Abundance, horizontal and vertical distribution of the invasive ctenophore Mnemiopsis leidyi in the central Balitc Sea, November 2007

The distribution and abundance of the invasive ctenophore Mnemiopsis leidyi in the Bornholm Basin, an important spawning ground of several fish stocks, and in adjacent areas in the central Baltic Sea was studied in November 2007. The study showed that M. leidyi were relatively small (body length 18.6 ± 7.6 mm) and they were patchily distributed over a large part of the investigated area. Specimens were found on 68 and 59% of stations sampled with a Bongo net (n=39) and an Isaac-Kidd midwater trawl (n=51), respectively. Vertically, the highest densities of M. leidyi occurred at 40 to 60 m around the halocline. Horizontally, the highest abundances were found north and west of Bornholm, but relatively high densities were also observed in the Slupsk Furrow. The mean abundance was 1.58 ± 2.12 ind. m -2 , the peak abundance was 8.92 ind. m -2 , and the average and peak population density were 0.03 ± 0.05 and 0.28 ind. m -3 , respectively. The abundances are low compared to densities recently observed in other areas of the Baltic region (e. g. Limfjorden, Åland Sea) and the estimated predation impact on zooplankton by M. leidyi was negligible in November 2007. However, because of the ctenophore’s wide distribution in the central Baltic Sea, its ability for rapid population growth, and its potential influence on fish stocks by competing for food and by preying on fish eggs and newly hatched larvae, close monitoring of the future development of M. leidyi in the Baltic Sea is strongly recommended.


Introduction
Several studies have recently described occurrences of the invasive lobate ctenophore Mnemiopsis leidyi A. Agassiz, 1865 in the Baltic Sea (Javidpour et al. 2006;Haslob et al. 2007;Janas and Zgrundo 2007;Kube et al. 2007;Lehtiniemi et al. 2007;Tendal et al. 2007). The ctenophore was accidentally introduced with cargo ballast water into the Black Sea in the early 1980s, where its explosive population growth coincided with a breakdown of, in particular, the anchovy fishery (Vinogradov et al. 1989;Mutlu 1999;Kideys 2002). Subsequently, it spread into adjacent waters of the Mediterranean and to the Caspian Sea (Kideys and Niermann 1993;Shiganova 1993;Shiganova et al. 2001). In autumn 2006 M. leidyi was observed, nearly simultaneously, at several locations in northern Europe, including the western Baltic Sea (Faasse and Bayha 2006;Javidpour et al. 2006;Boersma et al. 2007;Oliveira 2007;Tendal et al. 2007). Early in 2007, observations of M. leidyi in Danish waters and the Baltic Sea commenced to become numerous, and in many cases there were indications of mass occurrences (Tendal et al. 2007;Lehtiniemi et al. 2007). M. leidyi has been documented to overwinter in low abundances (1-4 ind. m -3 ) in the southern Baltic Sea at temperatures below 10 ºC (Kube et al. 2007). This has given the ctenophore an opportunity to extend its distribution range from the southwestern to the central Baltic Sea between autumn 2006 and spring 2007 (Kube et al. 2007). In August 2007, very high densities were registered in the central parts of a shallow Danish fjord system (Limfjorden) where some localities reached values above 200 ind. m -3 , and in one case (Skive Fjord) the density was measured to be 867 ± 121 ind. m -3 (Riisgård et al. 2007). These values considerably exceeded abundances reported from the northern Baltic Sea during spring and summer 2007 (Lehtiniemi et al. 2007) and even most values reported from the Black Sea during the period when the zooplankton and fish stocks collapsed in this area (Purcell et al. 2001b;Shiganova et al. 2001).
The strong focus on the mainly carnivorous M. leidyi is due to its predation impact on zooplankton and ichthyoplankton which may be detrimental to fish populations (Purcell et al. 2001b;Kideys 2002). Furthermore, M. leidyi has a great invasive potential due to its ability of rapid growth and reproduction, and its high tolerance of broad ranges of temperatures and salinities (Kremer 1994;Shiganova 1998;Purcell et al. 2001b). Individual growth rates and egg production in M. leidyi varies with size, food availability and temperature (Kremer 1994;Purcell et al. 2001b) and Kremer (1994) concluded that three factors act in a hierarchy to determine the abundance of M. leidyi: temperature, food availability and mortality (predation), in decreasing order of importance.
Most of the recent studies on M. leidyi in the Baltic Sea are restricted to coastal areas and report on independent, purely qualitative catches or sightings obtained by various methods on different dates over a broad time scale. Exceptions are the studies by Lehtiniemi et al. (2007) who described abundances of M. leidyi in the Gulf of Finland, the Åland Sea and the Gulf of Bothnia, and by Haslob et al (2007) who investigated M. leidyi abundances on a station grid in the Bornholm Basin. The Bornholm Basin is of great importance for several fish species in the Baltic Sea, especially for the eastern Baltic cod Gadus morhua Linnaeus, 1758, as this area presently represents the only spawning ground that supports successful reproduction of this stock Köster et al. 2005). Haslob et al. (2007) observed that the vertical distribution of M. leidyi showed an overlap with the water layers where cod eggs and, to a lesser degree, sprat eggs are neutrally buoyant. They hypothesized that this overlap may result in reduced recruitment success of these fish stocks due to predation by M. leidyi on fish eggs, as well as by competition for zooplankton food, influencing the survival success of fish larvae.
The present study investigated the horizontal and vertical distribution as well as abundances of M. leidyi in the Bornholm Basin and surrounding areas in November 2007. To our knowledge, this study represents the first extensive and systematic investigation on the distribution of M. leidyi in the central Baltic Sea on such a broad scale.

Materials and Methods
Samples of M. leidyi were collected during a cruise with the Danish research vessel 'DANA' from November 1st to 18th, 2007. The investigation covered an area from approximately 14.15° E to 18.15° E and 54.30° N to 56.15° N. The sampling was conducted with three different gears to obtain a coverage of the horizontal and vertical distribution of M. leidyi: an Isaac-Kidd midwater trawl (IKMT), a Multinet (5 nets), and a Bongo net. Corresponding ambient hydrographic conditions (salinity, temperature, oxygen concentration) were recorded at the sampling stations with a CTD (Sea-Bird plus).
Sampling with the Isaac-Kidd midwater trawl (mesh size 4 mm in the cod end) was conducted to investigate the horizontal and vertical distribution of M. leidyi. Hauls were taken on a regularly-spaced station grid with a distance of approximately 13.5 nautical miles between sampling stations. The duration of a haul was 30 min, and the gear was equipped with a flowmeter to measure the volume of filtered water. On stations with a bottom depth shallower than 35 m, one haul was made with the gear undulating between the surface and 5 m above the sea floor.

115
On stations deeper than 35 m, a two-layer sampling strategy was adopted to investigate the vertical distribution of M. leidyi. Thus, two hauls were taken; during the first haul, the gear was undulating between 5 and 25 m above the sea floor, while the second haul covered the remaining part of the water column, undulating between 25 m above the bottom and the surface. The decision of conducting two hauls at stations deeper than 35 m was based on the depth range at which the halocline was found in the study area (35-50 m), as previous studies by Haslob et al. (2007) and Kube et al. (2007) found M. leidyi exclusively below the halocline. Altogether, 91 IKMT hauls were taken on 51 stations.
To get a finer resolution of the vertical distribution of M. leidyi than obtained with the two-layer sampling strategy using the IKMT, hauls with an opening/closing Multinet (mesh size 335 μm, aperture 0.25 m 2 ) were taken on three stations. The vertical distribution of M. leidyi was resolved in steps of 5 or 10 m. The gear was towed, undulating in each depth range for approximately 5 to 6 min.
To assess abundances of M. leidyi, hauls with a Bongo net were taken on a total of 39 stations. The Bongo net (60 cm mouth diameter) was equipped with 500 μm mesh size nets, cod ends with 335 µm mesh size, and a flowmeter. This gear was towed in a double oblique haul integrating the entire water column, from 5 m above the bottom to the surface. The numbers of specimens from the two nets were combined. Counts of M. leidyi were standardised to individuals per m 2 by accounting for the filtered water volume and the maximum depth of the tow.
Specimens of M. leidyi were sorted from the samples, counted and the body oral-aboral length of each individual was measured to the nearest 1.0 mm with the aid of a sliding caliper. All samples were analysed within 30 min after capture. Individuals caught with Bongo and Multinet were generally less damaged than individuals captured with the IKMT, probably due to the different mesh sizes and haul durations.
A comparison of M. leidyi length distributions caught with the IKMT and with the Bongo net reflected that the catchability for smaller individuals (<10 mm) was lower with the IKMT (Figure 1), due to the relatively large mesh size of this gear. Generally, the catchability of the IKMT gear was smaller than the catchability of the Bongo net. However, we assume that the catchability of the IKMT is equal amongst hauls, and that IKMT stations are comparable with each other. A direct calculation of abundances per filtered volume from IKMT catches would largely underestimate true abundances of ctenophores. Assuming that the Bongo net has a catchability of 1, we compared Bongo catches to IKMT catches at those stations that were sampled with both gears. Only individuals >10 mm were considered in this analysis. One outlier was excluded (Bongo n=61, IKMT n=3), which may be due to a small scale patchiness in the distribution of M. leidyi. The analysis resulted in a significant linear relation between IKMT and Bongo catches (P<0.001, r 2 =0.72), and we used this relationship to convert IKMT data to abundance estimates (ind. m -2 ).

Results
While the size distributions of M. leidyi in the catches using the Multinet and the Bongo net were similar (cod end mesh size was 335 µm in both gears), the size distribution of M. leidyi retained in the IKMT and Bongo nets differed due to the different mesh sizes. M. leidyi between 6-59 mm (average 18.6 mm ± 7.6 SD) and 8-43 mm (average 21 mm ± 5.7 SD) were retained in the Bongo net and the IKMT, respectively ( Figure 1). Although the IKMT was able to catch some M. leidyi <10 mm, the catchability of this gear for smaller specimens was lower in comparison to the Bongo net. Thus, for the comparison of IKMT catches to Bongo net catches, only specimens >10 mm were considered.
The IKMT catches showed highest numbers of M. leidyi to occur in the deep hauls ( Figure 2) where 70% of specimens were caught, whereas 30% were caught in the shallow hauls. A peak in the deep-haul density of M. leidyi was observed in the area north of Bornholm (Figure 2b), The results from the three Multinet stations support the observation from the IKMT catches that highest densities of M. leidyi mainly occurred at larger depths ( Figure 3). In compari-son with the CTD data, the vertical distribution of Multinet catches shows that the highest M. leidyi densities were found close to the halocline, and at two stations (A and C) M. leidyi was not observed above 35 m, but it was found at low oxygen levels down to 0.29 ml l -1 .
Generally, M. leidyi was widely distributed in the study area being present in 68% of the Bongo net stations and 59% of the IKMT stations (Figures 2 and 4). The horizontal distribution of M. leidyi sampled with the Bongo net revealed highest abundances north and west of Bornholm, and relatively high densities were observed in the Slupsk Furrow which connects the Bornholm Basin to the Gdansk Deep further to the east ( Figure 4). However, no M. leidyi were observed in the southwestern part of the investigation area and in the northeastern part, in the region around the Middelbanke with very shallow water. The maximum abundance of M. leidyi caught at the Bongo net stations was 8.92 ind. m -2 (0.28 ind. m -3 ), with mean densities of 1.58 ± 2.12 ind. m -2 (0.03 ± 0.05 ind. m -3 ).
The horizontal distribution and abundance estimates obtained from the IKMT catches ( Figure 5) are largely in concordance with the results from the Bongo net samplings. Again, highest abundances were observed in the area north of Bornholm. However, abundances in the Slupsk Furrow were somewhat lower than observed by the Bongo net sampling, while the IKMT sampling suggested a second, smaller abundance peak at the southeastern edge of the Bornholm Basin, an area that was not covered by the Bongo net sampling. The observation that no M. leidyi occurred in the southwestern and northeastern parts of the investigated area was also confirmed by the IKMT sampling. However, the maximum values of abundance estimates from the IKMT sampling were lower compared to the Bongo net sampling (maximum 4.42 ind. m -2 for the IKMT and 8.92 ind. m -2 for the Bongo net sampling). It should be kept in mind, though, that the abundances obtained by means of the IKMT samplings only included ctenophores >10 mm body length.
Using equations given by Riisgård et al. (2007), the population filtration rate of M. leidyi with mean body length (18.8 mm) and peak density (0.28 ind. m -3 ) found in the present study was estimated to be F pop = 2.3×10 -3 m 3 d -1 , and subsequently the half-time of zooplankton (copepods) was found to be t 1/2 = 300 d.

Hydrographic influence on distribution
Temperature appears to be of overriding importance in determining conditions suitable for population increase and expansion of M. leidyi (Kremer 1994;Mutlu 1999;Purcell et al. 2001b;Shiganova et al. 2001;Purcell and Decker 2005). In its native environment, reproduction occurs from spring through autumn at temperatures ≥12 °C, reproduction peaks at temperatures of 24-28 °C (Purcell et al. 2001b), and in the Sea of Azov Shiganova et al. (1998,2001) did not find any specimens when the temperature dropped below 4 °C. Lehtiniemi et al. (2007) found the highest densities of M. leidyi eggs (90 eggs m -3 ) around the halocline at 80-60 m depth where the temperature ranged between 4.5-5 ºC. This indicates that although the optimal spawning temperature for M. leidyi in the Black Sea or in its native environment is high, M. leidyi may have adopted to the Baltic Sea environment and spawn now at lower temperatures. Besides, higher abundances of M. leidyi than previously recorded in the Black Sea and the Caspian Sea have been registered in a shallow Danish fjord system, at relatively low temperatures (11-19 °C) (Riisgård et al. 2007).
M. leidyi does not seem to be limited by salinities >2 PSU, but low winter temperatures, particularly in combination with low salinities, may prevent survival throughout the winter in some locations, e.g. the Sea of Azov (Purcell et al. 2001b). Data by Kube et al. (2007) show that M. leidyi occurred in the central Baltic Sea during the winter months below the halocline where temperatures do not drop below 4°C. Therefore, it seems reasonable to suggest that M. leidyi may use the Bornholm Basin as a refuge during winter, and to expand its distributional range in spring when temperature and food availability are increasing. It has been shown that M. leidyi also overwinters in the Gulf of Finland (Lehtiniemi et al. 2007), and it is therefore possible that other areas in addition to the Bornholm Basin act as winter refuges. Lehtiniemi et al. (2007) suggested that the distribution and further expansion of M. leidyi in the northern Baltic Sea may be restricted by low salinity or low oxygen concentrations as they did not find the ctenophore at salinities <5 PSU, or on localities where oxygen was depleted at water depths below 60-70 m. In the present study, we found individuals of M. leidyi at oxygen levels down to 0.29 ml l -1 which is in accordance with laboratory experiments showing 100% survival in a 96 h incubation at an oxygen concentration of 0.5 ml l -1 (Purcell et al. 2001a). This suggests that in the Baltic Sea, salinity may exert a stronger influence than oxygen on the expansion and distribution of the ctenophore.
In its native environment, M. leidyi encounters wide ranges of temperatures (2-32 °C) and salinities (2-38 PSU) (Purcell et al. 2001b). Nevertheless, both ctenophore size and sizespecific egg production were less at low temperature (9 °C) and salinity (7 PSU) than at warmer temperatures and higher salinities tested in the laboratory (Purcell and Decker 2005). Thus, even though M. leidyi has been reported to tolerate low salinities, it probably has a preference for higher salinities. Accordingly, it can not be precluded that in the Baltic Sea, M. leidyi is actively or passively (via drift) distributing to areas or depths with higher salinities and/or temperatures where growth and reproduction are optimal. This possibility receives some support from the depth-stratified IKMT and Multinet samplings which showed a clear peak in abundances of M. leidyi in the deeper water layers of the Bornholm Basin where higher salinities are found. However, in contrast to the studies by Haslob et al. (2007) and Kube et al. (2007) who found M. leidyi exclusively below the halocline, the stratified samplings in the present study showed that in November 2007, the ctenophore was also distributed above the halocline, and in shallower, less saline coastal areas. The preference of the ctenophore for high saline water may possibly reflect an active search, and if this is true, an outbreak of M. leidyi in the central Baltic Sea may hitherto have been restricted by low salinities. This hypothesis receives further support by findings of Riisgård et al. (2007), who observed high densities of M. leidyi at salinities of 27-28 PSU.

Densities in the Bornholm Basin
In May 2007, Haslob et al. (2007)  Compared to these high figures, the levels registered in the Bornholm Basin in November 2007 seem modest. However, as shown by Lehtiniemi et al. (2007), abundances may increase remarkably within a few weeks in some areas, which confirms the great potential of M. leidyi to reproduce and reach outbreak levels within a short time.

Potential influence of the Baltic sprat stock on M. leidyi population dynamics
It is striking that high abundances of M. leidyi have been observed in areas west and north of the central Baltic (Riisgård et al. 2007;Lehtiniemi et al. 2007), while the abundance in the Bornholm Basin and the surrounding areas is low. An explanation for this could be competition for food (zooplankton) caused by the currently very large stock of sprat Sprattus sprattus (Linnaeus, 1758). Among other possible factors, Purcell et al. (2001b) identified low competition for food by heavily overfished zooplanktivorous fish stocks to be one of the main reasons for the population explosion of M. leidyi in the Black Sea. In the Baltic, however, the sprat stock has increased considerably since the late 1980s and has been on a high level in recent years (Figure 6, Möllmann and Köster 2002). Möllmann et al. (2005) have shown a decrease in Baltic sprat condition due to strong intra-specific competition for zooplankton at large stock size. 1975 1980 1985 1990 1995 2000 2005 Total biomass (Ч1000 t) Accordingly, sprat may act as an important competitor for zooplankton in the central Baltic Sea, especially as sprat are migrating into the Bornholm Basin during spring and early summer for spawning (Aro 1989), i.e. periods likely to provide the best conditions for M. leidyi development. In addition to that, feeding sprat concentrate at depths between 50 m and the oxygen depleted deep water (Köster and Schnack 1994), i.e. the same depth range where the highest abundances of ctenophores have been found (Figure 3, Haslob et al. 2007).
Thus, the competition for food exerted by the presently large Baltic sprat stock may so far have prevented the population of M. leidyi to reach outbreak levels in the central Baltic Sea.

Predation impact of M. leidyi on zooplankton and cod eggs in the central Baltic Sea
Based on the calculations of population filtration rate and the half-time of zooplankton (t 1/2 = 300 d), the predation impact of M. leidyi on zooplankton in the central Baltic Sea in November 2007 is regarded negligible. Nevertheless, it may become significant (t 1/2 = 20-30 d) if the M. leidyi population density increases 10 to 15 times which may be realistic during the productive period. In this regard, the fact that M. leidyi was found around the entire Bornholm Basin in November 2007 is of major concern. Besides, high abundances were reported from fjord areas in Denmark (Riisgård et al. 2007) and from the northern Baltic Sea (Lehtiniemi et al. 2007). It is possible that these areas will act as donor areas, allowing M. leidyi to spread into the important fish spawning ground in the Basin centre. Such an expansion from coastal to offshore waters was observed in the Black Sea. Here abundance and biomass of M. leidyi was initially highest in inshore areas, which warm up earlier and/or are subjected to higher eutrophication. Subsequent increases of temperatures in offshore waters caused the expansion of populations into the central parts of the Black Sea (Mutlu 1999).
Despite the low present predation impact by M. leidyi on zooplankton, the ctenophore may exert a significant predatory influence on Baltic cod eggs at relatively low abundances due to an overlap in their vertical distribution in the Bornholm Basin. Haslob et al. (2007) observed highest densities of M. leidyi in the Bornholm Basin at the same depth range where cod eggs are found, a result that was corroborated by the present study.
The recruitment success of eastern Baltic cod is strongly dependent on increased egg survival due to inflows of high saline, oxygen rich water from the North Sea ). As discussed above, an outbreak of M. leidyi in the central Baltic Sea could so far have been restricted by low salinities. Thus, higher survival of cod eggs and larvae after inflow events may be counteracted by an intensified predation impact of M. leidyi whose abundances may simultaneously increase due to rising salinities.

Conclusion and recommendation
The abundances of M. leidyi in the central Baltic Sea observed in the present study are low compared to the mass occurrences recently reported from neighboring areas. However, the fact that M. leidyi is now spread over vast areas of the central Baltic Sea is of major concern due to the ctenophore's ability to survive through the winter in the Baltic Sea, its high potential for explosive population development and its possible negative influences on fish stocks. Therefore, close monitoring of the future development of M. leidyi in the Baltic Sea is strongly recommended.