The Neva Estuary, which is located at the top of the Gulf of Finland (Figure 1), receives water from the Neva River, the most full-flowing river of the Baltic Region, whose flow averages 2492 m³ s^–1 (78.6 km³ year^–1). It is brackish-water, non-tidal, shallow, with horizontal and vertical gradients of salinity. The morphometric and hydrochemical features of the Neva Estuary have been previously described in many publications (e.g. Golubkov and Alimov 2010; Golubkov and Golubkov 2020).

At present, heavy nutrient and organic matter loading, mainly from the Neva River and St. Petersburg City (the largest megalopolis in the Baltic Region with > 5 million citizens), are the most significant environmental problem for the Neva Estuary. Eutrophication, organic pollution and biological invasions are the most serious threats to the environment of the Neva Estuary (Telesh et al. 2008; Golubkov et al. 2017, 2018, 2020b). NIS have a significant impact on the Estuary environment, many of which have become the dominant species in macroinvertebrate and fish communities over the past two decades (Orlova et al. 2004; Golubkov and Litvinchuk 2015; Maximov 2015; Demchuk et al. 2021).

Samples were collected in the middle part of the Neva Estuary (Figure 1). The salinity of waters near the surface in this part of the Estuary ranges from 1 to 3 PSU and of near bottom waters up to 5 PSU; the depth ranges from 7 m at sampling station 7 and up to 24.5 m at station 5. Zoobenthos was sampled at seven stations (Figure 1) with a modified Van Veen grab (20 × 20 cm) in early August 2014–2020. Samples were sieved in a 0.25 mm mesh with filtered fresh water and preserved in 4% formaldehyde in plastic zip-bags. Large crustacean S. entomon (Isopoda) were collected by trawling. Benthic animals for SIA were taken to the laboratory alive. Fish were collected with hook and line and frozen in the laboratory.

Figure 1.

The upper and middle parts of the Neva Estuary with indication of sampling stations (1–7). Black lines: isobaths of 5, 10 and 20 m. Areas with dots indicate dense reeds. C1, C2 – gates for vessels; D1–D6 – waters gates in the St. Petersburg Flood Protection Facility. Anchor marked passenger and cargo ports. Red rectangles – the location of the Neva Estuary. Two-letter country codes are given according to ISO 3166-1 alpha-2 (ISO 2021).

Invertebrates were picked out of the samples under a stereomicroscope, identified, counted and weighed to the nearest 0.1 mg. Abundance and biomass of animals (wet weight, shells of molluscs included) were estimated as an arithmetic mean ± SEM (standard error of the mean) from seven replicates and re-calculated per 1 m² of bottom area.

Live animals for SIA were identified under stereomicroscope and separated by species into vessels containing filtered water. Animals were maintained alive during 2 days to allow gut clearance. Faecal material was removed periodically to prevent coprophagy. After 2 days, animals were dried at 60 °C for 48 h. Small conspecific individuals were homogenised in an agate mortar to make a composite sample. Samples consisting of muscle tissue were used in the case of large animals (S. entomon and fish) as was recommended by Keough et al. (1996). Homogenised animal tissue from composite samples were put into small tin capsules and weighed using a Mettler Toledo MX 5 balance with an accuracy of ± 1 μg. At least three replicates of each type of organism or material were prepared and analysed.

The SIA was performed according to standard methods (Keough et al. 1996) using a Thermo Delta V Plus isotope mass spectrometer (Thermo Scientific, United States) equipped with an element analyser at the Joint Usage Center “Instrumental Methods in Ecology” of the A.N. Severtsov Institute of Ecology and Evolution of RAS (Moscow, Russian Federation). Isotopic composition of C and N in organic matter was expressed in δ-notation relative to the international standard (vPDB for carbon and the atmospheric N₂ for nitrogen): δ (‰) = (R_sample/R_standard – 1) × 1000, where R = ¹³C/¹²C or ¹⁵N/¹⁴N. Samples were analysed with reference gas calibrated against IAEA (Vienna, Austria) reference materials USGS 40 and USGS 41. The drift was corrected using an internal laboratory standard (casein). The standard deviation of δ¹³C and δ¹⁵N values in the laboratory standard (n = 8) was < 0.2‰. We did not conduct a lipid extraction from animal tissues because it may bias the estimation of trophic links of consumers (Tarroux et al. 2010).

The SIAR v.4.2 package (Parnell and Jackson 2021) running in R Software (version 3.5.1; R Development Core Team 2021; www.r-project.org/) was used for estimating trophic links of the consumers, based on their δ¹³C and δ¹⁵N values. SIAR (Stable Isotope Analysis in R) is a mixing model, based on hierarchical Bayesian methods, used to estimate contributions of basal sources, taking into account uncertainty and variability of input data, as well as isotopic enrichment factors (Parnell et al. 2013).

Estimates of the diet of individual organisms were obtained for two-isotope models using the ‘SIARsolo’ command. The mixing models were run using iterations – 500,000, burn-in – 50,000 and thinning by 15, without using concentration dependencies. Model solutions were presented using credibility intervals (95%, 75%, 25% and 5%) of probability density function distributions (Parnell et al. 2010). To compare the proportions of each source for a group on the boxplots ‘siarproportionbygroupplot’ command was used. The sensitivity analysis was carried out on fractionation factors as recommended by Inger et al. (2006). Values for trophic fractionation of C and N isotopes were chosen from the C and N range, given in Michener and Kaufman (2007). When an acceptable solution could be found, the simulation results varied by a maximum of 5% for the average contribution to the diet. An acceptable solution was found with trophic enrichment ratios of 1.63‰ for δ¹³C and 3.56‰ for δ¹⁵N from Parnell et al. (2013).

Alien polychaete M. arctia dominated in the study area during the period of research (Figure 2). Potamothrix hammoniensis (Michaelsen) and Limnodrilus hoffmeisteri Claparède (Oligochaeta), as well as larvae of Chironomus plumosus (Linnaeus) (Diptera, Chironomidae), were subdominant. Indigenous crustacean S. entomon (Isopoda) and M. affinis (Amphipoda) that dominated in zoobenthos before 2000s (Maximov 2003; Golubkov and Alimov 2010) were rare.

Figure 2.

Mean biomass (g WW/m²) ± the standard error of the mean (SEM) of the dominant zoobenthic groups and their portions (%) in the total biomass of zoobenthos in 2014–2020.

The biomass of polychaete worms significantly decreased in 2014–2020 (Figure 3A). On the contrary, the biomass of oligochaetes showed a positive tendency during this period (Figure 3B). The biomasses of other groups, as well as the total biomass of zoobenthos, did not show any tendencies. Significant negative relationships (R = –0.44, p = 0.002, n = 46) were observed between the biomass of oligochaetes and polychaetes (Figure 4).

Figure 3.

Changes in the biomasses of Marenzelleria arctia (A) and oligochaetes (B) in 2014–2020. Vertical bars are ± SEM.

Figure 4.

Relationships between the biomass of oligochaetes and polychaetes in the Neva Estuary in 2014–2020.

The isotopic signatures of key benthic macroinvertebrates and fish species varied in a wide range, reflecting differences in the use of different resources and in the trophic level (Figure 5).

Figure 5.

Isotopic signatures (δ¹³C and δ¹⁵N values, mean ± SEM) of common zoobenthic species and fish in the Neva Estuary. Cp – Chironomus plumosus, Lh – Limnodrilus hoffmeisteri, Ph – Potamothrix hammoniensis, Mo – Monoporeia affinis, Mar – Marenzelleria arctia, Se – Saduria entomon, Ab – Abramis brama, Rr – Rutilus rutilus.

According to the SIAR modelling, the previously highly dominant invertebrate predator in natural communities, the isopod S. entomon, mainly fed on amphipod M. affinis and the larvae of Ch. plumosus (Figure 6A). M. affinis has also been an important resource for roach and bream (Figure 6B, C). In turn, S. entomon, played a significant role in the feeding of bream (Figure 6C). All of these predators poorly consumed alien polychaetes M. arctia, which dominated amongst non-predatory invertebrates.

Figure 6.

The proportion of the use of various prey by the predatory macroinvertebrate Saduria entomon (A) and the fish Rutilus rutilus (B) and Abramis brama (C) in the Neva Estuary according to the SIAR model. The dark grey, grey, light grey and white are 95%, 75%, 55% and 5% credibility intervals. The numbers indicate the average percentages in the diet for 95% probability. Cp – Chironomus plumosus, Lh – Limnodrilus hoffmeisteri, Ph – Potamothrix hammoniensis, Mo – Monoporeia affinis, Mar – Marenzelleria arctia, Se – Saduria entomon.

In accordance with the data on the biomass of consumers and SIA, polychaete worms form the dominant food chain in the food web in the middle part of the Estuary (Figure 7), which apparently plays an important role in energy flow and functioning of the benthic macroinvertebrate community. This food chain is actually a closed branch of the Estuary food webs, since Marenzelleria is hardly eaten by secondary consumers. The former dominant food webs, associated with native crustaceans, are now poorly developed.

Figure 7.

Food web and the share of various carbon resources in the diet of the main consumers in the ecosystem of the Neva Estuary. The use of autochthonous and allochthonous carbon is given according to Golubkov et al. (2019). The thickness of the arrows is proportional to the biomass of consumers.

An important goal of invasion biology is to identify environmental characteristics that may make a region particularly receptive to invasions (David et al. 2017). While some ecologists consider abiotic conditions that are favourable for survival of exotic species (Moyle and Light 1996; Lahdes and Karjala 2007), others focus on their biotic interactions of invaders with native biota (e.g. Ptáčníková et al. 2015; Vanderploeg et al. 2015). Moreover, some studies have shown that if abiotic factors are appropriate for an exotic species, this species is likely to invade successfully, regardless of the native species already present (Moyle and Light 1996; Marchetti et al. 2004; Golubkov et al. 2020b). However, low species and functional diversity also can open an ecosystem for alien species, because unsaturated niche space makes communities more vulnerable for invasions (David et al. 2017). The highest numbers of non-indigenous species were found in the transit water systems in the Baltic Sea: in the Gulf of Finland, the Szczecin Lagoon and the Vistula Lagoon (Zettler et al. 2014). These water systems have wide gradients of environmental characteristics and relatively low species diversity.

Salinity gradient and hypoxia events are the main driving forces of zoobenthic succession in the Baltic Sea area (Rumohr et al. 1996; Bonsdorff 2006; Zettler et al. 2014). Due to low salinity and periodic hypoxia stress, benthic communities in the northern Baltic Sea are species poor and are dominated by few species that currently affect ecosystem functioning (Villnäs et al. 2012). At the same time, freshwater and euryhaline species increase biodiversity in the freshened parts of the Baltic. Amongst them are species that benefit from organic and nutrient enrichment (Balushkina and Golubkov 2018; Golubkov et al. 2019).

All these trends are actual for the zoobenthic community in the middle part of the Neva Estuary. Historical data show that, at the beginning of the last century, benthic communities were rather species-poor and dominated by indigenous crustaceans, Monoporeia affinis and Saduria entomon (Golubkov and Alimov 2010). Later, due to the increase in organic pollution from St. Petersburg, the oligochaetes, Potamothrix hammoniensis and Limnodrilus hoffmeisteri and the larvae of Chironomus plumosus, which are typical in fresh or slightly brackish polluted waters, also became common. This community existed until the mid-1990s, when, after a series of hypoxic events, the biomass of zoobenthos, especially of hypoxic sensitive crustacean, as well as the energy flow in benthic communities, decreased many times (Golubkov and Alimov 2010; Golubkov et al. 2010; Maximov 2015). The restoration of benthic communities was associated with the development of the alien polychaete Marenzelleria arctia, firstly recorded in 2009, which became the dominant species of zoobenthos (Maximov 2011).

When introduced, a species may persist only if it is able to pass through environmental and biotic filters (David et al. 2017). Biotic filters include resource availability, competition with native species and predators. A decrease in the diversity and abundance of communities due to disturbances makes ecosystems less resistant to invasions, since it reduces competition with native species and frees up niche space (Shea and Chesson 2002). The current warming of the climate mainly contributes to the introduction of relatively thermophilic species from middle latitudes to more northern latitudes (e.g. Holopainen et al. 2016). However, in the case of M. arctia, the opposite direction of invasions was observed. This species is native to the coastal zone of the Arctic seas, where it is mainly found in river mouths (Sikorski and Buzhinskaya 1998; Sikorski and Bick 2004). It possibly entered the Baltic Sea from the White Sea with the ballast waters of ships through the White Sea-Baltic Canal, which includes a system of lakes and artificial canals (Figure 1). However, the construction of the canal was completed back in 1933. The naturalisation of M. arctia in the eastern Gulf of Finland only seven decades later was apparently facilitated by the temporary decline of native zoobenthos species due to hypoxic-anoxic events.

In subsequent years, in 2014–2020, the biomass of M. arctia decreased (Figure 3). However, the biomass of benthic crustaceans remained practically unchanged and remained very low. This means that, despite a significant decrease in the invader biomass, the community did not return to its original state and confirms the opinion that, even in those cases where there may be a reasonable expectation of decline, an introduced species can leave a trace long after its influence has weakened (Simberloff and Gibbons 2004). The reason for this may be associated with a change in the configuration of food chains after the introduction of an alien species into the native community.

Changes in food web structure following invasion might, in most cases, be mainly related to changes in trophic group abundances rather than to species extinctions, as suggested by a meta-analysis on aquatic ecosystems (Gallardo et al. 2016). The invasion of alien M. arctia also did not lead to the extinction of native benthic species. Instead, according to SIA results, it formed its own food chain with low connectivity to the rest of the Estuary food web (Figure 7).

In the Neva Estuary, M. arctia, as well as dominant oligochaete and chironomid species, mostly used allochthonous wastewater-derived carbon as a basal resource for their production (Figure 7; Golubkov et al. 2019). Oligochaetes L. hoffmeisteri and P. hammoniensis (Oligochaeta) and larvae of Chironomus spp. are important indicators of polluted waters and benefit from organic pollution (Saether 1979; Rodriguez and Reynoldson 2011). They reach high abundance in the eutrophic and polluted waters of the Baltic Estuaries (Wolnomiejski and Witek 2013; Balushkina and Golubkov 2018; Kornijów et al. 2019; 2021). On the contrary, non-indigenous M. affinis, in addition to allochthonous carbon, largely uses autochthonous carbon produced by phytoplankton (Figure 7; Golubkov et al. 2019). This species consumes phytoplankton and phytodetritis, which is deposited on the seabed throughout the Baltic Sea (Lopez and Elmgren 1989; Lehtonen and Andersin 1998).

M. arctia consumes surface sediments and suspension around their burrows (Renz and Forster 2013). Therefore, polychaetes and the amphipods likely compete for food resources, as both are deposit-feeding animals partly sharing similar food resources (Kotta and Ólafsson 2003). Interspecific and intraspecific competition for food was considered to be the main mechanism regulating the population size of M. affinis in the field (Wenngren and Ólafsson 2002; Kotta and Ólafsson 2003). Hence, it can be assumed that the development of the abundant polychaete population after the hypoxia events of the early 2000s prevents the recovery of the Monoporeia population. Moreover, an innate tendency of M. affinis to move up upon disturbance from burrowing activities of neighbours may result in enhanced susceptibility to predators (Lopez and Elmgren 1989). Thus, the successful competition of Marenzelleria with Monoporeia may be facilitated by the fact that the population of the latter, unlike polychaetes, is under the influence of S. entomon and fish (Figure 7).

It has often been observed that NIS populations decline after an initial period of high abundance (David et al. 2017). In the Neva Estuary, we also observed a progressive decline of M. arctia biomass during 2014–2020 (Figure 3A). Significant negative relationships were found between the biomass of oligochaetes and polychaetes in the Estuary (Figure 4). The reason may lie in interspecific competition for resources between polychaetes and oligochaetes, since both groups mainly use allochthonous carbon for their production (Figure 7; Golubkov et al. 2019). However, consuming organic matter from different layers of sediment can prevent food competition (Kornijów et al. 2021). L. hoffmeisteri and P. hammoniensis (Tubificinae, Oligochaeta) ingest food particles from deep sediment layers and defecate faeces at the sediment surface (Davis 1974; Rodriguez and Reynoldson 2011; Kornijów et al. 2021), while Marenzelleria spp. collect particles at the surface around burrows and defecate faeces deep into sediments (Renz and Forster 2013). Therefore, the activity of worms of these zoobenthic groups causes differently directed fluxes in bottom sediments. The bioturbation activity of polychaetes causes the transport of solid and dissolved organic matter from the surface to the deep layers of bottom sediments. They enrich the deep sediment layers with organic matter and stimulate microbial particles reworking in them (Renz and Forster 2013). On the contrary, the bioturbation activity of oligochaetes causes an upward transfer of old-age substances contained deep in bottom sediments, which reduces the content of organic matter on their surface (Davis 1974; Kornijów et al. 2021). Accordingly, this upward conveyor can negatively affect the feeding conditions of M. arctia if the oligochaete biomass increases. On the other hand, this can hardly lead to the complete elimination of Marenzelleria from the system, due to the sedimentation of phytoplankton and other organic particles from the water column. Most likely, this should help stabilise the biomass of polychaetes at some equilibrium level.

SIAR modelling showed that the main invertebrate predator, S. entomon and fishes do not feed on M. arctia. As a result, the food chain leading to polychaete worms practically does not interact with other food chains of the benthic food web (Figure 7). The probable reason is that Marenzelleria spp. burrow deep into bottom sediments (Renz and Forster 2013). Their burrows probably allow these species to effectively avoid predation by surface invertebrate predators and fish. Instead, surface-dwelling amphipod M. affinis is a preferred food item for Saduria (Ejdung and Elmgren 2001). This crustacean is also a favourite food of fish in the Neva Estuary (Kudersky et al. 2007). Subsurface-dwelling larvae of Ch. plumosus and burrowing oligochaetes are less important sources of carbon for secondary consumers (Figures 6, 7). The vertical distribution of zoobenthos is an important trait of its suitability for benthic-feeding fish (e.g. Moss and Timms 1989; Persson and Svensson 2006; Weigel and Bonsdorff 2018). In an estuary of the southern Baltic, the Darss-Zingster Bodden Chain, the only recognisable remains of M. viridis in bream, carp and flounder were the bristles (Winkler and Debus 1996). These were usually taken in with detritus. Whole Marenzelleria in the guts of these fish were never found. However, they have been found in small numbers in the guts of pikeperch. It probably consumed polychaetes during their nocturnal migration into the water column (Winkler and Debus 1996).

A hypothetical model for the position and functional role of M. arctia in the bottom food web is given in Figure 8. The invasion of M. arctia has created a new offshoot in the estuarine food webs. The lack of the top-down control obviously contributes to significant development of this food chain. Unlike native food webs, this food chain does not provide energy transfer from autochthonous and allochthonous organic matter to the upper trophic levels and obviously negatively affects demersal fish production. This means that NIS can not only fill in the free nodes in food webs, but also reconstruct food chains, diverting resources that were previously consumed by other species and directing them along chains that are not always useful for humans. On the other hand, earlier there were no burrowing polychaetes in the Neva Estuary and the invasion of Marenzelleria formed a new functional group in its ecosystem (Maximov et al. 2015). Their activity leads to bioturbation and bioirrigation of bottom sediments and stimulates the decomposition of organic matter by worms and microorganisms (Figure 8). However, all these processes deserve further detailed research.

Figure 8.

A hypothetical model for the position and functional role of Marenzelleria arctia in the bottom food web of the Neva Estuary. Arrows indicate directions of organic matter transfer.