PAR and Turbidity
In situ data from the AWIPEV-COSYNA underwater observatory close to the estuary of the Bayelva river were analyzed over time and used as a proxy for the development of the underwater light conditions in Kongsfjorden. Residual calculations used for the statistical analysis of changes over time are shown in Figure 11 and associated absolute values are presented in Appendix 8.
Turbidity significantly (p < 0.001) increased over time with an average numerical increase of 0.104 FTU units per year (slope per week = 0.002 FTU * average numbers of weeks per year = 52). Even more prominent as the absolute average increase in turbidity per year, however, was the change in the extreme values of turbidity. Starting in 2016, the positive residuals in turbidity significantly increased until 2020 with maximal values in 2019 of up to 70 FTU. In 2021, lower values, similar to the pre-2016 phase were observed. Contrary to the turbidity, the average PAR values per week significantly decreased over time (p < 0.01) from 2017 to 2021 with a numerical value of - 0.29 µmol m-2 s-1 per year. Similar to turbidity, not only the absolute numerical PAR values per week changed but also the seasonal character of the phases with lower photon fluence rates seemed to change. This became especially prominent in 2020 when PAR values lower than the expected mean were measured over the entire year.
4. Discussion
The present investigation indicates that Arctic kelp species are differently affected by the strong changes in environmental conditions that are prevailing along Svalbard’s coasts. Within the relatively short time period of 25 years, a community which used to be dominated by ‘Digitate Kelps’ transformed to an Alaria esculenta dominated kelp forest. This change was reflected in high biomasses, leaf area indices and adult densities of A. esculenta at 5m and 10m, depth levels from which the other two kelp species, Saccharina latissima and ‘Digitate Kelps’, retreated over time. However, our time series also showed that all investigated kelp and kelp-like species, including A. esculenta , decreased their depth distribution and abundance along the depth gradient throughout the years. Additionally, the biomass maximum and the whole kelp forest progressively shifted upwards to the uppermost depth level at 2.5m. This remained the only habitat in which all three prevailing kelp species still had a balanced age structure between juveniles and adults of different age classes characterizing a mature kelp forest while this relationship decreased with depth.
Svalbard experiences considerable impacts of global warming and Kongsfjorden is an Arctic fjord with numerous glaciers in transition (Bischof et al., 2019b). In July and August 1997 temperatures varied around 4°C at Hansneset in the water column down to 20m (Hanelt et al., 2001). In contrast, between 2016 and 2021 the monthly median ocean temperature at 11m depth was 6.1°C in August at Ny Ålesund on the southern coast of Kongsfjorden (Gattuso et al., 2023). At the same site the average water temperature in the surface layer (upper 10m) even reached a maximum of 8.4°C in summer 2020 (https://dashboard.awi.de/?dashboard=2847). Over the past decades warming air and water temperatures led to a severe decline of the seasonal sea ice extent and thickness so that in recent winters only the northern part of the inner bay was covered by thin sea ice (Pavlova et al., 2019; Payne & Roesler, 2019; Maturilli et al., 2019). This elongation of the open water period leads to an extension of the vegetation period which theoretically promotes kelp forest depth extension as long as it is accompanied by improved water transparency (Castro de la Guardia et al., 2023). However, our results indicate that this potentially positive effect of less ice scouring and reduced sea ice cover for kelp communities gets overshadowed by the counteracting effects of sediment plumes occurring as a consequence of increasing glacial melt (Niedzwiedz & Bischof, 2023; Payne & Roesler, 2019). Geyman et al. (2022) showed that glaciers on Svalbard (including Kongsfjorden area), retreated substantially over time as a response to warming summer temperatures. The increasing subglacial meltwater discharge of sea-terminating glaciers is suspected to be the main source of the increasing sedimentation in Kongsfjorden (Svendsen et al., 2002). With their long-term analysis of satellite images Konik et al. (2021) revealed that Kongsfjorden experiences the phenomenon of “coastal darkening” as water transparency considerably decreased between 1997-2019. We were able to confirm this trend with in situ measurements from the AWIPEV-COSYNA underwater observatory and provide evidence that the turbidity of the water column has increased over time at this coastal site while light availability for macroalgal photosynthesis decreased. The observed lower turbidity values in 2021 may have occurred due to the comparatively colder spring and summer temperatures in the marine Kongsfjorden ecosystem (https://dashboard.awi.de/?dashboard=2847). However, the location of our sensors at the outflow of the Bayelva river is not geographically close to our study site Hansneset and can therefore only serve as a proxy for the general trend of decreasing light levels with increasing glacial melt throughout Kongsfjorden.
Experimental studies have shown that most kelp species present in Kongsfjorden are capable of coping with increasing water temperatures if they do not surpass 10°C (Diehl & Bischof, 2021; Franke et al., 2021; Tom Dieck, 1993). Thus, the observed increase in summer seawater temperatures (Payne & Roesler, 2019) seemingly do not directly account for the observed changes in kelp structure. However, laboratory studies with early life stages by Zacher et al. (2019) indicated a potential competitive advantage of A. esculenta over L. digitataunder future Arctic warming. A. esculenta outcompeted L. digitata due to higher growth rates when the two species were co-cultivated at ambient (5°C) and elevated (9-10°C) summer temperatures but not at 15°C where A. esculenta gets close to its upper temperature limit (Zacher et al., 2019). The assumed North Pacific origin of the brown algae order Laminariales together with their relatively recent introduction to the Arctic after the last glaciation might be the reason for the generally high temperature tolerances in Arctic kelps (Lüning 1990; Tom Dieck 1993; Adey et al., 2008).
A decreasing annual light budget and the direct and indirect effects of sedimentation are most likely the main abiotic factors causing the observed changes in community dynamics and upwards shift of the kelp forest (Fragkopoulou et al., 2022; Smith et al., 2022). Ecophysiological studies showed that an increase in turbidity and sedimentation can have negative effects on photosynthetic rates of adult kelps (Roleda et al., 2008), germination capacity of spores as well as recruitment success of juvenile kelps (Roleda, 2016; Zacher et al., 2016) and thus on the overall productivity of Arctic kelps. A. esculenta spore germination and sporophyte recruitment thereby were less susceptible to sediment loading than L. digitata and S. latissima (Zacher et al., 2016). Similarly, Niedzwiedz & Bischof (2023) reported that under the current abiotic conditions in Kongsfjorden, with low underwater light availability and enhanced temperatures, A. esculenta is in advantage. In contrast to S. latissima , A. esculenta exhibited low compensation irradiance together with low dark respiration rates and a high carbon content independent of temperature treatments (3°C and 7°C), which support our in situ data (Niedzwiedz & Bischof, 2023). Even though experiments by Diehl & Bischof (2021) suggested that S. latissima is especially able to acclimate to an increase in temperature and nutrients as well as a decrease in salinity caused by Arctic glacial melt water, this did not seem to be a competitive factor shaping the current kelp forest at out study site.
Between 1996 and 1998, Hop et al. (2016) extensively investigated the macroalgal distribution at five sites along the axis of Kongsfjorden by combining quantitative destructive samplings with video transects. The authors report that the gradient in abiotic environmental conditions was reflected in highest macroalgal biomass at the outer fjord locations Kapp Mitra and Kapp Guissez while deepest macroalgal coverage was recorded in the inner fjord at Hansneset (Hop et al., 2016). Since than other studies on macroalgal communities in Kongsfjorden were mostly qualitative, except for Bartsch et al. (2016). According to the hydroacoustic mapping study from 2007, Kruss et al. (2017) showed that the coastline along the southern shore of Kongsfjorden was nearly exclusively covered with large macroalgae until 15m water depth, while further down to 30m macroalgal cover became much less. A video survey performed in summer 2009 (Schimani et al., 2022) also indicated dense kelp forests in the center of Kongsfjorden, including Hansneset, down to 30m. The latter authors assumed that kelp distribution along the fjord axis is controlled by the exposure to glacial melt and available hard substrata (Schimani et al., 2022).
Kelp forests were also investigated in other Svalbard fjords and macroalgal communities showed to be individually shaped by site specific physio-chemical conditions. A hydroacoustic investigation supported by underwater videos compared Isfjorden on the warm west coast of Svalbard and Storfjorden on the colder east coast (Wiktor et al., 2022). Macroalgal communities in both fjords were similar but macroalgal bottom coverage in water depths above 6m was considerably less in Storfjorden than in Isfjorden, assumingly due to higher ice scouring pressure in the colder Arctic fjord (Wiktor Jr et al., 2022). The observed pattern in Storfjorden may be similar to the Hansneset kelp forest from 1996/98 (Hop et al., 2012), whereas the current state (this study) might correspond more to reports from the warmer Isfjorden, also opening to the west coast, with highest macroalgal coverage at shallow water depths. In 2021 the variation between the single collected replicates from 5m and below was small, indicating a homogenous and undisturbed macroalgal community. In contrast, at 2.5m the samples were largely different, as one replicate was dominated by A. esculenta , one by ‘Digitate Kelps’ and the third was mixed but contained fewer adult kelps. This heterogeneity of replicates in the shallow subtidal indicates a heterogenous community exposed to ice scouring. Overall, the observed increase in macroalgal biomass at 0m and 2.5m between 1996/98 and 2012/13 as well as the present investigation provides additional evidence that the reduction of ice scouring pressure continued in 2021 (Bartsch et al., 2016; Hop et al., 2012).
In Hornsund, at the southern tip of Svalbard, kelp forest communities along the fjord axis reflect a strong gradient in abiotic conditions with varying distance to the glacier front (Ronowicz et al., 2020). Compared to Hansneset, total kelp biomass at the Hornsund sites were much lower at 5m and similar at 10m, except for the glacier free site where kelp biomass interestingly increased with depth (Ronowicz et al., 2020). In contrast to our study, ‘Digitate Kelps’ and S. latissima were prominent in the glacially exposed kelp forests of Hornsund whereas A. esculenta was only present at a site characterized by high water transparency (Ronowicz et al., 2020). Along the coastline of the Eastern Canadian Arctic Filbee-Dexter et al. (2022) reported a positive correlation between the elongation of the open water period and kelp biomass. In contrast to the biomass distribution along the depth transect at Hansneset the latter authors observed an increase in kelp biomass with depth from 5m to 15m across their 55 sites indicating that the influence of ice scouring at lower depths was much more pronounced than in Kongsfjorden. However, our findings of decreasing kelp biomass along the depth gradient were congruent to Smith et al. (2022) who reported a similar pattern for Laminaria hyperborea kelp forests in the U.K. which was strongly shaped by decreasing underwater light availability.
The Arctic kelps investigated in our study possess a differential strategy in biomass accretion and thereby carbon allocation to perennial structures of holdfast and stipe and annual formation of blades. While adult ‘Digitate Kelps’ invested most biomass in their holdfast and blades, S. latissima and A. esculenta individuals expressed highest stipe biomasses. These ecological differences between the three prevailing kelp species have diverse consequences. When abundance and dominance relationships of macroalgal species at Hansneset change over time, the 3D structure of the kelp forest and therefore the habitat conditions for associated species shift accordingly. Epifaunal biodiversity is highest in kelp holdfasts compared to blades and lowest on stipes, but even though species richness is consistent between ‘Digitate Kelps’, S. latissima and A. esculenta the most commonly associated species vary in a kelp specific manner (Włodarska-Kowalczuk et al., 2009). Consequently, the continuous alteration in biotic and abiotic factors at our study site has indeed already influenced the fauna inhabiting the kelp forest as species abundances, taxonomic composition as well as biomass distribution varied over time (Paar et al., 2016; Niklass, 2022). Paar et al. (2016) showed that the biomass and secondary production of associated macrozoobenthos is strongly associated with macroalgal depth distribution as both parameters were highest in the upper most sublittoral in 2012/13, which represented an inverted pattern compared to 1996/98. At greater depths, where dominant kelps are absent, other macroalgae species likeDesmarestia aculeata , Ptilota spp. and Phycodrys rubens also strongly promote epifaunal communities (Lippert et al., 2001; Włodarska-Kowalczuk et al., 2009), indicating their often overlooked important role for Arctic benthic ecosystems. Furthermore, higher trophic levels like fish are influenced by the structure and bottom coverage of habitat forming macroalgae (Brand & Fischer, 2016).
The Arctic flora is characterized by only a small amount of endemic Arctic species compared to those with a wider cold-temperate to Arctic distribution. Wilce (2016) identified only 21 of 161 macroalgal species to be Arctic endemics. In Kongsfjorden, this relation is even smaller as has been outlined by Hop et al. (2012) who noticed that one half of the macroalgal species encountered were truly Arctic to cold-temperate while the other half had even wider distribution ranges and there were only four Arctic endemic species including the kelp Laminaria solidungula . Especially this kelp may be negatively impacted by warming waters (Tom Dieck, 1992; Roleda, 2016). We observed that this rare species which had only been present with very small individuals at Hansneset (Bartsch et al., 2016; Hop et al., 2012) was not encountered anymore in our quantitative 2021 samples. However, in other Svalbard fjords L. solidungula is still present (Ronowicz et al., 2020; Wiktor Jr et al., 2022). In future, macroalgal species distribution ranges are predicted to shift northwards with an increasing number of Atlantic species potentially spreading into the warming Arctic while Arctic species retreat (Fredriksen et al., 2019; Krause-Jensen & Duarte, 2014; Kortsch et al., 2012; Weslawski et al., 2010).
Kelp forests contribute strongly to the net primary production and the coastal carbon cycle as carbon is fixed in their biomass through photosynthesis (Krause-Jensen & Duarte, 2016; Pessarrodona et al., 2022; Smale et al., 2022). Pessarrodona et al. (2022) highlighted their ecological importance by stating that the global net primary production of subtidal seaweed forests even exceeds coastal phytoplankton productivity. Smale et al. (2016) reported a wide average carbon standing stock of 721 g C m-2 at 5m for Laminaria hyperborea kelp forests along the coast of the U.K. Their conversion of biomass to carbon stock was based on the assumption that carbon content makes up ~30% of kelp DW which included also the holdfast and stipe DW (Smale et al., 2016). In contrast, we investigated blade carbon content (Figure 10) and showed that the carbon content differs in a kelp species specific manner (A. esculenta 34%, ‘Digitate Kelps’ 27%, S. latissima 32%). When applying the same calculation as Smale et al. (2016) to the overall kelp DW collected at Hansneset (Appendix 4) we estimated 489.6 g C m-2 at 2.5m and 190.2 g C m-2 at 5m. Both values are lower compared to the average of the U.K. kelp forests (Smale et al., 2016). But most importantly our study showed that carbon and nitrogen allocation strategies significantly vary between kelp species as has already recently been shown for Alaria marginata and S. latissimafrom Alaska (Umanzor & Stephens, 2022) and cold-temperate kelp species (Gilson et al., 2021). Consequently, the contributions of each kelp species to the overall carbon standing stock in kelp forests can vary according to the relative carbon content in biomass. In this respect the observed change from a ‘Digitate Kelps’ forest into an A. esculenta kelp forest reveals a higher potential for carbon allocation in A. esculenta blades compared to ‘Digitate Kelps’ blades. This may have even wider consequences for the whole carbon budget of the surrounding waters as there is a continuous release of dissolved organic carbon (DOC) from kelps into the surroundings (Weigel & Pfister, 2021) which may thereby have changed the DOC budget of Kongsfjorden within the last decade.
In Arctic fjords detached macroalgal detritus is transported to deeper locations where they may support secondary production or biological carbon sequestration (Cui et al., 2022; Schimani et al., 2022). Even though not significantly different, mean blade biomass of ‘Digitate Kelps’ (21.4 g DW) from Hansneset was nearly double compared to A. esculenta (12 g DW) in 2021 (Figure 8). Considering that kelp blades decay over the seasons and storm waves fragment or even detach the thalli (Krumhansl & Scheibling, 2012) it is likely that the change in species dominance leads to less carbon being exported for local carbon sequestration. Together with the kelp forest retreat at our study site, this might resemble the predicted decline in kelp forest contribution to marine carbon cycles under the negative impacts of increasing water turbidity (Blain et al., 2021). However, the potential contribution of kelp forests in general to natural carbon sequestration remains a controversially discussed topic in current research (Hurd et al., 2022; Krause-Jensen et al., 2022; Smale et al., 2022; Pedersen et al., 2020).