In a comfort zone and beyond—Ecological plasticity of key marine mediators

Abstract Copepods of the genus Calanus are the key components of zooplankton. Understanding their response to a changing climate is crucial to predict the functioning of future warmer high‐latitude ecosystems. Although specific Calanus species are morphologically very similar, they have different life strategies and roles in ecosystems. In this study, C. finmarchicus and C. glacialis were thoroughly studied with regard to their plasticity in morphology and ecology both in their preferred original water mass (Atlantic vs. Arctic side of the Polar Front) and in suboptimal conditions (due to, e.g., temperature, turbidity, and competition in Hornsund fjord). Our observations show that “at the same place and time,” both species can reach different sizes, take on different pigmentation, be in different states of population development, utilize different reproductive versus lipid accumulation strategies, and thrive on different foods. Size was proven to be a very mutable morphological trait, especially with regard to reduced length of C. glacialis. Both species exhibited pronounced red pigmentation when inhabiting their preferred water mass. In other domains, C. finmarchicus individuals tended to be paler than C. glacialis individuals. Gonad maturation and population development indicated mixed reproductive strategies, although a surprisingly similar population age structure of the two co‐occurring species in the fjord was observed. Lipid accumulation was high and not species‐specific, and its variability was due to diet differences of the populations. According to the stable isotope composition, both species had a more herbivorous diatom‐based diet in their original water masses. While the diet of C. glacialis was rather consistent among the domains studied, C. finmarchicus exhibited much higher variability in its feeding history (based on lipid composition). Our results show that the plasticity of both Calanus species is indeed impressive and may be regulated differently, depending on whether they live in their “comfort zone” or beyond it.


| INTRODUC TI ON
The fundamental way by which organisms cope with climate change is through ecological plasticity, which encompasses any type of environmentally induced change (e.g., morphological, physiological, behavioral, phenological). Zooplankton can respond to global environmental changes phenotypically (with alterations in their physiology or behavior) or evolutionarily (with a shifting genetic composition of populations) (Dam, 2013;Kelly et al., 2017). It is still not known whether marine species will have the capacity to adjust to ongoing changes through phenotypic plasticity in the short term or adapt in the longer term (Byrne et al., 2020). Although ecological plasticity refers only to the phenotypic changes that are expressed in the lifetime of a single organism, some of these changes can potentially be adaptive (Ghalambor et al., 2007) and evolve over time and space (Kelly et al., 2017;Sasaki & Dam, 2019). Therefore, phenotypic plasticity plays a key role in adapting to new and changing environments (Chevin et al., 2010;Pfennig & Ehrenreich, 2014;Pfennig et al., 2010).
Copepods of the genus Calanus are the key components of zooplankton in the Arctic and northern Atlantic waters (Aarflot et al., 2017;Carstensen et al., 2019;Jaschnov, 1970), and they play a crucial role in marine food webs as the main mediators between the microbial system, phytoplankton, and higher trophic levels. Due to their high lipid content (Falk-Petersen et al., 2009;van der Hoop et al., 2019;Mayzaud et al., 2016;Scott et al., 2000), they are responsible for the sustainability of large stocks of fish, seabirds, and marine mammals in the Arctic region ; Stempniewicz et al., 2007). Additionally, due to the various centers of distribution of particular species, they are highly valued as biological indicators of the hydrographical-ecological regimes and consequently of the effects of ongoing climate changes (Choquet et al., 2017;Gabrielsen et al., 2012;Kwasniewski et al., 2010;Møller & Nielsen, 2019). Although specific Calanus species are morphologically very similar, they have different life strategies and exhibit different ecological traits, which at the same time are highly adaptable to the extremely variable environmental conditions encountered by Calanus copepods during their lifespan (Bandara et al., 2019;Falk-Petersen et al., 2009;Feng et al., 2018;Kvile et al., 2018). Therefore, Calanus spp., owing to their high ecological plasticity, can reasonably be assumed to be able to cope with some extent with ongoing environmental changes (Byrne et al., 2020), but specific species may respond differently to warmer Arctic conditions (Falk-Petersen et al., 2009;Kjellerup et al., 2012;Scott et al., 2000). However, due to problems with proper Calanus species identification, knowledge about their ecological plasticity and functioning in various oceanic conditions is still very limited.
In recent decades, the most common method for discriminating the two species was based on prosome length measurements (Arnkvaern et al., 2005;Kwasniewski et al., 2003;Weydmann & Kwasniewski, 2008). However, this method was criticized recently because of an overlap in the size of the two species as an effect of their substantial size plasticity in response to food availability, water temperature, life cycle length, or even to predation pressure (Leinaas et al., 2016;Lindeque et al., 2006;Parent et al., 2011;Renaud et al., 2018). Genetic identification of Calanus species using either mitochondrial or nuclear markers (Leinaas et al., 2016;Lindeque et al., 2006;Parent et al., 2011;Renaud et al., 2018) was successful and offers a very promising, but still labor-intensive, expensive, and rather unsuitable approach for large field surveys. The use of mitochondrial markers for the molecular separation between the two species has been challenged because of questions about possible hybridization between them in the northwest Atlantic (Parent et al., 2012), but subsequently developed nuclear markers for Calanus species identification  can be used to detect events of potential recent hybridization .
Nonetheless, large-scale genetic surveys throughout the northeast Atlantic and Arctic oceans (Choquet et al., 2017(Choquet et al., , 2018 and novel studies of mitochondrial genomes of C. finmarchicus and C. glacialis (Weydmann et al., 2017) suggest that hybridization is unlikely.
Consequently, all new studies clearly show that if Calanus is to be distinguished to the species level, this must be supported by genetic analysis (Choquet et al., 2018;Renaud et al., 2018). The proper identification of Calanus to the species level is definitely not only a matter of taxonomical curiosity; the knowledge of whether one or the other species prevails and what is the extent of its ecological plasticity is a principal ecological factor in understanding and predicting the future functionality of north polar marine ecosystems.
An increasing temperature and volume of Atlantic waters flowing into the Arctic is causing shifts in hydrographic conditions (Polyakov et al., 2005(Polyakov et al., , 2020Walczowski et al., 2012). Shifts in the positions of water masses will result in changes to the distribution of associated zooplankton communities (Beaugrand et al., 2002;Chust et al., 2014;Hays et al., 2005). Therefore, along with the progressive climate warming in the Arctic, a regime shift from larger lipid-rich Arctic C. glacialis species toward smaller Atlantic C. finmarchicus species is expected (Aarflot et al., 2017;Kjellerup et al., 2012;Møller & Nielsen, 2019). Due to the commonly observed difference in the size and thus the amount of lipids between C. finmarchicus and C. glacialis, the two species were believed to support different arctic food webs (Renaud et al., 2018;Weslawski et al., 2009). The decreased availability of C. glacialis is expected to affect the breeding success of the most numerous seabirds, little auks (Jakubas et al., 2020;Stempniewicz et al., 2007). However, the predicted higher abundance of C. finmarchicus in the Arctic (Reygondeau & Beaugrand, 2011;Slagstad et al., 2011) is expected to favor fish stocks Renaud et al., 2018;Stempniewicz et al., 2007). Additionally, the changes in the timing of Calanus reproduction and development rate have critical significance for their availability to predators (Balazy et al., 2019;Bandara et al., 2019;Søreide et al., 2008).
The aim of this study was to verify the actual differences between C. finmarchicus and C. glacialis with regard to their morphology (size, pigmentation), life cycle phenology (according to population demography and gonad maturation), and feeding strategies (variability in diet composition and differences in lipid accumulation). These differences were studied both in the waters where each species originate from (the Atlantic domain of the Polar Front in the case of C. finmarchicus and the Arctic domain of the Polar Front in the case of C. glacialis) and in the waters in which they coexist (Hornsund fjord, Spitsbergen), both in the fjord main basin and in its glacial bay. The hypothesis of this study is that the traits of the Calanus species (e.g., size, pigmentation, population demography, reproductive readiness, lipid accumulation, diet) will differ depending on whether specific species (C. finmarchicus vs. C. glacialis) inhabits the preferred original water mass (comfort zone) or exist in suboptimal conditions (due to, e.g., temperature, turbidity, competition). We assume that both species will exhibit high plasticity of the traits studied; however, the extent of the plasticity will differ between the species, with C. finmarchicus expected to be more of a generalist and C. glacialis expected to be more of a specialist (as defined by Dam, 2013). This study is the first to incorporate many aspects to improve taxonomical and ecological species recognition, which is the main prerequisite for predicting future ecosystem shifts in the Northern Hemisphere.

| MATERIAL S AND ME THODS
The study was performed in the Polar Front region in the southern part of the West Spitsbergen shelf in July 2018 ( Figure 1). It was designed to collect Calanus either from the water mass they originate from (C. finmarchicus from the Atlantic Water (AT station) domain, from the West Spitsbergen Current that carries warm and saline Atlantic Water and C. glacialis from the colder and fresher Arctictype Sørkapp Current (AR station)), or from the Hornsund fjord, where two species co-occur, both in the main fjord basin (F station) and in the glacial bay (G station) (Figure 1a). The station representing the Atlantic water domain (AT) was characterized by the highest water temperature, salinity, and chlorophyll fluorescence ( Table 1).
The lowest water temperature, salinity, and chlorophyll fluorescence were observed at the station located in the glacial bay (G). However, the glacial bay was characterized by extremely high concentrations of particles (Table 1).
At each sampling station, two tows of a WP2 net (180 µm mesh size) were performed thorough the water column (70 m at G station, 200 m at F station, 170 m at AR station) or through the upper 200 m at the deep-water AT station. The sample from the first haul of the net was immediately fixed in a formaldehyde-borax solution for the analysis of Calanus abundance, developmental stage composition, and gonad maturation stages. The samples from the next net hauls were put in buckets filled with seawater close to the in situ temperature. Then, the fifth copepodite stages of specific species (distinguished as small and large size modes) were photographed and preserved for further analyses (Table S1, Figure 1b).
Overall, 1,252 photographs of Calanus were taken and analyzed to determine the prosome length, the area of the lipid sac, and the pigmentation. Methods for specific morpho-ecological traits were as follows:

| Size and species identification verification
The prosome length of 1,252 photographed Calanus CV copepodites was measured in ImageJ/Fiji free software for image analyses from the tip of the cephalosome to the distal lateral end of the last thoracic somite. The medians and quartiles of prosome length measurements are presented as box plots. Additionally, 991 copepodites were measured from the formaldehyde-fixed samples, among which 372 were CV copepodites, for which kernel density size distribution curves were fitted.
Genetic species identification was used to test the accuracy of the applied morphological method of species discrimination for further analyses. Genetic identification was performed for each individual using its antennae and a set of six nuclear insertion-deletion markers (InDels) in a multiplex polymerase chain reaction (PCR) following the protocol described in Choquet et al. (2017).
Accuracy of the morphological identification was calculated as the the number of all assessments. Overall, the accuracy was 93%, but the accuracy was also calculated separately for each species and sampling location ( Table 2). The species discrimination was 100% accurate at the fjord station (F) and almost fully accurate at the AR and AT stations. In the Atlantic domain, 93% accuracy was obtained. In the Arctic domain, one individual with a size of 2.93 mm initially identified as C. glacialis, turned out to be C. finmarchicus. A high percent of the species misidentification (22%) occurred only in the glacial bay G station. All the C. glacialis were properly identified, but many of the small copepodites, assumed to represent C.
finmarchicus, turned out to be small C. glacialis, with individuals as small as 2.39 mm.

| Red pigmentation
The quantification of astaxanthin content was based on highperformance chromatography (HPLC) according to the methods described in Stoń-Egiert & Kosakowska, 2005. Astaxanthin was isolated from previously lyophilized and weighed Calanus individuals by mechanical grinding in 90% acetone and sonication (2 min, 20 kHz, Cole Parmer, 4710 Series) for 2 hr in darkness. Then, after clarification, the extract was subjected to chromatographic analysis. The HP1200 system (Agilent, Perlan Technologies) was equipped with a C18 LichroCART™LiChrospher™ 100 RP18e (Merck) analytical column (dimensions 250 × 4 mm, particle size 5 μm, and pore size 100 Å). Pigments were separated in a gradient mixture of methanol, 1 M ammonium acetate, and acetone.
Calibration was conducted with commercially available standards (The International Agency for 14C Determination DHI Institute for Water and Environment in Denmark), which allowed for the qualitative assessment of astaxanthin (based on retention time and similarity with the absorbance spectrum of the standards) and quantitative assessment (based on response factor values obtained during the calibration procedure).
Moreover, the pigmentation was analyzed by the visual examination of photographs taken of live individuals by a color coding scheme, depending on whether there was a full (>50%, red), mid (10%-50%, zebra), slight (<10%, piece), or no (0%, transparent) coloration of antennae and prosome. The photographs were also used to calculate the average pigmentation, where the rate of pigmentation was scored for each part of the body from 0 to 3 (e.g., antennae, prosome) and 0 to 1 (swimming legs, urosome) and summed, considering that approximately 10% of the pigmentation may be contributed by the urosome, 10% from the legs, 20% from the prosome, and 60% from the antennas.

| Population development
The relative abundances of each of the Calanus copepodite stages were used to describe the population age structure. The abbreviations (CI-AF) refer to six successive copepodite stages of Calanus, that is, CI, CII, CIII, CIV, and CV refer to the first five copepodite stages, and AF to adult females. The species identification of particular developmental stages was based on the size discrimination assessed for the populations in the Hornsund fjord (Weydmann & Kwasniewski, 2008). The age structure of specific species in the water domains studied was tested by Fisher's exact test in R (fisher.test(data)), an independence test to determine whether there is a significant relationship between two categorical variables.
The gonad maturation stages of Calanus females were determined as described by Niehoff (2007) and Niehoff and Runge (2003).
In our case, we distinguished stage G3, characterizing females that are preparing for spawning (multiple layers of oocytes in both anterior and posterior diverticula); stage G3.5, indicating females that are ready to spawn (ventral layer of oocytes is already colored, but not yet fully brown); and G4, representing females that would spawn within hours (oocytes undergoing final maturation in the most ventral layer in the gonads). The spent stage (S) characterizes females that had already finished reproduction.

| Lipid content
Lipid content was determined as the mass of lipids that was extracted for the fatty acid composition analyses and estimated by the measurements of the lipid sac area. The lipid sac area was manually measured by contouring the sac perimeter by hand in all the photographed copepods. Furthermore, the specific equations derived from the reliable calibration of the individual lipid contents of Arctic copepods (Vogedes et al., 2010) were applied to calculate the total lipid content. Then, the percentage of the lipid sac area (fulfillment) was calculated as a function of the total area of the prosome.
To verify whether the lipid content differed between the two species, the unpaired two-sample Mann-Whitney test was performed, and to verify whether the lipid content differed among study locations and species, the nonparametric Kruskal-Wallis test was performed in R (Package "stats"). The output of the Kruskal-Wallis test indicated whether there is a significant difference between groups, but to know which pairs of groups are different, the function pairwise.wilcox.test was used to calculate pairwise comparisons between groups.

| Diet
To verify the source of stored lipids in of the two species occurring at the same time in diverse water masses and in fjord waters, where F I G U R E 2 (a) Box plots of the mean size (prosome length measured from the photographs of alive individuals) of the Calanus CV that were selected for further analyses, basing on the two size modes representing the two species: smaller C. finmarchicus and larger C. glacialis.
(b) Box plots of the mean size, with Calanus CV species tested genetically. C) Density plots of the prosome length of Calanus CV from the preserved samples they coexist, a combination of stable isotope (δ 13 C, δ 15 N) and fatty acid composition analyses was performed.
Stable isotope analysis was performed following the protocol of Lebreton et al., (2012). Samples were analyzed using an elemental analyzer (Flash EA 1112, Thermo Scientific, Milan, Italy) coupled to an isotope ratio mass spectrometer (Delta V Advantage with a ConFLo IV interface, Thermo Scientific, Bremen, Germany). The results are expressed in the δ unit notation as deviations from standards (Vienna Pee Dee Belemnite for δ13C and N2 in air for δ15N) following the formula: where R is 13C/12C or 15N/14N, respectively. Calibration was per-

| Size
A clear unimodal size distribution of Calanus copepods from the AT domain (Figure 2c), with an average prosome length of 2.6 mm, enabled the easy selection of C. finmarchicus for further analyses (Figure 2a). At the AT station, all copepods were assumed to be C. clearly represented specific species (Figure 2aBC). In the glacial bay, C. glacialis was found to be either large (3.1 mm) or much smaller than the assumed size limit (2.4-2.9, on average 2.7 mm), so the samples assigned as C. finmarchicus at station G ( Figure 2a) were substantially (approximately 40%) contaminated by the smaller sized C. glacialis (Table 1).

| Pigmentation
The amount of astaxanthin pigment was the highest in the Calanus

| Population development
The age structure of C. finmarchicus at the AT station did not differ significantly from the demographic structure found on the shelf (AR) and in the fjord (F) (Fisher's exact test, p < .001). However, approximately half of the population was represented by the CV life stage at the AR and F stations, while a high proportion of females and the first copepodite stages was found at the AT station ( Figure 4).
Additionally, the relative proportions of the three youngest life stages of C. glacialis were similar between the shelf (AR) and the fjord (F); however, the proportion of CV relative to CIV was higher in the fjord. The demographic structure of C. glacialis was advanced in the glacial bay (G) but was not significantly different from the structure in the fjord (Fisher's exact test, p < .001). Surprisingly, the greatest degree of synchronization in population development between the two species was observed in the fjord (Fisher's exact test, p = .36).
The analysis of the gonad maturation stages indicated that the females were either almost ready to spawn (G3 and G3.5), actively spawning (G4), or they had finished reproduction (S). All females actively produced eggs at the AT station ( Figure 5), half of them on the shelf (AR station) and only a quarter in the fjord (F station). In the glacial bay (G station), there were only a few females, and none of them reproduced ( Figure 5).

| Lipid content
Because of the larger body size of C. glacialis, the total amount of lipids per individual, either extracted or measured from photographs, was higher in C. glacialis than in C. finmarchicus in each of the oceanographic domains (Mann-Whitney U tests, p < .001) ( Figure 6). The lipid content and the percentage of lipids in total body area were lower at F I G U R E 4 The relative proportions of various life stages (CI-CVIF) of specific Calanus species in specific water domains ("?" is used when the species identification by size is uncertain)

| Diet
The The isotopic composition is presented on the trace plot, as an output of stable isotope mixing model ("simmr" library in R). Colors refer to specific species in specific studied domains (AT-Atlantic, AR-Arctic, F-fjord, G-glacial bay) especially high in C. glacialis at the F station (12%) and the lowest in C.
finmarchicus at the AT and F stations. The amount of 18:2n6, regarded as a tracer of freshwater/terrestrial input, was generally very low (less than 0.5%) (Tables S1, S2).
According to the ratios of stable isotopes, the lowest δ 15 N values were recorded in both Calanus species in their original water masses (in C. finmarchicus at the AT station and in C. glacialis at the AR station) ( Figure 7b). The median and most variable values of δ 15 N were found in C. finmarchicus in the Arctic domain (AR) and inside the fjord (F). The highest δ 15 N, associated with the lowest δ 13 C, was recorded in C. glacialis in the fjord, both in the main basin (F) and in the glacial bay (G), which was very similar in the small Calanus size fraction at the G station.

| D ISCUSS I ON
Many organismal traits exhibit phenotypic plasticity in response to environmental variation. Therefore, it is crucial to provide ground truth data on possible plasticity in the basic morphological and ecological traits of the key high-latitude zooplankton species expected to make up a new Arctic regime in different ways. Such knowledge would be of great importance to verify and extend many existing models aimed at predicting the future of the Arctic ecosystem based on Calanus ecology (Banas et al., 2016;Bandara et al., 2019;Feng et al., 2018;Ji et al., 2012;Melle et al., 2014;Renaud et al., 2018).
To be able to pinpoint differences in ecological plasticity, it is important to know whether individuals from studied locations represent a panmictic unit or genetically distant units (then observed differences could rather be due to local adaptation). Despite extensive research effort, population genetic differentiation in Calanus finmarchicus and C. glacialis is still under debate with conclusions ranging from a lack of population genetic structure in the North Atlantic for C. finmarchicus (Provan et al., 2009) and in the Arctic for C. glacialis (Weydmann et al., 2016) to a large-scale structure in the North Atlantic for C. finmarchicus (Unal & Bucklin, 2010) and between the Pacific and the Arctic Oceans for C. glacialis (Nelson et al., 2009). Preliminary results from recent advances in Calanus genomics using a capture enrichment-based approach in three population along the Norwegian coast showed weak or lack of genetic structure in C. finmarchicus and stronger genetic structure in C. glacialis . Although more populations need to be examined using such genome-wide approach for a conclusion on Calanus population genetic structure, it is reasonable to assume that the small-scale sampling locations in this study belong to a panmictic population for both species, while the term "population" in further discussion refers to a group of individuals that developed at same location, and therefore, same conditions.

| Size
For decades the species identification between C. finmarchicus and C. glacialis based on size separation was a convenient simplification, and according to this study, it is still a reliable method in the studied fjord, where two distinct size modes agreed well with previously recognized size ranges of both species (Weydmann & Kwasniewski, 2008). However, such size-based identification has been shown to be unreliable in some areas (Choquet et al., 2018;Gabrielsen et al., 2012;Lindeque et al., 2006;Renaud et al., 2018), which was also the case in our study of shelf and glacial regimes.
C. glacialis individuals were occasionally observed to be smaller than assumed. Size reduction in C. glacialis has frequently been recorded and mostly explained by the considerable plasticity of this species in relation to temperature differences (Choquet et al., 2018;Leinaas et al., 2016), the shorter length of its life cycle , or even by predator pressure (Bandara et al., 2019;Berge et al., 2012). However, assuming that in turbid glacial waters, their visual susceptibility to predation is already greatly reduced, the size reduction in C. glacialis should rather be considered as an effect of suboptimal living conditions, including extreme hydrographic dynamics and worsened food quality.
Unfortunately, the most intriguing, trimodal size distribution of Calanus observed on the Arctic side of the Polar Front (AR station) was not tested thoroughly, and only the two ends of the size spectrum were investigated further. Therefore, it remains an open question whether the numerically dominating medium-size mode (2.75 mm) was represented by small C. glacialis or by large C. finmarchicus. The prevailing abundance of the smaller size mode of Calanus at this station could suggest that advection of the Atlantic water across the Polar Front is higher than we assumed and thus leads to the predomination of C. finmarchicus. However, high size variability of genetically proven C. glacialis was observed at this station, so instead, we suggest that this mid-size fraction could represent a population of C. glacialis coming from the northern Barents Sea.
Interestingly, a few large individuals, proved genetically to be C.
glacialis, were also found in the Atlantic sector of the Polar Front

| Pigmentation
In line with our expectations, both species exhibited pronounced This observation, supported by the recent test of various morphological traits of Calanus species identification methods (Choquet et al., 2018), suggests that the red pigmentation of the antenna is not a species-specific feature, as was proposed by Nielsen et al. (2014).
The red pigmentation of Calanus appears to instead be locationspecific. Following our hypothesis about the "comfort zone," the red pigmentation of C. glacialis in Hornsund suggests that the environmental conditions in the fjord are more similar to the Arctic and preferable for C. glacialis, which agrees with its general dominance in the Hornsund fjord (Gluchowska et al., 2016;Trudnowska et al., 2014;Weydmann & Kwasniewski, 2008). In the Arctic, C. finmarchicus individuals tend to be paler compared with C. glacialis, either in Svalbard (Choquet et al., 2018) or in Greenland .
However, which species synthesizes this carotenoid pigment and at what intensity to protect the accumulated lipids from oxygenation are also strongly dependent on many factors other than water mass, including the depth of their occurrence and their diet (Mojib et al., 2014;Sommer et al., 2006). Copepods can adjust their level of red pigmentation quickly, even within a season, depending on the prevailing threat, UVR, predators (Hansson, 2000), or parasites. According to our observations, they can switch from red to pale within hour (e.g., when in a bucket in the fridge without any light), but they can become again red almost immediately when exposed to a light source. However, not all of individuals could return red pigmentation, even after long exposure to light. The fact that some individuals stayed pale may be explained by their need to have special precursors of carotenoids in their diet (Mojib et al., 2014;Sommer et al., 2006). Indeed, both species had similar stable isotope compositions in their "comfort zones" (both exhibiting high levels of red pigmentation) and variable stable isotope and lipid compositions, where their redness was variable.

| Population development
The observed synchronization in the population age structure between the two species in the fjord was surprising in light of the documented differences in reproductive timing and strategies of specific Calanus species (Falk-Petersen et al., 2009;Søreide et al., 2010), which is observed even when they co-occur in the same region (Arnkvaern et al., 2005;Niehoff et al., 2002;Swalethorp et al., 2011).
The similar age structures of both species may indicate that favorable feeding conditions occurred in the fjord prior to our sampling, which enabled both species to simultaneously develop and to adopt a similar strategy. Additionally, the transport of copepodites from the shelf could play a role, as the Calanus age structure in fjords is frequently found to be related to various advection rates (Kwasniewski et al., 2003;Trudnowska et al., 2020;Willis et al., 2006). This could be especially the case for C. finmarchicus, whose demography was very similar between the fjord and shelf and which is probably mostly advected to the fjord. The small-scale difference between fjord and shelf populations in demographic structure of C. glacialis, especially regarding the relative proportions between younger life stages and the most lipid-rich fifth stage (CV), could be of great importance for CV-selective predators such as little auks (Balazy et al., 2019;Jakubas et al., 2020;Welcker et al., 2009) and may explain why birds tend to forage on the shelf rather than in the fjord in the time window of their main feeding requirements (Jakubas et al., 2013).
Actively reproducing C. finmarchicus in Atlantic waters, in contrast to less active reproduction in the Hornsund fjord and shelf populations, which seemed to adopt the lipid accumulation strategy more, is possibly the other indication that the reproductive strategies of Calanus are very plastic (Daase et al., 2013;Kvile et al., 2018).
Most likely mixed reproductive strategies will enable Calanus to actively and flexibly optimize their growth and development to adapt to the expected variability in the timing and quality of the pulses of primary production (Ejsmond et al., 2018;Ji et al., 2013).

| Lipids
One of the key and most spectacular features of Calanus is its ability to accumulate large lipid reserves during a short grazing season in late spring and summer (Daase et al., 2013;Søreide et al., 2010;Wassmann, 2011). In our study, in most cases copepods were very full of lipids (approximately 60% of the body area was filled by the lipid sack), which means that they were well into the lipid accumulation process for a winter diapause phase. The lipid content and lipid fulfillment were highly variable, indicating strong interindividual differences. Moreover, lipid accumulation was not species-specific, which agrees perfectly with the recent study by Renaud et al. (2018). Interestingly, no correlation was found between red pigmentation and the lipid content or the percentage of lipids in the total body area (Pearson correlation, p > .005).
In the glacial bay, individuals of the large Calanus size mode were extremely full of lipids in contrast to ones from the small size mode, in which very small lipid sacks were observed. Whereas, in the fjord, C. glacialis exhibited decreased lipid accumulation perhaps as an effect of the competitive regime with C. finmarchicus or because of the different niche partitioning, which is in line with the different lipid and stable isotope compositions of both species in the fjord.

| Diet
Traditionally, Calanus copepods have been regarded as herbivores; however, their omnivore preferences (Cleary et al., 2017;Levinsen et al., 2000;Søreide et al., 2008;Yeh et al., 2020) and their ability to adjust to shifts in the timing and availability of their prey (Banas et al., 2016;Forest et al., 2011;Freese et al., 2016) can vary greatly. In addition, the complex hydrography (i.e., mixing of Atlantic, Arctic, and coastal water masses) around Svalbard significantly expands the pool of resources and niche diversity in terms of their habitat and diet (Kortsch et al., 2019). According to the δ 15 N stable isotope and lipid composition, both species had a mostly herbivorous, diatom-based diet in their original water masses (C. finmarchicus at the AT station and C. glacialis at the AR station).
However, the overall lipid composition differed significantly between the two species, regardless of the region. The differences in the feeding patterns of Calanus were found not only to be species-specific (Scott, Kwasniewski, Falk-Petersen, & Sargent, 2000, but also to be related to seasonal and environmental changes (Forest et al., 2011;Mayzaud et al., 2016;Melle et al., 2014;Yeh et al., 2020). Such striking differences in the diets of the co-occurring species is a question to consider whenever we deal with the advected versus local populations or if they are occupying different water layers and thus are exposed to various food sources. Such water column partitioning in utilizing different available resources by grazing at different depths was recently explained as a way to minimize the competition between the two Calanus species (Schmid & Fortier, 2019 While the diet of C. glacialis was rather similar among the studied domains, just with a more herbivorous tendency on the shelf and a slightly more omnivorous tendency in the fjord, C. finmarchicus exhibited much higher variability and differences in its feeding history. The populations found in the fjord and on the shelf showed elevated concentrations of the marker of Phaeocystis flagellates (Leu et al., 2006), which is indeed often dominant in the Svalbard fjords during summer (Kubiszyn et al., 2017;Piwosz et al., 2009).
As expected, the copepods in the glacial bay were depleted in 13 C, which is a signature of the terrestrial input to their diet (Dalsgaard et al., 2003) or starvation (Mayor et al., 2011). Additionally, in the photographs, we could see that they were loaded with mineral particles. The more omnivorous diet of both Calanus species in the glacial bay (highest δ 15 N) was probably caused by the decreased availability of phytoplankton due to diminished light conditions and the higher importance of microzooplankton in turbid waters (Halbach et al., 2019;Piwosz et al., 2009).
All these differences signify that information about variation in diet flexibility and feeding histories is another interesting trait to follow in the species-specific ecology of Calanus and is worth considering to understand how they individually utilize resources.

| Ecological consequences
Rapid biogeographical shifts in plankton have already occurred in the North Atlantic Ocean (Beaugrand et al., 2009), and now, analogous processes are occurring in the Arctic (Wassmann, 2011 An experimental study of the effects of temperature and food on C. finmarchicus and C. glacialis implies that the adaptation of C. glacialis to future warmer ocean conditions in the Arctic by spawning early will no longer be beneficial when C. finmarchicus matches the timing of gonad maturation with blooms (Kjellerup et al., 2012). Moreover, C. glacialis is probably not capable of exploiting longer periods of food availability in contrast to C. finmarchicus (Freese et al., 2016;Swalethorp et al., 2011). At the same time, the northward expansion of C. glacialis and its greater developmental rate are predicted to be an effect of increased primary production and decreased extent of ice in the central Arctic . While C. finmarchicus biomass and production may prevail in the future in northern regions due to its increased movement via advection, higher growth rate and shorter life cycle (Renaud et al., 2018;Scott et al., 2000;Weydmann, 2018), C. glacialis may also exhibit large spatial variability in its phenology, energy allocation, reproduction, and adaption to a wide range of environmental conditions (Banas et al., 2016;Bandara et al., 2019;Daase et al., 2013;Feng et al., 2018).
Our results show that the plasticity of both Calanus species is indeed impressive and may be differently applied, depending on the conditions in the water mass in which they reside, namely if they live in their "comfort zone" or beyond it. Our observations imply that when present at the same place and time, they can reach different sizes, take on different pigmentation, accomplish different states of population development, utilize different reproductive versus lipid accumulation strategies, and thrive on different foods. C.
finmarchicus was mostly "the guest" in the studied subregimes and it showed much more flexibility in most of the studied ecological traits, which suggests that it has strong potential to colonize newly available Arctic habitats under progressing climate change. In agreement with our hypothesis, the extent of plasticity was higher in C.
finmarchicus, which seems to be more of a generalist than C. glacialis.
Nevertheless, models are optimistic not only for the future of C. glacialis  but also for C. hyperboreus, which is expected to survive, albeit by becoming more "boreal-like" (Kvile et al., 2018).
Looking at the uncertainties raised in our study, we suggest that the role of ecological plasticity is probably not only a matter related to the species but also to the population. Additionally, the question whether Calanus will ever reach a tipping point with regard to its high ecological plasticity, regardless of the species, remains unanswered, hopefully forever. It is also possible that Calanus will be able to adjust to climate change through adaptive responses, provided that evolutionary changes occur at a rate similar to climate change (Dam, 2013).

ACK N OWLED G M ENTS
Special thanks go to Pierre Richard, Directeur de Recherche CNRS Emérite, Université de La Rochelle, for stable isotope analyses.

CO N FLI C T O F I NTE R E S T
None declared.