Introduction

The benthic biota plays an important role in marine ecosystems and in the global carbon cycle (Vardaro et al. 2009), yet, the dynamics of deep-sea benthic fauna in the vast abyssal and bathyal areas of the world oceans remain poorly understood (Laguionie-Marchais et al. 2013). The assessment of temporal changes in the benthic communities requires long-term observations with regular sampling (Glover et al. 2010; Larkin et al. 2010). For the deep sea below 2000 m water depth, regular biological observations are very scarce and only include a handful of sampling sites in the Mediterranean Sea, the Gulf of Mexico and several areas in the Atlantic and Pacific oceans (summarised in Larkin et al. 2010).

Temporal studies conducted at the Porcupine Abyssal Plane Sustained Observatory (PAP) in the North-East Atlantic (4850 m) since 1989 (Billett and Rice 2001) showed interannual changes in meiobenthos (Gooday and Rathburn 1999), macrobenthos (Galéron et al. 2001; Soto et al. 2010) and megafaunal assemblages (Billett et al. 2001, 2010). In the North-East Pacific (4100 m), the Station M has been sampled regularly since 1991 (Smith and Druffel 1998) exhibiting interannual changes in the density and taxonomical structure of macrofauna (Ruhl et al. 2008; Laguionie-Marchais et al. 2013) and the density of megafaunal echinoderms (Ruhl and Smith 2004). Correlations between the surface primary production, the fluxes of organic matter and metazoan community structure were demonstrated for both PAP and Station M (Galéron et al. 2001; Ruhl et al. 2008). Not only inter-annually but also seasonally varying food supply may affect the benthos as it was shown for the nematodes (Soltwedel et al. 1996) and polychaetes (Soto et al. 2010) in the NE Atlantic and for other benthic groups including arthropods and annelids in Station M (Smith et al. 2009).

In the Arctic, regular studies are conducted mostly in the shallow-water areas, e.g. in the White Sea (Sukhotin and Berger 2013) or in the Bering Strait area (Grebmeier et al. 2015). The Long-Term Ecological Research observatory HAUSGARTEN located in the transition zone between the northern North Atlantic and the Arctic Ocean represents the only time-series study in the Arctic deep sea including regular benthic surveys (Soltwedel et al. 2005, 2015). Interannual variations in meiofaunal abundance were reported during 5 years of investigation (2000–2004 and, separately, 2005–2009) along a bathymetric transect at HAUSGARTEN (Hoste et al. 2007; Gorska et al. 2014). Megafauna studies at HAUSGARTEN showed an overall decrease of megafaunal density during 2002–2007 at 2400–2600 m depth. This observation was correlated with the increase in bottom-water temperatures and decrease in the total organic content and microbial biomass in the surface sediments (Bergmann et al. 2011). Later, differences in megafaunal assemblages at HAUSGARTEN were reported with an overall increase in density for 2015 observations (Müller et al. 2016; Taylor et al. 2017).

Temporal changes in macrofaunal structure are usually assessed by comparing the general community characteristics (i.e. abundance, biomass, diversity), while the detailed species composition is rarely assessed (Soto et al. 2010; Laguionie-Marchais et al. 2013), especially in poorly studied Arctic deep-sea macrofauna with many species remaining undescribed (Bluhm et al. 2015; Oug et al. 2016). Moreover, the identifications of species from different samplings are often (or usually) done by different researchers leading to flawed comparisons of community characteristics because of incorrect taxonomic classifications (Bluhm et al. 2011). Therefore, it is impossible to treat large community data at species level.

Macrofauna investigations at HAUSGARTEN observatory were carried out by Wlodarska-Kowalczuk et al. (2004), Budaeva et al. (2008) and Vedenin et al. (2016), providing data from 2000, 2003 and 2012. The latter two studies have been conducted at comparable water depths (approximately 2500 m) while the study of Wlodarska-Kowalczuk et al. (2004) focused on depth-related gradients in community structure (300–3000 m depth range).

In the present study, we aimed to compare the samples taken at the HAUSGARTEN area in 2003 (Budaeva et al. 2008) with those collected in 2012 (Vedenin et al. 2016) since they have been obtained at approximately the same depth and during the same time of the year. This approach removes the depth and season factors from the analysis of temporal variation in the community structure. We carefully re-examined collections of material from both expeditions and revised species identifications by the same taxonomical specialists. Our goal was to reveal the temporal differences in the entire structure of macrobenthos and in the distribution of certain species over spatial scales of ~ 2 km and ~ 25 km.

Materials and methods

Sampling

Sediment samples for macrofauna investigations were taken during the RV Polarstern expeditions ARK-XIX/3c in July/August 2003 and ARK-XXVII/2 in July 2012 at the HAUSGARTEN observatory in the eastern part of the Fram Strait. Three stations (S1, S1-a and HGIV-a) were visited in 2003. The sampling was replicated in 2012 with one additional station (HGIV) taken near the northern station. The distances between stations were ~ 25 km (between HGIV and S1) and ~ 2 km (between S1 and S1-a and between HGIV and HGIV-a); the exact position of stations is shown in Fig. 1; coordinates and depth are shown in Table 1. All stations were located at water depths around 2500 m. An USNEL box corer (0.25 m2) was used for sampling; three replicated per station were taken in 2003 and only one core per station in 2012 (Table 1). Each sample was washed through a sieve of 0.5-mm in mesh size. In 2003, the entire cores were fixed in 4% formalin; in 2012, subcores of 0.125 m2 (one core per station, one subcore per core) were fixed in 10% buffered formalin. For additional information on sampling procedures, see Budaeva et al. (2008) and Vedenin et al. (2016).

Fig. 1
figure 1

Sampling area with box-corer stations taken during RV “Polarstern” expeditions ARK-XIX/3c (2003) and ARK-XXVII/2 (2012)

Full size image
Table 1 Stations where box-corer samples were taken during RV “Polarstern” expeditions ARK-XIX/3c (2003) and ARK-XXVII/2 (2012)
Full size table

Prior to statistical evaluations, the collections of macrofauna organisms from Budaeva et al. (2008) and Vedenin et al. (2016) were re-examined to verify original identifications. A combined species list for two datasets was composed based on the reviewed taxonomical identifications of species from the two collections.

Statistics

Densities and biomasses for all species were taken from original studies by Budaeva et al. (2008) and Vedenin et al. (2016) with adjustments regarding the revised identification of selected species.

We used untransformed data as a measure of species abundance and biomass. The similarity between the samples was estimated by using the Bray–Curtis similarity index. Hierarchical cluster analysis and nonmetric multidimensional scaling analysis (n-MDS) were used for examining the multivariate similarities. The results of the clusterisation were verified by ANOSIM analysis to reveal clusters significance level (Clarke and Warwick 2001). Temporal variability between the 2003 and 2012 samples was estimated by comparing the community characteristics over the different spatial scales: ~ 25 km (comparison of all samples) and ~ 2 km (comparison of the station pairs HGIV–HGIV-a and S1–S1-a). We estimated the species richness as the total number of the species in each sample. The species-individuals accumulation curves for the total fauna were plotted to estimate the level of saturation in the samples. Species diversity was measured in each sample using Pielou evenness (Pielou 1966), Hurlbert rarefaction index for 100 individuals (Hurlbert 1971) and Simpson index (Simpson 1949). The similarity percentage routine (SIMPER) was used to reveal species with highest contributions to between-years dissimilarities (Clarke and Warwick 2001). The Kruskal–Wallis test was used to identify the reliability of differences in abundance between the 2003 and 2012 samplings for certain taxa (Kruskal and Wallis 1952).

For the species present in either 2003 or 2012 samples, an algorithm estimating the likelihood of accidental presence or absence of a species in either sampling was applied. If we assume the constancy of the species distribution in two given samplings A and B, the probability of the species absence in every core of the sampling B would be (1 ‒ PA)N(B), where N(B) is the number of cores in sampling B and PA is the species occurrence (i.e. the percentage of cores where the species occurred) in sampling A. Using this ratio, we estimated the likelihood of the accidental absence of any species in either of the samplings. The number of cores required for species finding in sampling B was calculated by the equation: n = [log(α)] [log(1 − PA)]−1, where α is the likelihood of species finding in sampling B assumed as 0.95 (Azovsky 2018; Mheidze and Mirvis 1975; Kozlov 2014).

Statistical analyses were performed in Primer V6 (Clarke and Warwick 2001) and Past 3 (Hammer et al. 2003) software.

Results

Standardisation of datasets from 2003 and 2012

All the macrofauna taxa named differently in the previous publications are shown in the Table 2. Incorrect identifications were found among 53 taxa from 2003 samples and among 11 taxa from 2012 samples. Several taxa were excluded from the further statistical analysis: the foraminiferan Crithionina hispida (previously identified erroneously as the sponge Tetractinomorpha gen.sp. A), the pelagic amphipod Gammarus wilkitzkii (previously identified erroneously as Podoceridae gen.sp. or nonspecifically as Amphipoda gen.sp.), and the echinoid Pourtalesia jeffreysi (previously identified nonspecifically as Echinoidea gen.sp.) now considered as megafauna (Table 2).

Table 2 List of reidentified taxa from 2003 and 2012 box-corer samples taken during RV “Polarstern” expeditions ARK-XIX/3c and ARK-XXVII/2. Taxa with previously wrong names are shown
Full size table

After combining the datasets, several previously identified characteristics including abundance, biomass and species richness changed. The species number increased from 59 to 64 taxa in samples from 2003, and from 51 to 52 taxa in samples from 2012. Abundance and biomass values for the 2003 samples decreased significantly due to the exclusion of several taxa. New values of abundance, biomass, diversity and species number for each macro taxon are shown in Tables 3 and 4. The complete list of species densities from each sample (both 2003 and 2012) is shown in the Electronic Supplementary Material 1 (ESM 1). Most of the taxa identified to only generic or family level (e.g. Abyssoninoe sp., Desmosoma sp., Bonelliidae gen.sp., Harrimaniidae gen.sp.) are probably new to science.

Table 3 Number of species and relative proportions of biomass and density of the major taxa in the studied area in 2003 and 2012 box-corer samples taken during RV “Polarstern” expeditions ARK-XIX/3c and ARK-XXVII/2
Full size table
Table 4 Species number, density, biomass, Pielou evenness, Hurlbert rarefaction index for 100 individuals and Simpson diversity index calculated for each box-corer sample taken during RV “Polarstern” expeditions ARK-XIX/3c (2003) and ARK-XXVII/2 (2012)
Full size table

Comparison of community characteristics in 2003 and 2012

General characteristics of samples

The species-individuals accumulation curves were calculated for four combined sets of samples (northern 2003, northern 2012, southern 2003 and southern 2012 samples separately) (Fig. 2). The curves for 2012 were distinctly steeper than the curves for 2003. However, neither of them reached the saturation point. Most of the species, including the dominants (Galathowenia fragilis and Myriochele heeri), were present in both 2003 and 2012. In total, 76 macrofaunal species were found. Among the 39 species common for both years, 23 species were present only in the samples from 2003, while 14 species were present only in the samples from 2012 (ESM 1). The mean level of Bray–Curtis similarity was high (about 51 among all samples, 59 among the 2003 samples and 48 among the 2012 samples).

Fig. 2
figure 2

The species-individuals accumulation curves for the northern (HGIV and HGIV-a) and southern (S1 and S1-a) stations built separately for 2003 and 2012 samplings. Samples were taken by box-corer during RV “Polarstern” expeditions ARK-XIX/3c (2003) and ARK-XXVII/2 (2012)

Full size image

Despite the high taxonomical similarity among all samples, there was a difference in species composition between the northern stations (HGIV and HGIV-a, spatial scale ~ 2 km) and the southern stations (S1 and S1-a, spatial scale ~ 2 km). Biomass and abundance data showed same results. The 2003 samples from stations S1 and S1-a were very close to each other in the plot (mean similarity value over 77), the samples from station HGIV-a were equally similar (Fig. 3). In 2012, the stations S1 and S1-a were also closer to each other in the plot than to HGIV and HGIV-a, but the similarity value was lower than in 2003. The dendrogram shows the main division of the samples into two groups marking the spatial scale of ~ 25 km: all samples from the northern stations HGIV and HGIV-a clustered together while all samples from the southern stations S1 and S1-a formed another cluster. Within each of the two main clusters, the samples were separated by the sampling year (Fig. 3). ANOSIM test showed rather low values for comparison between the years (R—0.384, p—0.017) and high values for comparison between northern and southern stations (R—0.949, p— 0.002). The presence of the two main clusters was caused by slightly different composition of the dominant species: in the two northern stations, the oweniids Myriochele heeri and Galathowenia fragilis were dominant, while at the southern stations the bivalve Mendicula ockelmanni was equally abundant (ESM 1).

Fig. 3
figure 3

Dendrogram of Bray–Curtis similarity (a) and n-MDS plot (b) for the cores from 2003 and 2012 samplings. Samples were taken by box-corer during RV “Polarstern” expeditions ARK-XIX/3c (2003) and ARK-XXVII/2 (2012)

Full size image

Density and biomass values for most of the macrotaxa remained within the 2003 SD-range (Table 3). However, the density of crustaceans in 2012 was significantly lower compared to 2003 (Table 3). Comparisons of the integral community characteristics including total values of biomass, density, species richness, Pielou evenness, Hurlbert rarefaction and Simpson indices showed that most of the parameters remained in the same range (Table 4). However, Pielou evenness in the samples from 2012 was distinctly higher among all samples, and the Simpson diversity index was higher at the scale of ~2 km, i.e. in the pairs S1 – S1-a and HGIV – HGIV-a, respectively (Fig. 4a, b).

Fig. 4
figure 4

Mean values with standard deviation for the community integral parameters from 2003 and 2012 samples. a Pielou evenness for the scale of ~ 25 km; b Simpson diversity index for the scale of ~ 2 km. Samples were taken by box-corer during RV “Polarstern” expeditions ARK-XIX/3c (2003) and ARK-XXVII/2 (2012)

Full size image

Differences in taxonomical structure

A comparison of the samples using the SIMPER analysis showed significant differences in the species composition between the years 2003 and 2012. The species responsible for these differences are listed in the Table 5. Several species were significantly less abundant in 2012 than in 2003, including the bivalve Mendicula ockelmanni, the polychaete Myriochele heeri and the crustaceans Diastylis polaris and Pseudosphyrapus serratus. Most of the polychaetes and the bivalve Bathyarca frielei were more abundant in 2012 than in 2003. The polychaetes Chaetozone cf. jubata and Abyssoninoe sp. showed the greatest variation in density at the scale of ~ 25 km. In 2012, their abundance was significantly higher at all stations (p values 0.004) (Fig. 5a, b). The abundance of the cumacean D. polaris and the bivalve Yoldiella annenkovae in 2012 decreased (less significantly, p values 0.013) at the same scale (Fig. 5c, d). At the scale of ~ 2 km the density of the bivalve M. ockelmanni decreased: in S1–S1-a pair (from 75 to 33 ind. 0.25 m−2) and in HGIV–HGIV-a pair (from 4 to 0 ind. 0.25 m−2) (Fig. 6a). The results of the Kruskal–Wallis test confirmed the significant difference in density of these five species between 2003 and 2012. Chi-square and p values are shown in the Table 6. The abundance of other taxa shown in the Table 5 did not change significantly.

Table 5 Results of the SIMPER analysis of box-corer samples taken during RV “Polarstern” expeditions ARK-XIX/3c (2003) and ARK-XXVII/2 (2012). Comparison between the cores taken in 2003 and 2012
Full size table
Fig. 5
figure 5

Mean values with standard deviation of species density (ind 0.25 m−2) from 2003 and 2012 samples for the scale of ~ 25 km. Samples were taken by box-corer during RV “Polarstern” expeditions ARK-XIX/3c (2003) and ARK-XXVII/2 (2012)

Full size image
Fig. 6
figure 6

Mean values with standard deviation of species density (ind 0.25 m−2) from 2003 and 2012 samples for the scale of ~ 2 km. Samples were taken by box-corer during RV “Polarstern” expeditions ARK-XIX/3c (2003) and ARK-XXVII/2 (2012)

Full size image
Table 6 Results of the Kruskal–Wallis test for the taxa with the different density values between the years 2003 and 2012 from box-corer samples taken during RV “Polarstern” expeditions ARK-XIX/3c and ARK-XXVII/2
Full size table

At least two species (the gastropod Pseudosetia cf. griegi and the isopod Eugerda arctica) were found only in 2003 and six species (polychaetes Jasmineira schaudinni, Polycirrus sp. and Spiochaetopterus typicus, echiurans Bonelliidae gen.sp., nemerteans and enteropneusts Harrimaniidae gen.sp.) were found only in 2012 not randomly. The number of samples required to find these species is less than the number of samples collected for both groups (Table 7). Thus their presence/absence should unequivocally be interpreted as a result of temporal changes in the community composition.

Table 7 Species occurrence, likelihood of not finding a species, number of cores taken and number of cores required for finding a species calculated for the species present only in 2003 and only in 2012 samples
Full size table

Discussion

Temporal changes in general community characteristics

Temporal changes in the macrofaunal abundance, biomass and diversity were previously reported from the PAP and Station M observatories (Smith et al. 2009; Soto et al. 2010). The density and biomass, including the total values and the values determined for selected taxa (Polychaeta, Bivalvia, Arthropoda, etc.) varied significantly in relation to the flux of particulate organic carbon to the seafloor (Galéron et al. 2001; Ruhl et al. 2008). Specifically, seasonal increases in macrofaunal abundance and changes over a longer 10-year period in response to the altering food supply were demonstrated for Station M. At the same time, no changes were recorded across the two-year span (Smith et al. 2008, 2009). In our study, the total density and biomass as well as the abundance and biomass of macrofaunal taxa (e.g. Polychaeta, Mollusca, Echinodermata) showed no significant differences over the nine-year period. The only exception was Crustacea—their density decreased significantly.

Bergmann et al. (2011) reported an almost twofold decrease in megafaunal density during the period of five years (2002–2007) at the HAUSGARTEN area. After 2007, the total megafaunal density increased again approaching 2002-level in 2011 (Müller et al. 2016). This decrease concurred with an overall increase of the surface and near-bottom-water temperatures during 2002–2007 (Beszczynska-Möller et al. 2012), significant changes in plankton communities (e.g. the overall rise in pteropods numbers accompanied by a switch in dominance of species from Limacina helicina toward Limacina retroversa in 2004–2005) and subsequent changes in organic matter availability (potential food) at the seafloor (Soltwedel et al. 2015). In a study with similar two-point dataset but on megafauna, Taylor et al. (2017) reported that the overall increase of megafaunal densities coincided with the sediment-bound phytodetrital matter during 2004–2015. In our study, no overall increase in density was observed. The absence of complete 2003 and 2012 environmental data as well as lack of macrofaunal data between 2003 and 2012 does not allow to trace possible temporal changes similar to those found for megafaunal communities in the same area, and underlines the importance of time-series studies at higher temporal resolution. The slight increase in diversity in all 2012 samples (shown in Figs. 2, 4) could probably be explained by an overall increased organic matter availability at the seafloor in more recent years (Soltwedel et al. 2015). However, another possible explanation might be that certain taxa (e.g. polychaetes Polycirrus sp. or echiurans Bonelliidae gen.sp., see ESM 1) were overlooked in 2003 (see explanation in Table 5).

Temporal changes in taxonomical structure

Temporal versus spatial heterogeneity

Budaeva et al. (2008) reported a two-levelled horizontal heterogeneity in the 2003 samples. The first level represented by species assemblages occupied an area of several kilometres, the second was represented by certain species patches about 150 cm2 size. It was also shown that there is no spatial variability between the samples taken at the distance less than ~ 400 m from each other (Budaeva et al. 2008). Thus, all the significant differences between 2003 and 2012 samplings should be interpreted as temporal. The structure of the dendrogram and MDS plot (Fig. 3) suggests that the differences between the same stations in time exceed the inner spatial heterogeneity of each of the three stations. However, the temporal differences do not exceed the spatial variability in the scale of ~ 25 km (the distance between the northern and the southern sampling sites).

Temporal changes in selected species

Compared to the data provided by Wlodarska-Kowalczuk et al. (Wlodarska-Kowalczuk et al. 2004) the composition of the dominant species (i.e. Myriochele (= Galathowenia) fragilis, and M. heeri) as well as their density and biomass (and that for entire macrofauna) remained within the same range over the 12-year period. Macrofauna densities found between 2000 and 3000 m water depth on the slopes of the Yermak Plateau (~ 100 km north from our study area) and off North-East Greenland in 1997 revealed similar abundance and biomass values for the dominant species (e.g. G. fragilis) at the HAUSGARTEN sites (summarised by Degen et al. 2015). Unlike G. fragilis and other oweniid polychaetes, other dominant species like the bivalve Mendicula ockelmanni showed much lower densities in 2012 compared to 2003. In fact, this species is completely absent in the species list published by Wlodarska-Kowalczuk et al. (2004).

Comparing macrofauna data at HAUSGARTEN between 2003 and 2013, the most severe changes in density occurred among the cirratulid polychaete Chaetozone cf. jubata. A similar high temporal variability in the abundance of cirratulid polychaetes (i.e. Chaetozone spp. and/or Tharyx spp.) was previously shown for the PAP site (Soto et al. 2010). The increase of Chaetozone spp. is commonly associated with increased organic matter input. However, the cirratulids (especially the deep-sea species) are poorly studied in terms of taxonomy and biology, and any conclusions based on the changes in cirratulid density should be done with caution (Chambers and Woodham 2003).

At HAUSGARTEN, about half of the total number of species was unique for either 2003 or 2012 samples. Among ten species unique for 2012, at least six were absent in 2003. However, it must be admitted that some of those taxa become almost unidentifiable after fixation, and therefore some specimens had a good chance to be overlooked or classified as undeterminable fragments while sorting the sediments. These specimens include the polychaetes Polycirrus sp., echiurans Bonelliidae gen. sp., nemerteans and harrimaniid enteropneusts (Table 7). With the exception of these species, only two species are left to be unique for 2012. Overlooking of these species may also explain the different shapes of the species-individuals accumulation curves with the steeper 2012-curves (Fig. 2). The poor state of preservation after fixation is a common problem for some taxa. It has been noted for different fragile species of Polychaeta (Zhirkov 2001), Nemertea (Buzhinskaja 2010), Echiura (Zenkevitch 1966), Enteropneusta (Osborn et al. 2011) and other groups. Their identification in the samples fixed together with the sediment is therefore often complicated or even impossible. The only way to avoid this problem is treating the samples onboard before fixation, which is rarely possible.

Data quantity and quality

Unlike the HAUSGARTEN area, other deep-sea observatories (e.g. PAP or St. M) have been regularly sampled for macrofauna for more than 20 years. Numerous replicates of macrofaunal samples are taken at PAP and St. M since 1989, in some years even seasonally, allowing to reveal correlation between different environmental factors, organic input and macrofaunal density and biomass (Galéron et al. 2001; Ruhl et al. 2008; Smith et al. 2009; Soto et al. 2010). Unfortunately, due to lack of taxonomical expertise, deep-sea macrofauna is usually identified only to family or even class level (Galéron et al. 2001; Smith et al. 2009; Soto et al. 2010). Bluhm et al. (2011) reported that the differences in species identification by different investigators may often coincide with regional differences when studying the spatial structure of macrobenthos. The number of renamed taxa shown in Table 2 is a good example illustrating the importance of careful re-identification to the species level by the same specialist, when looking for temporal variations in the community structure.

As a conclusion, this study provides first data on temporal dynamics of macrobenthos in the Arctic deep sea. Only a two-point dataset on HAUSGARTEN macrobenthos is available which allows only a first glimpse into temporal variability. Another limitation in the interpretation of results from macrofauna studies at HAUSGARTEN is the lack of background data for more than half of the stations. Previously Vedenin et al. (2016) showed correlation in certain species abundance with latitude (and, possibly, with the food availability). However, in our case, environmental data are only available for stations HGIV and S1, i.e. for 5 out of 13 samples (Klages et al. 2004; Soltwedel 2013), which impedes correlating the macrofaunal variations with the environmental parameters over the different spatial scales (~ 2 km and ~ 25 km) investigated.

Conclusions

Investigations on temporal dynamics in benthic communities generally focus on changes in main community characteristics and/or the changes in certain taxa. In this study, we focused on the complete species lists which allowed us tracing changes in density of certain species between two sampling events in 2003 and 2012. Over the nine-year period no variability in total density, biomass and diversity were found. However, while significant differences in abundance were detected for several common species (e.g. the bivalves Mendicula ockelmanni and polychaetes Chaetozone cf. jubata), other macrofaunal taxa (e.g. Bivalvia and Polychaeta) showed no such differences. Besides these discrepancies, at least two species (Pseudosetia cf. griegi and Eugerda arctica) were unique for 2003 samples and two species (Jasmineira schaudinni and Spiochaetopterus typicus) were exclusive for the 2012 samples. We interpret both density changes and presence/absence of certain species as temporal changes in the community structure.

The comparison of the complete species lists from stations using cluster analysis and multidimensional scaling revealed significant temporal variations at the scale of ~ 2 km (i.e. in the station pairs S1–S1-a and, separately, HGI–HGIV-a). However, the differences in time did not exceed the variability in the scale of ~ 25 km, i.e. the distance between the northern and the southern station pairs.

Our study confirmed the importance of a reliable identification of species. A careful re-identification of every single specimen from previous samplings is needed before comparing the whole species lists from different sampling events. In this study, only two datasets were compared which allowed a first brief look into the long-term temporal variability of macrofauna. Continued sampling at HAUSGARTEN and simultaneously collected environmental data will allow assessing the long-term dynamics of macrobenthos and the driving factors for changes.