Combining data from different sampling methods to study the development of an alien crab Chionoecetes opilio invasion in the remote and pristine Arctic Kara Sea

Size composition revealed by different sampling methods

Different methods of collection yielded diverse size distribution of crabs. Although the size composition of adult male and female snow crabs usually differs, we discuss their aggregate composition in order to compare the data from trawling with the video data, for which no sex differentiation is possible. The carapace width (CW) of crabs caught during SIO trawling ranged from 4 to 117 mm (Fig. 2A). Mixture analysis of CWs identified 9 distinct size groups from the bulk of SIO crabs (7 groups from the analysis and 2 were added manually to decrease noise during the analysis) (Table 1). The majority of crabs were of small size, with CW mode at 14 mm (while the mean was 16 mm) and another abundant group at CW 10 mm.

The mixture analysis of crabs caught by MMBI trawling resulted in only 5 size groups (1 of which was added manually) (Table 1). Overall, crabs caught by this method were of larger size, with the minimum carapace width of 22 mm and maximum of 120 mm. A large portion of crabs was within the 52 mm size group (mean was 57 mm) (Fig. 2B). The analysis merged smaller crabs into one group with large standard deviation (36 ± 7 mm). However, the mixture analysis identified more distinct large size groups (over 52 mm) in MMBI samples, than from SIO trawling and video sampling.

The smallest crab identified on the video had seven mm CW and the largest 127 mm. Mixture analysis yielded 4 size groups for this method (Table 1). Most frequently crabs were within two broad size groups with CW 15 ± 3 mm and 46 ± 11 mm (Fig. 2C). The mode CW was 15 mm, however for this method the mean CW differed at 32 mm.

The boxes in Fig. 2 (interquartile range containing 50% of the data) of Blagopoluchiya Bay are very similar for both methods, and include CW below 20 mm. The mode, median and means for the trawling and video samplings in Blagopoluchiya Bay are also very similar (Table 2). However, the maximum size observed in video (50 mm) was larger than in the Sigsbee trawling sample (38 mm), while the minimum was very similar (7 and eight mm respectively). Both methods yielded large number of crabs: 735 by trawling and 388 caught on video. The SIO trawling sample revealed more distinct and sharp size groups than the video sample (Fig. 3). However, the mixture analysis identified only one additional size group in the SIO trawling sample (Table 1).

The two methods had similar results for the south of the Kara Gates Strait (Gates 2, Fig. 4; Table 2). The central 50% of crabs had CW between 43 and 57 mm. Trawl sampling collected 20 crabs, and we observed 186 crabs on the video. While the minimum CW of crabs from trawling and video were very similar, 27 and 22 mm respectively, the video detected much larger crabs (127 mm) than the trawling (71 mm). However, we caught large crabs (up to 117 mm) while trawling in the nearby location (Gates 1, 2, Figs. 1, 4 Table 2).

We detected a very low number of crabs in Haug Bay using both methods (4 and 7 respectively). However, their sizes were very different. In trawl sample all crabs were 15–16 mm; in the video sample the sizes ranged from 27 to 50 mm, while the majority was within 42–47 mm.

The most distinct difference was observed in Abrosimov Bay (Figs. 4A, 4B Table 2). A large proportion of the trawl catch were small crabs (minimum and mode at four mm CW), and the central half of crabs were between 4 and 32 mm. On the video we could observe only larger crabs (minimum nine mm), and the central half of the crabs were much larger, between 34 and 48 mm. The largest crabs in both methods were very similar (50 mm trawling, 58 video). We measured twice as many crabs (88) on the video than in the trawling sample (43).

Sedov and Tsivolka Bays were sampled using different methods and cannot be compared directly. Trawling sample in Sedov Bay brought 28 crabs with 45 mm maximum and 10 mm minimum CW. On the Tsivolka Bay video we identified 215 crabs from 9 to 58 mm CW.

At the 53 MMBI trawling stations the minimum CW ranged from 22 to 72, and the maximum from 39 to 120. The modes and means ranged from 32 to 94 and 34 to 75 respectively. Most often, the minimum CW was 47 mm, maximum was 57 mm; mode and mean were 52 mm. Between 2 and 61 crabs were caught at the MMBI sampling stations, most commonly 3 crabs per station, but 12 crabs on average.

At the 53 MMBI trawling stations, the maximum, mode and mean CW sizes only weakly correlated with the depth (Table 3). In both SIO trawling and video samples the maximum CW sizes had strong correlation with the depth, although the sample size was very small.

When we mapped the maximum CW size for each station, we observed a trend in the MMBI samples, where larger crabs tend to be found in the south and northwards along the Yamal Peninsula. Towards the center of the western Kara Sea the maximum CW of the crabs decreases (Fig. 5). Similar mapping of ovigerous female findings did not show any observable trends.

Sex ratio and sex related differences in size composition

Trawling allowed identifying and comparing crabs of different sexes. The male to female ratios were strikingly different between SIO (bays of the Novaya Zemlya Archipelago) and MMBI (west of the Yamal Peninsula) (Fig. 1): 0.8 and 3.8 respectively. There were only 8 ovigerous and 366 non ovigerous females in the SIO samples (ratio 0.02), while there were 72 ovigerous and only 67 non ovigerous females in the MMBI samples (ratio 1.07).

The central half of the CW size distribution of crabs differed between samples for all sexes, except for ovigerous females (Fig. 6). Ovigerous females were within a narrow size range of 44 and 58 mm in the SIO samples, and between 42 and 72 in the MMBI samples (Table 4). The non ovigerous females from SIO trawling samples ranged from 11 to 47 mm. The MMBI non ovigerous females were larger and ranged from 22 to 63 mm. Overall, male sizes also differed between these two sampling methods and area of collection. The CW of 293 crabs caught by SIO Sigsbee trawl ranged from 11 to 117 mm, and of the 523 crabs caught by MMBI large trawl ranged from 23 to 120 mm. Small mesh in the Sigsbee trawl allowed us to collect 190 juvenile crabs (less than 11 mm).

Abundance estimation

Population density of crabs was calculated from the video transects (Table 5). Overall 3,132 frames were analyzed, resulting in 3,776 m² of bottom inspected. The maximum density was observed near the Kara Gate Strait and in Blagopoluchiya Bay (0.87 and 0.55 crabs/m² respectively). In the other bays (located between the Kara Gate Strait and Blagopoluchiya Bay) the population density of crabs was several times or an order of magnitude lower, reaching the minimum value of 0.01 crabs/m² in Haug Bay (Table 5).

Advantages and disadvantages of the applied methods to study the snow crab population in the Kara Sea

The three methods discussed here revealed different aspects of the Chionoecetes opilio population size structure in the Kara Sea. The Sigsbee trawl used by SIO has small mesh and catches crabs as small as four mm CW, which is the size of recently settled crabs (Conan et al., 1996). However, it also has a small opening, and some large and agile crabs can escape. The video recording of the same area shows that large crabs are present, although not always caught. The large bottom trawl is able to catch large crabs (22–120 mm CW), but does not retain younger crabs, due to its large mesh. We do however know that at least at some of the MMBI stations juvenile crabs were present. In some cases a similar to the SIO Sigsbee trawl was used, but the data was not dully recorded, and is thus omitted from the results. The combined use of this trawling gear could provide the full picture of the existing size groups in a population.

It is easy to observe larger crabs on the video, although some smaller crabs can also be spotted (up to seven mm CW) (Fig. 7A). The Video Module floats over the bottom with very little impact. Due to the muddy sediments in the studied area, every sudden movement of large agile organisms (crabs, fish) creates a cloud and can easily be spotted on the video. In all of the recorded footage, there were very few cases of such clouds: in most of them it was a fish, and sometimes crabs would run forward, and stop, therefore still recorded by us (Fig. 7C). It is safe to say that larger crabs (CW 30 mm and above) are quantitatively recorded on the video. However, crabs smaller than approximately 30 mm are probably substantially underestimated. Snow crabs are known to borrow in the sediments, especially in younger stages (Conan et al., 1996; Dionne et al., 2003). In some cases with good visibility, an outline of submerged crabs could be seen on the surface of the muddy sediments (Fig. 7B). Although, it is possible that a few crabs were not counted due to low visibility and deep burrowing. Therefore, the crab densities calculated from the video footage can be largely underestimated.

The highest density of crabs recorded in 2016 was in Blagopoluchiya Bay (0.87 crabs/m²). The crabs were small in both the Sigsbee trawl and on the video (majority of crabs with CW below 20 mm). Even though the mixture analysis identified two distinct small sized groups (8–9 mm and 14–15 mm) for both sampling methods, the trawling sample had much sharper differences between the groups, and the larger groups were much more distinct (Figs. 3A, 3B, Table 1). Such differences could be due to possible errors in size measurements of filmed crabs. In the video samples, the CW was measured by a ruler, which has lower precision. In addition, the measurements were recalculated based on the distance between the two laser points that were also measured by a ruler. The crabs were not always in absolutely plain position towards the camera, and the visibility often did not permit to see the edges of carapace clearly. Therefore, there could be some additional noise in CW measurements from the video footage in comparison to the direct measurements by calipers of live organisms.

However, the most important error in the identification of size groups using video samples was due to sexual dimorphism. The size groups (instars) of Ch. opilio are extensively described in the literature and the young crabs (pre puberty molting <20–30 mm) seem to have similar size groups across their range of habitat (Ito, 1970; Kon, 1980; Sainte-Marie, Raymond & Brêthes, 1995; Ernst et al., 2012). These instars are in accordance with those observed for young crabs in the Kara Sea in 2014 (Zalota, Spiridonov & Vedenin, 2018) and in 2016 (present study, Table 1). After puberty molting (which was shown to be at 37–40 mm for males and 17 mm for females, in the Gulf of St. Lawrence) crabs’ growth rate and possible skipping of molts in females is strongly affected by temperature (Sainte-Marie, Raymond & Brêthes, 1995; Alunno-Bruscia & Sainte-Marie, 1998; Dawe et al., 2012), which is less than 0 and −1 °C in most areas of the Kara shelf (Polukhin & Zagretdinova, 2016); see also (Zalota, Spiridonov & Vedenin, 2018). Further growth and survival success of larger crabs may also be affected by low food availability for benthic predators in the Kara Sea (Zenkevich, 1963; Kulakov et al., 2004). As they age further, they molt approximately once a year, or even rarer until they reach their terminal molt (males at CW (postmoult) as small as 40 mm up to 150 mm; females 30-95 mm) (Ito, 1970; Robichaud, Bailey & Elner, 1989; Comeau et al., 1998; Sainte-Marie & Hazel, 1992; Sainte-Marie, Raymond & Brêthes, 1995; Alunno-Bruscia & Sainte-Marie, 1998).

Taking into account these errors, caused by aggregating males and females in video samples, it is not surprising that crabs larger than 20 mm blur into one large size group in the video data, while crabs caught by the large MMBI trawl and measured more accurately can be separated into at least 4 size groups (Table 1, Figs. 2B, Figs. 2C). Even finer size structure in the MMBI samples can be seen if crabs over 20 mm are separated according to sex (Table 1). Males have 7 distinct size groups over 20 mm CW, whereas, females’ groups are more blurred. This can be due to the differences in molting and growth rates. Since video data cannot provide information on sexual dimorphism, the work done to study exact differences in the size structure of immature and mature crabs in the Kara Sea is not presented in this paper. Nevertheless, the obtained data still permits to identify general size groups and to approximate their relative quantities.

Development of snow crabs’ invasion and its possible role in Blagopoluchiya Bay

Blagopoluchiya Bay appears to have very different size structure of the snow carb population compared to all other sampled areas. Most of the crabs, both caught in the trawl and in the video footage, were less than 20 mm, and form 2 high frequency groups at around 10 and 15 mm CW (Figs. 3A, 3B, 4; Table 1). These size groups correspond to the age of less than 2 years. In the Gulf of St. Lawrence it takes 16 to17 months for crabs to achieve CW 10 mm, and another 16 months to achieve size 20 mm through multiple molting events (Sainte-Marie, Raymond & Brêthes, 1995). Therefore, the majority of these crabs could not have settled much earlier than 2014. That year we caught crabs no bigger than five mm in the vicinity of that bay (station 51 in (Zalota, Spiridonov & Vedenin, 2018). Indeed, that was the first year when the crabs have been observed across the entire western Kara Sea and in most cases they were young, with high abundance of just settled crabs (Zalota, Spiridonov & Vedenin, 2018).

There were a few larger crabs in Blagopoluchiya Bay that were caught on camera. Their sizes are not big enough to assume that they actively migrated from other areas. This suggests that there were earlier successful settlings of crabs, but the proportion of larger crabs is almost negligible (Figs. 3A, 3B). The density of the young crabs in the bay is very high (0.87 crabs per m²) and in some cases we observed up to 8 crabs in one video frame (approximately 0.5 m²). Such high densities suggest that a recent combination of favorable oceanographic and sea ice conditions in this area facilitated a massive settling and survival of snow crabs. The larvae that settled in Blagopoluchiya Bay were likely transported by the Eastern Novaya Zemlya current.

This current originates from the Barents Sea water, entering the Kara Sea off the northern coast of the Novaya Zemlya and is directed to the south-west along the eastern coast of this archipelago (Pavlov & Pfirman, 1995). Previously, the bays of the Novaya Zemlya, especially in the north, such as Blagopoluchiya Bay, have been blocked by ice longer than most of the western Kara Sea (AARI, 2007–2014). Although, a narrow Northern Novaya Zemlya Polynya occurred from time to time (Gavrilo & Popov, 2011). Since 2011, changing sea conditions of the 2000–2010s manifested in an abrupt sea ice decrease in June (NOAA, Snow and Ice, Regional Sea Ice, 1979–2018). Although, there was more ice in the spring of 2014 (in comparison to 2011), the sea ice cover along the Novaya Zemlya Archipelago was the first to retreat and an extensive polynya was formed (AARI, 2007–2014). Early sea ice decay could have facilitated seasonal development of phyto- and zooplankton, and hence favorable conditions for feeding and successful settling of crab larvae.

When the first snow crabs were settling, they experienced low predator pressure in the Kara Sea benthos. Adult snow crabs that are highly cannibalistic were not yet present, and the native crab species, Hyas araneus, are very rare in that area (Anisimova, Ljubin & Menis, 2007); authors’ observations). In addition, predatory demersal fishes have substantially lower diversity and abundance in the Kara than in the Barents Sea (Dolgov et al., 2009; Dolgov, Benzik & Chetyrkina, 2014; Dolgov & Benzik, 2016).

Initially snow crabs could have settled in the southern bays and in the central western part of the Kara Sea (in addition to larvae transportation adults could have reached it by active migration) before they reached the northern bays of the Novaya Zemlya Archipelago. Hence, we can observe such difference in the population size structure of Blagopoluchiya Bay (narrow range of size groups) in comparison to the southern areas (all size groups are present).

Indication of formation of spatial population structure

Overall, there seems to be a pattern in the snow crabs’ size distribution across the western Kara Sea. In the southern Kara Sea these patterns are not related to depth since there is no strong correlation between the CW and the depth across all MMBI stations (large bottom trawl). These stations are positioned on a broad, slightly sloping shallow shelf (46 to 195 m). However there is a visually observable trend of maximum sizes prevailing in the Kara Gates Strait and northwards along the Yamal Peninsula (Fig. 5). This closely resembles prevailing current path of the Barents Sea waters, known as the Yamal Current (Pavlov & Pfirman, 1995; Zatsepin et al., 2010a). The area along the Yamal Peninsula and the Novaya Zemlya Archipelago has higher benthic biomass rates than in the center of the western Kara Sea (Antipova & Semenov, 1989; Denisenko, Rachor & Denisenko, 2003; Kulakov et al., 2004; Kozlovskiy et al., 2011). The decrease of the maximum CW from the Yamal Peninsula towards the center of the western Kara Sea could be due to lower food availability further away from the Barents Sea influence, and thus the crabs have insufficient nutrition to achieve larger sizes. However, this hypothesis needs further confirmation.

There could also be behavioral separation, where smaller sized crabs are forced to move to a territory with less food to escape cannibalism, which is very common among this species (Conan et al., 1996; Comeau et al., 1998). No such trends can be observed along the Novaya Zemlya Archipelago, probably due to low sampling effort. There are reports of difference in the habitat preferences of different sized snow crabs in their native habitat areas (Comeau et al., 1998; Ernst et al., 2012). The bays of the Novaya Zemlya have the potential to act as a nursery for smaller, more vulnerable specimens, as it has been observed in the Gulf of St. Laurence (Comeau et al., 1998). Whether this separation exists or will ever exist in the Kara Sea is hard to say at this point. However, the vicinity of deep trough along the Novaya Zemlya Archipelago could attract larger crabs and lead to size related migration out of the bays.

The recruitment of crabs at the early stages of population establishing in the Kara Sea might be mostly due to the inflow of the larvae from the Barents Sea (Zalota, Spiridonov & Vedenin, 2018). Here we present findings of substantial number of ovigerous females, most of which have been found along the Yamal Peninsula (with no apparent spatial or depth distribution patterns), and none in the bays. It is hard to say whether this was due to sampling gear limitation to catch representative sample of larger crabs or a reflection of the real picture. All ovigerous females had CW larger than 40 mm (Fig. 6). This corresponds to the size of female’s terminal (sexual maturity) molt reported in the literature (starting from 35 mm) (Sainte-Marie, Raymond & Brêthes, 1995; Alunno-Bruscia & Sainte-Marie, 1998). Crabs of these sizes had a low catchment rate in the SIO trawling samples along the Novaya Zemlya Archipelago. In most cases the video samples suggest that crabs with CW larger than 40 mm prevail in the vicinity of most sampled bays (Fig. 4). Therefore, it is likely that reproducing crabs are present in most sampled areas. It is safe to say, that at present the snow crab has a reproducing population in the Kara Sea.

The spatial structure of the snow crab population in the Kara Sea is still in process of formation. The data of 2016 indicate that this process may lead to a quite complex system, which is based on local recruitment, transport of larvae from the Barents Sea and across the western Kara shelf, formation of nursery grounds, and an active migration of adults and their concentration in particular shelf areas with appropriate feeding conditions. This system on the other hand can’t be static as it is influenced by changing advection of the Barents Sea water and its interaction with the water of river discharge origin (Zatsepin et al., 2010b), sea ice regime, trophic conditions and predation pressure on juvenile crabs.