Thompson et al., 2000; Johnsen et al., 2002), light intensity falls off with depth to zero, and nutrient diffusion and liquid water are also extremely limited, which could greatly reduce or prevent microbial activity at depths where no light penetrates and phototrophic metabolism no longer occurs. This suggests that microbial community composition may shift in response to the habitat transition from the surface snow to the deep ice. Indeed, field investigations report a very common phenomenon that colorful surface snow is dominated by the red-pigmented Chlamydomonas nivalis in many glaciers around the world (Thomas & Broady, 1997; Müller et al., 1998; Stibal et al., 2007) but deep ice is not (Thomas & Broady, 1997). These results point to another important microbial process, postdeposition, which may regulate community changes through the metabolic activity of microorganisms in glacial system.

In this minireview, the spatial and temporal variations of microorganisms in surface and deep snow, and through depth profiles of glaciers are discussed. Because processes in glacial system are extremely complex and only limited data are available, it is not possible to address all of the factors relevant to climatic and environmental effects on glacial ecosystems. We have focused the discussion within the context of two possible mechanisms: aeolian deposition and postdepositional selection, which drive changes in microbial assembly in the glacier. To evaluate postdeposition effects on the microbial community structure in glaciers, we compared the bacterial diversity composition in the surface snow with that in the deep snow in the Kuytun 51 Glacier (84°24′E, 43°43′N), Tianshan Mountains, China (Xiang et al., 2009). We examined how the microbial deposition associated with dry/wet cycles influences the distribution of microorganisms in the glacier by analysis of the microorganisms from Malan (102 m deep, 35°50′N, 90°40′E, Yao et al., 2006) and Vostok ice cores in Antarctica (from 1500 to 2750 m depth, 78°28′S, 106°48′E; Abyzov et al., 1998). Finally, we explored our current gaps in knowledge and suggested future directions for research on microorganisms in glaciers in relation to climatic and environmental changes.

Difference in microbial community at the different habitats in glacial surface snow

The glacier surface usually appears as an apparent transition from snow to firn (the mixed snow–ice environment) to solid-ice/water-flow from the top to the middle to the terminus (ablation) zones (Watanabe et al., 1984; Hoham & Dual, 2001; Takeuchi, 2002; Takeuchi & Kohshima, 2004; Takeuchi & Li, 2008), which causes the strong gradients in light intensity, temperature, nutrients, and water availability on glacial surfaces (Takeuchi et al., 1998, 2001; Takeuchi & Kohshima, 2004; Takeuchi & Li, 2008). In the ablation zone of many glaciers, the unique water-filled cylindrical melt-holes (termed cryoconite holes) are scattered on the glacier surface, are rich in organic matter, and provide niches for complex organisms and biologically mediated chemical reactions (Stibal et al., 2006, 2007; Fountain & Tranter, 2008), while at the snow zone, dry snow accumulates on the surface and contains extremely low organic matter, but has relatively high-albedo/low-light density (Takeuchi & Kohshima, 2004; Takeuchi & Li, 2008). Such strong gradients result in significant heterogeneity of microbial communities on the surface of glaciers worldwide (Table 1). Bacterial phylogenetic analysis of 16S rRNA gene clone libraries from the East Rongbuk Glacier showed the heterogeneity of bacteria diversity at the different glacial habitats on the glacier surface (Liu et al., 2007), and investigations from the southern Yala Glacier in the middle Himalaya Mountains, Nepal, indicated an altitudinal pattern of algae across the glacial surface (Yoshimura et al., 1997). Seven species of algae occurred in the bottom of the glacier (5100–5200 m a.s.l.), the dominant species being Cylindrocystis brebissonii. Eleven species of algae appeared in the mixed snow–ice environment (5200–5300 m a.s.l.), with Mesotaenium berggrenii being the dominant species. Only four species of algae occurred in the snow environment in the upper part of the glacier (5300–5430 m a.s.l.), with Trochiscia sp. being the dominant species. Takeuchi et al. (1998) also observed a similar altitudinal pattern of snow algae across the AX010 Glacier, eastern Himalaya Mountains, Nepal (4950–5380 m a.s.l.). Across the Patagonian Tyndall Glacier, Chile, Takeuchi & Kohshima (2004) found five species of known snow algae: M. berggrenii, C. brebissonii, Ancylonema sp., Closterium sp., Chloromonas sp., and a new species, Oscillatoria sp. (cyanobacteria). The distribution of algae and cyanobacteria varied both along the surface and with altitude. Mesotaenium berggrenii, C. brebissonii, Closterium sp., and Ancylonema sp. were dominant at the low part of the glacier (370–940 m a.s.l.), whereas Oscillatoria sp. (cyanobacteria) and Chloromonas sp. were dominant in the snow–ice environment (940–1300 m a.s.l.). A new algal species prevailed in the upper part of glacier (1300–1500 m a.s.l.). Results from Takeuchi (2001) showed that although C. nivalis was found on the whole surface of the Gulkana Glacier, Alaska, it dominated the upper part of the glacier. Ancylonema nordenskioldii, M. berggrenii, Raphidonema sp., and Oscillatoria sp. (cyanobacteria) prevailed in the middle of the glacier, whereas A. nordenskioldii, C. brebissonii, and Oscillatoria sp. (cyanobacteria) were common at the terminus of the glacier. These results indicate a strong physical–chemical gradient effect on the distribution of the algal and cyanobacterial taxa on glacial surfaces, which provides indirect evidence of a postdeposition effect on the algal and cyanobacteria community on the glacial surface.

Table 1.   Summary of indicated algae and cyanobacteria in the different glaciers*
NicheDominant algae and cyanobacteria with altitude
Yalaa 28°C14′N, 85°C36′EAX010b 27°C42′N, 86°C34′ETyndallc 51°C15′S, 73°C15′WGulkanad 63°C16′N, 145°C25′W
  • *

    The dominant algae and cyanobacteria were identified by microbial morphological characteristics under optical microscope (Nikon E600) with 0.5% erythrosin cell stain. For statistic analysis, five snow and ice samples were randomly collected by using a stainless-steel scoop at depths of 1–2 cm at each niche, and the specific algal and cyanobacteria cell numbers were counted under microscope in three to six replicates for each subsample (see the detailed methods in references).

  • a

    Yoshimura et al. (1997).

  • b

    Takeuchi et al. (1998).

  • c

    Takeuchi & Kohshima (2004).

  • d

    Takeuchi (2001).

Snow5300–5430 masl5300–5430 masl1300–1500 masl1650–1800 masl
Trochiscia sp.Trochiscia sp.new algal speciesChlamydomonas nivalis
Snow-ice5200–5300 masl5200–5300 masl940–1300 masl1500–1650 masl
Mesotaenium berggreniiOscillatoria sp. (cyanobacteria)Oscillatoria sp. (cyanobacteria)
Chloromonas sp.
Ancylonema nordenskioldii,
Mesotaenium berggrenii,
Raphidonema sp.
Oscillatoria sp. (cyanobacteria)
Solid ice5100–5200 masl5100–5200 masl370–940 masl1270–1500 masl
Cylindrocystis brébissoniiCylindrocystis brébissoniiMesotaenium berggrenii, Cylindrocystis bréissonii,
Closterium sp., Ancylonema sp.
Ancylonema nordenskioldii,
Cylindrocystis brébissonii,
Oscillatoria sp. (cyanobacteria)

Great variability in algal and cyanobacterial biomass is also significant in different habitats on the glacial surface, which indicates both aeolian deposition and postdeposition effects on the quantitative distribution of microorganisms (Fig. 1). Studies of the Himalayan Yala Glacier have shown a decreasing occurrence of the total algal and cyanobacterial biomass with increasing altitude (Yoshimura et al., 1997). Data on the Gulkana glacial algal biomass in Alaska have also shown a decreasing presence at the upper part of the glacier at 1600–1800 m a.s.l., but an increasing trend with altitude in ice to snow–ice at 1270–1600 m a.s.l. (Takeuchi, 2001). The discrepancy in the altitudinal distribution of total algal and cyanobacterial biomass between the Himalayan glacier and the Alaska glacier may be due to the difference in the microbial load on the glacier surface and microbial ecological conditions between the two glaciers. The frequent coverage of snow during the summer monsoon on the Himalayan Yala Glacier reduces the temperature and light intensity in summer, and thus affects algal and cyanobacterial growth in snow, resulting in a lower biomass at the upper part of the glacier than at its middle-bottom. The algal and cyanobacterial biomass is affected by the limitation of available meltwater in the glacial snow environment, whereas algae and cyanobacteria are often exposed to light and meltwater in the snow and ice habitat and are less affected by the snow cover (Yoshimura et al., 1997; Takeuchi et al., 1998). The lower algal biomass in the ice zones at the bottom compared with the middle-upper part of the Gulkana Glacier (Fig. 1) could be attributed to wash-out of unicellular algae from the microbial community, because the filamentous cyanobacteria form aggregates with particles and organic matter, which helps trap them in the ice, whereas unicellular organisms do not form such aggregates and so are readily washed out of the ice (Takeuchi, 2001).

Figure 1.

 Variations in the total biomass of phototrophs in the different habitats in the top surface snow on Gulkana Glacier (1200–1800 m) and Yala Glacier in Himalayas (5100–5500 m, Yoshimura et al., 1997). Error bar: SD. The algal and cyanobacterial biomass was calculated using an optical microscope (Nikon E600) and represented by the total algal and cyanobacterial volume collected on the surface snow/ice per unit area (adapted from Takeuchi, 2001).

Difference in bacterial phylogenetic composition between surface and deep snow

Colorful snow cover, frequently present on glaciers around the world, is caused by red-pigmented Chlorophyta (green algae), usually C. nivalis (Thomas & Broady, 1997; Müller et al., 1998; Takeuchi et al., 2006; Stibal et al., 2007). An extensive survey of 45 sites at Tongariro National Park in the North Island and the Southern Alpine of New Zealand from 1991 to 1995 showed the presence of Chlamydomonas sp. only in the top 15 cm of snow (Thomas & Broady, 1997). This suggests the important role of photosynthesis in the shift in microbial communities observed during the transition through the snow depth profile. Cyanobacteria were reported in Svalbard glaciers (Stibal et al., 2006), polar glaciers (Mueller & Pollard, 2004; Priscu et al., 2005; Fujii et al., 2008), and other mountain glaciers around the world (Takeuchi et al., 2001; Takeuchi & Kohshima, 2004; Segawa et al., 2008). These results support the conclusion that cyanobacteria and snow algae are the common phototrophs in the glacial system, and their primary productivity sustains cold-tolerant heterotrophic organisms in the surface snow (Aitchison, 2001; Hoham & Duval, 2001). Further investigations of the phylogenetic composition of bacteria between the surface and deep snow using PCR amplification, cloning, and sequencing of 16S rRNA genes from the Kuytun 51 Glacier samples confirmed the concept of postdeposition effect on the microbial communities in the glacier ice (Xiang et al., 2009). Bacterial phylogenetic data from the Kuytun 51 Glacier showed that phototrophic cyanobacteria occurred on the whole glacial surface and accounted for 5–26% of the total clones in five of eight 16S rRNA gene clone libraries, but they were below the limit of detection in three libraries from deep snow at 25–295 cm below the top snow cover (Fig. 2). Further statistical analysis of microbial communities from the glacier surface, deep snow, and deep ice demonstrated the significant heterogeneity of the microbial community composition (Fig. 3). The dramatic community diversity shift between the surface and deep snow/ice coincides with the transition of microbial habitats from the light-rich conditions in the surface to the low-light intensity environments in the deep snow, and presents direct evidence of postdeposition effect on the community succession of microorganisms in the glacier.

Figure 2.

 Clonal frequency of the main bacterial phylogenetic groups in each of eight clone libraries established using PCR amplification, cloning, and sequencing and blast analysis of 16S rRNA genes from the surface and deep snow in the Kuytun 51 Glacier (data adapted from Xiang et al., 2009).

Figure 3.

 Hierarchical cluster analysis showing the overall phylogenetic distances among the clone libraries from surface, deep snow, and deep ice. The surface and deep snow bacterial sequence data were from our results from the Kuytun 51 Glacier (Xiang et al., 2009), and the deep ice core sequence data from the Malan Glacier (Xiang et al., 2004). Distances were estimated with the weighted UniFrac algorithm (Lozupone & Knight, 2005). A sequence jackknifing technique was applied to each cluster to determine the sensitivity of the relationships to sample size. UniFrac distance indicated difference among the bacterial communities. The bacterial population structures were significantly different between the Malan and Kuytun Glacier (P<0.01).

On glacial surfaces, microorganisms are often exposed to light, nutrients, water, and a temperature of around 0 °C in summer, which facilitates the growth of phototrophic cyanobacteria. By contrast, in the deep snow and ice, where the temperature abruptly drops to −8 to −56 °C (Salamatin et al., 1998; Thompson et al., 2000; Johnsen et al., 2002), the snow cover limits the ambient light intensity and thus reduces the growth rate of cyanobacteria, which are light-dependent species (Pandey et al., 1995; Takeuchi & Kohshima, 2004). Therefore, it is not surprising that a shift in microbial communities from the prevalent phototrophs in the surface snow to the dominant heterotrophs in the deep snow occurs in the Kuytun 51 Glacier. Furthermore there are no visible Chlamydomonas in the deep snow at the Tongariro National Park in the Southern Alpine of New Zealand (Thomas & Broady, 1997). This suggests that some species such as phototrophic cyanobacteria present in the surface may not frequently spread to the deep snow or ice, where they are buried by the overlying snow when postdeposition processes are strong in summer season.

Influences of aeolian activities on the distribution of microorganisms in ice cores

Variations in the physical–chemical and biological parameters of ice cores from glaciers show the strong aeolian deposition effects on the distribution of microorganisms in the glacier ice (Figs 4 and 5). Analysis of the deep ice cores from both Malan Glacier and an Antarctic glacier (Vostok station) showed that high biomass was usually found in the dirty layers, whereas low biomass was found in the clean ice layers (Figs 4 and 5). This suggests an association of microbial load with dust deposition in the glacier ice. However, analysis of the 13-m-long Muztag Ata ice core showed that high biomass was found in the clean ice layers, whereas low biomass presented in some dirty layers (Xiang et al., 2006). Low biomass was also found in the dirty layers (at 87 m depth) in the Vostok BH5 ice core (at 30–130 m depth; Priscu et al., 2008b). This suggests that microbial load might not always be associated with the dust deposition in the glacier.

Figure 4.

 Bacterial biomass, mineral particles, and δ18O in the Malan ice core. (a) Presence of highly dirty layers along the depth profile (adapted from Yao et al., 2006). (b) Total bacterial biomass (adapted from Zhang et al., 2002). Total bacterial cells were estimated using Olympus BH-2 microscopic counts of acridine orange-stained cells. (c) Eleven-point running average of δ18O value (adapted from Yao et al., 2006). The δ18O value was measured by Finnegan MAT-252 gas stable isotope ratio mass-spectrometer. The upper 24-m ice core was annually dated using seasonal δ18O variations and annual visible dust layers (Wang et al., 2003a, b), and the bottom ice core was dated by δ18O variations (Yao et al., 2006). H1 and H2, high biomass 1 and 2, respectively; L1 and L2, low biomass 1 and 2, respectively; W1 and W2, warm periods 1 and 2, respectively; C1 and C2, cold periods 1 and 2 (as indicated by the grey-shaded areas), respectively.

Figure 5.

 Bacterial biomass, mineral particles, and δ18O in the Vostok station ice core at 1500–2750 m depth. (a) Values of oxygen isotope ratio (18O/16O, adapted from Jouzel et al., 1993 in reference Abyzov et al., 1998). (b) Total microbial biomass (adapted from Abyzov et al., 1998). Total microbial cells was estimated using luminescent microscopic (LYUMAM-1-2) counts of fluorescent dye fluorescamine-stained cells. (c) Mineral particle concentration along the depth profile (adapted from Petit et al., 1990 in reference Abyzov et al., 1998). Total microparticle concentrations were measured using a Coulter counter. The high biomass presented during the extremely cold period is indicated in the grey-shaded area.

Previous investigations of ice cores showed different effects of temperature on the distribution of microorganisms in glacier ice. Both Malan and Vostok station ice core data showed a negative correlation between microbial abundance and temperature (Abyzov et al., 1998; Yao et al., 2006), but the recent reports from shallow ice cores (a 25-m-long ice core, 49°47′N, 87°43′E) from the Sofiyskiy Glacier in the Altai Mountains of Russia (Uetake et al., 2006), Guoqu Glacier in the Geladaindong Mountain regions (a 47-m-long ice core, 33°34′N, 91°10′E; Yao et al., 2008), and Muztag Ata Glacier (a 13-m-long ice core, 38°17′N, 75°04E, Xiang, 2005) showed a positive relationship between microbial biomass and temperature. The discrepancy in the long and shallow ice core data could be attributed to the different microbial response to the global temperature changes on different time scales. Malan and Vostok station ice cores recorded the historical period from ad 1170 to 1999 (Yao et al., 2006), and 100–242 thousand years before (kyr bp, Abyzov et al., 1998), respectively. High biomass as indicated by the grey-shaded areas in Figs 4 and 5 could be related to the strong aeolian activities during the extremely cold period (Peltier, 1950; Thornbury, 1954) that led to strong transportation and high influx of microorganisms to the glacier surface, while microbial abundance in the warm summer seasons, as indicated in the shallow ice cores from the Muztag Ata, Sofiyskiy, and Guoqu Glaciers (Xiang, 2005; Uetake et al., 2006; Yao et al., 2008), could be a result of high microbial influx and microbial activities and growth in the surface snow in summer.

Possible mechanisms controlling the distributions of microorganisms along glacial depth profiles

The community shifts as a result of microbial metabolism in the surface snow as observed from the Kuytun 51 Glacier raise two questions about the mechanisms responsible. The first is how much of the microbial community in the deep ice was a result of postdeposition processes that occurred when meltwater was available in the surface snow in the warm summer season. In the case of the Kuytun 51 Glacier, it was clear that strong microbial activity in (or near) the surface snow caused a significant shift in the composition of microbial communities along the depth profile (Figs 2 and 3). The phototrophic cyanobacteria were very common across the glacier surface, but rarely in the deep snow (Fig. 2). Lack of phototrophic cyanobacteria in the deep snow is most likely due to low photosynthesis with decreasing light intensity with increasing snow depth. However, phototrophic cyanobacteria and algae were reported throughout a 25-m-long ice core extracted from the Russian Altai Glacier (Uetake et al., 2006), and in the deep snow in the East Rongbuk Glacier on the northern slope of the Himalayas (Liu et al., 2006), in the Yala Glacier on the southern slope of the Himalayas, Langtang region of Nepal (Yoshimura et al., 2000), and in the Guoqu Glacier at Mount Geladaindong in the north of the Tibetan Plateau (Yao et al., 2008). Differences in the distribution of phototrophs in the deep snow/ice are possibly attributable to the different microbial influx and community shift in response to changes in seasonal temperature and light intensity on the glacier surface during the deposition period. It is most likely that the postdeposition process through microbial activity in snow/ice has greater influence on the microbial community in the temperate Southern Alpine of New Zealand (Thomas & Broady, 1997) and in subcontinental-maritime glaciers (e.g. Kuytun 51 Glacier, Xiang et al., 2009) than in the continental-maritime (subpolar) glaciers (e.g. East Rongbuk Glacier, Liu et al., 2006; Geladaindong Glacier, Yao et al., 2008). In the extremely cold-based continental-maritime and continental glaciers, the low light intensity and temperature in the surface snow not only greatly limit the growth of phototrophs in the surface snow but also reduce microbial growth in near-surface snow. The phototrophs may therefore be preserved in the deep snow/ice in the high mountain glaciers. High biotic influx and seasonal change in the microbial activity near the surface snow may partially explain the frequent presence of phototrophic cyanobacteria and algae in the deep ice in the temperate (maritime) glaciers (e.g. Russian Altai Glacier, Uetake et al., 2006; Himalayan Yala Glacier, Yoshimura et al., 2000). In the temperate glaciers such as Akkem Glacier in the Russian Altai Mountains and Himalayan Yala Glacier, the relatively nutrient-rich conditions with higher concentrations of organic matter in the Yala Glacier than in the Arctic Devon Glacier (Takeuchi & Li, 2008), and relatively mild environments with lower-albedo/higher-light-density in the Yala Glacier than in the inland Gulkana Glacier of Alaska (Takeuchi, 2002) were able to support highly diverse phototrophs in summer. A large amount of phototrophs might be not completely decomposed by bacteria near the surface snow and thus could be buried by the overlying snow and preserved in the deep ice. However, the question of how much of the community composition is dictated by postdeposition processes still remains undecided.

Only recently has there been a growing recognition that microbial processes through microbial activity and growth under icy environments can be an important driver in glacial ecosystems (Stibal et al., 2006, 2007; Foreman et al., 2007; Fountain & Tranter, 2008; Hodson et al., 2008). The metabolic activities of the microorganisms under icy environments (Thomas & Broady, 1997; Paerl & Priscu, 1998; Priscu et al., 2005; Stibal et al., 2007) could drive the dynamics of carbon, nitrogen, and energy sources under the glacier ecosystem, and the material made available might then support the growth and survival of cold-tolerant microorganisms, which would drive the dynamics of microbial communities in glacial systems (Fig. 8). The Kuytun 51 Glacier data clearly indicated that microbial communities may turn over in the surface snow when meltwater and nutrients are accessible to the microorganisms deposited onto the glacier surface (Figs 2 and 3). Phylogenetic analysis of bacterial community structure from more ice cores will be helpful to improve understanding of microbial community dynamics.

Figure 8.

 Scheme showing the proposed mechanisms driving microbial community shifts in glaciers and interaction among environmental drivers, microbial physiology, community composition, and ecosystem processes.

The other question is how the two mechanisms (aeolian deposition and postdepositional selection) interactively regulate the distribution of microorganisms in a glacier. It is most likely that the two mechanisms are in effect simultaneously, each having a different dependence on glacial depth, climatic changes in temperature, light intensity, and wind speed during the deposition periods. The first occurs before deposition: microorganisms are transported by wind with organic matter, aerosol, and dust (dry deposition) onto the glacier surface (see Figs 6, 7a, b and 8) and thus are deposited and preserved in the deep ice as observed from Figs 4 and 5. Microorganisms are also transported by aerosol, dust, and snow or snow alone (wet deposition, Fig. 8) to the glacier surface. The wet deposition can be an abrupt event, as documented by a snow fall in Changchun city in northeastern China, associated with a serious dust storm that occurred 2000 km away in north China during April 5–7, 2000 (Hu et al., 2001), and resulted in a high concentration of bacteria associated with dirty snow to 8.9 × 104–1 × 105 CFU mL−1. Microbial deposition may be occasionally associated with dust events (Fig. 6). Microorganisms are extremely small (e.g. the mean estimated diameter of bacteria in the Muzta Ata Glacier is ∼0.45 μm), and can be transported easily and thus freely deposited onto glacier surfaces. The biomass of microorganisms in the atmosphere fluctuates (Figs 6 and 7a), but this is not always associated with changes in dust content (Fig. 6). The disassociation of microbial particle transportation in the atmosphere can be partially explained by the biological ice nucleators, and thus microorganisms act as tiny catalysts of ice nucleation and incorporate the moisture vapor, dust, and aerosol into the tiny microbial assemblages and drop onto the glacier surface (Christner et al., 2008).

Figure 6.

 Rate of arrival of bacteria (red points) and dust (blue points) blown from African desert sources to an air collector on Barbados during 1996–1997 (adapted from Prospero et al., 2005 in Rohde et al., 2008).

Figure 7.

 Seasonal changes in bacterial biomass and mineral particle density in the surface snow and the atmosphere over the Tateyama Mountain, Japan. (a) Cumulative numbers of bacteria and mineral particles supplied from the atmosphere to the snow surface. (b) Total bacterial biomass and numbers of small mineral particles in the surface snow. Bacterial biomass and mineral particle density were estimated by microscopic counts of DAPI-stained cells and particles (adapted from Segawa et al., 2005).

The second mechanism, occurring after deposition, depends on ambient temperature in the surface and deep snow/ice and snow accumulation rate during the processes of snow compaction. Following deposition onto a glacier, the microbial processes can drive a community shift, as observed from the Kuytun 51 Glacier, which could be strongly affected by ecological selection and microbial adaptation to the extreme environmental conditions (Fig. 8). In the deep ice, the atmospheric gases, marine salts, and nutrients in terrestrial dust grains that become trapped in the ice could serve as nutrients for microbial metabolism (Price, 2000, 2007). However, the community may remain little changed in the deep ice in the continental (polar) and subcontinental (subpolar) glaciers because of a rapidly decreasing metabolic rate with decreasing temperature (Price & Sowers, 2004). The size of the biomass pool and community composition probably depends on the current meteorological and climatological factors rather than simply on glacial depth, which means that biomass and perhaps microbial community structure on the surface of a glacier still reflect the climatic and environmental conditions before burial, despite considerable changes in microbial community structure after deposition. The aeolian deposition most likely determines the initial size of the microbial biomass pool and the species composition as a function of depth, while postdeposition processes usually modify this pattern because the availability of nutrients and water in melting snow initiates microorganism activity in icy environments (Thomas & Broady, 1997; Paerl & Priscu, 1998; Priscu et al., 2005; Stibal et al., 2007), thus altering the community structure during warm summer seasons following a large deposition onto the glacier (Thomas & Broady, 1997; Xiang et al., 2009). The products of the two processes remain in the snow and ice strata.

Approaches to understanding microbial deposition and glacial ecosystem

The effect of dry/wet deposition, microbial deposition with dust and aerosol (dry deposition) associated with snow fall (wet deposition) on biogeochemical processes and microbial responses, is much more difficult to determine, although there is some limited evidence of microbially mediated nutrient cycling (Jones, 1999; Hodson et al., 2005; Hodson, 2006) and biogeochemical effects on the glacial ecosystem (Skidmore et al., 2008). It is therefore necessary to apply a multitude of approaches to disentangle the various effects of microbial deposition on the glacial ecosystem in order to identify the main processes that control biogeochemical activities after deposition on the glacial ecosystem.

One approach is the characterization of the microbial deposition itself, in terms of monitoring the dynamics of microbial deposition onto the glacier and its effect on the status and activities of microorganisms and the glacial ecosystem. Manipulation of microbial deposition can be performed under field conditions (Segawa et al., 2005). In a field experiment, investigation of microorganisms deposited onto the glacier involves the use of a mobile fixed trap for collection of atmospheric samples that arrive at the collector (Segawa et al., 2005). Data on dynamics of microorganisms in the surface snow make a comparison of microorganisms in the atmosphere with those in snow possible (Segawa et al., 2005), which can help to elucidate the postdeposition effect on the distribution of microorganisms in the glacier ice.

Stable isotopes and the recent development of pulse–amplitude modulation fluorometry show great promise for elucidating the relationships between microbial activity and biogeochemical cycling in cold environments (Fernández-Valiente et al., 2007; Stibal et al., 2007). Analysis of δ13C, δ18O, and δ15N of microbial mats in Antarctica has been very useful in advancing our understanding of the nature of microbial processes under extremely cold and oligotrophic environments (Fernández-Valiente et al., 2007; Priscu et al., 2008a) and postdeposition effects on the distribution of microorganisms in the deep ice. Pulse–amplitude modulation fluorometry has also been used to examine the physiological state and photosynthetic activity of the snow alga C. nivalis in a snowfield in Svalbard (Stibal et al., 2007). Studies on the activities of microorganisms in glacier snow using labeled carbon and nitrogen sources (with 15N and 13C) can be applied in conjunction with temperature manipulation (Novis et al., 2007) to examine the effects of seasonal temperature and environmental changes on the ecological processes of the glacial system.

Application of both the currently available tools and impending advances in molecular approaches to the glacial ecosystem will enable a better understanding of microbial diversity and how the microbial community changes in response to temperature changes at seasonal and long-term time scales and factors such as nutrient content in the current glacier surface snow. The clone-library sequencing method and molecular biology fingerprinting techniques such as membrane lipid analyses (phospholipid fatty acids, Kytöviita & Fritze, 2008), denaturing gradient gel electrophoresis (Zhang et al., 2006), terminal restriction fragment length polymorphism (Bhatia et al., 2006; Lazzaro et al., 2008; Twin et al., 2008), and amplified fragment length polymorphism (Gunde-Cimerman et al., 2008) are being used to characterize the diversity of microorganisms in glacial and alpine environments. Real-time quantitative PCR (qPCR) is being used to increase our understanding of the dynamics of microorganisms in ice (Segawa et al., 2005). These investigations have shown differences in the diversity of microbial communities in glacier ice, which reflects the influence of climatic and environmental conditions (e.g. temperature and nutrient sources) on the microbial communities in the glacier. Phylogenetic analysis of community diversity needs to be related to data on microbial function through analysis of microbial metabolic activities and traditional culture techniques for isolation of the special functional bacterial groups such as phototrophic C. nivalis and cyanobacteria, which could improve our understanding of how ecological function changes and consequent microbial community shifts in the glacial surface snow during the deposition period. Sequencing of metagenomic libraries from glacial ice is being used to produce a large amount of information about biological diversity and to advance our understanding of the nature of organisms in glacial ice (Lee et al., 2008; Pipers et al., 2008). DNA hybridization techniques such as FISH (Lazzaro et al., 2008) and microarray-based transcription profiling using specific molecular probes can be used to accurately identify biological diversity at a very fine taxonomic and ecological functional level, which can be helpful to qualitatively determine the contribution of the specific species to the microbial community pool in the ice. Techniques such as live/dead stains and fluorescence-labeled live-cell-based qPCR have made it possible to distinguish between living and dead cells in ice (Miteva et al., 2004; Wang & Levin, 2006; Xiang et al., 2009), to identify specific living phylogenetic groups (Wang & Levin, 2006), and thus can be explored to quantitatively estimate the dynamics of the dominant live species in the ice. The above molecular tools and techniques need to be extended for identification of the main functional groups that control the size of the pool of living biomass in ice. Microbial taxonomic diversity studies in combination with the measurements of metabolic activities of microorganisms can be helpful to improve our understanding of microbial processes in the glacial ecosystem.

Heuristic and mathematical models to simulate the microbial processes that are very difficult to measure in the field or in the laboratory enable predictions and identification of key processes for future research. These models have shown that both aerobes and strict anaerobes can coexist within an isolated ice lattice and can metabolize at finite rates even at very low temperatures (−40 °C) and even in a diffusion-limited microenvironment (Price, 2000; Price & Sowers, 2004; Rohde & Price, 2007). These mathematical models provide the framework for advanced experimental tools that are complementary to field and laboratory experiments for addressing the numerous effects of microbial postdeposition on glacial ecosystem processes.

Finally, to achieve a full understanding of the effects of global climatic and environmental changes on the size of the microbial population pool and community diversity and biogeochemical cycling of the glacial ecosystem, additional multidisciplinary studies such as microbial analysis in combination with physical–chemical parameters (e.g. ice core dating and texture characteristics) are required to explore the nature of glacial ecosystem. International cooperation, as demonstrated so successfully in the Vostok and GISP2 ice core projects, will be essential for a complex and integrated investigation using microbiological, geochemical, geophysical, and geographical techniques.

Acknowledgements

We extend our special thanks to Dr P. Buford Price for great efforts to improve this paper. We would also like to thank Dr Shiro Kohshima and the anonymous reviewers for their helpful comments on improvement of this paper. This work was supported by the NSF project of China (Grant 40471025 and 40571038).

Ancillary