The distinctly different past and present characteristics and activities of the studied glaciers at the two catchments manifested as distinct habitat templates within a range of extremely rough to benign conditions. These conditions were condensed into three groups by the GIm: extremely harsh (MC1 and MC2), harsh (KC1–KC3), and benign (MC3, MC4, and KC4). The MC is representative of a high degree of glacierization, but also of a glacier (Pasterze glacier) that has rapidly decreased in size during the last few decades. The latter accounts for the two distinct sections of the MC: (1) a highly unstable proglacial, comprising the glacier source (MC1) and MC2, which consists mainly of unconsolidated moraine material, and (2) the consecutive silted floodplain area (Krainer and Poscher 1992). The ongoing rapid loss of ice masses during the study period (Fischer 2010) led to a strong glacier influence at the proglacial (e.g., low water temperature and a low GIm) that was distinctly modified along the silted floodplain area (e.g., high water temperature and high GIm). Such silted floodplain areas can often be ascribed to large melt water streams (Krigstom 1962; Hodgkins et al. 2009).
In contrast, the KC is a typical example of a less glacierized catchment where the accumulation area now predominates after the wide and effective retreat of the ablation area. It is also representative of the fragmentation of a once contiguous glacial surface, as is similarly expected for many glaciers in response to predicted climate changes (Jiskoot and Müller 2012). This fragmentation, resulting partly in cliff glacier remnants, was responsible for habitat conditions that contradicted a longitudinal gradient of amelioration along the KC main channel sites, particularly during the melting period. Cliff remnants at KC3 intruded upon a proposed amelioration gradient, as their summer-time effluents caused high current-water velocities at KC3. These findings suggest that the downstream development of glacier streams as a consequence of ongoing glacier retreat will abrogate any downstream longitudinal increase in benignity in glacier rivers (see Milner et al. 2001) when a stream accompanying the glacier remnants develops in parallel.
One of the basic assumptions in this study, namely, that the two catchments cause significant differences in the MMB because of differences in their harshness – defined according to the percentage of glacierization, glacier retreat patterns, and stream ages – was not confirmed. Nevertheless, single community descriptors (e.g., algal, protist, nematode, and rotifer species numbers; rotifer and nematode H′ diversity; and nematode maturity), might better capture the harshness of the catchments, as evidenced by the lower values in the MC, indicative of higher overall catchment harshness. Structural alterations in the macrozoobenthos were shown to occur with decreasing glacerization, with an increase in the numbers of individuals and species with ongoing glacier retreat in alpine areas (Füreder 2007). Our study comparing the KC and MC is the first to analyse similar causalities for species number of the micro- and meiozoobenthos, comprising protists, rotifers, and nematodes. Thus, as for the macrofauna, we also found evidence of an increase in the taxonomic richness with decreasing glacierization, at least to the extent occurring between the two catchments (from 60% to 25%).
Between-catchment comparisons, as discussed above, are generally rare in glacier river studies. More often, the focus is on the distribution of community patterns (abundances, richness) as a function of varying habitat harshness along three main axes: (1) the longitudinal decrease in harshness with increasing distance from the glacier margin, which should typically be paralleled by an increase in stream age; (2) a seasonal decrease in harshness outside the glacier ablation period (1 and 2: Ward 1994; Milner et al. 2001); and (3) a lateral component enabling the existence of less extreme rhithral, glacio-rhithral, and krenal habitats (Tockner and Malard 2003; Uehlinger et al. 2003). Most of the studies referring to these axes have reported an increase in structural complexity with respect to the diversity and/or abundances of their studied organisms together with a decreases in harshness (Battin et al. 2001; Lods-Crozet et al. 2001a,b; Robinson et al. 2001; Burgherr et al. 2003; Bürgi et al. 2003; Malard et al. 2006; Eisendle 2008). However, observations refuting these basic schemes have also been made, for example, a report of a decrease in fungal species numbers in the river downstream (Gessner and Robinson 2003) and the results of this study (but see below).
In this study, taxon numbers of benthic fungi and of most of the other groups decreased (see Table 3) downstream from the MC glacier source to MC2, and from KC2 to KC3 although by different amounts. MMB abundances also partly decreased in the river downstream along the MC and KC main channel sites. These observations deviate from the abovementioned scheme in glacier river research, in which a longitudinal downstream increase in taxonomic richness and abundances in relation to downstream habitat amelioration and succession progress (stream age). However, these deviations are consistent with the particular catchment characters in this study, as already discussed herein, namely, the instability of MC2 and the lengthened, glacier-accompanied main channel course of the KC. Finally, to answer the question whether harshness or stream age has greater determining relevance for the MMB, comparisons among KC3 and KC4, as well as KC1 and MC2, and KC4 and the sandur complex lead to the conclusion that benignity and harshness are more important than stream age, both within and between catchments.
In its original form, the GI was introduced to improve the relational analysis of glacier influence on macrozoobenthic patterns (Ilg and Castella 2006). The GI integrates within-stream conditions that change with varying glacier influence (water temperature, EC, BI, and turbidity). Despite several shortcomings, the index is relatively easy to apply and, in most cases it distinguishes complex habitat conditions, relating them, more or less strongly, to community patterns (Ilg and Castella 2006; Brown et al. 2010). The reasonable reach grouping confirmed the utility of the modified version developed for this study. More importantly, perhaps, ours is the first application of the GI to the MMB and it revealed well-established relationships for most of the micro- and meiobiota characters. These relationships were often stronger than those previously reported for the few macroinvertebrate structural parameters (richness, abundances) examined thus far (Brown et al. 2010). Specifically, the GIm clearly indicated the positive response of certain community parameters (e.g., taxonomic richness of nematodes, diatoms, and other algae; abundances of bacteria, flagellates, algae, and nematodes) to decreasing glacier influence, which in itself entails complex and reach specifically altered conditions between and among catchments.
Predictions on the nature of the MMB changes along glacier-dependent river systems, however, should be generalized only with caution, based on the varying results described herein for the different groups. A summary of the results obtained for relation patterns between the MMB and several abiotic parameters clearly showed, for example, that an increase in water temperature is insufficient to promote an increase in the abundance or taxonomic richness of most groups when it is accompanied by high-current velocity and low-bed instability. The predicted summer-time shift to precipitation-driven effluents in alpine catchments as discussed in Milner et al. (2009) might therefore have similar repressive effects on MMB richness and density, due to their flashy, erosive, and less predictable run-off character. A shift of water sources might also bear serious consequences for catchment functionality if there is a loss of benign habitats in response to a loss of groundwater recharge function during the ablation period (Tockner and Malard 2003). It is particularly these habitats that, regardless of their bed sediment texture, are important local hot spots of diversity and thus of functionality.
Selected community aspects
This study adds new and important information regarding the autotrophic component of glacier-fed rivers. Our results suggest that high-diatom species richness is not restricted to the epilithon or to those habitats with diminished glacial influence and are thus in contrast to previous studies (see Rott et al. 2006). In particular, the number of diatom species at silted floodplains highlights the need for greater attention to these habitats in future research. However, it is not only diatom species numbers that seem to be distinctly underrepresented when exclusively determined on epilithic substrates, as indicated by other algal species number among KC reaches (compare Bürgi et al. 2003). A comparison of the photoautotrophic components among the MC and KC reaches showed the delayed succession of a diverse autotrophic community not limited to diatoms in riverine habitats shaped by rapid deglacierization. Finally, ANFs, although widely neglected, may be important primary producers in glacier-fed river habitats. As a subject of cryoconite research, their general ecological importance within glacier biomes has already been demonstrated (Hodson et al. 2008; Anesio et al. 2009). Based on their distribution patterns, as determined in our study, ANFs can be considered as a constant and seemingly adaptive component within carbon cycling processes in glacier-fed river habitats driven by varying glacier influences.
Despite consistently strong differences in the thermal and hydrological regimes of the two catchments, there were no significant corresponding differences in the resemblance patterns of the bacterial and flagellate microbiota. Specifically, microbiota densities might be positively rather than negatively affected by a huge loss of ice masses. This conclusion is based on the abovementioned growing appreciation of glaciers as important habitats of microbial organisms, thereby fulfilling important ecological functions on a regional as well as a global scale (Kastovska et al. 2007; Mindl et al. 2007; Hodson et al. 2008; Anesio et al. 2009; Stibal et al. 2012; Viles 2012). Bacteria have been shown to rapidly alter their genetic expression patterns, changing from swimming to sessile life forms (Hall-Stoodley et al. 2004). This could be advantageous in shifts from glacial to riverine benthic habitats, such as by facilitating colonization at recently deglacierized sites, in favoring downstream drift, and in the recolonization of unstable habitats.
The distribution patterns of the fungal CFU counts and fungal species numbers did not show a uniformly increasing trend in the shift from harsh to benign conditions or from high to low glacierization, suggesting their independence from varying glacial influence and thus from glacially driven hydrological and thermal regimes. The taxa from the benthic substrate clearly differed from those described in glacier river studies on fungi, in which leaf-litter traps were used (Gessner and Robinson 2003). The MC and KC taxa partly resembled the spectra of terrestrial taxa related to understory vegetation (De Bellis et al. 2007), corroborating the assumed role of fungi in connecting aquatic habitats and their catchments (Jobard et al. 2010) but also implying changes in the benthic fungal community in response to those in the surrounding vegetation (Walther et al. 2005; Parmesan 2006; Leonelli et al. 2011). The observed fungal patterns might reflect the importance of fungi as an essential nutrient resource in high-altitude reaches, particularly for smaller invertebrates (e.g., nematodes and rotifers), by providing suitable high-energy and high-quality “mouthfuls” of food. This, together with the reliance of the fungal taxa on their surroundings, would add a further, previously unrecognized component to both the intrinsic relationships between potential resources and their consumers and the alterations thereof by ongoing changing environmental conditions.
Among those invertebrates for which species composition was investigated, habitat conditions and deglacierization stages contributed less to the similarity patterns of rotifers than of nematodes. This finding can be explained by the clear predominance of bdelloid rotifers, a group particularly well adapted to hardy conditions and with a number of pioneer species (Wallace et al. 2006; Hodson et al. 2008; Webster-Brown et al. 2010). Both the behavioral adaptations of monogononts searching for shelter in deeper interstitial layers (Schmid-Araya 2000) and the diurnal migrations of monogononts and nematodes aimed at high-flow avoidance (Smith and Brown 2006) would account for the maintenance of low, but permanent densities of these groups despite frequent high-flow conditions. These activities would also provide a behavioral advantage for these organisms in the future in response to climate change, when they are potentially confronted with an increased glacial melt and the prolongation thereof as well as less predictable precipitation flood events.
Although less abundant at particularly harsh sites, nematodes were more diverse than rotifers at each reach and catchment, with the highest abundances at the extremely differently structured benign sites. Therefore, nematodes might be able to well cope with a variety of reach conditions, a conclusion supported by the variety of running waters harboring nematodes (Eisendle 2009 and references therein). The nematode resemblance patterns clearly differentiated between the different KC and MC habitat types, and the nematode maturity was related to the GIm-established reach grouping. Furthermore, the observed nematode feeding types (data not shown; Yeates et al. 1993), were in good accordance with the respective food resources of each investigated reach. Together, these results support the usefulness of nematodes as indicators of glacially induced habitat templates as well as the development of benthic resource variety along glacier-fed rivers. Finally, we tentatively predict the loss of Hofmaenneria hazenensis (Mulvey 1969) (first recorded for the Arctic), with the loss of glacial influence, because in this study it inhabited only those river sites with distinct surface glacier influences (all except KC4).