The autumn plankton community of an Arctic fjord: impact of temperature and salinity on the functional response of two copepod species “to watch”

Abstract

Here, we describe the plankton community of a freshwater impacted Arctic fjord, Kangerluarsuk Ungalleq, west Greenland, during early autumn 2023. Small phytoplankton (< 10 μm) dominated the autotrophs, and protozooplankton (ciliates and dinoflagellates) dominated the grazers, accounting for 98% of zooplankton biomass in the upper part of the water column. Protozooplankton was mainly constituted by aloricate ciliates, while the cyclopoid Oithona similis was the most abundant copepod. Calanus spp. contributed most to copepod biomass, especially in the cold, salty bottom water inside the midway sill. We also investigated the impact of temperature rise and salinity decrease on two copepod species from the fjord, Acartia longiremis and Eurytemora americana, which we consider potential benefactors of climate change. E. americana is a non-indigenous species in Greenland, and this was, to our knowledge, the first observation. Neither of the species altered their functional response (fecal pellet production in relation to food concentration) in low salinity, which indicate high freshwater tolerance. Temperature had a large effect on A. longiremis’s functional response, where the Q₁₀ for maximum ingestion was 8.6. In contrast, E. americana showed a much weaker response, with a Q₁₀ for maximum ingestion of 1.6. Our results suggest that A. longiremis could benefit from temperature rise if food is sufficient, while E. americana does not pose a threat to the native Arctic copepod community.

INTRODUCTION

Due to polar amplification, the Arctic has endured an up to four times faster temperature rise as a consequence of climate change than the global average (Rantanen et al., 2022). This is bound to affect marine environments, where it has already caused increased water temperatures (Koenigk et al., 2020), decreased sea ice coverage (Comiso et al., 2008) and increased freshwater inputs (Peterson et al., 2002; Yamamoto-Kawai et al., 2009). Though uncertainty about the severity, climate prediction models generally agree that these changes will continue in the foreseeable future (Peng et al., 2020; Khosravi et al., 2022). Arctic marine planktonic food webs are characterized by a relatively low biodiversity compared to the rest of the northern hemisphere (Ibarbalz et al., 2019), which make these systems particularly vulnerable to environmental changes (Johannessen and Miles, 2011) and easier for non-indigenous species to colonize (Chan et al., 2019; Møller and Nielsen, 2020). Temperature rise is a driver of the poleward movement of species, observed all over the globe and especially pronounced in marine systems (Melbourne-Thomas et al., 2022), where temperate species are now finding advantageous habitats in the Subarctic and Arctic region (Beaugrand et al., 2009; Neukermans et al., 2018), and the native Arctic species are moving further north or being outcompeted (Fossheim et al., 2015).

Due to their size and at times long water residence times, fjords can integrate and accumulate even short term changes in external forcing (Jackson et al., 2018). Therefore, fjords can be viewed as micro-climates, useful sentinels for how climate change and anthropogenic activities are influencing marine ecosystems. Fjords are glacially formed estuaries which subsequently can be very deep and have several sills that can influence exchange with the open sea. They are recipients of freshwater from rivers or marine terminating glaciers, which influences their mixing dynamics by regulating the strength and position of pycnoclines and supply of nutrients. The Arctic fjords provide various crucial ecological services, among which are nursery areas for juvenile fish (Olsen et al., 2010) and feeding grounds for marine mammals and birds (Stempniewicz et al., 2017). In this light it is important to study and follow developments in fjord ecosystems and the impacts of human activities.

Research on the biology in Arctic fjords is generally biased towards the spring bloom, and while this indisputably is a defining period, our understanding of seasonal succession should be improved (Marquardt et al., 2016). The importance of the period outside the spring bloom was demonstrated by Søreide et al. (2022), who found the highest zooplankton density during autumn in Billefjorden, Svalbard. Similarly Arendt et al. (2013) found copepod biomass to peak in August, and copepod biomass in September to exceed that of both May and June. As copepods are important food sources for fish larvae and juveniles (Bouchard and Fortier, 2020), periods of high copepod biomass are driving early-stage fish growth (Castonguay et al., 2008). In contrast to the spring bloom, where Calanus spp. constitute the majority of zooplankton biomass (Arendt et al., 2013; Madsen et al., 2008), the autumn zooplankton community is often dominated by protozooplankton (Levinsen and Nielsen, 2002) and smaller copepods (Arendt et al., 2013; Madsen et al., 2008). Although autumn is an understudied season in the Arctic, there is evidence that climate change is triggering changes in the zooplankton community (Ardyna et al., 2014; Balazy et al., 2021). As Arctic copepod communities are generally shifting from large, fatty to smaller species (Møller and Nielsen, 2020), knowledge on current and historical autumn communities can also provide insides into future zooplankton configurations outside the autumn season.

In this paper, we have two main objectives: (i) to address the knowledge gap of Arctic marine plankton ecology outside the spring bloom, and (ii) to investigate phenotypic responses to changes in temperature and salinity of two copepod species, that we consider potential benefactors of climate change. For objective 1, we describe the plankton community in a terrestrial-freshwater impacted fjord, Kangerluarsuk Ungalleq, Greenland, during early autumn. As many glaciers are retreating, they potentially shift from being marine-terminating to land-terminating (Stuart-Lee et al., 2021; Meire et al., 2023). This can fundamentally change the base of the food web, by inducing a shift in the plankton community towards smaller species of both phyto- and zooplankton (Meire et al., 2023). Therefore, increasing our baseline understanding of fjords like Kangerluarsuk Ungalleq is important for understanding the emerging ecosystems in the warming Arctic. For objective 2, we identify the copepods A. longiremis and E. americana as species that could become increasingly important in the future, and test experimentally how they respond to acute temperature rise and decreased salinity. A. longiremis is frequently found in warmer environments (Peters et al., 2013) and has before been important during a warm Arctic autumn (Debes et al., 2008). E. americana is a non-indigenous species, which has to our knowledge not previously been found in Greenlandic waters, but has become dominant in another non-native region (Berasategui et al., 2023).

MATERIALS AND METHODS

Study area

The survey in this study was conducted in the west Greenlandic fjord Kangerluarsuk Ungalleq 27 km north of Sisimiut (Fig. 1A). The fjord is narrow (ca. 1 km), 25 km long, and has two basins separated by a midway sill with a depth of 13 m. The inner basin is up to 54 m deep. The outer basin is up to 66 m deep and a sill with a depth of 9 m is located at the mouth of the fjord (Fig. 1C). The fjord is influenced by tides with a tidal range of 2–5 m (3 m on the sampling day). It receives freshwater from a hydropower plant located at the inner eastern end of the fjord, fed with water through a tunnel from Lake Tasersuaq. This power plant started operating in 2010 and provides Sisimiut with electricity. No previous scientific literature exists from the fjord.

Sampling

Sampling was conducted on September 5^th, 2023, along a fjord transect. Depth was recorded continuously (Fig. 1C) using an echosounder and seven stations selected for sampling (Fig. 1B). At each station (st.), a CTD-probe with a fluorometer (AML-6 LGR, Chlorophyll A&B Blue Excitation Sensor powered by Turner designs) was deployed to characterize the water column structure and properties. The sensors included temperature, salinity, dissolved oxygen, chlorophyll fluorescence and turbidity. Water samples were taken at the surface (1 m), and just above the bottom using a 5-L Niskin bottle. Two vertical tows with a 200 μm WP-2 plankton net of diameter 70 cm were taken to collect mesozooplankton. The first tow was through the surface layer from the bottom of the upper pycnocline (5–10 m) to the surface, while the second tow was throughout the whole water column. The position of the pycnocline was identified in-situ from CTD-data. To account for under-sampling of smaller zooplankton by the 200 μm WP-2 net, we inferred corrections for the smaller copepod stages (Fig. 2; supplementary material Fig. S1). These corrections were based on data from a supplementary sampling, September 6^th, 2024, where we took vertical tows at st. 2, 4 and 6 with both 200 μm and 45 μm WP-2 nets. For species and stages retained by both mesh sizes, we corrected the data from 2023 by under-sampling factors, which are the geometric means of the 200:45 μm WP-2 abundance fractions, for the given species and stage (Fig. 2). For species and life stages only retained on the 45 μm WP-2 (supplementary material Fig. S1), we corrected the abundance by assuming a similar relative abundance of life stages between 2023 and 2024. The size below which under-sampling occurred was species specific (Fig. S1) and ranged between 400 and 500 μm prosome length. The correction with the small mesh size increased the average water column copepod biomass by 3%, while copepod abundance was increased 62%.

Hydrography and nutrients

The CTD data was processed in MATLAB (version 23.2), by averaging into 0.5 m bins and excluding upcast data. Nutrient samples were taken from surface and bottom water samples into 30 mL acid cleaned polyethylene bottles and kept cold and dark until frozen at −18°C within 5 hours of sampling. After ca. 7 weeks the samples were defrosted and analyzed for nitrite (NO₂⁻), nitrate (NO₃⁻), phosphate (PO₄³⁻) and silicate (SiO₄⁴⁻) using an AMS SmartChem200 discrete analyzer. Nitrite, phosphate and silicate were measured following the reagent concentrations for the manual method in Hansen and Koroleff (2007), with the inclusion of a 150 g L⁻¹ solution of Sodium dodecylsulfate as a dispersant for phosphate analysis. Nitrate was measured following the vanadium chloride reduction technique (Schnetger and Lehners, 2014). Community reference material for nutrients in seawater supplied by Kanso Technos Co. LTD, Japan, was used for assuring accuracy and assessing precision.

Plankton community structure and distribution

The depth distribution of phytoplankton was characterized along the transects by the chlorophyll fluorescence measured by the CTD probe. Water samples for determination of fractionated chlorophyll a (chl a) concentrations were taken from the 1 m water samples in dark plastic containers. Immediately at arrival in the laboratory, the content of the containers was gently mixed, and 200 mL triplicate sub samples filtered through 0.2 μm GF/F and 10 μm Whatman filters, to measure the concentrations of total and > 10 μm phytoplankton fractions, respectively. After filtration, the filters were put in 10 mL plastic tubes and chl a extracted in 10 mL 96% ethanol and stored in −18°C. The samples were kept dark at all times. After ca. 9 weeks, chl a concentrations were measured on a Turner Fluorometer, calibrated against a pure chl a standard (DHI). Carbon biomass of the phytoplankton was calculated using a carbon:chl a ratio of 43 (Nielsen and Hansen, 1999).

Water samples for microzooplankton (>10 μm) were also taken from the 1 m samples and stored in 300 mL amber glass bottle. The samples were fixed in-situ by adding Lugol’s solution to 2% final concentration and kept in darkness at room temperature. After ca. 8 weeks a well-mixed 50 mL subsample from each station was allowed to settle overnight in a Utermöhl settling chamber after which microzooplankton was counted under an inverse microscope (Nikon Diaphot 200). All cells were counted and divided into taxonomic groups and size-classes. Biomass of microzooplankton was calculated from their volume and the volume to carbon ratios from Menden-Deuer and Lessard (2000):

where M is carbon biomass (Pg C cell⁻¹) and P_vol is cell volume (μm³).

After a vertical tow for mesozooplankton, the net was thoroughly rinsed with seawater and the content of the cod-end transferred into 500 mL plastic containers. Immediately, when back on land the samples were fixed with buffered formalin to 4% final concentration. After ca. 10 weeks these samples were rinsed with 0.45 μm filtered seawater, and zooplankton identified and counted using a stereomicroscope. A subsample of each mesozooplankton species or genera including copepodite stages were taken for length measurements and the average length of each subsample used for biomass calculation. Biomass of all identified mesozooplankton taxa was calculated based on length weight relations from the literature (supplementary material, Table S2).

Estimated clearance rate, ingestion and production

We estimated clearance rate, ingestion and production of zooplankton in the surface layer (within and above the upper pycnocline). Ciliates, dinoflagellates and copepods were considered for this exercise, and we assumed that they were homogeneously distributed in the surface layer. Volume specific clearance rate was calculated by assuming maximum clearance rates following Hansen et al. (1997).

where C_max is the maximum specific clearance rate at 20°C (10⁵ h⁻¹), P_vol is grazer volume (μm³ ind⁻¹) and a is a coefficient (see Table I). C_max was adjusted to the ambient temperature observed in the fjord, by applying a Q₁₀ of 2.8 (Hansen et al., 1997).

Community clearance rates were calculated as.

where F is the fraction of water cleared by either ciliates, dinoflagellates or copepods per day (L L⁻¹ day⁻¹), C_max,s is specific clearance rate at 6°C (day⁻¹), P_vol,s is again grazer volume (L ind.⁻¹) and A_s is grazer abundance (ind. L⁻¹). The index, s, denotes the size-classes of a respective zooplankton type. Note that the units are converted into L and days.

Ingestion, I (μg C L⁻¹ day⁻¹), was calculated by multiplying F with the biomass of the respective zooplankton’s prey. We assumed 100% ingestion of encountered food items and that ciliates were feeding on phytoplankton < 10 μm, dinoflagellates were feeding on phytoplankton > 10 μm and copepods were feeding on ciliates, dinoflagellates and phytoplankton > 10 μm.

The biomass production of ciliates, dinoflagellates or copepods (μg C L⁻¹ day⁻¹) was calculated by multiplying the ingestion with a gross growth efficiency of 0.33 (Hansen et al., 1997).

Impact of temperature and salinity on functional response—Experimental setup and procedure

The copepods A. longiremis and E. americana were selected as species of interest for the experimental part of the study. E. americana can be distinguished from other Eurytemora spp. by the presence of two long terminal setae on the 5^th pair of swimming legs (Fig. 3). We tested their phenotypic responses towards climate change by investigating functional responses of fecal pellet production (FPP) in relation to food concentration at different temperatures and salinities. FPP is here considered a proxy for ingestion, where the assimilation efficiency is 67% (Besiktepe and Dam, 2020; Kiørboe et al., 1985). Animals were collected in Kangerluarsuk Ungalleq and Ulkebugten at Sisimiut, by horizontal tows in the upper 10 m of the water column, using a 200 μm WP-2 plankton net. After a tow the cod end was emptied into two 20-L cooling boxes holding ambient surface seawater. In the laboratory, A. longiremis and E. americana females were carefully sorted and each species transferred to 5-L aerated buckets, where they were kept as stocks at 5°C with plenty of food. Experiments were performed in 250 mL Pyrex bottles in quadruplicates, with three copepods per bottle. The functional response was tested by using six different diatom (Thallassiosira weissflogii) concentrations, 250, 500, 1 000, 2000, 4 000, 8 000 cells mL⁻¹, corresponding to 32.8, 65.5, 131, 262, 524, 1 048 μg C L⁻¹ (Dutz et al., 2008). E. americana experiments included the full spectrum of food concentration, while A. longiremis experiments only ranged from 32.8 to 524 μg C L⁻¹. The copepods were acclimatized for 24 h in the target food concentration, temperature and salinity prior to experiment start. After acclimatization, the copepods were examined, medium refreshed and dead individuals removed before the experiments were run for 24 h. During experiments, bottles were rotated manually 2 times daily to keep the food in suspension. The experiments were terminated by filtering the content of each bottle onto a 40 μm filter, which was rinsed into a petri dish, where fecal pellets were counted to determine FPP. Pellets minimum three times longer than wide were counted, and other fragments neglected. Due to shortage of animals, copepods were examined after experiment termination, and healthy individuals used for further experiments in the same food concentration. If mortality had occurred, dead individuals were replaced from the copepod stocks before acclimation (Table II).

Salinity effects were tested in 29 PSU (observed salinity in the fjord) and 18 PSU at 6°C (observed temperature in the fjord). Temperature effects were tested at 6 and 12°C, and because no salinity effects were observed, the data from both 18 and 29 PSU were pooled in the analysis of temperature effects. Low temperature incubations were done in an AEG fridge equipped with a DYNAMICAIR-unit and high temperature incubations were done in a winecooler (Cavecool Chill Ruby—34 bottles—Dual zone). Temperatures were monitored with temperature loggers (HOBO Pendant MX Water) at ten-minute intervals. Experimental temperatures were 5.9 ± 0.4 and 11.8 ± 1.0°C for A. longiremis in cold and warm treatments, respectively, and 6.0 ± 0.9 and 11.9 ± 1.6°C for E. americana in cold and warm treatments, respectively (mean ± SD). The diatom Thalassiosira weissflogii was chosen as food source and cultured in two 1-L bottles with constant aeration and light. When an aliquot of the T. weissflogii culture was taken for experiments, it was replaced with the same amount of 0.45 μm filtered seawater and silicate enriched B1 medium to keep the culture in exponential growth. Experimental food concentrations were prepared by diluting the T. weissflogii culture with either 29 or 18 PSU 0.45 μm filtered surface water. Dilution factors were determined by counting the concentration of the T. weissflogii culture fixed with Lugol’s in a Sedgwick-Rafter counting chamber under a stereomicroscope. A subsample of fecal pellets from each copepod species was measured under a stereomicroscope to establish specific FPP (μg C (μg C cop)⁻¹ day⁻¹) (Table II). Pellets were assumed cylinder shaped and the volume:C equation from Swalethorp et al. (2011) adopted. Copepod carbon content was calculated from prosome length of adult females from Kangerluarsuk Ungalleq, for E. americana following the Eurytemora affinis equation from Lloyd et al. (2013) and for A. longiremis following the Acartia clausi equations of Cataletto and Fonda Umani (1994) (Table II). If mortality had occurred during experiments (Table II), the dead copepod was assumed not to have contributed to fecal pellet production. The specific ingestion (μg C (μg C cop)⁻¹ day⁻¹), was then calculated based on specific FPP by assuming an assimilation efficiency of 67% (Kiørboe et al., 1985).

Statistical analysis of functional response experiment

To describe the functional responses, experimental data was fitted with a Holling type II model, using the expression from Helenius and Saiz (2017):

where I is specific ingestion (μg C (μg C cop)⁻¹ day⁻¹), I_max is maximum specific ingestion (μg C (μg C cop)⁻¹ day⁻¹), K_m is half saturation (μg C L⁻¹) and T_w is prey concentration (μg C L⁻¹). Due to a generally high ingestion and subsequently low number of observations in the low ingestion-region, the choice of whether to fit with a Holling type II or III (Helenius and Saiz, 2017) became somewhat arbitrary. We chose type II because of its relative simplicity and previous reports for Acartia spp. and Eurytemora spp. (Houde and Roman, 1987; Barthel, 1983, respectively). Q₁₀ for maximum specific ingestion, I_max, was calculated for the investigated temperature range, 6–12°C, following the equation from Grote et al. (2015):

where I_max,12 and I_max,6 are maximum specific ingestion at temperatures T₁₂ and T₆, respectively.

Graphical illustrations of experimental results and statistical analyses were performed in the software Rstudio (version 4.1.1). Effects of temperature and salinity on the functional responses were tested statistically, both by applying an analysis of covariance (ANCOVA) and comparing 95%-confidence intervals of I_max and K_m between treatments, using confint2() function in R. Non-overlapping confidence intervals were regarded different with statistical significance, and this difference will be referred to as significant further on in the paper. The ANCOVA included food concentration as a numerical variable and either of the climate variables, temperature or salinity, as a categorical variable. In addition, the interaction between food and a climate variable was included. Model assumptions were examined visually using qq-plots and none of the models showed obvious violations.

RESULTS

Hydrography and nutrients

The hydrography in the fjord was influenced by the presence of an inner sill which separates the fjord into two basins and restricts the deepwater flow along the fjord. This was evident from a large temperature difference in bottom waters between basins (Fig. 4B), where the main basin held bottom water of ~ 6°C, while the inner basin bottom waters were ~ 0°C. The salinity of bottom waters in each basin was also different, with the main basin having a salinity of < 31 PSU and the inner basin a salinity of > 32 PSU (Fig. 4A). Surface water temperatures along the fjord were ~ 6°C at all stations, and salinities ranged from 20 to 29 PSU, with the lowest salinities observed in the inner fjord. There was an upper pycnocline throughout the whole transect (Fig. 4C), sustained by a plume of freshwater entering the inner fjord.

Combined nitrate and nitrite concentrations ranged between 0.04 and 4.13 μmol L⁻¹, while phosphate concentrations ranged between 0.06 and 1.01 μmol L⁻¹ (Fig. 5C). The lowest concentrations were found in surface waters with on average 0.28 μmol L⁻¹ NO₃⁻ + NO₂⁻ and 0.14 μmol L⁻¹ PO₄³⁻. In contrast, silicate concentrations were comparable in surface and bottom waters and ranged between 0.90 and 8.51 μmol L⁻¹ (Fig. 5). The Redfield ratio of N:P:Si, 16:1:15 (Brzezinski, 1985), indicate that there was a consistent surplus of SiO₄⁴⁻ and PO₄³⁻ compared to the nitrogen sources (Fig. 5).

Plankton community

The maximum concentration of chl a was found in the upper 10 m (Fig. 4D) at all stations. The mismatch between surface and bottom Si:P Redfield ratios (Fig. 5A) indicate that primary production was not driven by silicate-using algae, i.e. diatoms. This was corroborated by the chl a fraction > 10 μm, which includes diatoms and only constituted a minor part of primary producers, between 8 and 18% of surface phytoplankton, which ranged from 63 ± 18 μg C L⁻¹ to 125 ± 14 μg C L⁻¹ (mean ± SD at sampling stations, n = 3; Fig. 6A). The surface layer microzooplankton community was dominated by aloricate ciliates (Fig. 6B; Table III), which accounted for 64–89% abundance and 83–97% biomass. Biomasses and densities of microzooplankton taxa can be seen in (Table III).

Ciliates, dinoflagellates and phytoplankton > 10 μm were considered potential food for copepods. Surface layer biomass of copepod food ranged between 40 and 77 μg C L⁻¹ (Fig. 6C) and was dominated by ciliates, 38 ± 16 μg C L⁻¹, followed by phytoplankton > 10 μm, 9.9 ± 2.5 μg C L⁻¹ and dinoflagellates, 2.8 ± 1.5 μg C L⁻¹ (mean ± SD of all stations, n = 7 for ciliates and dinoflagellates and n = 21 for phytoplankton).

The copepod community in the surface layer was dominated by the cyclopoid Oithona similis, with A. longiremis, and Pseudocalanus spp. also found in substantial numbers (Fig. 7A; Table IV). The density of copepods in the surface layer ranged between 442 and 1933 ind. m⁻³ and O. similis constituted 29–83%, while the experimentally investigated species A. longiremis and E. americana constituted 0.0–53% and 0–0.2%, respectively. Further information on abundance and biomass for other taxa from the fjord can be seen in Table IV. At st. 1 and 2 the average density of copepods in the whole water column, 3 942 and 5 302 ind. m⁻³, respectively, was much higher than in the surface layer (Fig. 7B and 7A), 1933 and 442 ind. m⁻³, respectively. This illustrates a sub-pycnocline peak in the number of copepods, and O. similis accounted for 89% of the water column abundance at both stations. At st. 3–7, the average density of copepods in the water column was much lower than at st. 1–2 and below that in the surface layer, indicating an aggregation of especially the smaller copepod species in and above the upper pycnocline. The comparison between surface layer and water column copepod densities (Fig. 7A and 7B) also illustrate a higher contribution of Pseudocalanus spp. below the pycnocline, especially at st. 3 and 5, where they accounted for 48 and 54% of the water column abundance, respectively. A. longiremis and E. americana constituted 1–32% and 0–0.2% of water column copepod abundance, respectively. The copepod biomass in the surface layer ranged between 0.21 and 1.35 mg C m⁻³ (Fig. 7C), and while O. similis was still the largest contributor, 12–58%, its small size resulted in the other species increasing in relative importance (Fig. 7; Table IV). The average copepod biomass in the water column was highest at st. 2 and 3 (Fig. 7D), 4.9 and 5.5 mg C m⁻³, respectively, with Calanus spp. accounting for 33 and 78%, respectively. The biomass contribution from Calanus spp. was much larger in the whole water column than the surface layer at st. 2 and 3, which indicate an aggregation at depth in the cold-water bulk inside the midway sill (Fig. 4B). A. longiremis comprised 0.2–35% of surface layer copepod biomass but had a lower relative contribution to biomass in the whole water column (Table IV). E. americana was only found sporadically at the stations closest to the ocean, though also in Ulkebugten ca. 20 km from the fjord.

Chaetognaths contributed most to non-copepod mesozooplankton biomass. Even though abundance was low, its large size resulted in an average biomass comparable to Calanus spp. below the pycnocline (Table IV).

Overall, the zooplankton community in the surface layer was dominated by protozooplankton. Especially aloricate ciliates, which on average accounted for 99% of the estimated surface layer secondary production (Fig. 8; Table I).

Impact of temperature and salinity on functional response

At the observed environmental conditions during our survey, 6°C and 29 PSU (Fig. 4B and 4A), A. longiremis had a significantly lower I_max than E. americana, 0.17 and 0.40 μg C (μg C cop)⁻¹ day⁻¹, respectively (Fig. 9C and 9D; Table V). The two species did not differ in their ability to utilize low food concentrations, K_m, but estimates of this parameter were highly variable in all experiments, as can be seen by the wide confidence intervals (Table V). The future climate scenario of decreased salinity did not significantly affect either I_max or K_m for any of the species (Figs 9C and 8D; Table V). Furthermore, salinity in combination with food concentration only explained 9 and 13% of the variability in ingestion for A. longiremis and E. americana, respectively (Table V).

Temperature increase had a more pronounced effect, as both A. longiremis and E. americana had a higher I_max at 12°C compared to 6°C (Fig. 9A and 9B), though only conclusively significant for A. longiremis. At 12°C, A. longiremis had a higher I_max than E. americana, 0.60 and 0.53 μg C (μg C cop)⁻¹ day⁻¹, respectively, though not significantly different. While A. longiremis’s I_max increased 263% from 6 to 12°C, it increased 33% for E. americana. In the tested temperatures, A. longiremis had an I_max-Q₁₀ of 8.6, while E. americana had an I_max-Q₁₀ of 1.6 (Table V). Making the most conservative comparison of Q₁₀’s between the species based on the I_max confidence intervals, the Q₁₀ of A. longiremis was significantly higher than for E. americana. The half saturations were not significantly affected by temperature for neither species but again, conclusions on this parameter become somewhat dubious due to wide confidence intervals. Temperature effects on the two copepod species were also evident in the ANCOVA, explaining 65% of the variability in specific ingestion for A. longiremis and 25% for E. americana (Table V).

DISCUSSION

Hydrography

Kangerluarsuk Ungalleq is an interesting study area, because it has characteristic features such as terrestrial freshwater inputs, an isolated inner basin and a main basin influenced by shelf water, in a relatively confined space. The cold and salty bottom water in the inner basin (Fig. 4) was likely formed by brine rejection during winter as observed in Storfjorden, Svalbard (Haarpaintner et al., 2001), and the sustained deep density gradient indicate that bottom waters here were more isolated compared to the main basin. This implies that vertical mixing was low in the inner basin, and therefore that primary production was sustained by horizontal inputs from the main basin and land, which was further emphasized by st. 2 and 3 having the highest bottom concentrations of nitrogen (Fig. 5). The generally low nitrogen concentrations in the surface layer, below the Redfield ratio compared to phosphate and silicate at all stations, indicates that phytoplankton was nitrogen limited, as also suggested for late summer in Young Sound in August (Rysgaard et al., 1999) and Disko Bay (Nielsen and Hansen, 1999). In contrast to more populated regions, the freshwater from remote Greenlandic catchments is typically low in concentrations of fixed nitrogen while having phosphate and silicate concentrations comparable or higher than seawater (Skovsholt et al., 2020). It should be noted that in early autumn, one can expect primary production in surface waters to be supported by remineralization of nitrogen (Svensen et al., 2019), but we did not measure ammonium during our survey.

Plankton community

Protozooplankton as the dominant grazer during autumn (Fig. 8) is in accordance with Levinsen and Nielsen (2002), who found protozooplankton to constitute ca. 75% of September zooplankton biomass in Disko Bay. In contrast, Seuthe et al. (2011) found mesozooplankton biomass to exceed protozooplankton, but as they acknowledge, they do not consider mixotrophic species like Laboea strobila, which are included in both the present study and Levinsen and Nielsen (2002). Our estimated protozooplankton clearance rate of 0.54 day⁻¹ in the surface layer (Table I) suggests strong grazing pressure when compared to Coello-Camba et al. (2015), who report an experimentally derived growth rate of 0.3 day⁻¹, for an Arctic fjord phytoplankton community at 6°C. Therefore, protozooplankton grazing together with nitrogen limitation, were likely the main regulators of phytoplankton growth during early autumn in the fjord. The copepod biomass found in our study is comparable to September observations from the upper 50 m in Disko Bay, Greenland (Hansen et al., 1999; Madsen et al., 2008), but abundance lower by an order of magnitude compared to September samples from Godthåbsfjorden (Arendt et al., 2013). Many studies highlight the importance of the harpacticoid Microsetella norvegica for copepod communities in Arctic fjords (Arendt et al., 2013; Svensen et al., 2018), but during our survey in 2023 we did not observe a single individual. In our correction-sampling in 2024, M. norvegica was abundant and under-sampled, but post-nauplii stages were consistently retained by the 200 μm WP-2 net, so we conclude that M. norvegica was not present in the fjord during autumn 2023. However, small copepods are important in the Arctic (Stuart-Lee et al., 2024), and our 62% increment in copepod abundance after the sampling correction emphasizes the need for including small mesh sizes, when sampling Arctic mesozooplankton.

The aggregation of Calanus spp. in the deep parts of st. 2 and 3 was likely caused by entrapment in the cold water inside the midway sill, where they might have entered diapause, while most individuals in the main basin were carried out of the fjord to the deeper waters off the shelf. The ecological role of Calanus spp. in the fjord during autumn could therefore be different from the other copepod species, because they do not feed during diapause (Hirche, 1983). This would push the relative estimated secondary production even further towards protozooplankton. However, the presence of Calanus spp. is still important, because their size and large lipid storage make them high value prey for fish larvae compared to the smaller copepods (Silberberger et al., 2021). Like Calanus spp., Chaetognaths mainly contributed to the mesozooplankton biomass below the surface layer (Table IV). Chaetognaths feed on mesozooplankton (Falkenhaug, 1991) and have previously been reported to affect copepod biomass in an Arctic fjord (Patuła et al., 2023), so they might have been a strong predator on inactive, energy rich Calanus spp. Our experimentally tested copepod species held quite different roles in the autumn food web. A. longiremis constituted a minor but consistent part of the copepod community, similar to previous observations in coastal Arctic (Madsen et al., 2008; Nielsen and Andersen, 2002). E. americana was only sporadically present in 2023, and we did not observe any individuals during our correction sampling in 2024. The species has been observed along the eastern coasts of USA and Cananda as well as Iceland (Moon et al., 2016), so it could potentially have come to Greenland through advection, or by anthropogenic introduction.

Our estimated carbon budget (Fig. 8) was based on zooplankton samplings from the surface layer, and thus does not represent the whole water column. We expect the relative contribution from copepods to zooplankton carbon turnover to be higher below the surface layer, as copepod biomass was on average higher there, whereas microzooplankton biomass is expected to follow the chl a profiles and decrease towards the bottom (Nielsen and Andersen, 2002). However, as the surface layer held most primary production (Fig. 4D) this is likely where most secondary production occurs, and therefore the best representative layer. The assumption that the copepods were feeding at maximum clearance rate in the fjord was not supported by our experimental results, where the half saturations of the functional responses in 6°C for both species were below the food concentration in the fjord, 51 μg C L⁻¹ (Fig. 6C). However, the shape of a functional response curve is highly dependent on prey organisms (Saage et al., 2009), and we expect the average prey in the fjord to be less accessible compared to our experiment, hence clearance rates pr. μg available food might be higher in the fjord than in our laboratory experiments.

Experiments and future perspectives

What we test experimentally in this study is the acute response to environmental changes. De Juan et al. (2023) showed that copepods can have an immediate response to temperature rise, while over several generations they can adapt and return to their pre-stressed state. Therefore, our results are more environmentally relevant for extreme and sudden events like marine heatwaves, which have become more frequent and intense in the Arctic throughout the past decades, especially in the autumn (Huang et al., 2021), than they are for long term, gradual changes in temperature and salinity.

The lack of response to salinity changes for both copepod species in the experiments, (Figs 9C and 8D) was expected, as both A. longiremis (Dutz and Christensen, 2018) and E. americana (Hoffmeyer et al., 2009) are known to be euryhaline. This was also demonstrated by mostly finding A. longiremis in the brackish surface layer in our survey (Table IV). A. longiremis responded strongly to changes in temperature with an I_max-Q₁₀ of 8.6, which is high for copepods (Tyrell and Fisher, 2019). Ingestion and egg production are strongly correlated for copepods (Kiørboe et al., 1985), so in isolation, the high I_max-Q₁₀ of A. longiremis should be a competitive advantage at community level in a warming environment. However, as I_max but not K_m was significantly different for A. longiremis between 6°C and 12°C, a temperature induced increase in ingestions is food dependent. Two field studies showed opposing trends in the relationship between temperature and the success of A. longiremis. Dvoretsky and Dvoretsky (2023) found a significant positive correlation between temperature and A. longiremis abundance in the Russian Barent Sea, temperature regime 2–12°C, while Winans et al. (2023) found strong reductions in A. longiremis biomass in abnormally warm years in Puget Sound, USA, temperature regime 7–16°C. Under the average observed copepod-food concentration in the fjord during our survey, 51 μg C L⁻¹ (Fig. 6C), A. longiremis ingestion in our experiments increased 102% from 6°C to 12°C (Fig. 9B), while at half and double food concentration the increase was 57% and 151%, respectively. This illustrates the high dependence of ingestion on food availability, and indicates that whether a high I_max  -Q₁₀ is an advantage for A. longiremis depends on its ability to fulfill increased metabolic needs. Kiørboe et al. (1985) and Morata and Søreide (2015) demonstrated that respiration also varies with food availability, for Acartia tonsa and Calanus glacialis, respectively. Therefore, the metabolic aspects should be resolved experimentally, in order to quantify niches where A. longiremis could be successful. Generally, our results suggest that a future temperature rise will benefit A. longiremis in productive areas, whereas it will be a detriment in oligotrophic environments. Whether or not climate change gives A. longiremis a competitive advantage also depends on how other autumn species respond. A strong competitor to A. longiremis is O. similis, which was the most abundant copepod during our survey, and has already been proposed to become increasingly important in the Arctic (Balazy et al., 2021). Based on our results, the non-indigenous E. americana does not pose an immediate threat to the Arctic copepod communities, but we suggest that future monitoring efforts screen for this species. We further suggest that Arctic monitoring expand the focus on fjords, to increase our spatial and temporal baseline understanding, and the possibility of identifying environmental changes. Lastly, we advise to include a small mesh size, e.g. 45 μm, when sampling mesozooplankton in the Arctic.

CONCLUSION

We found the fjord plankton community in early autumn to be defined by the smaller size fractions. Small phytoplankton dominated the primary producers, and zooplankton was mainly comprised protozooplankton, both in terms of biomass, estimated grazing and secondary production. The two most important copepods in the fjord were O. similis and Calanus spp., the former was the most abundant and the latter accounted for most biomass, especially in the cold, salty bottom water inside the midway sill. We showed that the small, freshwater tolerant copepod, A. longiremis, increased grazing instantaneously in response to higher temperature, while the non-indigenous E. americana did not have the same immediate response. Whether temperature rise can be a competitive advantage for A. longiremis depends on its ability to satisfy increased metabolic demands, and how its competitors like O. similis adapt. Our results do not suggest that E. americana poses a threat to the native community of Arctic copepods.

ACKNOWLEDGEMENTS

We would like to thank Bo Lings, Alex I. Rump, Gilbert Nimskov and Jacob E. Vedelsby for their assistance with sampling. Professor Colin Stedmon is thanked for help with nutrient measurements and discussion on interpretation of the hydrography and nutrient distributions.

FUNDING

This work was supported by the European Union’s Horizon 2020 research and innovation program under grant agreement no. [869383], ECOTIP-Ecological tipping cascades in the Arctic Seas and the Research Council of Norway through the project Climate Narratives no. [324520].

DATA AVAILABILITY

Data from both the field study and the experiments can be acquired by contacting the corresponding author.

REFERENCES