Biological Invasions

, Volume 17, Issue 9, pp 2605–2619 | Cite as

Understanding winter distribution and transport pathways of the invasive ctenophore Mnemiopsis leidyi in the North Sea: coupling habitat and dispersal modelling approaches

  • Carmen David
  • Sandrine Vaz
  • Christophe Loots
  • Elvire Antajan
  • Johan van der Molen
  • Morgane Travers-Trolet
Original Paper


The invasive ctenophore Mnemiopsis leidyi has been reported in various coastal locations in the southern North Sea in the past years. Since 2009, International Bottom Trawl Surveys have recorded this species each winter in open waters. As this species, well-known for its dramatic disturbance of ecosystems, was expected not to be able to overwinter offshore it is crucial to understand its distribution dynamics. Two modelling methods, a quantile regression and a particle tracking model, were used (1) to identify habitats where the invasive ctenophore M. leidyi could survive the North Sea cold winters and (2) to investigate the dispersal of individuals between these different habitats, emphasizing favorable areas where sustainable populations could have been established. Temperature was found to be the crucial factor controlling the winter distribution of M. leidyi in the North Sea. High abundance predictions in winter were associated with low values of temperature, which characterise south-eastern coastal areas and estuaries influenced by riverine runoff. A retention-based M. leidyi population was indicated along the northern Dutch coast and German Bight and a transport-based population offshore from the western Danish coast. Individuals found in the open waters were transported from southern coasts of the North Sea, thus the open water population densities depend on the flux of offspring from these areas. This study provides the first estimates of the overwinter areas of this invasive species over the cold winters in the North Sea. Based on the agreement of habitat and dispersal model results, we conclude that M. leidyi has become established along south-eastern coasts of the North Sea where the environment conditions allows overwintering and it can be retained for later blooms.


Mnemiopsis leidyi North Sea Overwinter refuges Habitat modelling Particle tracking 


Over recent decades, gelatinous species have become a major ecological concern worldwide due to increased abundance in their native habitats that affects the local ecosystems (Mills 2001) and potential spreading to other areas either by natural expansions or by human activities (Graham and Bayha 2007; GESAMP 1997; Ghabooli et al. 2011; Lehmann and Javidpour 2010). A feared invader, the ctenophore Mnemiopsis leidyi Agassiz, 1865 has become well known for its predatory capabilities (Costello et al. 2012) and ecological consequences (Shiganova 1998; Finenko et al. 2006). It is known for large populations in native areas (Kremer 1994) and explosive invasions in Eurasian seas (Purcell and Arai 2001; GESAMP 1997). Listed among the 100 ‘World’s Worst’ invaders by the International Union for Conservation of Nature (IUCN), M. leidyi is often associated with fishery collapses (Purcell and Arai 2001), competing for resources with local populations (Bilio and Niermann 2004; Mills 1995) and feeding on pelagic fish eggs and larvae (Purcell and Arai 2001; Kremer 1979; Purcell et al. 1994; Monteleone and Duguay 1988).

Native from estuaries along the coastline of North and South America, M. leidyi arrived in Europe in the 1980’s in the Black Sea and by the early 1990’s, it had already invaded the seas of Azov, Marmara and eastern Mediterranean (Shiganova et al. 2001), and later in 1999, it was reported for the first time in the Caspian Sea (Ivanov et al. 2000). Populations developed to very high densities, up to 12 kg m−2 in coastal waters and 3 kg m−2 in open waters (GESAMP 1997), and was reported to disrupt the functioning of these basins’ ecosystems (Shiganova and Malej 2009; Shiganova and Bulgakova 2000; Kideys et al. 2000; Kideys 2002). In 2006, this invasive species was observed in the North Sea, first on the Dutch and Danish coasts (Faasse and Bayha 2006; Boersma et al. 2007), and in the western Baltic Sea (Javidpour et al. 2006; Janas and Zgrundo 2007), raising many concerns about the potential impact on these commercially important ecosystems. Faasse and Bayha (2006) suggested that it might have been present in Dutch waters for several years but misidentified and only the recent increase in sea surface temperature could have driven the development of large ctenophore swarms. In the following years, its presence was signalled along the North Sea continental coastline from France to Sweden (van Ginderdeuren et al. 2012; Antajan et al. 2009, 2014a) and recently it was recorded in open water during winter.

Showing wide tolerance to physical parameters (i.e. temperatures between 2 and 32 °C, salinities between 2 and 38) (Purcell et al. 2001), M. leidyi is close to the reported limits of its occurrence in the North Sea considering low temperatures, especially in winter. Even though the persistence of M. leidyi through winters with temperatures below 1 °C has been recorded in its native habitat in Narragansett Bay, Rhode Island (Costello et al. 2006), temperatures below 4 °C were believed to prevent overwintering in the central and north Baltic sea (Lehtiniemi et al. 2011). While the geographical distribution of the species is limited by the physical and chemical properties of the environment, the abundance is controlled mainly by food availability and natural predators (Shiganova 1998). The absence of major predators in the Black Sea at the moment of introduction is believed to have permitted its explosive invasion with ecological consequences along the trophic chain (Shiganova and Bulgakova 2000). On the contrary, presence in the North Sea of their natural predators (Hosia and Titelman 2011; Purcell and Cowan 1995) and competitors (Hamer et al. 2011) also found in their native habitats could limit population growth and explain the lower abundances recorded in this region. Still the concern of future abundance explosion remains, as individuals have high reproduction ability when food concentrations and temperatures are high (above 10 °C) (Costello et al. 2006; Reeve et al. 1989).

The spatial extent of ctenophore populations and the bloom periods in the North Sea remain yet unclear. Regular surveys have been conducted in coastal areas, but inadequate sampling and preservation techniques may have underestimated the abundance or even overlooked its presence. M. leidyi presence in the North Sea appeared as a surprise, when in winter 2009 it was unexpectedly found offshore during the International Bottom Trawl Survey (IBTS). Here, we sought to address critical knowledge gaps on the distribution of M ledyi in North Sea, focusing on the winter period. The present study tries to reveal some of the above stated uncertainties by complementary modelling tools and to address the questions: (1) as M. leidyi resisted the cold North Sea winter so far, which areas support winter populations and represent potential winter refugia? (2) Since M. leidyi was recorded in offshore waters, how does large-scale circulation affect the spatial distribution of ctenophore?

The first part of the study focuses on modelling the potential winter habitat of M. leidyi in the North Sea, taking into account environmental and trophic features for the years 2009–2011. We aim to identify overwinter refuges and the environmental and biotic factors associated with regions of high abundance. The second part of this study will investigate the dispersal patterns of the populations between the potential habitat areas previously defined using a Lagrangian particles tracking model, Ichthyop (Lett et al. 2008), coupled with a hydrodynamic model, General Estuarine Transport Model (GETM) (Burchard and Bolding 2002).

Materials and methods

Study area and International Bottom Trawl Survey data

The North Sea is a relatively shallow sea with a large cyclonic gyre that rotates around the basin (Brown et al. 1999). The northern part is deeper, subject to strong oceanic influences, characterised by seasonal stratification of the water column. The southern North Sea is shallower (20–50 m) and remains mixed for most of the year. The continental coastal zone has a mean depth of 15 m, being under strong influence of rivers inputs (Mackinson and Daskalov 2007).

The North Sea is sampled biannually during the IBTS, a European program fundamentally dedicated to compute abundance’s indices of the main fish stocks. For the most recent years, the French Research Institute for Exploration of the Sea (IFREMER) had extended their winter sampling to the gelatinous zooplankton, monitoring M. leidyi distribution among other species. Data were collected during winters (January–February) 2009–2011 from a total of 347 stations with a Midwater Ring Net (black conical net with an opening diameter of 2 m, 13 m long, with mesh size of 1.6 mm reared with 500 µm mesh for the last 1 m) (Fig. 1). Gelatinous species were sorted, identified and measured live on board. Recurrent patches of M. leidyi (max. 308 ind. 1000 m−3) were observed off the Dutch and the Danish coasts and were composed of lobate individuals, and their oral-aboral length most often of 2 to 3 cm (range between 1 and 7 cm) (Antajan et al. 2014b). Environmental parameters were recorded at each station using CTD and Niskin bottles and showed no variation with depth. At stations where M. leidyi was present, temperature was between 1.2 and 6.8 °C, salinity between 27.2 and 35.5, and chlorophyll a concentration was <0.2 µg l−1.
Fig. 1

Distribution of Mnemiopsis leidyi (ind. 1000 m−3) in the North Sea in winter (January–February) 2009–2011; data from the 3 years are presented in red, blue and green circles with the size of circles relative to the ctenophore abundance at the sampling stations; bathymetry of the study area is shown at 20 m interval isobaths

Statistical modelling of potential habitat

A Quantile Regression model (QR) (Koenker 2005) was used to estimate M. leidyi distribution associated with North Sea environmental conditions. The statistical model was built on IBTS data of winters 2009–2011. Normality of the data was tested with a Shapiro test and normalized by log transformation when necessary. Variables were selected a priori from all available data by Spearman’s rank correlation coefficient, used to identify redundant variables. Among 39 environmental variables available, 20 variables were used to fit the model (Table 1). Quadratic terms for all selected variables were added to the dataset.
Table 1

Environmental variables collected during International Bottom Trawl Survey in North Sea in winter (January–February) 2009–2011 and used for fitting the quantile regression model; physical and chemical variables used represent sub-surface values

Physical and biogeochemical variables

Biological variables (ind. 5000 m−3)

Potential predator/competitor abundance

Potential prey abundance (fish larvae)

Depth (m)

Beroe spp. (Gronov, 1760)

Sardina pilchardus (Walbaum, 1792)

Temperature (°C)

Cyanea lamarckii (Péron and Lesueur, 1810)

Clupea harengus (Linnaeus, 1758)


Pleurobrachia pileus (Müller, 1776)

Trisopterus luscus (Linnaeus, 1758)

Chlorophyll a (µg l−1)


Microstomus kitt (Walbaum, 1792)

NO2− (µg l−1)



NH4+ (µg l−1)



PO4− (µg l−1)



Organic matter (µg l−1)



SiOH (µg l−1)


Two models were developed: one predicts the species response to the physical and chemical environment and the other the response to the biotic environment. They were coupled to give the final model. Initially two full models containing all variables were fitted and then tested for the quantiles: 75th, 80th, 85th, 90th and 95th. Starting from the initial full models, terms were removed by backward elimination based on average p values across the range of quantiles, until arriving at a model where all terms remained significant (p < 0.05) for at least one quantile. This selection procedure may result in overfitting the model (Vaz et al. 2008). Whether the selection procedure resulted in an appropriate model was assessed by comparing the final model with three alternative models of varying complexity levels at the same quantile. The testing was performed with an ANOVA analysis of variance (Stahle and Wold 1989), a widely used statistical method for testing linear hypotheses extended to quantile regression (Chen et al. 2008).

Model evaluation was done by cross-validation on the same dataset. Following Vaz et al. (2008) two separate validation tests were performed: the Correct Classification test and the Spearman’s rank correlation test. The Correct Classification test defines the proportion of observed values in the validation dataset that falls below those predicted. A bootstrap procedure was used to estimate the mean and the confidence limits of the Correct Classification statistic of the model as it produces more robust validation of the model. The model is validated if the selected quantile is less than the upper confidence limit of the bootstrapped mean Correct Classification statistic. The Spearman’s rank correlation test assesses how well the relationship between two variables can be described using a monotonic function. The test was preferred to compare the observed and predicted values as its correlation coefficient (rs) does not assume a linear relationship between the variables. The test was considered passed for a p value <0.05 and a rs > 0.1 (Legendre and Legendre 1998). Modelling and statistical analyses were performed in the software R (R Core Team 2013). Mapping was realized with ArcGIS 10 (©ESRI).

Modelling dispersal using coupled hydrodynamic and particle tracking models

The General Estuarine Transport Model (GETM, see was used to generate flow fields for the particle tracking model. GETM is a three-dimensional hydrodynamic model that uses the General Ocean Turbulence Model for the vertical turbulence structure. The geographic domain of the model extends between 48.5° and 60°N latitude and between 5°W and 16°E longitude. Spherical coordinates are used in the horizontal plane, while general coordinates are used in the vertical (equidistant in shallow water and contracted near the surface and the sea bed in deeper water). Model resolution on the horizontal is 6 nautical miles and it has 25 vertical layers. The shallow-water equations are solved on an Arakawa C grid. Lateral boundary conditions are represented by a land mask on the horizontal numerical grid (Burchard et al. 2012). Elevations and currents, derived from a shelf-wide barotropic model driven by Topex Poseidon satellite altimetry data, were imposed at the open boundaries using a Flather boundary condition. The model allows for drying and flooding of tidal flats. The model has been validated for the North Sea and was used previously, among others, for eutrophication assessment (Lenhart et al. 2010). Generated fields were saved every hour, providing physical parameters and currents data.

A Lagrangian particle tracking model was coupled offline with the hydrodynamic model to simulate gelatinous plankton dispersal in the North Sea. Ichthyop (see is an Individual Based Model (IBM), initially designed to study the effects of environmental factors on the dynamics and dispersal of fish eggs and larvae (Lett et al. 2008). The method relies on tracking the positions of particles within water velocity fields generated by three-dimensional hydrodynamic models. For this study input velocity fields come from the GETM oceanic model archived for the years 2007–2008. Transport of particles relies only on advection (or drift) as no diffusion term was introduced. The IBM model was run with a time step of 2400 s, based on an explicit Runge–Kutta iteration scheme. Simulation output from the Ichthyop model providing position of each particle at each time step were saved in netcdf files and further analysed in the software R (R Core Team 2013).


At the basin scale, M. leidyi individuals are considered to be dispersed as passive particles. Particles were released at midnight on the 1st day of each month from February 2007 to January 2008 and tracked until 1st February 2008. The end of the simulations was chosen to correspond with the IBTS sampling period. Simulations consisted of releasing 30,000 particles each month over the domain sampled by IBTS survey. Particles were released homogenously on the horizontal plane and over the entire water column without any preferential depth. The release domain was divided into areas according to their importance for the ctenophore overwintering following habitat model results. Using abundance predicted by the QR model, homogenous areas were delimited and characterized as high potential, intermediate or unfavourable winter refuges for M. leidyi. The resulting areas are compared notably through their final density in particles. Particles transported out of the study domain are not discussed. All particles found in the study domain at the end of simulations were considered, including beached particles. The flux of particles is analysed between the areas of origin and the areas of arrival. For the latter, the area of origin of particles was defined as identical if the arrival area was the same as the release area, adjacent if the particle comes from neighbouring areas or distant if it comes from any other areas. The area of origin of particles was analysed in each arrival area and used to classify the arrival areas. A hierarchical classification was performed on the relative abundance of particles (defined relatively to the arrival area as identical, adjacent or distant) using the Ward method and Euclidian distance (Legendre and Legendre 1998). The Ward method was chosen as its criterion minimizes the total within-cluster variance (decrease in variance for the cluster being merged). The temporal dynamics of the flows of particles was also studied by looking at contribution of origin areas to final particle distribution per season.


Habitat model

The Quantile Regression model (QR) prediction of the potential habitat of M. leidyi in North Sea environmental winter conditions gave satisfactory results explaining 63 % of variance in the dataset. The model, significant for the 90th quantile, retained five variables: temperature and its quadratic term, the abundance of two gelatinous species, Cyanea lamarkii and Pleurobrachia pileus, and the abundance of one group of fish larvae, Syngnathidae. Among selected variables, temperature had a negative relationship with M. leidyi abundance, while all the biological variables showed a positive linear relationship (Supplementary material 1). The relationship with temperature was described by a second degree polynomial which predicts lower abundance at intermediate temperatures, and higher abundance at the lowest and highest temperature observed during the survey.

The selected 90th quantile was confirmed by the Correct Classification test with the upper confidence limit of the bootstrapped correct classification statistic of 92 %. Model evaluation based on the Spearman’s rank correlation coefficient gave a positive correlation of r = 0.44 between observed and predicted values, within the 95 % confidence intervals [0.36; 0.53], calculated by 1000 bootstrap datasets.

The final QR model was then compared to alternative QR models of varying complexity: (1) “null” (intercept only) model, (2) only simple terms of the five retained variables and (3) maximal model having both simple and quadratic terms of the five retained variables (Supplementary material 1). Tested with an ANOVA analysis of variance, the final QR model was significantly different from the “null” model (F = 37, p < 0.001). When only simple terms were used, the model differed significantly from the final QR model (F = 15, p < 0.001); emphasizing both linear and quadratic relationships. The final model was not significantly different from the maximal model (F = 1.7, p = 0.15).

The model prediction map indicated high values in the south-eastern North Sea, with maximum abundance of 2.76 ind. m−3 located along the North Dutch, German and Danish coasts (Fig. 2). At Zeeland estuarine area and along East Anglia coasts intermediate abundances were predicted. Following the contour lines of the 70th quantile of predicted abundance on the potential habitat map, the prediction map was divided into 10 homogeneous areas. One area corresponding to the Dutch, German and Danish coasts was further divided into 5 areas to make the distinction between coast and offshore waters. As the main records of M. leidyi are from this area (Boersma et al. 2007; Faasse and Bayha 2006; Oliveira 2007), it was important to add this distinction to better understand the dispersal of individuals in these areas and the connectivity between areas. This characterization resulted in 14 polygons of unequal surface area varying from 2470 km2 in A1 to 85,693 km2 in A14, and separating low, moderate and high predicted abundance areas (Fig. 2; Table 2). The resulting areas were classified according to their averaged potential abundance as: (1) high potential overwintering areas A3 and A5–9; (2) intermediate areas A10, A11 and A13; (3) unfavourable areas A1–2, A12 and A14 (Table 2).
Fig. 2

Interpolated predicted probabilities of M. leidyi abundance in winter conditions resulted from the habitat model; the black lines divide the study domain into 14 areas (A1–A14) according to the predicted probabilities of M. leidyi abundance, following the 70th quantile of prediction

Table 2

Winter habitat suitability for M. leidyi in the 14 areas according to the habitat model predictions


Surface (km2)

Mean (ind. km−2)


Habitat suitability















High potential










High potential





High potential





High potential





High potential





High potential


























Surface = size of area, Mean = averaged potential abundance predicted by quantile regression model, SD = standard deviation of potential abundance within each area; habitat suitability for each area was defined considering Unfavourable an area having the Mean < 0.5 ind. km−2, Intermediate for Mean values between 0.5 and 1 ind. km−2 and High potential for Mean values >1 ind. km−2

Particle tracking

At the end of the passive particle tracking simulations, on the 1st of February 2008, particles present in each of the 14 areas previously defined (Fig. 2) were analysed according to their release area and released date (season) (Supplementary material 2). Particle density at each release event was homogenous horizontally over the entire domain of 0.08 ind. km−2.

Highest cumulated densities, reaching more than ten times the release density, were found in coastal areas A5 and A7, and in oceanic areas A10 and A11 (Fig. 3). In A5 this was mainly due to retention (particle originated from the same area A5) and inflow from adjacent areas A4, A6 and A9, while in A7 it was due to small contributions from many areas, adjacent as well as the farthest A1–A3, A12 and A13. Intermediate densities in the range 0.4–0.6 ind. km−2 characterised the majority of areas, among which the coastal A6 had high values mainly due to retention, and the oceanic A12 had high values due to adjacent areas contributions. Lower values of 0.2 ind. km−2 were found in A2 and A13, while the lowest densities were found in A8 and A9.
Fig. 3

Cumulated density of particles at 1st February 2008 (from the 12 simulations/release dates) in each arrival areas A1–A14 and their origin of release; legend shows the release areas A1–A14

The classification of areas on the origin of particles gave six clusters of similar areas (Figs. 4, 5). For the following description of results and discussion, two similar terms, retention and concentration, need to be clarified to avoid confusion. Retention defines particles retained in the release area, while concentration refers to aggregation of particles, whatever were the areas of origin of the particles.
Fig. 4

Relative contribution of particles on 1st February 2008 (from the 12 simulations/release dates) in each arrival areas A1–A14 according to their release area defined as identical if the arrival area was the same as the release area, adjacent if the particles come from neighbouring areas or distant if they come from any other areas

Fig. 5

Hierarchical cluster of arrival areas based on relative contribution of particles according to the three groups of particles origin (identical, adjacent and distant); map in the top right shows colour coding of the 14 areas according to the obtained clusters

The first cluster grouped the areas A1 and A14 and was dominated by retention with very low contribution from adjacent areas. No particle from the distant areas arrived here. In the second cluster, which grouped coastal areas A5 and A6, retention explained half of the present particles whereas the other half mostly came from adjacent areas. We note a very low inflow from the rest of the domain. The third cluster was composed of areas A2–A4 where retention dominated over inflow from adjacent areas. Areas A12 and A13 formed the fourth cluster in which retention decreased (<30 %) and the dominant contribution came from adjacent areas. The areas A9, A10 and A11, grouped into fifth cluster, were characterised by more than half of particles inflow from adjacent areas and the other half divided almost equally from identical and distant areas. The sixth cluster associated areas A7 and A8 which had the highest contribution from distant areas (approximately 50 %), moderate to low inflow of particles from adjacent areas and low retention rates for A8.

Considering the release season of the particles and the duration of their drift (until the 1st of February 2008), particles released in spring (March, April, May) had the longest drift and those released in winter (December, January, February) the shortest (except February 2007). The majority of arrival areas showed similar proportion of particles released in each of the four seasons (Fig. 6), except arrival areas A6, A8 and A9. Area A6 had higher number of particles released in winter, with the highest contribution of particles released in February 2007. Thus longer dispersal time of particles in this area indicated long-term retention. Areas A8 and A9 were dominated by particles released in autumn–winter, suggesting shorter dispersal time of particles, and leading to the idea of a fast transit in these areas.
Fig. 6

Relative contribution of particles at 1st February 2008 in each arrival areas A1–A14 according to their release dates, cumulated by season: spring (Mar–May), summer (Jun–Aug), autumn (Sep–Nov) and winter (Feb 2007, Dec 2007 and Jan 2008)


Mnemiopsis leidyi is described as a coastal and estuarine species, both in native and invasive seas. Most observations in European northern seas were made in coastal areas and were mainly recorded during spring and autumn which correspond to the reproductive seasons (Riisgård et al. 2007; van Ginderdeuren et al. 2012; Faasse and Bayha 2006). Unlike most surveys conducted in coastal waters, IBTS represented an opportunity to investigate the geographical limits of M. leidyi distribution in North Sea, considering offshore waters, and how these limits relate to the environment variability. During IBTS in January–February 2009–2011, the presence of alive ctenophore in offshore areas raised questions regarding overwintering potential in the North Sea, but mostly about the spatial and the temporal limits of its distribution.

Combining the Geographical Information Systems (GIS) with multivariate statistical tools (Hirzel et al. 2002), the quantile regressions model the upper bound of the species-environment relationship, estimating how the environment is limiting the distribution of a species (Vaz et al. 2008; Planque et al. 2007). The crucial factor controlling the winter distribution of M. leidyi in the North Sea was found to be the temperature. No cause-effect is implied by the model results, but rather a co-realisation of different factors and relationship types. The quadratic and negative nature of this relationship associated high abundance predictions in winter with low values of water temperature which characterise coastal areas and estuaries influenced by rivers runoff. Found at its lowest tolerance limits, M. leidyi has survived the cold North Sea winters so far at temperatures below 2 °C. Here after the use of term “overwintering” refers to the habitat characteristics that would permit ctenophore survival during winter environmental conditions when other threats endangering the life or existence of animals, e.g. predation, natural mortality, competition, starvation are not considered. Thus “survival” in the context of habitat model predictions refers strictly to limitation by environmental characteristics. Our results indicated potential overwintering areas along the south-eastern coasts of the North Sea (A3, A5–A9) where individuals would be able to survive the cold winters until spring warming provides again favourable conditions for population reproduction and growth. To establish a permanent population however, M. leidyi needs to reproduce which requires better environmental conditions than survival only. In its native areas, the reproduction of the ctenophore is known to be temperature dependent and restricted under 10 °C (Costello et al. 2006), resulting in blooms usually in early summer (Sullivan et al. 2001). Lehtiniemi et al. (2011) made the distinction between survival and reproduction of M. leidyi in the Baltic Sea considering threshold values of 3 and 10 °C, respectively. During IBTS surveys M. leidyi was observed at temperatures below 7 °C and only lobate individuals were collected, which confirmed that it was not the period of reproduction. Under current environmental conditions, M. leidyi may not be able to reproduce in large numbers in coastal and offshore waters of the North Sea, but this may change with predicted climate change scenarios (Van der Molen et al. 2014).

Besides temperature, habitat model formulation takes potential trophic relations into consideration, predicting positive association of M. leidyi with the scyphomedusa C. lamarkii, the ctenophore P. pileus and with Syngnathidae juveniles, again without inferring any causal relationship, rather the spatial co-occurrence. Syngnathidae and P. pileus are native zooplanktivores and can compete for food with M. leidyi (Mutlu et al. 1994) whereas C. lamarckii is known to be able to feed on M. leidyi (Riisgård et al. 2007).

Establishment of planktonic species in an area however, depends on additional features, among which oceanic physical processes may be important descriptors. The persistence of a population could be explained by retention, sustained by local productivity, or by constant inflow of individuals from potentially productive surroundings (Costello et al. 2012). A possible source population was suggested along the northern Dutch coast (A5) and in the German Bight (A6) by both habitat model and particle tracking. Here retention could be regarded as the controlling factor for local population sustainability (Costello et al. 2012). Individuals would survive the winter conditions and would be able to provide offspring to offshore areas in the following year. Winter survival in shallow areas was described for the ctenophore in Narragansett Bay, Rhode Island, followed by reproduction in spring and offspring supply to the open-bay population (Costello et al. 2006). Spread of offspring is an important factor influencing an invasive species’ spread rate, as reduced reproduction can lead to slower rates of geographical range expansion (Lockwood et al. 2013). However, the environmental conditions in which M. leidyi may reproduce and reach high densities are as narrow in the North Sea as they are in the Baltic Sea (Lehmann and Javidpour 2010). The same number of particles were released every month in our simulations, thus results showed dispersal possibilities without reflecting seasonal variability in population size. Using estimated reproduction rates for ctenophores along south-east coasts related to spring–autumn months to release particles could provide more realistic estimates of future geographical expansion. Then more realistic spread rate would be obtained, which rely on the scale of transport processes and duration.

In our simulations, transport of particles after 1 year resulted in highest densities offshore of the Frisian Islands (A5) and western Danish coastal areas (A7), where habitat modelling identified high potential overwintering areas. The western Danish coast was identified as concentration area as the high density of particles found here was mainly due to external inflow of particles. A different situation was encountered along Frisian Island (A5), where retention was the dominating process. Retention resulted in moderate particle density in German Bight (A6), a potential overwintering area according to habitat model. All coastal areas in the south-eastern North Sea (A2–A7) had high retention rates; yet the mechanism involved may be different. Retention offshore of Zeeland estuarine area (A3) may be related to estuarine shelter. High beaching rate occurred along the estuarine area which was not represented by the hydrodynamic model. ‘Particle beaching’ refers to particles transported out of the study area limit. Therefore high beaching rate in this area may be related to the estuarine shelter and thus possibly explain the ctenophores population recorded here (Faasse and Bayha 2006). Low resolution of the hydrodynamic model near coasts might have overlooked a self-sustainable population here, whereas the habitat model indicated a potential overwintering area. Results from a similar study in the Scheldt estuary concluded that the estuaries possess enough retention capability to keep an overwintering population, and enough exchange with coastal waters of the North Sea to seed offshore populations (Van der Molen et al. 2014). Another potentially important area for a sustained ctenophore population overlooked by our study is the Wadden Sea as it was not represented in the hydrodynamic model nor by the potential habitat map, even though M. leidyi presence was recorded here (Faasse and Bayha 2006; van Walraven et al. 2013).

More offshore, particles were concentrated over the Oyster Grounds area (A10–A11), where intermediate overwintering conditions were indicated by habitat model. Thus the moderate ctenophore abundance predicted in winter in this area could be compensated with transport of individuals in spring from retention-based areas. The real measure of reproductive population connectivity requires though an understanding of who is surviving and why (Cowen and Sponaugle 2009). The question remains of whether individuals concentrated here would find suitable conditions for reproduction or would be transported alive closer to the coast to further participate in reproduction?

The offshore areas A8 and A9 are viewed as passage corridor with lowest particles densities and mainly released during latest simulations, having short drift duration. Environmental winter conditions in these areas predicted high potential overwintering, but the northward flow through these areas suggested that the individuals would be transported out of the study area.

Low particle densities were found on Flemish Banks (A2) and Netherlands’ coasts south of Texel Island (A4) and low overwintering potential was predicted by the habitat model. A sustainable population here would be dependent on constant inflow of individuals from other areas, yet our simulations results do not support such a mechanism. Since observations of ctenophore exist on these coasts (Faasse and Bayha 2006), they may provide a seasonal habitat for estuarine populations but were not predicted by the habitat model to allow overwintering at sea. Dover Strait (A1) and offshore of eastern England, including the Dogger Bank (A14) had even less inflow of particles. This is logical, because the residual circulation is such that source areas for A1 and A14 are outside the area in which particles were released. On the contrary, along the East Anglian coast (A13) habitat model indicated that the wintery North Sea environmental conditions would permit overwintering and the currents would allow a limited level of retention (Fig. 5). However, up to now no M. leidyi individuals have been found on the English coast despite several plankton surveys (S. Pitois, Our results indicated the East Anglian coast as a potential area of invasion for M. leidyi, accessible by currents, suggesting two hypotheses: either this species has not yet arrived here, or there are other environmental variables impeding the development of M. leidyi in the area.

The present study investigated transport patterns of M. leidyi in the North Sea, emphasising connectivity patterns between presumably established populations. Individual exchanges were simplified in particle tracking simulations by representing passive transport processes without physiological processes (growth, reproduction, mortality) which are not known for M. leidyi in the North Sea. Using a particle tracking model independent of biological variables provided a general insight of dispersal patterns, and can be useful to evaluate the routes of any planktonic invader (Lehmann and Javidpour 2010). More knowledge from field and laboratory experiments describing physiological processes of the ctenophore in North Sea conditions, applied to simulations, would provide more species-specific results concerning population dynamics. Future investigation of explicit transport of individuals from sustainable populations should emphasize reproduction season and population dynamics for realistic population size predictions. In addition to the spatial dimensions inherent in transport, dispersal involves a survival probability, and thus food availability and predation are important (Pineda et al. 2007). Results of a parallel study, investigating M. leidyi bloom potential in the North Sea using, among others, adult and juvenile food fields from a biogeochemical model to predict M. leidyi survival and reproduction ability, showed broad agreement with the results presented here (Collingridge et al. 2014).

Absolute levels of particle exchanges are difficult to determine with Lagrangian models, when parameters like mortality are sensitive and give variable results, but still relative features can be addressed. The area of particle release seems to be a major determinant in transport success (Parada et al. 2003), even though ideally locomotion behaviour should be considered in dispersal models (Guizien et al. 2006). For M. leidyi no vertical migration behaviour has so far been reported, but water stratification, strong halocline (Kube et al. 2007; Schaber et al. 2011), surface sheer stress (Mianzan et al. 2010) or predator avoidance (Titelman et al. 2012) were noticed to influence its vertical distribution. In the Baltic Sea and American coasts agglomerations of ctenophores were observed either close to bottom, confine within surface mixed layer (Mianzan et al. 2010), or below the halocline (Kube et al. 2007; Schaber et al. 2011). Given that most of the southern North Sea is shallow and well mixed (Lelievre et al. 2012; Munk et al. 2009), no preferential depth was selected during dispersal simulations and particles were released in the entire water column. Thus, simulations focused on the horizontal distribution of particles and the importance of the release area. Caution should be taken when interpreting the results from simple advection models, as particle exchange rates may be overestimated, because models often fail to account for the decrease in particle concentrations resulting from mortality (Cowen et al. 2000). Including beached particles into the results may have overestimated retention and concentration rates, but eliminating them would have biased the connectivity between areas. Individuals may successfully arrive in a new area and die after a certain time, thus contributing to the overall transport. Low beaching rates were found offshore of the Frisian Islands, confirming the high retention rates. High beaching rates along French, Belgian, south Dutch and western Danish coasts could considerably reduce the abundances of living organisms previously estimated. However, estuaries encountered on these coasts were not resolved by the hydrodynamic model, so in reality beached particles could in fact represent surviving individuals within the estuarine areas.

A good prediction was obtained of the potential overwintering areas; still the prediction relied on the IBTS data which contained few presence observations of M. leidyi related to sample area (presence recorded in 44 stations of a total 347). Moreover, due to vessel constraints IBTS sampling area and thus the analysis and prediction did not include coastlines and estuaries. Including more observations, particularly along the south-eastern costal area, would improve our understanding of M. leidyi overwintering and its behavior close to the coastline.

The association with the lowest temperature values in which the ctenophore was recorded so far in the European Seas indicates how tolerant this species proves to be to the environmental characteristics, which should be considered as winter refuge areas rather than environmental preferences. Retention/concentration patterns identified from hydrodynamic simulations partly matched the distribution of predicted abundance of ctenophore in winter, which in turn reflect the real distribution found during 3 years of sampling. The current study does not allow sorting the importance of dispersal versus environmental conditions for driving M. leidyi observed distribution. Conversely they should be considered complementary as dispersal informs on the possibility for individuals to be retained in a particular area whereas environmental conditions affect ctenophore survival. Furthermore, the association of environmental features with abundance of ctenophore provides just a snapshot of a system in transition. Since it was suggested that M. leidyi might have been present in North Sea waters for several years and that only the recent increase in water temperature could have driven the development of large ctenophore swarms (Faasse and Bayha 2006), the limitations of M. leidyi population development in the current environmental conditions are believed to change in the future with predicted climate change scenarios (Van der Molen et al. 2014).

The novelty of this study relies on coupling two modelling methods. One identified areas where North Sea cold winters conditions would allow overwintering of M. leidyi and the other investigated dispersal of individuals between different areas. Habitat model predicted the presence of ctenophore related to environmental factors like water temperature and occurrence of other species; particle tracking models investigated transport patterns related to flow regimes. A retention-based, self-sustained M. leidyi population was identified along the northern Dutch coast and German Bight and a transport-based local population offshore from western Danish coast. The possibility of a local population in the Zeeland area was suggested, even though it was not proved by our study. This study provides the first estimates of the area of this invasive species in the cold winters in the North Sea. Our results implied that living ctenophores found in open waters in winter were transported from the southern coasts of the North Sea, thus the open water population densities depended on reproduction success in these areas. Based on the agreement of habitat and dispersal model results, we conclude that M. leidyi has become established along southeastern coasts of North Sea. To understand M. leidyi’s long term impact on the North Sea we need to predict population size fluctuations, thus an ecosystem approach is advisable, which includes, beside food availability, interaction with predators and competitors.



This work was partly funded by the EU under the InterReg IVa-2seas (MEMO project) program. Carmen David was further supported by IFREMER. The authors gratefully acknowledge the captain and the crew of the R/V Thalassa for help during IBTS sampling and the scientific staff that have participated to ctenophore collection on board.

Supplementary material

10530_2015_899_MOESM1_ESM.docx (182 kb)
Supplementary material 1 (DOCX 182 kb)
10530_2015_899_MOESM2_ESM.docx (19 kb)
Supplementary material 2 (DOCX 18 kb)


  1. Antajan E, Delesmont R, Dewarumez J-M, Lefebvre A, Luczak C, Warembourg C (2009) Surveillance écologique et halieutique du site de production électronucléaire de Gravelines Novembre 2007–Octobre 2008Google Scholar
  2. Antajan E, Bastian T, Raud T, Brylinski J-M, Hoffman S, Breton G, Cornille V, Delegrange A, Vincent D (2014a) The invasive ctenophore Mnemiopsis leidyi A. Agassiz, 1865 along the English Channel and the North Sea French coasts: another introduction pathway in northern European waters? Aquat Invasions 9(2):167–173CrossRefGoogle Scholar
  3. Antajan E, Loots C, Travers-Trolet M, Vaz S (2014b) Recurrent winter observations of Mnemiopsis leidyi swarms in the Southern North Sea. In Report of the joint ICES/CIESM workshop on Mnemiopsis science (JWMS), vol 14. ICES CM 2014/SSGHIEGoogle Scholar
  4. Bilio M, Niermann U (2004) Is the comb jelly really to blame for it all? Mnemiopsis leidyi and the ecological concerns about the Caspian Sea. Mar Ecol Prog Ser 269:173–183CrossRefGoogle Scholar
  5. Boersma M, Malzahn AM, Greve W, Javidpour J (2007) The first occurrence of the ctenophore Mnemiopsis leidyi in the North Sea. Helgoland Mar Res 61(2):153–155. doi: 10.1007/s10152-006-0055-2 CrossRefGoogle Scholar
  6. Brown J, Hill AE, Fernand L, Horsburgh KJ (1999) Observations of a seasonal jet-like circulation at the central North Sea cold pool margin. Estuar Coast Shelf Sci 48(3):343–355. doi: 10.1006/ecss.1999.0426 CrossRefGoogle Scholar
  7. Burchard H, Bolding K (2002) GETM: a general estuarine transport model; scientific documentation. European Commission, Joint Research Centre, Institute for Environment and SustainabilityGoogle Scholar
  8. Burchard H, Bolding K, Umlauf L (2012) General Estuarine transport model—source code and test case documentation. Baltic Sea Research Institute Warnemunde, Germany and Bolding & Burchard ApS, Asperup, DenmarkGoogle Scholar
  9. Chen K, Ying Z, Zhang H, Zhao L (2008) Analysis of least absolute deviation. Biometrika 95(1):107–122. doi: 10.1093/biomet/asm082 CrossRefGoogle Scholar
  10. Collingridge K, van der Molen J, Pitois S (2014) Modelling risk areas in the North Sea for blooms of the invasive comb jelly Mnemiopsis leidyi A. Agassiz, 1865. Aquat Invasions 9(1):21–36CrossRefGoogle Scholar
  11. Costello JH, Sullivan BK, Gifford DJ, Keuren DV, Sullivan LJ (2006) Seasonal refugia, shoreward thermal amplification, and metapopulation dynamics of the ctenophore Mnemiopsis leidyi in Narragansett Bay, Rhode Island. Limnol Oceanogr 51(4):1819–1831CrossRefGoogle Scholar
  12. Costello JH, Bayha KM, Mianzan HW, Shiganova TA, Purcell JE (2012) Transitions of Mnemiopsis leidyi (Ctenophora: Lobata) from a native to an exotic species: a review. Hydrobiologia 690(1):21–46. doi: 10.1007/s10750-012-1037-9 CrossRefGoogle Scholar
  13. Cowen RK, Sponaugle S (2009) Larval dispersal and marine population connectivity. Annu Rev Mar Sci 1:443–466. doi: 10.1146/annurev.marine.010908.163757 CrossRefGoogle Scholar
  14. Cowen RK, Lwiza KMM, Sponaugle S, Paris CB, Olson DB (2000) Connectivity of marine populations: open or closed? Science 287(5454):857–859. doi: 10.1126/science.287.5454.857 PubMedCrossRefGoogle Scholar
  15. Faasse MA, Bayha KM (2006) The ctenophore Mnemiopsis leidyi A. Agassiz 1865 in coastal waters of the Netherlands: an unrecognized invasion? Aquat Invasions 1(4):270–277. doi: 10.3391/ai.2006.1.4.10 CrossRefGoogle Scholar
  16. Finenko GA, Kideys AE, Anninsky BE, Shiganova TA, Roohi A, Tabari MR, Rostami H, Bagheri S (2006) Invasive ctenophore Mnemiopsis leidyi in the Caspian Sea: feeding, respiration, reproduction and predatory impact on the zooplankton community. Mar Ecol Prog Ser 314:171–185CrossRefGoogle Scholar
  17. GESAMP (IMO/FAO/UNESCO-IOC/WMO/WHO/IAEA/UN/UNEP Joint Group of Experts on the Scientific Aspects of Marine Environmental Protection) (1997) Opportunistic settlers and the problem of the ctenophore Mnemiopsis leidyi invasion in the Black Sea. Report and Studies GESAMP, vol 58Google Scholar
  18. Ghabooli S, Shiganova TA, Zhan AB, Cristescu ME, Eghtesadi-Araghi P, MacIsaac HJ (2011) Multiple introductions and invasion pathways for the invasive ctenophore Mnemiopsis leidyi in Eurasia. Biol Invasions 13(3):679–690. doi: 10.1007/s10530-010-9859-8 CrossRefGoogle Scholar
  19. Graham WM, Bayha KM (2007) Biological invasions by marine jellyfish. In: Nentwig W (ed) Ecological Studies, vol 193. Springer, Berlin, pp 239–255. doi: 10.1007/978-3-540-36920-2_14 Google Scholar
  20. Guizien K, Brochier T, Duchêne J-C, Koh B-S, Marsaleix P (2006) Dispersal of Owenia fusiformis larvae by wind-driven currents: turbulence, swimming behaviour and mortality in a three-dimensional stochastic model. Mar Ecol Prog Ser 311:47–66. doi: 10.3354/meps311047 CrossRefGoogle Scholar
  21. Hamer HH, Malzahn AM, Boersma M (2011) The invasive ctenophore Mnemiopsis leidyi: a threat to fish recruitment in the North Sea? J Plankton Res 33(1):137CrossRefGoogle Scholar
  22. Hirzel AH, Hausser J, Chessel D, Perrin N (2002) Ecological-niche factor analysis: how to compute habitat-suitability maps without absence data? Ecology 83(7):2027–2036CrossRefGoogle Scholar
  23. Hosia A, Titelman J (2011) Intraguild predation between the native North Sea jellyfish Cyanea capillata and the invasive ctenophore Mnemiopsis leidyi. J Plankton Res 33(3):535–540. doi: 10.1093/plankt/fbq106 CrossRefGoogle Scholar
  24. Ivanov VP, Kamakin AM, Ushivtzev VB, Shiganova T, Zhukova O, Aladin N, Wilson SI, Harbison GR, Dumont HJ (2000) Invasion of the Caspian Sea by the comb jellyfish Mnemiopsis leidyi (Ctenophora). Biol Invasions 2(3):255–258. doi: 10.1023/a:1010098624728 CrossRefGoogle Scholar
  25. Janas U, Zgrundo A (2007) First record of Mnemiopsis leidyi A. Agassiz, 1865 in the Gulf of Gdansk (southern Baltic Sea). Aquat Invasions 2(4):450–454CrossRefGoogle Scholar
  26. Javidpour J, Sommer U, Shiganova T (2006) First record of Mnemiopsis leidyi A. Agassiz 1865 in the Baltic Sea. Aquat Invasions 1(4):299–302CrossRefGoogle Scholar
  27. Kideys AE (2002) Fall and rise of the Black Sea ecosystem. Science 297(5586):1482–1484PubMedCrossRefGoogle Scholar
  28. Kideys AE, Kovalev AV, Shulman G, Gordina A, Bingel F (2000) A review of zooplankton investigations of the Black Sea over the last decade. J Mar Syst 24(3–4):355–371. doi: 10.1016/s0924-7963(99)00095-0 CrossRefGoogle Scholar
  29. Koenker R (2005) Quantile regression, economic society monographs. Cambridge University Press, CambridgeCrossRefGoogle Scholar
  30. Kremer P (1979) Predation by the ctenophore Mnemiopsis leidyi in Narragansett Bay, Rhode Island. Estuar Coast 2(2):97–105. doi: 10.2307/1351633 CrossRefGoogle Scholar
  31. Kremer P (1994) Patterns of abundance for Mnemiopsis in US coastal waters: a comparative overview. ICES J Mar Sci 51(4):347–354. doi: 10.1006/jmsc.1994.1036 CrossRefGoogle Scholar
  32. Kube S, Postel L, Honnef C, Augustin CB (2007) Mnemiopsis leidyi in the Baltic Sea: distribution and overwintering between autumn 2006 and spring 2007. Aquat Invasions 2(2):137–145. doi: 10.3391/ai.2007.2.2.9 CrossRefGoogle Scholar
  33. Legendre P, Legendre L (1998) Numerical ecology. Elsevier, AmsterdamGoogle Scholar
  34. Lehmann A, Javidpour J (2010) Potential pathways of invasion and dispersal of Mnemiopsis leidyi A. Agassiz 1865 in the Baltic Sea. Hydrobiologia 649(1):107–114. doi: 10.1007/s10750-010-0233-8 CrossRefGoogle Scholar
  35. Lehtiniemi M, Lehmann A, Javidpour J, Myrberg K (2011) Spreading and physico-biological reproduction limitations of the invasive American comb jelly Mnemiopsis leidyi in the Baltic Sea. Biol Invasions 14(2):341–354. doi: 10.1007/s10530-011-0066-z CrossRefGoogle Scholar
  36. Lelievre S, Vaz S, Fox C (2012) The effectiveness of the continuous underway fish egg sampler compared with standard vertical plankton hauls for sampling the eggs of demersal species in the southern North Sea. Fish Oceanogr 21(2–3):109–124CrossRefGoogle Scholar
  37. Lenhart H-J, Mills DK, Baretta-Bekker H, van Leeuwen SM, der Molen JV, Baretta JW, Blaas M, Desmit X, Kühn W, Lacroix G, Los HJ, Ménesguen A, Neves R, Proctor R, Ruardij P, Skogen MD, Vanhoutte-Brunier A, Villars MT, Wakelin SL (2010) Predicting the consequences of nutrient reduction on the eutrophication status of the North Sea. J Mar Syst 81(1–2):148–170. doi: 10.1016/j.jmarsys.2009.12.014 CrossRefGoogle Scholar
  38. Lett C, Verley P, Mullon C, Parada C, Brochier T, Penven P, Blanke B (2008) A Lagrangian tool for modelling ichthyoplankton dynamics. Environ Model Softw 23(9):1210–1214. doi: 10.1016/j.envsoft.2008.02.005 CrossRefGoogle Scholar
  39. Lockwood JL, Hoopes MF, Marchetti MP (2013) Invasion ecology, 2nd edn. Wiley, UKGoogle Scholar
  40. Mackinson S, Daskalov G (2007) An ecosystem model of the North Sea to support an ecosystem approach to fisheries management: description and parameterisation. Science Series Technical Report, Cefas Lowestoft, 142 (142):14–18Google Scholar
  41. Mianzan H, Martos P, Costello J, Guerrero R (2010) Avoidance of hydrodynamically mixed environments by Mnemiopsis leidyi (Ctenophora: Lobata) in open-sea populations from Patagonia, Argentina. Hydrobiologia 645(1):113–124. doi: 10.1007/s10750-010-0218-7 CrossRefGoogle Scholar
  42. Mills CE (1995) Medusae, siphonophores, and ctenophores as planktivorous predators in changing global ecosystems. ICES J Mar Sci 52(3–4):575–581CrossRefGoogle Scholar
  43. Mills CE (2001) Jellyfish blooms: are populations increasing globally in response to changing ocean conditions? Hydrobiologia 451(1):55–68. doi: 10.1023/a:1011888006302 CrossRefGoogle Scholar
  44. Monteleone DM, Duguay LE (1988) Laboratory studies of predation by the ctenophore Mnemiopsis leidyi on the early stages in the life history of the bay anchovy, Anchoa mitchilli. J Plankton Res 10(3):359–372CrossRefGoogle Scholar
  45. Munk P, Fox CJ, Bolle LJ, Van Damme CJG, Fossum P, Kraus G (2009) Spawning of North Sea fishes linked to hydrographic features. Fish Oceanogr 18(6):458–469. doi: 10.1111/j.1365-2419.2009.00525.x CrossRefGoogle Scholar
  46. Mutlu E, Bingel F, Gucu AC, Melnikov VV, Niermann U, Ostr NA, Zaika VE (1994) Distribution of the new invader Mnemiopsis sp. and the resident Aurelia aurita and Pleurobrachia pileus populations in the Black Sea in the years 1991–1993. ICES J Mar Sci 51(4):407–421CrossRefGoogle Scholar
  47. Oliveira OMP (2007) The presence of the ctenophore Mnemiopsis leidyi in the Oslofjorden and considerations on the initial invasion pathways to the North and Baltic Seas. Aquat Invasions 2(3):185–189. doi: 10.3391/ai.2007.2.3.5 CrossRefGoogle Scholar
  48. Parada C, Van Der Lingen CD, Mullon C, Penven P (2003) Modelling the effect of buoyancy on the transport of anchovy (Engraulis capensis) eggs from spawning to nursery grounds in the southern Benguela: an IBM approach. Fish Oceanogr 12(3):170–184. doi: 10.1046/j.1365-2419.2003.00235.x CrossRefGoogle Scholar
  49. Pineda J, Hare JA, Sponaungle S (2007) Larval transport and dispersal in the coastal ocean and consequences for population connectivity. Oceanography 20(3):22–39. doi: 10.5670/oceanog.2007.27 CrossRefGoogle Scholar
  50. Planque B, Bellier E, Lazure P (2007) Modelling potential spawning habitat of sardine (Sardina pilchardus) and anchovy (Engraulis encrasicolus) in the Bay of Biscay. Fish Oceanogr 16(1):16–30. doi: 10.1111/j.1365-2419.2006.00411.x CrossRefGoogle Scholar
  51. Purcell JE, Arai MN (2001) Interactions of pelagic cnidarians and ctenophores with fish: a review. Hydrobiologia 451(1–3):27–44CrossRefGoogle Scholar
  52. Purcell J, Cowan JJ (1995) Predation by the scyphomedusan Chrysaora quinquecirrha on Mnemiopsis leidyi ctenophores. Mar Ecol Prog Ser 129:63–70. doi: 10.3354/meps129063 CrossRefGoogle Scholar
  53. Purcell JE, Nemazie DA, Dorsey SE, Houde ED, Gamble J (1994) Predation mortality of bay anchovy Anchoa mitchilli eggs and larvae due to scyphomedusae and ctenophores in Chesapeake Bay. Mar Ecol Prog Ser 114:47–58CrossRefGoogle Scholar
  54. Purcell JE, Shiganova TA, Decker MB, Houde ED (2001) The ctenophore Mnemiopsis in native and exotic habitats: US estuaries versus the Black Sea basin. Hydrobiologia 451(1–3):145–176CrossRefGoogle Scholar
  55. R Core Team (2013) R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria.
  56. Reeve MR, Syms MA, Kremer P (1989) Growth dynamics of a ctenophore (Mnemiopsis) in relation to variable food supply. I. Carbon biomass, feeding, egg production, growth and assimilation efficiency. J Plankton Res 11(3):535–552. doi: 10.1093/plankt/11.3.535 CrossRefGoogle Scholar
  57. Riisgård HU, Bøttiger L, Madsen CV, Purcell JE (2007) Invasive ctenophore Mnemiopsis leidyi in Limfjorden (Denmark) in late summer 2007—assessment of abundance and predation effects. Aquat Invasions 2(4):395–401CrossRefGoogle Scholar
  58. Schaber M, Haslob H, Huwer B, Harjes A, Hinrichsen HH, Koster FW, Storr-Paulsen M, Schmidt JO, Voss R (2011) The invasive ctenophore Mnemiopsis leidyi in the central Baltic Sea: seasonal phenology and hydrographic influence on spatio-temporal distribution patterns. J Plankton Res 33(7):1053–1065. doi: 10.1093/plankt/fbq167 CrossRefGoogle Scholar
  59. Shiganova TA (1998) Invasion of the Black Sea by the ctenophore Mnemiopsis leidyi and recent changes in pelagic community structure. Fish Oceanogr 7(3–4):305–310CrossRefGoogle Scholar
  60. Shiganova TA, Bulgakova YV (2000) Effects of gelatinous plankton on Black Sea and Sea of Azov fish and their food resources. ICES J Mar Sci 57(3):641CrossRefGoogle Scholar
  61. Shiganova T, Malej A (2009) Native and non-native ctenophores in the Gulf of Trieste, Northern Adriatic Sea. J Plankton Res 31(1):61–71. doi: 10.1093/plankt/fbn102 CrossRefGoogle Scholar
  62. Shiganova TA, Mirzoyan ZA, Studenikina EA, Volovik SP, Siokou-Frangou I, Zervoudaki S, Christou ED, Skirta AY, Dumont HJ (2001) Population development of the invader ctenophore Mnemiopsis leidyi, in the Black Sea and in other seas of the Mediterranean basin. Mar Biol 139(3):431–445CrossRefGoogle Scholar
  63. Stahle L, Wold S (1989) Analysis of variance (ANOVA). Chemometr Intell Lab Syst 6(4):259–272. doi: 10.1016/0169-7439(89)80095-4 CrossRefGoogle Scholar
  64. Sullivan BK, Keuren DV, Clancy M (2001) Timing and size of blooms of the ctenophore Mnemiopsis leidyi in relation to temperature in Narragansett Bay, RI. Hydrobiologia 451(1):113–120. doi: 10.1023/a:1011848327684 CrossRefGoogle Scholar
  65. Titelman J, Hansson LJ, Nilsen T, Colin SP, Costello JH (2012) Predator-induced vertical behavior of a ctenophore. Hydrobiologia 690(1):181–187. doi: 10.1007/s10750-012-1056-6 CrossRefGoogle Scholar
  66. Van der Molen J, Collingridge K, Pitois S, Vansteenbrugge V, Hostens K, Van Walraven L, Langenberg V, van der Veer HW (2014) Modelling aspects of Mnemiopsis leidyi in the North Sea. In: Report of the joint ICES/CIESM workshop on Mnemiopsis science (JWMS), vol 14. ICES CM 2014/SSGHIEGoogle Scholar
  67. van Ginderdeuren K, Hostens K, Hoffman S, Vansteenbrugge L, Soenen K, De Blauwe H, Robbens J, Vincx M (2012) Distribution of the invasive ctenophore Mnemiopsis leidyi in the Belgian part of the North Sea. Aquat Invasions 7(2):163–169CrossRefGoogle Scholar
  68. van Walraven L, Langenberg VT, van der Veer HW (2013) Seasonal occurrence of the invasive ctenophore Mnemiopsis leidyi in the western Dutch Wadden Sea. J Sea Res 82:86–92. doi: 10.1016/j.seares.2013.02.003 CrossRefGoogle Scholar
  69. Vaz S, Martin CS, Eastwood PD, Ernande B, Carpentier A, Meaden GJ, Coppin F (2008) Modelling species distributions using regression quantiles. J Appl Ecol 45(1):204–217. doi: 10.1111/j.1365-2664.2007.01392.x CrossRefGoogle Scholar

Copyright information

© Springer International Publishing Switzerland 2015

Authors and Affiliations

  • Carmen David
    • 1
    • 2
  • Sandrine Vaz
    • 1
  • Christophe Loots
    • 1
  • Elvire Antajan
    • 1
  • Johan van der Molen
    • 3
  • Morgane Travers-Trolet
    • 1
  1. 1.French Research Institute for Exploration of the Sea (IFREMER)Boulogne-sur-MerFrance
  2. 2.Alfred Wegener Institut Helmholtz-Zentrum fur Polar- und MeeresforschungBremerhavenGermany
  3. 3.Lowestoft LaboratoryCentre for Environment, Fisheries, and Aquaculture Science (CEFAS)LowestoftUK

Personalised recommendations