Skip to main content
NCBI home page
Ambio logoLink to Ambio
. 2018 May 5;47(8):884–892. doi: 10.1007/s13280-018-1050-y

Effects of an invasive polychaete on benthic phosphorus cycling at sea basin scale: An ecosystem disservice

Antonia Nyström Sandman 1,✉, Johan Näslund 1, Ing-Marie Gren 2, Karl Norling 3
PMCID: PMC6230331  PMID: 29730794

Abstract

Macrofaunal activities in sediments modify nutrient fluxes in different ways including the expression of species-specific functional traits and density-dependent population processes. The invasive polychaete genus Marenzelleria was first observed in the Baltic Sea in the 1980s. It has caused changes in benthic processes and affected the functioning of ecosystem services such as nutrient regulation. The large-scale effects of these changes are not known. We estimated the current Marenzelleria spp. wet weight biomass in the Baltic Sea to be 60–87 kton (95% confidence interval). We assessed the potential impact of Marenzelleria spp. on phosphorus cycling using a spatially explicit model, comparing estimates of expected sediment to water phosphorus fluxes from a biophysical model to ecologically relevant experimental measurements of benthic phosphorus flux. The estimated yearly net increases (95% CI) in phosphorous flux due to Marenzelleria spp. were 4.2–6.1 kton based on the biophysical model and 6.3–9.1 kton based on experimental data. The current biomass densities of Marenzelleria spp. in the Baltic Sea enhance the phosphorus fluxes from sediment to water on a sea basin scale. Although high densities of Marenzelleria spp. can increase phosphorus retention locally, such biomass densities are uncommon. Thus, the major effect of Marenzelleria seems to be a large-scale net decrease in the self-cleaning capacity of the Baltic Sea that counteracts human efforts to mitigate eutrophication in the region.

Electronic supplementary material

The online version of this article (10.1007/s13280-018-1050-y) contains supplementary material, which is available to authorized users.

Keywords: Benthic–pelagic coupling, Ecosystem services, Eutrophication, Invasive species, Nutrient cycling

Introduction

Effective integration of ecosystem services in marine management involves spatial planning and decision making, identification, mapping and quantification of the relationship between ecosystem characteristics and their associated functions and services (de Groot et al. 2010). However, there are few examples of ecosystem services mapping in marine environments and the quality of available data (such as habitat maps) is often inadequate (Maes et al. 2012).

Deep soft-substrate sediments (sand and mud) cover approximately 80% of the Baltic Sea seafloor (Al-Hamdani and Reker 2007) and these habitats, including the organisms living there, provide ecosystem services such as the remineralization of nutrients through decomposition of organic matter. Benthic fauna enhance transport of materials and affect fluxes of oxygen and nutrients (Carstensen et al. 2014), but these processes are species-specific (Karlson et al. 2007, 2011; Norling et al. 2007) and density-dependent (Hietanen et al. 2007; Norkko et al. 2012). The quantitative influence of benthic fauna on various processes in the nitrogen and phosphorus cycles is not fully understood (Carstensen et al. 2014). Marenzelleria (Polychaeta) was first observed in the southern Baltic Sea in 1985 (Bick and Burckhardt 1989) and was likely introduced via ballast water. Three species of Marenzelleria have been identified in the Baltic Sea: M. viridis, M. neglecta and M. arctia (Blank et al. 2008). In their native waters (North American East coast and Russian Arctic Ocean), Marenzelleria spp. are relatively uncommon and their populations are limited by predation (Virnstein 1977; Sarda et al. 1995). The highly invasive Marenzelleria species complex is now an important component of the benthic community throughout the Baltic Sea (Leppäkoski and Olenin 2000; Zettler et al. 2002). Densities of several thousand individuals m−2 are often found in soft seafloor substrates (e.g. Kotta et al. 2001; Kauppi et al. 2015) but this genus is rarely a major component of the total biomass (Kauppi et al. 2015; Gogina et al. 2016).

The introduction of Marenzelleria spp. into the highly eutrophied Baltic Sea has probably changed important benthic processes across the entire region but the specific large-scale consequences of its introduction on the benthic ecosystem are poorly understood. Norkko et al. (2012) and Maximov et al. (2015) suggested that Marenzelleria spp. can improve insufficient oxygen conditions on the seafloor and thus enhance long-term phosphorus retention. They also suggested that this effect is probably density-dependent. M. arctia have been shown to significantly increase the resource use efficiency of freshly deposited organic matter in a typical Baltic Sea sediment community (Karlson et al. 2011). This suggests that organic matter deposition in the benthic community likely decreases when Marenzelleria spp. are present, enabling more nutrients to be assimilated in secondary production or remineralized. All three Marenzelleria species are mobile as sub-adults and adults and disperse over large distances in the planktonic larval phase. Adult individuals burrow deeper and are less sensitive to oxygen deficiency in sediments than native species such as Limecola balthica, Monoporeia affinis and Pontoporeia femorata (Hietanen et al. 2007; Karlson et al. 2011; Renz and Forster 2013). As hypoxia increases the benthic efflux of phosphate (Conley et al. 2002; Reed et al. 2011), an increase in oxidized sediments due to high abundances of Marenzelleria spp. could increase phosphorus retention across the entire Baltic Sea. Oxygenation of sediments per se, does not guarantee phosphorus retention. Increased phosphorous remineralization due to a higher utilization rate of deposited organic matter could further decrease phosphorus retention. High rates of phosphorus release from sediments occur even under oxic conditions, which have been observed in the Baltic Sea (Conley et al. 1997; Lehtoranta and Heiskanen 2003; Kauppi et al. 2017). Marenzelleria spp. increased phosphorus retention in one experimental study (Bonaglia et al. 2013) while other have reported decreased phosphorous retention (Hietanen et al. 2007; Viitasalo-Frösén et al. 2009; Renz and Forster 2014), or no distinguishable effect due to large variation among replicates (Karlson et al. 2005; Norling et al. 2007; Urban-Malinga et al. 2013).

A decrease in phosphorus recycling due to Marenzelleria spp. could benefit ocean water quality by increasing the self-cleaning capacity of the sea. More of the phosphorus in the sea would be buried in the sediments (Norkko et al. 2012), counteracting eutrophication and its associated ecosystem damage. This suggests the possibility that certain targets on maximum phosphorus pools in the marine basin, such as the HELCOM (2013a) targets, may be obtained at a lower cost, as an increase in the ocean’s self-cleaning capacity would decrease the cost of necessary human interventions such as pollution abatement measures in agriculture, industry or sewage (Gren et al. 2013). In contrast, an increase in phosphorus recycling may raise the cost of achieving maximum phosphorus pool targets because of the need for more phosphorus reduction at the sources. Thus, a quantification of the possible Baltic-wide influence of Marenzelleria spp. on phosphorus cycling is important for several stakeholders. From a management perspective, it would be desirable to have more complete data of the effects of non-indigenous marine species on relevant spatial and temporal scales in order to calculate costs and benefits from preventive actions, such as regulations on exchange of ballast water in international shipping. Since eradicating established invasive marine species is next to impossible, we should aim to reduce the risk of introducing additional species and mitigate the negative effects of existing invaders. Additional measures will be needed to achieve the aims of the EU Marine Strategy Framework Directive, and it is important for decision makers to have information on associated costs and benefits. The topic of potential large-scale effects of Marenzelleria spp. on Baltic Sea nutrient dynamics is also relevant to several international environmental commitments, e.g. the HELCOM Baltic Sea Action Plan and the EU Water Framework Directive.

This study provides basin scale quantifications of the potential net contribution of current Marenzelleria spp. biomass densities on phosphorous fluxes from the benthic to the pelagic ecosystem compartments in the Baltic Sea.

Materials and Methods

A Random Forest model of Marenzelleria spp. biomass, based on relationships to environmental variables, was used to predict Marenzelleria spp. biomass across the entire Baltic Sea at a 200 m horizontal grid cell resolution. Random Forest is a machine-learning algorithm that utilizes many small classification or regression tree models to produce a combined result (i.e. the forest) for prediction (Breiman 2001). This method can quickly analyse very large datasets, automatically fit complex interactions between predictor variables, and estimate predictor variable importance. Cutler et al. (2007) provided a description of Random Forest algorithms and their ecological applications.

The model was fit with randomForest for R (version 4.6-12) (Breiman 2001; Liaw and Wiener 2002; R Core Team 2014) using biomass data of Marenzelleria spp. from benthic grab samples from across the entire Baltic Sea. Data from the Helsinki Commission (2013b, c) were complemented with data from Finland (OIVA database, Finnish Environment Institute), Denmark (MADS database, National Environmental Research Institute MADS, mads-en.dmu.dk) and Sweden (SHARK database, Swedish Meteorological and Hydrological Institute). A total of 18 412 samples taken between 2000 and 2015 were collated. In most of the Baltic Sea area, environmental monitoring did not differentiate between the different species of Marenzelleria, which are known to hybridize (Götting et al. 2011). Therefore, species level data were aggregated to genus. As the grab sampling area varied among samples, all biomasses are expressed in g m−2. Abundance information was only available for about 25% of the samples (mainly from archipelago areas in southern Sweden and Finland, where biomass data were scarce) and a conversion factor of 16.52 mg ind−1 (the mean individual weight in Swedish monitoring data) was used to estimate wet weight biomass. To reduce autocorrelation bias in densely sampled areas, a 100 m buffer zone was created around each sampling location and whenever these buffer zones overlapped only the most recent sample was used. If there were several samples from the same year, one sample was selected at random. This method yielded a total of 8962 samples for use in the modelling. A total of 7962 samples were used to train the model, and 1000 randomly selected samples were set aside for subsequent (external) validation.

Environmental predictors covering depth, slope, halocline, salinity, temperature, Secchi depth, seafloor currents and wave exposure were available. Collinearity among predictors was tested using variance inflation factors (VIFs; Quinn and Keough 2002), and predictors with VIF > 3 were excluded from further analysis to facilitate model interpretation. The final model included a spatial layer derived from geographical latitude/longitude coordinates, depth, seafloor slope, standard deviation of salinity, Secchi depth, near-bed current speed and wave exposure. Details of the predictors used are shown in Table 1.

Table 1.

Environmental variables used in the analysis (1–7) and for exclusion of areas (8–10)

Environmental variables Description Sources
1 Depth Interpolated depth, 200 m resolution EMODnet
2 Slope Seabed slope BALANCE
3 Salinity Standard deviation of seafloor salinity EUSeaMap
4 Secchi depth Modelled Secchi depth EUSeaMap
5 Seafloor currents Seafloor current speed EUSeaMap
6 Wave exposure Wave energy, log transformed EUSeaMap
7 Coordinate layer Latitude * longitude (decimal degrees)
8 Substrate Benthic marine landscapes, reclassified to rock, hard bottom complex, sand and mud + clay BALANCE
9 Photic zone Modelled photic zone BALANCE
10 Oxygen depletion Oxygen concentrations at seafloor, reclassified to < or > 2 ml l−1 BALANCE
Open in a new tab

The model was used to predict Marenzelleria spp. biomass for the entire Baltic Sea at 200 m horizontal resolution. Confidence intervals were calculated using randomForestCI 1.0.0 (Wager et al. 2014; Wager 2016) for R (R Core Team 2014). Spatially explicit phosphorus fluxes were calculated using two different approaches: (i) based on the biophysical model by Norkko et al. (2012), and (ii) based on measured fluxes at naturally varying densities of Marenzelleria spp. (Norling 2007, see Appendix S1 for a summary). The shallow coastal zone (assumed to be equivalent to the BALANCE model of the photic zone, Al-Hamdani and Reker 2007) was excluded from the phosphorus calculations, as phosphorus dynamics in this zone are more complex due to uptake from primary producers, larger abundance of hard substrates, and proximity to anthropogenic phosphorus sources. The differentiation between shallow coastal and deep-water sites is supported by Kauppi et al. (2017), where phosphorous flux dynamics was compared. Anoxic or hypoxic (suboxic) areas in the Baltic (i.e. with bottom oxygen concentrations < 2 mg l−1) have no long-term occurrence of macroscopic fauna, including Marenzelleria spp., and were also excluded.

Measured phosphorus fluxes from intact sediment cores at different densities of Marenzelleria spp. from Norling (2007, Appendix S1, Table S1, Fig. S1) were used to derive a linear relationship between Marenzelleria spp. biomass and phosphorus flux between sediment and water (Table S2). We used the slope of the relationship as a model for the additional effect of Marenzelleria spp. in the presence of other species. Norkko et al. (2012) found that the depth-integrated phosphorus content of the sediment was a sigmoidal function of faunal density, with a net release of phosphorus for densities below ca. 3269 ind m−2 and a net burial at higher densities. As the function was only defined for values greater than 1000 ind m−2, we extrapolated the phosphorous flux between 0 and 1000 ind m−2 using a quadratic polynomial function (y = − 0.0024x2 + 0.1207x). Since most samples lacked adequate sediment information (either missing or classified using very different methods), substrate was not used in the biomass model, although phosphorus cycling probably differs between sandy and muddy substrates. On sandy erosion bottoms, the phosphorus content in the upper 3 cm of the sediment was about half of that on silty-muddy accumulation bottoms (Viktorsson et al. 2013). In the model, we parameterized this and the typically lower average biomass in sand compared to mud (which is not accounted for in the biomass model, as substrate was not included as a predictor) through a phosphorus flux that was assumed to be 50% lower in sand (sediment type according to EUSeaMap, Table 1) compared to muddy sediments. For the linear function based on the Norling data (y = 12.613x), we assumed the phosphorous flux calculations to be representative of the summer, during which recently deposited organic material is available (Kauppi et al. 2017). The phosphorous flux values were therefore adjusted to be representative of 4 months during a typical year. The annual change in phosphorus flux was calculated for each grid cell and aggregated for each type of substrate and sea basin.

Results

The Marenzelleria spp. biomass model explained 57% of the sample variance. The normalized root-mean-square deviation (NRMSE) for observed versus predicted was 4% for training data and 7% for test data; r2 was 0.45 and 0.24, respectively (Table 2). The model is of acceptable quality according to benchmarks described by Bučas et al. (2013). The model predicted a total average biomass of 73.4 kton Marenzelleria spp. in the Baltic Sea (Table 3) with a 95% CI of 60–87 kton.

Table 2.

Statistics for the Marenzelleria spp. biomass prediction including root-mean-square error (RMSE) and normalized root-mean-square error (NRMSE)

Train Test
RMSE 5.746 4.917
NRMSE 0.041 0.069
Mean 1.515 1.191
r 2 0.447 0.239
Open in a new tab

Sample sizes were ntrain = 7962, ntest = 1000

Table 3.

Average biomass of Marenzelleria spp. and average net phosphorous (P) flux per year, based on measured experimental data and biophysical model data

Basins Biomass P-flux (measured data) P-flux (model data)
Average Min Max Average Min Max Average Min Max
Baltic Proper 21.1 15.7 26.9 2.2 1.6 2.8 1.5 1.1 1.9
Gulf of Riga 6.0 5.3 6.7 0.7 0.7 0.8 0.5 0.4 0.6
Gulf of Finland 15.9 14.5 17.4 1.6 1.5 1.8 1.1 1.0 1.2
Bothnian Sea 23.7 19.7 27.7 2.6 2.1 3.0 1.7 1.4 2.0
Bothnian Bay 6.6 4.9 8.3 0.6 0.4 0.7 0.4 0.3 0.5
Total 73.4 60.0 87.1 7.7 6.3 9.2 5.2 4.2 6.1
Open in a new tab

Sums are expressed in kton per basin. Min and max are based on 95% confidence intervals

The highest Marenzelleria spp. biomasses were predicted in the Gulf of Finland, Gulf of Riga, and along the Swedish Bothnian Sea coast. The maximum predicted wet weight biomass was 28.3 g m−2, similar to the 99th percentile obtained from the experimental data (29.7 g m−2; Fig. 1a). Maps of confidence intervals, which provide measures of uncertainty, show that areas with larger uncertainties are Western Bothnian Sea as well as Eastern and Western Baltic Proper (Fig. 1b).

Fig. 1.

Fig. 1

Open in a new tab

Baltic Sea maps of predicted Marenzelleria spp. biomass (a), its 95% confidence interval (b) and expected phosphorous release due to Marenzelleria spp. at a 200 m horizontal resolution, based on experimental data (c) and biophysical model data (d). White areas in the Baltic Sea are anoxic or hypoxic areas, which were excluded from the study as they have no long-term occurrence of macroscopic fauna. Maps of expected increases in phosphorus flux are available as electronic supplementary material in tiff format (Figs. S2 and S3). Figure colours used are taken from a palette developed for colour-blind people (http://jfly.iam.u-tokyo.ac.jp/color/)

The average net change in phosphorus flux from sediment to water due to the presence of Marenzelleria spp. in the Baltic Sea was 7.7 kton year−1 (95% CI of 6.3–9.2 kton year−1) based on measured fluxes in sediment (Table 3; Fig. 1c). The corresponding average net change in phosphorus flux was 5.2 kton year−1 (95% CI of 4.2–6.1 kton year−1) based on the biophysical model (Table 3; Fig. 1d). The basins with the highest estimated increase in phosphorus flux was the Bothnian Sea and the Baltic Proper (Table 3; Fig. 1c, d). High expected increases in phosphorus flux were observed in archipelago areas in Sweden and Finland and in the Gulf of Riga (Fig. 1c, d). Maps of expected increases in phosphorus flux are available as electronic supplementary material in geo-referenced tiff format (Figs. S2 and S3).

Discussion

Our results show that the current biomass densities of Marenzelleria spp. in the Baltic Sea increase the sea basin scale release of phosphorous from sediment to water. Although the activities of Marenzelleria spp. may potentially enhance long-term sedimentary phosphorus retention locally, the Marenzelleria spp. biomass densities needed to achieve this positive effect rarely occur. On the contrary, given the predicted biomasses, the recycling of phosphorus from Baltic Sea sediments to water is likely to increase, leading to higher dissolved concentrations and thus increased abatement costs for achieving internationally agreed reductions in phosphorus (HELCOM 2013a). Eutrophication is a severe problem in the Baltic Sea and all processes influencing nutrient cycling and availability need to be quantified. Although Marenzelleria spp., at high densities, could potentially provide a valuable ecosystem service by reducing phosphorous availability, in the current situation the opposite seems to be the case.

The invasion by Marenzelleria spp. followed a typical pattern for invasive species. The highest abundances occurred in 2005–2006 followed thereafter by decreased densities (Kauppi et al. 2015). As the data used in this study cover the time period 2000–2015, the temporal pattern was not resolved in our analysis. However, lower densities of Marenzelleria spp. than those found in this study would not contribute to a net burial of phosphorus according to our models. Gustafsson et al. (2012) estimated the release of phosphorus from Baltic Sea sediments at 229 kton year−1 during 1997–2006 based on physical variables. Compared to this amount, the presence of Marenzelleria spp. increases recirculation of phosphorus from the sediments by 2–4%. Our current results, with their assumptions and uncertainties, give an indication of the large-scale effects of one genus and the potential magnitude of its influence on these processes.

Research on larvae of Chironomidae (Diptera) in lakes (Gallepp 1979; Chaffin and Kane 2010) have demonstrated release of phosphorus from sediments and this release is directly related to larval density. The maximum density (6585 larvae m−2 = 28 g m−2) corresponded to a release of soluble reactive phosphorus of 2.8 g m−2 year−1 (Gallepp 1979), which is similar to our results (Fig. 1c, d). According to the function derived by Norkko et al. (2012), an abundance of 3269 ind m−2, corresponding approximately to a biomass of 54 g m−2, is necessary to reach the tipping point between net release and burial of phosphorous. Thus, if all the currently oxygen-depleted areas of the Baltic were to be oxygenized and colonized by Marenzelleria spp., a biomass of more than 54 g m−2 would be required, according to the model, to counteract the stimulating effect of Marenzelleria spp. on phosphorus efflux from the oxygenated sediments. The difference in Marenzelleria spp. densities that were applied in our model could be the main reason why the results from this study, that focus on larger scales, differ from the results of Norkko et al. (2012) who applied an equivalent model on a local and near-coastal scale. In coastal areas of Finland, Kauppi et al. (2015) found a maximum biomass of 57.73 g m−2 which is sufficient to cause a net burial of phosphorous. However, in the open sea, they found a maximum biomass of only 26.86 g m−2. Since we did not include the photic zone in the present study, the total effect of Marenzelleria spp. might differ from the results presented here. The open sea constitutes 90% of the area of the Baltic Proper and Gulf of Finland, and 80% of the area of the Gulf of Riga Open Sea sediment processes are thus likely to dominate the overall effect of Marenzelleria spp. on larger scales. The Bothnian Sea and Bothnian Bay have, in comparison to the other sea basins, generally sandier substrates with lower organic content (Borg and Jonsson 1996; Leipe et al. 2011), and a higher iron availability (Borg and Jonsson 1996). Therefore, the phosphorous flux estimates for these basins are more uncertain and may likely be overestimated.

Ecologically relevant data on phosphorous fluxes across substrates, species, sea basins, seasons and a gradient of Marenzelleria spp. densities are unavailable so our study results are dependent upon several assumptions which contribute varying degrees of uncertainty. Different Marenzelleria species likely have different impacts on the Baltic Sea ecosystem as they vary in bioturbation activity (Renz and Forster 2013). Norkko et al. (2012) based the model on measured fluxes from Quintana et al. (2013), who studied M. viridis. Using this function we also inherit any assumptions they made and assume that the three Marenzelleria species have the same bioturbation activity. This may not be true, as the species vary in their functional ecology (Renz and Forster 2013), but due to the lack of distribution data on the individual Marenzelleria species, this assumption was necessary. Renz and Forster (2014) found clear differences among the Marenzelleria species when expressing phosphorous fluxes and water exchange rates per individual. However, due to the lighter individual biomass of M. arctia, the differences were small or indistinguishable when comparing wet weight rates across species. For the predominant muddy and deeper areas of the Baltic Sea, M. arctia is the most common of the three sibling species (Blank et al. 2008; Kauppi et al. 2018). The function based on measured flux data of M. arctia is thus more representative for most of the Baltic and is similar to the situation in the Gulf of Finland (Isaev et al. 2016). The substrate maps available on Baltic Sea are of poor quality, but are generally more accurate in deeper, homogenous areas, which were the main focus of this study. We accounted for seasonality in the model calculations and our assumption agrees well with field measurements from a 33 m site (Kauppi et al. 2017). The Marenzelleria spp. biomass model explains a relatively modest 57% of the variance, but on larger scales this result was adequate. Relatively high natural variance in the model is also expected, due to both local population dynamics and patch bias resulting from the relatively small sampling area in benthic grab samples. Effects of the differences in functional ecology among the three Marenzelleria species on the study results are unclear. Renz and Forster (2013, 2014) demonstrated that effects on phosphate fluxes are similar (per wet weight) among all three species and suggest that interspecific differences are not significant in this case. Our experimental data were similar to the data used in other studies (Renz and Forster 2014; Kauppi 2017) when normalizing phosphorous fluxes to wet weight, further validating our results. An unaddressed variable, due to lack of applicable data, was the potential sorption of phosphorous to iron oxides. This could be of relatively greater importance in the Bothnian Sea and Bothnian Bay but may also interact with species-specific effects. Renz and Forster (2014) found a stimulation of anaerobic bacterial processes around the burrows of M. neglecta and M. viridis but not for M. arctia. A further source of variation in estimates for the Bothnian Sea and Bothnian Bay results from relatively low primary production in these basins and production highly limited by the long winter. Therefore, smaller amounts of organic matter are deposited. Nonetheless, the final results in this study were similar regardless of using theoretical modelling or based on experimental data assuming comparability across species. We are confident that the estimates are valid, especially for the Baltic Proper and the Gulf of Finland. The results show that the potential effects on Baltic Sea phosphorous fluxes are of significance for eutrophication management in the Baltic Sea region. We recommend that all national monitoring programmes in the Baltic Sea should determine Marenzelleria spp. samples to species level when appropriate, as an identification key is available for use on intact adult and subadult individuals (width about > 1.2 mm) (Bick 2005). Further research, especially on substrate mapping and ecologically relevant measurements of phosphorous flux at varying biomasses of Marenzelleria spp. is needed to increase estimate accuracy and further reduce uncertainties. This would provide data helpful for improving eutrophication management of the Baltic Sea.

Conclusions

Marenzelleria spp. cannot be eradicated from the Baltic Sea, so management options are limited to considering the effects of Marenzelleria spp. in eutrophication abatement plans and avoiding introduction and spread of other invasive non-indigenous species. The estimated yearly increase in fluxes of phosphorus from the sediments to the water in the Baltic Proper corresponds to about 12% of the yearly burial (Gren et al. 2013). This means a corresponding reduction in the self-cleaning capacity of the basin. Gren et al. (2013) found that a 10% decrease in the self-cleaning capacity of the Baltic Proper increases the cost of meeting HELCOM (2013a) targets, within 70 years, by approximately 45% or 6.42 × 109 Swedish kronor per year (ECB exchange rate 2nd January 2018 was 1 Euro = 9.8 SEK). The results of this study show the potential effects of one invasive genus on an ecosystem and illustrate the necessity of management programmes for preventing the introduction of potentially invasive non-indigenous species. This study covers only one aspect of the ecosystem services or disservices provided by Marenzelleria spp. Considering the large total biomass of Marenzelleria spp. in the Baltic Sea, the genus is also likely to influence other processes on large scale. Processes that may have been influenced includes nitrogen cycling (Hietanen et al. 2007), oxygenation (Bonaglia et al. 2013), re-activation of high priority contaminants from pulp fibre sediments in the Bothnian Sea and Bothnian Bay, release of contaminants in Baltic Sea Proper areas (Granberg et al. 2008), and food availability for fish such as cod (Winkler and Debus 1996). An ecological function (bioturbation by Marenzelleria spp.) that could be an ecosystem disservice in some situations might be a benefit in other conditions (Saunders and Luck 2016). The magnitude of these opposite impacts can be effectively weighted and considered, when appropriate knowledge is available, using cost–benefit analysis.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary material 1 (PDF 983 kb) (982.3KB, pdf)
Supplementary material 2 (TIFF 3076 kb) (3MB, tif)
Supplementary material 3 (TIFF 2768 kb) (2.7MB, tif)

Acknowledgements

The authors would like to thank the Swedish Research Council for Environment, Agricultural Sciences and Spatial Planning (FORMAS) and Swedish Environmental Protection Agency (Naturvårdsverket) for financial support (VALUES Project), and Stockholm Marina Forskningscentrum (Östersjöcentrum) for time and space at Askö Laboratory; Andrey Sikorski for taxonomic expertise; Mats Westerbom for help with the OIVA database; Joanna Norkko, Ragnar Elmgren, Anna-Stiina Heiskanen, Eva Roth, and Gunilla Ejdung for valuable comments on the study; Two anonymous reviewers and the editor, whose comments greatly improved the manuscript; HELCOM as well as laboratory and field staff contributing to the benthic monitoring of the Baltic Sea.

Biographies

Antonia Nyström Sandman

is a Research Scientist at AquaBiota Water Research. Her research interests cover various aspects of linkages between spatial distribution of marine organisms and habitats and marine management, including species distribution modelling and ecosystem services in a spatial context.

Johan Näslund

has previously worked with management of invasive alien species at the Swedish Environmental Protection Agency and is currently working as a Research Scientist at AquaBiota. His research interests include eDNA and invasive alien species management.

Ing-Marie Gren

is a Professor at Department of Economics, Swedish University of Agricultural Sciences, and her research interest is on design of environmental policy instruments and valuation of ecosystem services.

Karl Norling

is a Senior Analyst (PhD) at Swedish Agency for Marine and Water Management. His research interest focuses on understanding the importance of sediment living macrofauna for biogeochemical processes, as well as consequences of anthropogenic impacts (e.g. eutrophication, organic matter enrichment, hypoxia, contaminants, temperature and ocean acidification) for benthic community structure and function.

Contributor Information

Antonia Nyström Sandman, Phone: +46852230252, Email: antonia.sandman@aquabiota.se.

Johan Näslund, Email: johan.naslund@aquabiota.se.

Ing-Marie Gren, Email: ing-marie.gren@slu.se.

Karl Norling, Email: karl.norling@havochvatten.se.

References

  1. Al-Hamdani, Z., and J. Reker, ed. Towards marine landscapes in the Baltic Sea, Vol. #10. BALANCE Interim Report, 2007. http://balance-eu.org/. Accessed 04 March 2015.
  2. Bick A. A new Spionidae (Polychaeta) from North Carolina and a redescription of Marenzelleria wireni Augener, 1913, from Spitsbergen, with a key for all species of Marenzelleria. Helgoland Marine Research. 2005;59:265–272. doi: 10.1007/s10152-005-0002-7. [DOI] [Google Scholar]
  3. Bick A, Burckhardt R. First record of Marenzelleria viridis (Polychaeta, Spionidae) in the Baltic Sea, with a key to the Spionidae of the Baltic Sea. Mitteilungen aus dem Zoologischen Museum in Berlin. 1989;65:237–247. doi: 10.1002/mmnz.19890650208. [DOI] [Google Scholar]
  4. Blank M, Laine AO, Juerss K, Bastrop R. Molecular identification key based on PCR/RFLP for three polychaete sibling species of the genus Marenzelleria, and the species’ current distribution in the Baltic Sea. Helgoland Marine Research. 2008;62:129–141. doi: 10.1007/s10152-007-0081-8. [DOI] [Google Scholar]
  5. Bonaglia S, Bartoli M, Gunnarsson JS, Rahm L, Raymond C, Svensson O, Yekta SS, Brüchert V. Effect of reoxygenation and Marenzelleria spp. bioturbation on Baltic Sea sediment metabolism. Marine Ecology Progress Series. 2013;482:43–55. doi: 10.3354/meps10232. [DOI] [Google Scholar]
  6. Borg H, Jonsson P. Large-scale metal distribution in Baltic Sea sediments. Marine Pollution Bulletin. 1996;32:8–21. doi: 10.1016/0025-326X(95)00103-T. [DOI] [Google Scholar]
  7. Breiman L. Random forests. Machine Learning. 2001;45:5–32. doi: 10.1023/A:1010933404324. [DOI] [Google Scholar]
  8. Bučas M, Bergström U, Downie AL, Sundblad G, Gullström M, Von Numers M, Šiaulys A, Lindegarth M. Empirical modelling of benthic species distribution, abundance, and diversity in the Baltic Sea: Evaluating the scope for predictive mapping using different modelling approaches. ICES Journal of Marine Science. 2013 doi: 10.1093/icesjms/fst036. [DOI] [Google Scholar]
  9. Carstensen J, Conley DJ, Bonsdorff E, Gustafsson BG, Hietanen S, Janas U, Jilbert T, Maximov A, et al. Hypoxia in the Baltic Sea: Biogeochemical cycles, benthic fauna, and management. Ambio. 2014;43:26–36. doi: 10.1007/s13280-013-0474-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Chaffin JD, Kane DD. Burrowing mayfly (Ephemeroptera: Ephemeridae: Hexagenia spp.) bioturbation and bioirrigation: A source of internal phosphorus loading in Lake Erie. Journal of Great Lakes Research. 2010;36:57–63. doi: 10.1016/j.jglr.2009.09.003. [DOI] [Google Scholar]
  11. Conley DJ, Stockenberg A, Carman R, Johnstone RW, Rahm L, Wulff F. Sediment–water nutrient fluxes in the Gulf of Finland, Baltic Sea. Estuarine, Coastal and Shelf Science. 1997;45:591–598. doi: 10.1006/ecss.1997.0246. [DOI] [Google Scholar]
  12. Conley DJ, Humborg C, Rahm L, Savchuk OP, Wulff F. Hypoxia in the Baltic Sea and basin-scale changes in phosphorus biogeochemistry. Environmental Science and Technology. 2002;36:5315–5320. doi: 10.1021/es025763w. [DOI] [PubMed] [Google Scholar]
  13. Cutler DR, Edwards TC, Jr, Beard KH, Cutler A, Hess KT, Gibson J, Lawler JJ. Random forests for classification in ecology. Ecology. 2007;88:2783–2792. doi: 10.1890/07-0539.1. [DOI] [PubMed] [Google Scholar]
  14. de Groot RS, Alkemade R, Braat L, Hein L, Willemen L. Challenges in integrating the concept of ecosystem services and values in landscape planning, management and decision making. Ecological Complexity. 2010;7:260–272. doi: 10.1016/j.ecocom.2009.10.006. [DOI] [Google Scholar]
  15. Gallepp GW. Chironomid influence on phosphorus release in sediment–water microcosms. Ecology. 1979;60:547–556. doi: 10.2307/1936075. [DOI] [Google Scholar]
  16. Gogina M, Nygård H, Blomqvist M, Daunys D, Josefson AB, Kotta J, Maximov A, Warzocha J, Yermakov V, Gräwe U, Zettler ML. The Baltic Sea scale inventory of benthic faunal communities. ICES Journal of Marine Science. 2016;73:1196–1213. doi: 10.1093/icesjms/fsv265. [DOI] [Google Scholar]
  17. Götting M, Mikkat S, Verleih M, Bastrop R. Proteomic comparison of two invasive polychaete species and their naturally occurring F-1-hybrids. Journal of Proteome Research. 2011;11:897–905. doi: 10.1021/pr200710z. [DOI] [PubMed] [Google Scholar]
  18. Granberg ME, Gunnarsson JS, Hedman JE, Rosenberg R, Jonsson P. Bioturbation-driven release of organic contaminants from Baltic Sea sediments mediated by the invading polychaete Marenzelleria neglecta. Environmental Science and Technology. 2008;42:1058–1065. doi: 10.1021/es071607j. [DOI] [PubMed] [Google Scholar]
  19. Gren I-M, Savchuck OP, Jansson T. Cost-effective spatial and dynamic management of a eutrophied Baltic Sea. Marine Resource Economics. 2013;28:263–284. doi: 10.5950/0738-1360-28.3.263. [DOI] [Google Scholar]
  20. Gustafsson BG, Schenk F, Blenckner T, Eilola K, Meier HM, Müller-Karulis B, Neumann T, Ruoho-Airola T, et al. Reconstructing the development of Baltic Sea eutrophication 1850–2006. Ambio. 2012;41:534–548. doi: 10.1007/s13280-012-0318-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. HELCOM. Ministerial Declaration: Taking further action to implement the Baltic Sea Action Plan—Reaching good environmental status for a healthy Baltic Sea (online). HELCOM: Copenhagen, 2013a.
  22. HELCOM. HELCOM HUB—Technical Report on the HELCOM Underwater Biotope and habitat classification. Baltic Sea Environmental Proceedings 139, 2013b.
  23. HELCOM. Red List of Baltic Sea underwater biotopes, habitats and biotope complexes. Baltic Sea Environmental Proceedings 138, 2013c.
  24. Hietanen S, Laine AO, Lukkari K. The complex effects of the invasive polychaetes Marenzelleria spp. on benthic nutrient dynamics. Journal of Experimental Marine Biology and Ecology. 2007;352:89–102. doi: 10.1016/j.jembe.2007.07.018. [DOI] [Google Scholar]
  25. Isaev AV, Eremina TR, Ryabchenko VA, Savchuk OP. Model estimates of the impact of bioirrigation activity of Marenzelleria spp. on the Gulf of Finland ecosystem in a changing climate. Journal of Marine Systems. 2016 doi: 10.1016/j.jmarsys.2016.08.005. [DOI] [Google Scholar]
  26. Karlson K, Hulth S, Ringdahl K, Rosenberg R. Experimental recolonisation of Baltic Sea reduced sediments: Survival of benthic macrofauna and effects on nutrient cycling. Marine Ecology Progress Series. 2005;294:35–49. doi: 10.3354/meps294035. [DOI] [Google Scholar]
  27. Karlson AML, Almqvist G, Skora KE, Appelberg M. Indications of competition between non-indigenous round goby and native flounder in the Baltic Sea. ICES Journal of Marine Science. 2007;64:479. doi: 10.1093/icesjms/fsl049. [DOI] [Google Scholar]
  28. Karlson AM, Näslund J, Rydén SB, Elmgren R. Polychaete invader enhances resource utilization in a species-poor system. Oecologia. 2011;166:1055–1065. doi: 10.1007/s00442-011-1936-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Kauppi L, Norkko A, Norkko J. Large-scale species invasion into a low-diversity system: Spatial and temporal distribution of the invasive polychaetes Marenzelleria spp. in the Baltic Sea. Biological Invasions. 2015;17:2055–2074. doi: 10.1007/s10530-015-0860-0. [DOI] [Google Scholar]
  30. Kauppi L, Norkko J, Ikonen J, Norkko A. Seasonal variability in ecosystem functions: Quantifying the contribution of invasive species to nutrient cycling in coastal ecosystems. Marine Ecology Progress Series. 2017;572:193–207. doi: 10.3354/meps12171. [DOI] [Google Scholar]
  31. Kauppi L, Norkko A, Norkko J. Seasonal population dynamics of the invasive polychaete genus Marenzelleria spp. in contrasting soft-sediment habitats. Journal of Sea Research. 2018;131:46–60. doi: 10.1016/j.seares.2017.10.005. [DOI] [Google Scholar]
  32. Kotta J, Orav H, Sandberg-Kilpi E. Ecological consequence of the introduction of the polychaete Marenzelleria cf. viridis into a shallow-water biotope of the northern Baltic Sea. Journal of Sea Research. 2001;46:273–280. doi: 10.1016/S1385-1101(01)00088-0. [DOI] [Google Scholar]
  33. Lehtoranta J, Heiskanen A-S. Dissolved iron:phosphate ratio as an indicator of phosphate release to oxic water of the inner and outer coastal Baltic Sea. Hydrobiologia. 2003;492:69–84. doi: 10.1023/a:1024822013580. [DOI] [Google Scholar]
  34. Leipe T, Tauber F, Vallius H, Virtasalo J, Uścinowicz S, Kowalski N, Hille S, Lindgren S, et al. Particulate organic carbon (POC) in surface sediments of the Baltic Sea. Geo-Marine Letters. 2011;31:175–188. doi: 10.1007/s00367-010-0223-x. [DOI] [Google Scholar]
  35. Leppäkoski E, Olenin S. Non-native species and rates of spread: Lessons from the brackish Baltic Sea. Biological Invasions. 2000;2:151–163. doi: 10.1023/a:1010052809567. [DOI] [Google Scholar]
  36. Liaw A, Wiener M. Classification and regression by randomForest. R News. 2002;2:18–22. [Google Scholar]
  37. Maes J, Egoh B, Willemen L, Liquete C, Vihervaara P, Schägner JP, Grizzetti B, Drakou EG, et al. Mapping ecosystem services for policy support and decision making in the European Union. Ecosystem Services. 2012;1:31–39. doi: 10.1016/j.ecoser.2012.06.004. [DOI] [Google Scholar]
  38. Maximov A, Bonsdorff E, Eremina T, Kauppi L, Norkko A, Norkko J. Context-dependent consequences of Marenzelleria spp. (Spionidae: Polychaeta) invasion for nutrient cycling in the Northern Baltic Sea. Oceanologia. 2015;57:342–348. doi: 10.1016/j.oceano.2015.06.002. [DOI] [Google Scholar]
  39. Norkko J, Reed DC, Timmermann K, Norkko A, Gustafsson BG, Bonsdorff E, Slomp CP, Carstensen J, et al. A welcome can of worms? Hypoxia mitigation by an invasive species. Global Change Biology. 2012;18:422–434. doi: 10.1111/j.1365-2486.2011.02513.x. [DOI] [Google Scholar]
  40. Norling, K. Ecosystem functions in benthos: Importance of macrofaunal bioturbation and biodiversity for mineralization and nutrient fluxes. PhD-Thesis, Department of Marine Ecology, University of Gothenburg, Sweden, 2007.
  41. Norling K, Rosenberg R, Hulth S, Grémare A, Bonsdorff E. The importance of functional biodiversity and species specific traits of benthic fauna for ecosystem functions in marine sediments. Marine Ecology Progress Series. 2007;332:11–23. doi: 10.3354/meps332011. [DOI] [Google Scholar]
  42. Quinn GP, Keough MJ. Experimental design and data analysis for biologists. Cambridge: Cambridge University Press; 2002. [Google Scholar]
  43. Quintana CO, Kristensen E, Valdemarsen T. Impact of the invasive polychaete Marenzelleria viridis on the biogeochemistry of sandy marine sediments. Biogeochemistry. 2013;115:95–109. doi: 10.1007/s10533-012-9820-2. [DOI] [Google Scholar]
  44. R Core Team . R: A language and environment for statistical computing. Vienna: R Foundation for Statistical Computing; 2014. [Google Scholar]
  45. Reed DC, Slomp CP, Gustafsson BG. Sedimentary phosphorus dynamics and the evolution of bottom-water hypoxia: A coupled benthic–pelagic model of a coastal system. Limnology and Oceanography. 2011;56:1075–1092. doi: 10.4319/lo.2011.56.3.1075. [DOI] [Google Scholar]
  46. Renz JR, Forster S. Are similar worms different? A comparative tracer study on bioturbation in the three sibling species Marenzelleria arctia, M. viridis, and M. neglecta from the Baltic Sea. Limnology and Oceanography. 2013;58:2046–2058. doi: 10.4319/lo.2013.58.6.2046. [DOI] [Google Scholar]
  47. Renz JR, Forster S. Effects of bioirrigation by the three sibling species of Marenzelleria spp. on solute fluxes and porewater nutrient profiles. Marine Ecology Progress Series. 2014;505:145–159. doi: 10.3354/meps10756. [DOI] [Google Scholar]
  48. Sarda R, Valiela I, Foreman K. Life cycle, demography, and production of Marenzelleria viridis in a salt-marsh of southern New England. Journal of the Marine Biological Association of the United Kingdom. 1995;75:725–738. doi: 10.1017/S0025315400039138. [DOI] [Google Scholar]
  49. Saunders ME, Luck GW. Limitations of the ecosystem services versus disservices dichotomy. Conservation Biology. 2016 doi: 10.1111/cobi.12740. [DOI] [PubMed] [Google Scholar]
  50. Urban-Malinga B, Warzocha J, Zalewski M. Effects of the invasive polychaete Marenzelleria spp. on benthic processes and meiobenthos of a species-poor brackish system. Journal of Sea Research. 2013;80:25–34. doi: 10.1016/j.seares.2013.02.005. [DOI] [Google Scholar]
  51. Viitasalo-Frösén S, Laine AO, Lehtiniemi M. Habitat modification mediated by motile surface stirrers versus semi-motile burrowers: Potential for a positive feedback mechanism in a eutrophied ecosystem. Marine Ecology Progress Series. 2009;376:21–32. doi: 10.3354/meps07788. [DOI] [Google Scholar]
  52. Viktorsson L, Ekeroth N, Nilsson M, Kononets M, Hall POJ. Phosphorus recycling in sediments of the central Baltic Sea. Biogeosciences. 2013;10:3901–3916. doi: 10.5194/bg-10-3901-2013. [DOI] [Google Scholar]
  53. Virnstein RW. The importance of predation by crabs and fishes on benthic infauna in Chesapeake Bay. Ecology. 1977;58:1200–1217. doi: 10.2307/1935076. [DOI] [Google Scholar]
  54. Wager, S. randomForestCI: Confidence intervals for random forests. R package version 1.0.0. 2016.
  55. Wager S, Hastie T, Efron B. Confidence intervals for random forests: The jackknife and the infinitesimal jackknife. Journal of Machine Learning Research. 2014;15:1625–1651. [PMC free article] [PubMed] [Google Scholar]
  56. Winkler, H.M., and L. Debus. Is the polychaete Marenzelleria viridis an important food item for fish. In Proceedings of the 13th symposium of the Baltic marine biologists, 1996, 147–151.
  57. Zettler L, Daunys D, Kotta J, Bick A. History and success of an invasion into the Baltic Sea: The polychaete Marenzelleria cf. viridis, development and strategies. In: Leppäkoski E, Gollasch S, Olenin S, editors. Invasive aquatic species of Europe. Distribution, impacts and management. Dordrecht: Kluwer Academic Publishers; 2002. pp. 66–75. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary material 1 (PDF 983 kb) (982.3KB, pdf)
Supplementary material 2 (TIFF 3076 kb) (3MB, tif)
Supplementary material 3 (TIFF 2768 kb) (2.7MB, tif)

Articles from Ambio are provided here courtesy of Springer

RESOURCES