Response of benthic macroinvertebrate assemblages to round (Neogobius melanostomus, Pallas 1814) and tubenose (Proterorhinus semilunaris, Heckel 1837) goby predation pressure
- 264 Downloads
One of the main assumed impacts of invasive gobies is predation on benthic macroinvertebrates. Despite numerous dietary studies, however, quantitative evaluations of impact in European river systems are scarce. Here, we investigate the impact of tubenose (Proterorhinus semilunaris, Heckel 1837) and round (Neogobius melanostomus, Pallas 1814) gobies on macroinvertebrates in a lowland river (River Dyje, Czech Republic) by allowing and preventing gobiid access to rip-rap substrate naturally colonised by invertebrates at two sites (Site 1—tubenose goby only, Site 2—tubenose and round gobies). Gobies had a negative impact on invertebrates at both sites, with overall invertebrate density reduced by 15% (ca. 17.9 g m−2 per year) at Site 1 and 36% (ca. 23.6 g m−2 per year) at Site 2. Both species showed increased impact in summer and ingested larger invertebrates preferentially, resulting in an overall reduction in invertebrate body size. Tubenose gobies had a significant impact on Annelida, Gastropoda, Crustacea and Ephemeroptera nymphs, while tubenose and round goby together impacted Annelida, Bivalvia (Dreissena), Gastropoda, Crustacea, Ephemeroptera nymphs, Odonata nymphs and Chironomidae larvae. Our results confirm that round and tubenose gobies can have a significant negative impact on aquatic invertebrate density and community composition.
KeywordsInvasive species Gobies Macroinvertebrates Impact European rivers Diet
Ponto–Caspian gobies are currently amongst the most successful groups of invasive fish species, having colonised many European and North American inland freshwaters, primarily through transport in ship’s ballast water (Corkum et al., 2004; Roche et al., 2013). Once established, gobies tend to expand their new range up and downstream, either by secondary transport (e.g., via pleasure boats or in angler bait buckets; Adámek et al., 2007), natural migration (i.e., swimming) or by downstream drift of larvae (Janáč et al., 2012, 2013a, b). Gobies have also colonised a number of navigable (e.g., the Rivers Drava, Tisza and Váh; Danube drainage; Harka & Bíró, 2007) and non-navigable tributaries (e.g., the River Morava; Danube drainage; Jurajda et al., 2005) through natural migration or angler introductions.
Goby expansion in Europe has been facilitated to a large degree by the artificial rip-rap bank stabilisation commonly found along channelised rivers (Jurajda et al., 2005; Semenchenko et al., 2011), a riverine habitat that is usually poorly colonised by native fish. Furthermore, dietary plasticity (Števove & Kováč, 2013; Hôrková & Kováč, 2014) allows gobies to utilise food resources (i.e., macroinvertebrate taxa) in relation to their availability (Vašek et al., 2014; Všetičková et al., 2014). Many studies, particularly those from the Great Lakes of North America, have recorded dreissenids as the dominant dietary item in adult round goby diet (Ray & Corkum, 1997; Taraborelli et al., 2010; Brush et al., 2012). European studies (e.g., Polačik et al., 2009; Adámek et al., 2010; Brandner et al., 2013; Borcherding et al., 2013) tend to differ, however, recording round and tubenose gobies as preferring non-mollusc invertebrates, with Crustacea, Ephemeroptera, Trichoptera and Chironomidae dominant. Gobiids are generally assumed to have a detrimental impact on macroinvertebrate assemblages (e.g., Kuhns & Berg, 1999; Lederer et al., 2006), with subsequent knock-on impacts on native fish and food webs (Krakowiak & Pennuto, 2008; Lederer et al., 2008), making them of great interest to biologists. Several studies have attempted to examine the impact of gobies on macroinvertebrates in either standing or running waters, especially in North America (Barton et al., 2005; Lederer et al., 2006; Kipp & Ricciardi, 2012). These have tended to be one-off field experiments, however, that have compared macroinvertebrate assemblages between sites with different goby densities, usually during the summer months. To our knowledge, only one study has directly evaluated goby predation pressure by controlling their access to food (Kuhns & Berg, 1999), and none have examined and quantified predation pressure throughout the year.
In this study, we examine the impact of round and tubenose gobies on benthic macroinvertebrates in the River Dyje, using mesh bags that allow or prevent goby access to naturally colonised rip-rap. In doing so, we aim to determine any changes in macroinvertebrate community structure caused by the effective ‘elimination’ of goby pressure and compare this with impact from tubenose goby alone and tubenose and round gobies together.
Materials and methods
This study was performed at two sites on the River Dyje (min. flow rate 9.5 m3 s−1; Danube river basin; south-eastern Moravia, Czech Republic) between July 2012 and June 2013 (Fig. 1). Site 1 (Nové Mlýny), which has a tubenose goby population but no round gobies, is located on the left bank of the Dyje, 150 m downstream of the Nové Mlýny reservoir and hydropower plant (48°51′28.202″N 16°43′27.031″E; river km 45.9; width = 47 m, current velocity = 0.10–0.16 m s−1). Site 2 (Břeclav), which has populations of both round and tubenose gobies, is situated on the right bank of the Dyje, 1 km downstream of the town of Břeclav (48°44′20.419″N 16°35′26.077″E; river km 23.5; width = 44 m, current velocity 0.12–0.18 m s−1). The two sites were selected as they provide sufficient goby density and similar habitat availability (bankside rip-rap), i.e., a river bank slope of <20% and a shallow <1 m shoreline zone. Average tubenose goby density was estimated at around 1.59 ind.m−1 of rip-rap at Site 1 and 0.43 ind.m−1 at Site 2, with round goby density at Site 2 estimated at 1.71 ind.m−1 (Table 4). Relative goby density was calculated as catch per unit effort, i.e., relative number per metre of rip-rap sampled (for a fuller explanation, see Valová et al., 2015). Round gobies outnumbered tubenose gobies at Site 2 soon after establishment, with the drop in tubenose goby density apparently caused by competition with round goby. Multiple fish surveys have shown gobies dominating the fish assemblage all along the rip-rap bankside of the Dyje, with other (native) fish species rare in this habitat (Adámek et al., 2013; Valová et al., 2015). Our own sampling at the site during this study (unpublished data) confirmed this with an overall density of native fish species of approx. 0.27 ind.m−1 at Site 1 and 0.38 ind.m−1 at Site 2 (Table 4).
Mesh substrate bags
Two types of mesh substrate bag were used to evaluate macroinvertebrate response to gobiid predation, each containing approximately 20 litres loose volume (mean total weight 19 kg, range 18–20 kg) of rocks collected from the rip-rap bank (mean size 14 × 10 × 6 cm), representing an average area of 1.2 m2 (1.1–1.6 m2) per bag. Ten bags were installed at each site, five with a 4 × 4 mm mesh (herein termed the control) and five with a 20 × 20 mm mesh (herein termed the treatment). The 20 × 20 mm treatment mesh size allowed free access to both tubenose and round gobies without size limitation, while the 4 × 4 mm mesh size was chosen as a compromise allowing free macroinvertebrate movement into the bag but preventing access to gobies >30 mm SL (predetermined by comparing body width and depth of 30 mm gobies and the control mesh size). Access of gobies to the treatment bags was confirmed by the regular occurrence of both species [23–84 mm standard length (SL)] inside the bags during the experiment, the average number recorded while examining the treatment bags ranging between 1 and 2 per bag. While gobies <30 mm could theoretically pass into the control bags, it is highly unlikely that they had any significant impact as they feed mainly on zooplankton and small Chironomidae (Borcherding et al., 2013).
Each set of 10 bags was placed in a line up to 2 m from the bank at around 50-cm depth and exposed for one month [except December 2012–April 2013 (see below)] to ensure complete colonisation of the substrate by macroinvertebrates. Each month, the bags were carefully removed and immediately transferred to a 90 l container of river water where each stone was thoroughly washed and all macroinvertebrates were removed and collected. Each sample was then washed over a 500 µm sieve and transferred to a separate 1 l bottle containing 4% formaldehyde. After inspection, the cleaned stones were returned to the bag and replaced in the river at the same site. In total, 7 monthly samples were collected over the year (sampling impossible from December 2012 to April 2013 due to an exceptionally high water discharge of 100–150 m3 s−1). All macroinvertebrates were sorted in the laboratory to the lowest possible taxonomic level (genus or species) and summarised into family groups, which were then weighed as wet biomass.
The following taxonomic groups were considered further for statistical analysis: Annelida, Bivalvia, Crustacea, Ephemeroptera, Odonata, Trichoptera and Chironomidae (for specific species and taxa, see Appendix—Supplementary Material). Taxa at low abundance (Heteroptera, Coleoptera and Diptera other than Chironomidae) or otherwise of low importance (Hydrina and Tricladida—inedible or low edibility) were not considered further in the analysis.
In order to reduce weighting on those months with low abundance, we calculated a weighted average monthly impact (weighting for each taxa within each month given by the ratio of monthly abundance to maximum monthly abundance, where abundance = control + treatment). In order to demonstrate seasonal variance in gobiid impact, we calculated impact separately for summer 2012 (July and August), autumn 2012 (September, October and November) and spring 2013 (May and June). Only taxa with an abundance >15 individuals (treatment + control) within each season/site were analysed statistically. In addition, Bivalvia, Ephemeroptera and Trichoptera were analysed separately as genera within these higher taxa displayed distinct behavioural and microhabitat differences that could affect their role in goby diet. Two distinct sizes of bivalve (Dreissena) were recorded [small (<5 mm) and large (>5 mm)—based on proportion between width and depth]. Ephemeroptera consisted of three distinct sub-groups (Baetis, Potamanthus, Caenis), while Trichoptera consisted of five sub-groups (trichopteran pupae, Hydroptila, Hydropsyche, Neureclipsis, Ecnomus).
Differences in community structure were tested by Permutational Multivariate Analysis of Variance (PERMANOVA), using matrices based on Bray–Curtis distances (Anderson, 2001). PERMANOVA uses a multivariate analogue of Fisher’s F ratio to compare variability within groups versus variability between groups, P values being obtained using permutations (Anderson, 2001). In this study, 100 permutations were conducted for each PERMANOVA test. Permutations were constrained within each month in order to stress inter-bag (treatment and control) differences and suppress the effect of seasonal changes in macroinvertebrate fauna.
Differences in abundance and the biomass/abundance ratio (B/A; calculated as a proxy for mean body size as the wet weight of an invertebrate group divided by the abundance of the same group) of the most important taxa were tested using generalised estimating equations (GEE; Zuur et al., 2009) as an extension of the Generalised Linear Model for data with pseudo-replication and non-normal distribution. For analysing abundance, we used Poisson distribution and the logit link function, while Gamma distribution and the inverse link function were used for the B/A ratio. We used “exchangeable” correlation structure as each sampling (month) was considered independent and each sampling started on cleaned, non-colonised stones.
All statistical analysis were performed using the vegan and geepack packages in R software ver. 3.0.3 (R Core Team, 2015), with significance for all statistical tests set at P < 0.05.
Overall, a total of 106,995 (1.97 kg wet biomass) macroinvertebrates were recorded in the control and 81,185 (1.19 kg wet biomass) from the treatment bags at Site 1 (tubenose goby only). At Site 2 (tubenose and round goby), we recorded 47,229 macroinvertebrates (0.27 kg wet biomass) in the control and 28,775 (0.12 kg wet biomass) in the treatment bags. Although overall invertebrate abundance decreased by 15% at Site 1, the difference was not significant (GEE, df = 1, 68, P > 0.05). The 36% decrease recorded from the treatment bags at Site 2, however, was significant (GEE, df = 1, 68, P < 0.001). At both sites, we detected a significant difference in the composition of macroinvertebrate communities inhabiting the treatment and control bags (PERMANOVA; Site 1, df = 1, 68, F = 1, 254, P < 0.01; Site 2, df = 1, 68, F = 1, 743, P < 0.01).
Summary of results for generalised estimating equations testing the effect of control (fish excluded) and treatment (fish allowed access) substrates on the abundance of the most important invertebrate (higher taxonomic level) at Sites 1 and 2
Summary of results for Generalised Estimating Equations testing the effect of control (fish excluded) and treatment (fish allowed access) substrates on the abundance of the most important invertebrate (lower taxonomic level) at Sites 1 and 2
Summary of results for Generalised Estimating Equations testing the effect of control (fish excluded) and treatment (fish allowed access) substrates on the biomass/abundance ratio (a proxy for body size) for higher invertebrate taxa at Sites 1 and 2
There was little difference in treatment and control community composition between seasons and most taxa displayed similar seasonal cycles at both sites. A significant difference in overall community composition was only observed in summer at Site 1, (PERMANOVA; df = 1, 18, F = 2, 998, P < 0.05) and in spring at Site 2 (PERMANOVA; df = 1, 18, F = 7, 599, P < 0.001).
Here, we investigate the influence of two invasive goby species on the benthic macroinvertebrate assemblage at two river sites (Site 1 hosting tubenose goby only and Site 2 hosting tubenose and round goby together) by controlling access to stony substrate. Unlike previous studies, which have tended to be one-off summer investigations, our study was designed to follow the impact of gobiids on macrozoobenthos over the whole year.
In this study, we investigated the influence of goby feeding pressure on benthic macroinvertebrate populations by preventing round and tubenose goby access to test substrates and comparing these with controls.
Overall and seasonal estimated density of gobiids and native species (catch per unit effort; individuals per metre of rip-rap shoreline) at Sites 1 and 2 on the River Dyje in 2012/13 (Jurajda and Šlapanský, unpublished data)
Unexpected extremely high water discharge during the colder months prevented us (both practically and for safety reasons) from sampling over the complete yearly cycle as planned. As the majority of temperate fish species drastically reduce their food intake over the cold period (Lynch & Mensinger, 2013; Všetičková et al., 2014), however, and many macroinvertebrate species have hatched or are otherwise unavailable during this period, we believe that our study successfully covered the most important part of year as regards food intake, and hence adequately reflects predation pressure of Ponto–Caspian gobiids on macroinvertebrates in this (and similar) rivers.
Despite these marginal limitations, the method provides a good indication of the impact of a predator on local macroinvertebrate communities and has been used in previous studies as a surrogate for predation impact of an invasive species (Kuhns & Berg, 1999; Moorhouse et al., 2014).
At both sites, we detected significant differences in macroinvertebrate community composition, with overall invertebrate abundance decreasing by 15% at Site 1 and 36% at Site 2, indicative of predation pressure by tubenose and round gobies. In some taxa (Annelida, Crustacea, Ephemeroptera), we recorded a decrease in overall abundance of more than 50%. Previous studies have also recorded a negative goby impact on macroinvertebrate communities (Lederer et al., 2006; Kipp & Ricciardi, 2012); however, in each case, goby density was two times higher than that at our sites (average overall density at Site 1—1.59 ind.m−1 tubenose goby only, Site 2—2.14 ind.m−1 tubenose and round goby). This implies that gobies can have a significant impact on benthic macroinvertebrate communities, even at relatively low densities.
Ponto–Caspian gobies are generally considered to be generalist benthic feeders, capable of utilising most invertebrate prey (Adámek et al., 2007; Copp et al., 2008; Borcherding et al., 2013; Brandner et al., 2013). Some laboratory and field diet studies, however, have indicated that gobies can show strong taxa selectivity and preference for non-mollusc invertebrates (Polačik et al., 2009; Vašek et al., 2014; Všetičková et al., 2014). Our data also confirm that gobiid impact is taxa-selective; hence, the presence of gobiids could eventually lead to changes in the macroinvertebrate community of the Dyje, as has been observed in invaded river systems of North America (Kuhns & Berg, 1999; Lederer et al., 2008; Kipp & Ricciardi, 2012).
A number of previous studies have recorded crustaceans and aquatic insect larvae as a preferred food items in goby diet (Adámek et al., 2007; Vašek et al., 2014; Všetičková et al., 2014) and this was also confirmed in this study, with Annelida, Crustacea and Ephemeroptera all exhibiting a decrease in density at both sites over the year (Table 1). Furthermore, we recorded a significant decrease in Caenis at both sites, and Potamanthus at Site 2 (Table 2). Indeed, both taxa have previously been recorded as common prey of both gobiid species on the Dyje (Vašek et al., 2014). These taxa are both classed as ‘walking’ mayfly nymphs and their lack of mobility may have made them more vulnerable to predation. On the other hand, Baetis, a ‘swimming’ form of Ephemeroptera, showed a significant increase in number on the treatment substrates at Site 2, possibly reflecting their higher mobility and a greater ability to escape goby predation. Overall, Ephemeroptera were found at very low abundance at Site 1, thus preventing any general conclusions (Table 2).
We observed a significant impact on Gastropoda abundance at both sites (Table 1). Lederer et al. (2006) and Kipp & Ricciardi (2012) also recorded a negative relationship between goby density and Gastropoda abundance, both suggesting that the drop in abundance was due to goby predation. Gastropoda, however, were not recorded as favoured prey in studies by Brandner et al. (2013) and Vašek et al. (2014). Gobies are clearly able to consume and affect gastropod populations, with preference probably based on relative abundance (of both gastropods and more preferred prey) in the environment and the size and digestibility of the gastropods available, with tubenose goby presumably concentrating on smaller individuals (see Vašek et al. 2014) and round goby able to consume both small and large gastropods due to their larger gape size (Balshine et al., 2005; Bergstrom & Mensinger, 2009; Vašek et al., 2014). A similar explanation is likely to apply to consumption of Bivalves (below).
A significant decline in Bivalvia was only observed at Site 2 (Table 1), presumably due to the presence of round goby, whose greater gape size allows them to make better use of this dietary resource. Bivalvia at our sites consisted of mostly small (<5 mm) and large (>5 mm) Dreissena. We observed significant changes in the abundance of large Dreissena at both sites, with a significant increase in abundance at Site 1 and a significant decrease at Site 2 (Table 2). This was almost certainly due to the presence of round goby at Site 2. Due to their smaller gape size, tubenose goby are unable to utilise most Dreissena as food (Adámek et al., 2010; Všetičková et al., 2014), thereby preventing any obvious impact and allowing populations to grow. Round goby, however, are able to utilise most sizes of Dreissena (see Kuhns & Berg, 1999; Barton et al., 2005) and this is reflected as a negative impact at Site 2, despite their lower relative abundance at the site.
A number of recent studies have indicated that Dreissena are not a primary prey item of round goby when other, more preferred prey (e.g., amphipods) are available in suitable quantities (Polačik et al., 2009; Vašek et al., 2014). On the other hand, Ray & Corkum (1997) noted that, while round gobies were able to ingest larger Dreissena, they appear to prefer smaller mussels of up to 12 mm, i.e., mussels from both our “small” and “large” size classes. While we observed no significant difference in the numbers of small Dreissena between the control and treatment substrates at either site, both Dreissena size classes were always much less abundant at Site 2, making comparisons difficult (Table 2). While this could indicate predation pressure by round gobies, we did not have enough data from before the gobiid invasion to be sure. Despite this, our data do suggest that round goby are capable of reducing abundance of Dreissena in the Dyje, just as they have in the much denser populations in the Great Lakes (Diggins et al., 2002; Lederer et al., 2006).
Gobiid impact on Chironomidae was similar at both sites (Table 1), with a significant difference at Site 2 and a close-to-significant difference at Site 1. Several studies have recorded Chironomidae as a preferred prey item for both round and tubenose gobies (Adámek, et al., 2007; Broza et al., 2009; Krakowiak et al., 2012; Vašek et al., 2014), and this was confirmed by their clear decline in the treatment substrates. Round goby were clearly able to exert a strong overall impact on this invertebrate group, presumably due to their ability to consume greater amounts of food than tubenose goby due to their larger body size.
In Odonata, a significant decline was only observed at Site 2, again suggesting round goby was responsible for the decline (see also Bivalvia and Chironomidae in Table 1). While Odonata were not widely consumed by tubenose goby at Site 1, abundance was generally very low and the absence of effect may have been caused by stochastic factors. Adámek et al. (2010) and Všetičková et al. (2014) also noted that Odonata were taken least often in tubenose goby diet.
By taxa, changes were revealed as a significant increase in trichopteran pupae and a decrease in Hydroptila larvae at Site 1 and a decrease in Neureclipsis at Site 2 (Table 2). The net cases of trichopteran pupae probably reduce predation by round and tubenose gobies as they are effectively camouflaged and/or hard to detach. We suggest that small species, such as Hydroptila, are more sought after by tubenose goby due to their limited gape size. As round goby are less limited by gape size, they are able to consume all sizes of Trichoptera, including caseless Neureclipsis. It is possible that the lack of any observable difference in other trichopteran species was caused by extensive and quick recolonisation of the substrate bags by larvae from outside or, possibly, by the availability of other, more easily found and consumed prey.
Several taxa at both sites exhibited smaller body size in the treatment substrates, though only two (Annelida, Crustacea) showed a significant difference at both sites (Table 3). A similar trend was observed at both sites for Trichoptera, though the results were only significant at Site 1 and close-to-significant at Site 2. There was also a significant decrease in chironomid body size, though only at Site 2. Kipp & Ricciardi (2012) also recorded a negative relationship between goby density and prey body size in the St. Lawrence River, suggesting that gobies could shift dominance of one taxon in relation to others. Here, we suggest that the overall body size of annelids and A. aquaticus (Crustacea), the main impacted invertebrates on the Dyje, was reduced due to significant gobiid predation pressure on larger individuals, which represent a better energy resource in terms of cost/benefit ratio. Note, however, that the overall drop in trichopteran body size was strongly affected by the presence of high numbers of small individuals following hatching in spring.
We suggest that the lack of any difference in body size for other invertebrate taxa (e.g., Ephemeroptera), despite high predation impact, was due to seasonal and intraspecific variability throughout the year. In general, the Ephemeroptera assemblage was unevenly composed of small (Caenis and Baetis) and large nymphs (Potamanthus), with small dominant species. None of these species showed any relationship between body size and the presence or absence of gobies, suggesting no preferential predation on species or size, at least in these taxa. Only Odonata exhibited increased body size in the treatment bags at Site 1 (Table 3), though this was caused by its generally low abundance and the presence of several larger individuals.
The greatest decrease in invertebrate abundance was detected at both sites in summer, and in spring at Site 2, while the lowest impact was observed at both sites in autumn, and in spring at Site 1 (Figs. 2, 3; Table 4).
Seasonal variation in gobiid activity may be related to either temperate-dependant feeding or to reproduction activity. Adámek et al. (2010), Taraborelli et al. (2010) and Lynch & Mensinger (2013) all recorded lower food uptake by tubenose gobies during the spring spawning period, while others have recorded gobies ingesting less food in autumn and winter (e.g., Brush et al., 2012; Lynch & Mensinger, 2013; Všetičková et al., 2014). Hence, the lower predation pressure observed at Site 1 in spring may have been caused by reduced feeding during reproduction, the slight decline in impact in autumn is being caused by reduced feeding as water temperatures dropped. While the spawning season of round gobies (dominant at Site 2) also covers spring (May, June), there was a clear decrease in macroinvertebrate abundance at Site 2 at that time. While tubenose gobies are an annual species, with the whole population spawning at the same time (Valová et al., 2015), round gobies mature later (see Janáč et al., 2013b), hence there is always a cohort of young round gobies not focused on reproduction in spring and still feeding. Possible gobiid migration to deeper pools (Pennuto et al., 2010; Lynch & Mensinger, 2013; Valová et al., 2015) may also have played a role in decreasing predation pressure over autumn and winter. On the other hand, as our ‘spring’ data were sampled in 2013 (as opposed to 2012 for the other seasons), inter-year variability (in temperature, macroinvertebrate abundance/activity, fish abundance/activity, etc.) may also have had an effect on impact. Finally, while seasonal fluctuations in gobiid impact did appear to follow fluctuations in gobiid density (Figs. 2, 3; Table 4), this ‘relationship’ should be treated with some caution as the fish density fluctuations could have been affected by seasonal variability in gear efficiency (i.e., electrofishing efficiency decreases as water temperature decreases).
Clearly, there is a wide range of variables potentially involved in the degree of gobiid impact on macroinvertebrates and the complexity of potential interactions makes identification of specific drivers of impact at any one time difficult to determine. An overall impact (i.e., high impact in summer and lower in autumn) was identifiable, however, and we suggest the most likely cause to be changes in predation pressure over the course of the year resulting from seasonal variance in gobiid activity and/or density. We suggest that further, more detailed studies are required to identify the specific processes behind impact fluctuations.
Invertebrate abundance was reduced by 15% at Site 1 (tubenose goby only) compared with 36% at Site 2 (tubenose and round goby). Average tubenose goby density at Site 1 was around 1.59 and 0.43 ind.m−1 at Site 2, with round goby density at Site 2 at 1.71 ind.m−1 (i.e., total density at Site 2 = 2.14 ind.m−1). Note that, while the presence of round goby increased overall goby density at Site 2 by around 35%, actual numbers of tubenose goby were reduced by around 73%, presumably due to competition. As such, round goby were the main cause of the increase in overall predation pressure at Site 2, due not only to their greater numbers but also their size and wider dietary niche breadth (Vašek et al., 2014). Note, however, that there was almost double the number of macroinvertebrates at Site 1, probably due to differences in microhabitat and food availability. Site 1 was located 200 m downstream of a eutrophic water reservoir, which provides a rich source of drifting planktonic food. This was reflected in an increased abundance of filter feeders (e.g., Dreissena, Hydropsyche, Neureclipsis, Ecnomus) at Site 1. Site 1 was also typified by an abundance of rocky bottom substrate (stabilisation below the dam outlet), which provided ample habitat for Dreissena. The absence of such bottom substrate at Site 2 (rip-rap limited to the bankside) may have caused round gobies to increase predation on invertebrate prey as a substitute for Dreissena, which can form a large part of the diet in other areas (Ray & Corkum, 1997; Taraborelli et al., 2010; Brush et al., 2012). Furthermore, the presence of Dreissena beds at Site 1 will have increased productivity of the surrounding waters still more by adding structural complexity to the bottom substrate and increasing available food (pseudofaeces) for benthic invertebrates (Kuhns & berg, 1999; Diggins et al., 2002). A similar negative relationship between lack of dreissenids and increased predation on macroinvertebrates was also observed by Lederer et al. (2006).
Our study revealed a clear negative overall impact of gobies on the macroinvertebrate community of the Dyje, despite a relatively low gobiid density of up to 2.14 ind.m−1 (Site 2), with invertebrate density potentially reduced by 15% at Site 1 (tubenose goby only) and 36% at Site 2 (tubenose and round goby). This represents an estimated reduction in macrozoobenthos biomass of 17.9 g m−2 of macrozoobenthos per year by tubenose goby alone and 23.6 g m−2 per year by round and tubenose goby together, which compares well with the figure of 78–127 g m−2 estimated by Taraborelli et al. (2010) for Lake Ontario, where goby density was five to eight times higher than that on the Dyje.
Both goby species were highly taxa-selective and had a negative impact on invertebrate body size. Tubenose gobies had a significant impact on Annelida, Gastropoda, Crustacea and Ephemeroptera nymphs, while tubenose and round gobies together impacted Annelida, Bivalvia, Gastropoda, Crustacea, Ephemeroptera nymphs, Odonata nymphs and chironomid larvae. Round gobies appeared to concentrate on Annelida, Crustacea, Ephemeroptera and Chironomidae, along with larger prey items such as Odonata and Bivalvia (Dreissena). Although our study design did not allow us to distinguish between direct and indirect effects, we suggest that the impact of gobies on macroinvertebrates consists of both direct predation and indirect top-down regulation effects [e.g., an increase in the abundance/size of predatory macroinvertebrates (Odonata) or loss/reduction of functional groups], which could lead to a cascading effect on the whole aquatic invertebrate community (Lederer et al. 2006; Kipp and Ricciardi 2012). Greatest impact of both species was felt mostly in summer. These results do not necessarily mean that gobies are affecting native fish assemblages on the Dyje, however, as macroinvertebrate abundance and biomass is relatively high in the Dyje (4.9–10.4 g m−2; Sukop, 2010). As such, gobies, which have primarily colonised rip-rap bankside habitat largely unused by other species, are unlikely to be having a significant impact on the recruitment of aquatic invertebrates on the Dyje or reducing them to levels that could lead to competition with native species. As goby density is expected to increase on the Dyje (and elsewhere in the lowland rivers of South Moravia) over the coming years, however, there is likely to be an increase in their impact on the macroinvertebrate community, with subsequent knock-on impacts to native species, especially in less productive river systems.
This study was supported by the Grant Agency of the Czech Republic, Grant no. P505/11/1768. We thank Markéta Mrkvová and Zdeňka Jurajdová for help with fieldwork. We are much indebted to the representatives of the Moravian Angling Union and their Local Clubs (Rakvice and Břeclav) for allowing this research in their waters.
- Adámek, Z., S. Zahrádková, P. Jurajda, I. Bernardová, Z. Jurajdová, M. Janáč & D. Němejcová, 2013. The response of benthic macroinvertebrate and fish assemblages to human impact along the lower stretch of the rivers Morava and Dyje (Danube basin, Czech Republic). Croatian Journal of Fisheries 71: 93–115.CrossRefGoogle Scholar
- Anderson, M. J., 2001. A new method for non-parametric multivariate analysis of variance. Austral Ecology 26: 32–46.Google Scholar
- Brush, J. M., A. T. Fisk, N. E. Hussey & T. B. Johnson, 2012. Spatial and seasonal variability in the diet of round goby (Neogobius melanostomus): stable isotopes indicate that stomach contents overestimate the importance of dreissenids. Canadian Journal of Fisheries and Aquatic Sciences 69: 573–586.CrossRefGoogle Scholar
- Copp, G. H., V. Kováč, I. Zweimüller, A. Dias, M. Nascimento & M. Balážová, 2008. Preliminary study of dietary interaction between invading Ponto–Caspian Gobies and some native fish species in the river Danube near Bratislava (Slovakia). Aquatic invasions 2: 189–196.Google Scholar
- Harka, Á. & P. Bíró, 2007. New pattern in Danubian distribution of Ponto–Caspian gobies a result of global climatic change and or canalization. Electronic Journal of Ichthyology 1: 1–14.Google Scholar
- Lederer, A., J. Janssen, T. Reed & A. Wolf, 2008. Impact of introduced round goby (Apollonia melanostoma) on driessenids (Driessena polymorpha and Driessena bugensis) and on macroinvertebrates community between 2003 and 2006 in the littoral zone of Green Bay, Lake Michigan. Journal of Great Lakes Research 34: 690–697.CrossRefGoogle Scholar
- Lusk, S. & K. Halačka, 1995. The first finding of the tubenose goby Proterorhinus semilunaris, in Czcech Republic. Folia Zologica 44(1): 90–92.Google Scholar
- Lusk, S., V. Lusková & L. Hanel, 2010. Alien fish species in the Czech Republic and their impact on the native fish fauna. Folia Zoologica 59(1): 57–72.Google Scholar
- Moorhouse, T. P., A. E. Poole, L. C. Evans, D. C. Bradley & D. W. Macdonald, 2014. Intensive removal of signal crayfish (Pacifastacus leniusculus) from rivers increases number and taxon richness of macroinvertebrate species. Ecology and Evolution 4(4): 494–504.CrossRefPubMedPubMedCentralGoogle Scholar
- Prášek, V. & P. Jurajda, 2005. Expansion of Proterorhinus marmoratus in the Morava River basin (Czech Republic, Danube R. watershed). Folia Zoologica 54(1–2): 189–192.Google Scholar
- R Core Team (2015). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/.
- Všetičková, L., M. Janáč, M. Vašek, K. Roche & P. Jurajda, 2014. Non-native western tubenose gobies Proterorhinus semilunaris show distinct site, sex and age-related differences in diet. Knowledge and Management of Aquatic Ecosystems 414: 10.Google Scholar