The structure and diversity of freshwater diatom assemblages from Franz Josef Land Archipelago: a northern outpost for freshwater diatoms

Study area and sites

FJL is a large archipelago that occupies an area of 16,000 km² and consists of 192 islands. The archipelago was formed together with the Spitsbergen Archipelago by the uplift and broken margin of the Barents Shelf (Solheim, Evgeny & Heintz, 1998). FJL is basically a remnant of the basaltic plateau fractured by tectonic faults and covered by glaciers (over 85%), with the highest elevation at ca. 600 m. Most of the islands are small, only nine exceed 500 km² with large ice-free areas of 80–500 km² on only four islands.

FJL lies within the high arctic climatic zone, with an average annual air temperature of −13.2 °C. July and August are the only two months with average air temperatures above zero (+0.7 °C and +0.1 °C, respectively) (VNIIGMI-MCD data accessed at http://oopt.aari.ru). Periods with above freezing average daily temperatures last from 2 months in the southern portion of FJL to 1.5 months in the north. Rainfall averages 250–300 mm year⁻¹ with only 40 mm during warm months. On average, FJL is much colder than Spitsbergen, which has an annual air temperature −7 °C, and four months with air temperatures above zero (Førland et al., 2011), warmer than the Canadian high Arctic islands (e.g., average annual air temperature in Eureka, Nunavut −19.9 °C, additional data can be found in http://www.eureka-nt.climatemps.com/temperatures.php).

The hydrographical network of FJL is not well developed; most of the streams are intermittent, while all rivers are completely frozen in winter. There are over 850 lakes and ponds in FJL but only 10 of them are deeper than 1 m (i.e., not frozen to the bottom M Gavrilo field observation). Principal water sources are melting snow and glaciers. Thus, most of freshwater habitats are very harsh, having short growing season and are only temporary.

Approval to conduct field studies was granted by the Russian Federation Ministry of Education and Science Approval Ref. No. 14-368 of 06.05.2013 permission to the Russian Arctic National Park.

Diatoms analysis

During the 2013 National Geographic Pristine Seas expedition to FJL, 18 samples from freshwater habitats (2 lakes, 7 ponds, and 9 streams) were sampled from different islands across the archipelago (Fig. 1 and Table 1). Scrapings were done on rock surfaces, which consisted of a mixture of mosses and cyanobacteria biofilm. In the field, samples were fixed and preserved in 3% formaldehyde.

Diatom samples were processed by hot digestion using hydrogen peroxide (33% H₂O₂) and HCL (2 ml 1 M). Through a series of dilutions, the peroxide and acid were removed. Subsequently, the samples were dried onto cover glass and mounted in Naphrax^® (R.I. = 1.7, Brunel Microscopes Ltd., UK) on a microscope slide following the method described in Battarbee et al. (2001). Five hundred diatom valves per slide sample were identified and counted across random transects at 1,000× magnification using a Zeiss Axio Imager A1 microscope (Carl Zeiss Inc., Germany) equipped with a 100× objective (Zeiss Plan-Apo 1.4 numeric aperture) and differential Interference contrast optics. Scanning electron micrographs were taken using a FEI Quanta 2000-F-SEM (10 kV, WD 10 mm).

Diatom identifications were based on contributions by Lange-Bertalot & Metzeltin (1996), Metzeltin & Witkowski (1996), Lange-Bertalot & Genkal (1999), Krammer (2000), Lange-Bertalot (2001b); Lange-Bertalot (2001a), Krammer (2002), Krammer (2003), Van De Vijver, Beyens & Lange-Bertalot (2004), Antoniades et al. (2008), Kulikovskiy et al. (2010), Zimmermann, Poulin & Pienitz (2010), Lange-Bertalot, Bak & Witkowski (2011) and Levkov (2013).

Data analysis

Detrended correspondence analysis (DCA) was conducted to estimate the length of the gradient on the square-root-transformed diatoms abundance data. Due to the large length of the gradient (>5 SD), a Principal Coordinates Analysis (PCO) was used based on Hellinger distances to reveal the structure among the sampled diatom assemblages (Legendre & Legendre, 1998). Only diatoms with abundances >1% in at least one sample were included in the PCO, which was performed using CANOCO 5.0 (Ter Braak & Šmilauer, 2012).

Diatom assemblage diversity was calculated in terms of species richness (S) and the exponential of Shannon’s entropy [exp(H)], which is more appropriate than classical diversity indices for assessing statistically independent measures of alpha (α) and beta (β) diversity (Jost, 2007). For each habitat type (lentic and lotic), diversity (D) was partitioned into α, β, and gamma (γ; total diversity) components following Jost (2007). Beta diversity was used as a measure of the differences in the species composition between habitat types. Evenness was calculated using two diversity orders. The first, of order 1, is based on Shannon entropy (EF₁) and the second, of order 2 (EF₂) is based on Simpson’s diversity metric (Jost, 2010). Hence, EF₂ is an approach to the proportion of the dominant species in the assemblage, whereas EF₁ also takes into account all the common species in the assemblage.

In order to recognize a discontinuous subset of samples, a K-means clustering with the optimal number of groups chosen according to the Calinski-Harabasz pseudo-F-statistic was performed (Legendre & Legendre, 1998). The ecological indicator values for each diatom taxon were identified to assess the ecological preferences of individual species for each of the clusters. The Indval-index was employed to control the effects of unequal size of groups (De Cáceres & Legendre, 2009).

All analyses of diversity metrics were performed using the R statistical program (R v. 3.0.2.; http://www.r-project.org/) and additional functions from the R-packages “vegetarian” v. 1.2 , “vegan” v. 2.2-1 and “indicspecies” v. 1.7.4 to calculate Indval-index.

Diversity and distribution patterns

The reported number of diatom taxa per sample in our study was similar to previous Arctic surveys (Douglas & Smol, 1995; Laing, Pienitz & Smol, 1999; Weckstrom & Korhola, 2001; Antoniades & Douglas, 2002; Michelutti et al., 2003; Jones & Birks, 2004; Antoniades, Douglas & Smol, 2005). However, these surveys also sampled surface sediments, which could have increased the number of observed taxa per sample. Surface sediment samples often integrate a disparity of habitats over time. Our study also found many diatom morphotypes that could not be identified, illustrating that taxa still need to be described, and supporting the recognition of high diversity and richness of freshwater diatoms across the Arctic (Foged, 1981; Antoniades et al., 2008; E Pinseel et al., unpublished data). Many species (new Arctic taxa) are presently included within species complexes (e.g., A. minutissimum, N. cf. perminuta, E. cf. fogedii, P. marginulatum) or yet to be described (e.g., Psammothidium sp1).

The large compositional turnover among samples (5.3 sd in the DCA) indicates large differences in diatom assemblages among sites. The main source of diatom assemblage variability was related to habitat (PCO axis 1; lotic versus lentic). Nevertheless, two clusters (2 and 5; Fig. 4) had samples from both lotic and lentic habitats, similar to findings from Spitzbergen (E Pinseel et al., unpublished data). Small pond assemblages could be largely influenced by the arrival of taxa from streams (Antoniades & Douglas, 2002), and conversely streams with low velocity (discharge) could support lentic assemblages. Unfortunately, we do not have stream velocity or water residence time data to confirm these observations. However, characteristic taxa of stream habitats in FJL, such as R. petersenii and F. capucina (complex), were also common in FJL pond samples. This observation indicates that pond diatom assemblages in FJL may be influenced by streams, as observed in other Arctic regions (Antoniades & Douglas, 2002; E Pinseel et al., unpublished data). Therefore, the dominance of stream taxa in pond sample FJ07 suggests a large influence of stream species. The same reasoning can be applied to the pond samples in cluster 2 (FJ04, FJ06, and FJ15), but results are less evident. This cluster, showed higher richness and diversity, with a dominance of circumneutral/slightly acidic taxa (pH optima ca. 6.8–7.2) such as C. gandrupii, H. gallica, H. laevissima, P. marginulatum, N. homburgensis, and H. balfouriana (Antoniades, Douglas & Smol, 2004; Antoniades et al., 2008). On the PCO biplot, cluster 1 (Fig. 4, with only stream samples), was characterised by the presence of alkaliphilic taxa (optimum pH 8.0–8.3 in Antoniades et al. (2008)) such as, E. laevis, P. bioretii, E. fogedii (complex), A. copulata, and A. inariensis. Therefore, this second axis of diatom variability is likely correlated with a pH gradient, which is a main source of variability for diatom assemblages (Battarbee et al., 2001), with Algers, Mabel, and Wilkzec Islands potentially marking alkaline conditions with higher pH. The pH gradient should be, in part, related to the complex archipelago geology with intrusions of basalt between sedimentary bedrocks (Dibner & Fursenko, 1998). However, recent deglaciated areas tend to have increased alkalinity and pH in lakes and streams from bedrock and till exposure (Engstrom et al., 2000; Milner et al., 2007). Thus, lake age (Milner et al., 2007) could be an alternative hypothesis to explain the lower pH optima of the characteristic taxa from cluster 2, rather than changes in the site geological settings.

Interestingly, the samples from cluster 2 also showed higher diversity metrics, due to the higher equitability among the non-dominant taxa (EF₁). Less stable environments exposed to physical stressors (e.g., current flow) could impact dominance by minimizing the effects of competitive exclusion and promoting species coexistence (Hughes et al., 2007) maintaining species diversity (Chesson, 2000). In Antarctic streams, environmental instability (changes in stream velocity and successive periods of freezing and thawing) promote small cell size, maintaining higher diatom richness (Pla-Rabes et al., 2013), with diatom assemblages showing higher resilience (recovery to pre-disturbance conditions; e.g., flood event) (Stanish, Nemergut & McKnight, 2011). There are indicators to suggest that the biology at some FJL sites support high environmental variability. First, cluster 2 samples are characterised by the presence of aerophilic taxa (H. laevissima, H. paracontenta, and C. fusus), which indicate potential periods of lower water availability, and possibly even desiccation. Second, species from cluster 2 showed lower pH optima (6.8–7.2 in Antoniades et al., 2008) compared to cluster 1 (7.9–8.2 in Antoniades et al., 2008), which means lower buffering capacity in cluster 2 sites, and consequently higher environmental variability. Furthermore, lower alkalinity sites would tend to be older (Milner et al., 2007), and consequently have a higher chance of accumulating species (Adler et al., 2005). Interestingly, cluster 2 sites showed large expanses of ice-free areas (Khun and Wilton Islands), and more human influence (Tikhaya Bay, Cape Tegetthoff, Cape Flora). Nevertheless, our data set did not include samples from the largest ice-free areas located on the largest islands in the archipelago (Alexandre Land and Prince George Land), which could potentially be diversity hotspots due to large biogeographic sizes.

Biogeographical patterns

The general diatom flora observed in FJL is similar to other circumpolar localities (Cleve, 1898; Lange-Bertalot & Genkal, 1999; Antoniades & Douglas, 2002; Stewart, Lamoureux & Forbes, 2005; Antoniades et al., 2008). A comparison of photomicrograph documented taxa in this study (Appendix S2) with Antoniades et al. (2008) for the Canadian Archipelago shows that 67% of the taxa observed in FJL were present in both studies. This indicates a circumpolar distribution, along with some local species selectivity due to microhabitats and regions. In an earlier study, Foged (1981) demonstrated that many more species are present in Arctic microhabitats, but often not recorded due to their rare distributions. Interestingly, a small araphid species flock representing Staurosirella, Staurosira, and Pseudostaurosira, was scarce in FJL, with only S. pinnata abundant and characteristic of pond samples in cluster 2. This was also observed in Svalbard (E Pinseel et al., unpublished data). The Staurosirella, Staurosira, and Pseudostaurosira flocks typically dominate hard substrate habitats in low productivity ponds and lakes across the Canadian and western Russian Arctic (e.g., Metzeltin & Witkowski, 1996; Laing, Pienitz & Smol, 1999; Michelutti et al., 2003; Bouchard, Gajewski & Hamilton, 2004; Paull et al., 2008). In our study of the twenty most abundant taxa, Nitzschia and Psammothidium were the most common genera with no clear associations within species flocks. These genera, along with Achnanthidium, are indicative of moss and sediment microhabitats. Moss wetlands have a diverse flora with select genera like Luticola, Humidophila, and Eunotia having unique compositions (Douglas & Smol, 1995; Van De Vijver, Van Kerckvoorde & Beyens, 2003). Luticola taxa observed from FJL were similar to the Novaya Zemlya Archipelago (Lange-Bertalot & Genkal, 1999), while different from the Canadian Archipelago (Antoniades et al., 2008). Some common Arctic stream species, such as Hannaea arcus, occurred in low abundance at FJL, which could be related to the absence of specific habitat requirements such as high current velocity (Antoniades & Douglas, 2002). Likewise, taxa from the genera Humidophila and Placoneis appeared to be species rich in FJL, which has also been observed in Svalbard (E Pinseel et al., unpublished data). In locations with more productive freshwater habitats, taxa from the family Cymbellaceae (Cymbella, Cymbopleura, and Encyonema) are noted to be abundant, with greater species richness, and widely distributed across microhabitats (Douglas & Smol, 1995; Michelutti et al., 2003). Except for C. gandrupii, currently recognized circumpolar taxa do not dominate FJL assemblages, although species richness is clearly linked to circumpolar and regionally endemic taxa (Fig. 2).

Geographic dispersal, especially with respect to a potential circumpolar flora, requires mechanisms that are inherently controlled by climate. Freshwater organisms such as diatoms could be dispersed by air as well as organisms such as birds, fishes, or mammals (Kristiansen, 1996). Interestingly, Nansen (1897) documented a similarity in diatoms and microbes in melt water pools from the Bering Strait and Greenland, and along with other observations, proposed that drifting ice could pass across the North Pole. Cleve (1898) was the first to suggest a potential circumpolar distribution of diatom species. Arctic drift ice has preferential pathways throughout the Arctic region and has been found to transport viable freshwater algae such as diatoms (Cleve, 1898; Abelmann, 1992; Pfirman et al., 1997). Therefore, drifting ice could be an important mechanism for dispersal, hence the recognition of a circumpolar flora. Additionally, airborne diatoms could also reach different Arctic regions. For example, dust storms routinely transport particulates from northern Asia to western North America (Chin et al., 2007). In this regards, population genetic studies of diatoms across the Arctic can assist in answering questions of biogeographic distribution (Hamilton, Lefebvre & Bull, 2015). Specifically, population genetics may help identify potential vectors of migration like drifting ice trajectories, atmospheric wind patterns, and bird migration pathways.

Consequently, the concept of circumpolar diatom species is applicable to a broader region including northern boreal regions from the Palearctic and Nearctic areas. For instance, the presence of Boreozonacola hustedtii recently described from a Mongolian peat-bog (Kulikovskiy et al., 2010), has been found in Siberia (Lange-Bertalot & Genkal, 1999), Lapland (Lange-Bertalot & Metzeltin, 1996), Canada (Zimmermann, Poulin & Pienitz, 2010), the Rocky Mountains USA (Bahls, 2010), and in FJL.

Future scenarios of diversity

Franz Josef Land Archipelago is a “zakaznik” Russian protected area (nature refuge corresponding to IUCN IV). As in Antarctica, climate change and an increasing human footprint are the two main threats for conservation in FJL (Chown et al., 2012). However, due to the poor knowledge of diatom distribution and ecology, it is difficult to predict future changes in productivity and diversity. Current predictions are theoretical using knowledge from other regions and ecological theory.

Arctic biodiversity has been exposed to strong selection mechanisms in harsh and fluctuating environments (glaciation/deglaciation periods) for millennia. However, at present we consider rapid climate change as the main factor impacting biodiversity in polar ecosystems (Millennium Ecosystem Assessment, 2005). In addition to climate change, increasing human activity (e.g., shipping, mining, fishing, tourism) in the Arctic are new stressors that need to be assessed, as well as other stressors such as global atmospheric pollution and species invasions, which could push the Arctic ecosystems to unknown scenarios (Larsen et al., 2014). All these threats imply rapid changes in the local environments (habitat change) and associated environment connectivity (dispersal mechanisms); both factors will modify the FJL freshwater ecosystems and communities.

Rapid environmental change can lead to a reduction in species richness and diversity (extinction event), which is often followed by a transition to novel communities dominated by generalist species with broader niches (Blois et al., 2013). This scenario represents a possible negative impact on specialized circumpolar taxa, with lower densities in FJL and other Arctic regions. There is a decreasing diversity gradient from the equator to the poles, although it is weaker and patchier for freshwater compared to terrestrial ecosystems (Hillebrand, 2004). Paleolimnological records across the Arctic region indicate that recent climate warming has driven diatom compositional changes with an increase in local diatom richness. These changes have been related to an increase in the length of the growing season (ice-free period), which increase habitat availability and the structural complexity of diatom communities (e.g., development of complex of periphytic communities) (Smol et al., 2005; Lim, Smol & Douglas, 2008). An increase in air temperature, and consequently an increase in the length of the ice-free season (growing period) would facilitate the establishment of species from warmer regions (i.e., Lapland, low tundra in the Canadian Arctic), as has been observed in other ecosystems (Dornelas et al., 2014). The introduction of species would be accelerated by the expected increase in connectivity among Arctic regions due to an increase in the human footprint. In addition, an increase in temporal climate variability (seasonal and inter-annual) could promote species coexistence by reducing competitive exclusion through fluctuation-dependent mechanisms such as storage effect and relative nonlinearities of competition (Chesson, 2000). Our data show that sites with higher diversity are characterized by the presence of slightly acidophilic and aerophilic species (cluster 2), which indicate fluctuating environmental conditions.

However, to sustain local species diversity, the total surface area of available habitat must be maintained. Despite the lower surface area-taxa richness relationship in high latitude regions and in freshwater ecosystems, it is higher for islands (Drakare, Lennon & Hillebrand, 2006). Hence, in the short-term, FJL will increase aquatic freshwater habitats due to melting ice as reported from other similar arctic regions (e.g., E Pinseel et al., unpublished data). In the long-term, the expected decrease of water bodies in the Arctic related to climate warming such as changes in the permafrost (i.e., loss of thermokarst lakes) (Larsen et al., 2014) and a reduction of ice would lower habitat availability, despite the observed pole-ward transport of moisture (Zhang et al., 2012). Consequently, we would expect a short term increase in diversity due to a warmer climate, followed by a decrease due to habitat loss.

At the global scale there are predicted increases in rates of extinction, particularly in the polar regions (Millennium Ecosystem Assessment, 2005). However, there is evidence that biodiversity will remain constant or even increase over time (Vellend et al., 2013; Dornelas et al., 2014), including human impacted areas and isolated oceanic islands (Sax, Gaines & Brown, 2002). This apparent contradiction (species loss versus gain) is due to different scales of study and measures of diversity, which need to be taken into account when considering biodiversity patterns (McGill et al., 2015). For instance, an increasing trend in local α-diversity could parallel a decreasing trend in global β-diversity (Sax & Gaines, 2003; Dornelas et al., 2014; Magurran et al., 2015).

One of the main factors determining β-diversity is regional connectivity and habitat heterogeneity. The general expected increase in human footprint related activities across circumpolar regions (e.g., tourism, polar transport, fishing, and natural resource extraction) would increase the homogenization of communities across the Arctic, therefore reducing β-diversity. Conversely, if drifting sea-ice has been enhancing FJL diatoms α-diversity by facilitating propagule dispersion, then a reduction in sea-ice would increase β-diversity. This change in diversity could be offset by an increase in human visitation. Ice-floe trajectories over FJL archipelago show two main sources: one from the Ob and Yenisei rivers through the Kara Sea, and the other from the Lena River and New Siberian Islands (Novosibirskiye Ostrova Archipelago) through the Láptev Sea (Pfirman et al., 1997). FJL diatoms suggest connections with these northern Siberia areas. The continuous arrival of freshwater diatoms from northern Siberia could also stabilize populations of rare diatom species (low abundance) species, which would otherwise become extinct in a harsh and fluctuating environment.

In the short term, the α-diversity of FJL diatoms is expected to increase with climate warming in association with an increase in human activities. In contrast, β-diversity (circumpolar) is expected to decrease due to an increase in circumpolar connectivity resulting from stressors related to an increase in anthropogenic activities. However, in the longer-term, an expected loss of freshwater bodies and habitat diversity could reduce diatom populations and increase competitive exclusions and therefore local species extinctions (decreasing α-diversity; Fig. 6). This process would be worse for rare species, which are mainly of circumpolar distribution or regionally distributed. We could argue that Arctic species have adapted to past climate fluctuations and will be able to persist. However, the human footprint will compound the impact of isolation due to less sea-ice (extension and duration), which has occurred in past warmer periods, consequently producing a net loss in β-diversity (Fig. 6). Therefore, local α-diversity could not be replenished as fast as in previous periods with the return of a new colder climate period in association with a parallel increase in connectivity across the Arctic Ocean.

Our discussion is founded on the novelty of the recent warming (habitat change), which parallels an increase in Arctic human activity (dispersal facilitation). However changes in other environmental and dispersal factors such as bird migration pathways, and atmospheric and ocean circulation modes could also affect the distribution of circumpolar diatoms. Furthermore, the sensitivity and adaptability of diatoms to environmental change are also dependent on endogenous factors (e.g., genetic diversity, phenotypic plasticity, species traits, populations density, fitness, dispersal ability, persistence, etc.), which are likely to modulate a disparity in diatom responses to external threats and consequently their vulnerability and future distribution (Dawson et al., 2011).

Finally, the low frequency of small araphid species flocks such as Staurosirella, Staurosira, and Pseudostaurosira, also observed in Svalbard (E Pinseel et al., unpublished data), has interesting biogeographic implications. A better distributional record of these taxa across the circumpolar region is required including the need to explore more freshwater habitats in the FJL Archipelago, such as hard substrate habitats, ponds, and stream that are largely influenced by sea bird colonies, as well as other factors, to characterize the complete diatom diversity of this isolated and northern archipelago.

To conclude, there is an urgent need to explore the Arctic to understand, predict, and mitigate future environmental change scenarios to preserve this unique ecosystem. In this regard, monitoring programs and/or paleolimnological approaches using diatoms as environmental indicators could be used as early warnings of environmental changes.