Climate driven changes in atmospheric, ice and oceanic conditions are likely resulting in shifts in the biogeochemical processes and microbial communities at pan-Arctic scales. However, a comprehensive understanding of regional trends and potential disparities among regions is hampered by the scarcity of long-term datasets of observed or modelled conditions (CAFF, 2017; Juranek, 2022). A set of studies describing the different hydrographic functioning and ecosystem production have revealed a wide variety of properties and pathways in the carbon cycle across the Arctic (Dunton et al., 2006; Grebmeier et al., 2006; Hirche et al., 2006; Hop et al., 2006; Michel et al., 2006; Rysgaard and Nielsen, 2006; Schmid et al., 2006; Tremblay et al., 2006; Wassmann et al., 2006a). These carbon cycles vary in efficiency depending on the processes transforming organic matter into nutrients (carbon mineralisation), and nutrients into organic matter (primary productivity). Productivity on the Arctic shelves is generally considered to be limited by the concentrations of available nitrogen compounds, especially in the Pacific Arctic (Mills et al., 2018; Ko et al., 2020), rather than by irradiance limitation from sea-ice and photoperiod (Tremblay and Gagnon, 2009; Tremblay et al., 2015). Most of the nutrient supply is advected through inflow shelves (Le Fouest et al., 2013b; Torres-Valdés et al., 2013; Tremblay et al., 2015). River inputs into sub-Arctic inner shelves may also be an important local source of nutrients and organic matter from land (Dittmar and Kattner, 2003; Dunton et al., 2006; Carmack et al., 2016), and estimates suggest terrestrial sources with coastal erosion, contribute to 28%–51% of net primary productivity in the Arctic Ocean via the Transpolar Drift (Le Fouest et al., 2013a; Terhaar et al., 2021), but see contrasting results in Le Fouest et al. (2013a).

At the pan-Arctic scale, rising temperatures and prolonged open water seasons have driven substantial increases (>50% over two decades) in sub-Arctic primary productivity, with earlier and extended blooms of pelagic phytoplankton (Arrigo et al., 2008; Kahru et al., 2011; Kahru et al., 2016; Hill et al., 2013; Hill et al., 2018; Ardyna et al., 2014; Arrigo and van Dijken, 2015). In addition, the decrease in multi-year ice and increase in first-year ice extent are driving a temporary increase in ice-algae habitat (Lim et al., 2022). Production increased in most regions but with varying magnitude (Arrigo and van Dijken, 2015). The Barents Sea was the area that experienced the highest increase in primary production over the last two decades (Dalpadado et al., 2020; Lewis et al., 2020). Primary production has also increased in spatial extent in the Bering Sea during warm years (Nielsen et al., 2024). The Canadian Arctic Archipelago, Baffin Bay and the eastern Greenland shelf display an altered functioning, with changes in productivity spatial-temporal patterns (Michel et al., 2015; Blais et al., 2017; Mayot et al., 2020). Data are scarce for other shelves, but short term studies also revealed interannual variation in production and particulate organic carbon export between years with varying sea ice cover (Lalande et al., 2009; Bienhold et al., 2022).

In recent years, ice dynamics and water temperature no longer suffice to explain increases in primary production on inflow shelves (Lewis et al., 2020). It has been hypothesised that the increased transport of deep Atlantic waters into the Arctic might have increased dissolved inorganic nitrogen (DIN) supply, favouring higher primary production in downstream regions (Henley et al., 2020; Lewis et al., 2020). However, despite increasingly northward Atlantic water flow and subsequent increased vertical mixing, multiple studies point to a long-term decrease in nutrients across most of the pan-Arctic, with the exception of the pan-Arctic slope (Oziel et al., 2022). Oxygen and nutrients seem to have decreased within the halocline of the Eastern and Western Eurasian Basin (Polyakov et al., 2020a). Similarly, there has been a decline in nitrate concentrations in the Atlantic water inflow into the Barents Sea between 1990–2010 (Oziel et al., 2017), and no long-term (35 years) trend in nutrients from Atlantic water flowing into the Arctic north of Svalbard (Duarte et al., 2021). In the Chukchi and Beaufort seas, only oxygen concentration has increased (Polyakov et al., 2020a), while nitrate and phosphate concentrations have decreased by 79% and 29% respectively (Zhuang et al., 2021). This was hypothesised to result from the joint effect of decreased vertical mixing following the strengthened stratification due to the loss of multi-year sea ice, and the increased consumption by primary producers (Zhuang et al., 2021). At interior and outflow shelves, nitrogen discharge from river outputs is usually quickly utilised, for example, through efficient denitrification in the Laptev Sea (Sanders et al., 2022; Tuerena et al., 2022), and nitrate and other dissolved inorganic nutrient fluxes from land seem to have declined (Tank et al., 2023). Lack of observations currently greatly limits our understanding of nutrient transport and budgets (Torres-Valdés et al., 2016).

Changes in primary production can also be linked to changes in phytoplankton communities. Phytoplankton communities affect carbon flux attenuation and production regimes (Ardyna et al., 2011; Wiedmann et al., 2020), and vary with environmental conditions across all Arctic shelves and basins (Ardyna et al., 2011; Min Joo et al., 2012; Lee et al., 2019; Ibarbalz et al., 2023). For example, small algae, more efficient at assimilating nutrients and thus potentially more competitive compared to local larger species, have been observed in higher quantities in the Canada Basin and coastal Canadian Arctic Ocean (Li et al., 2009; Blais et al., 2017), and in the Atlantic water masses west of Svalbard (Lalande et al., 2013). In the Beaufort Sea, microbial communities have transitioned towards taxa characteristic of oligotrophic environments (Comeau et al., 2011). In addition, strong advection at inflow shelves has driven the northward shift of temperate species of coccolithophores (Winter et al., 2014; Neukermans et al., 2018; Oziel et al., 2020). In contrast, there is no evidence of temporal changes in microbial community composition or biomass in the North Water (northern Baffin Bay) or the Laptev Sea (Freyria et al., 2021; Bienhold et al., 2022). In the rest of the Arctic, time series are scarcer, but microbial communities, including phytoplankton, have been characterised by recent studies, setting baselines for the investigation of climate impacts (Demidov et al., 2021; Polyakova et al., 2021; Sukhanova et al., 2021; Liu et al., 2023; Bezzubova et al., 2024). Shifts in microbial communities are likely to drive changes in trophic functioning, affecting the diets, phenology and development of grazing species (Søreide et al., 2010; Daase et al., 2011). Changes in planktonic communities and new environmental conditions can cause harmful algal blooms (HAB), which have direct impacts on the food-webs and human consumption (Okolodkov, 2005; Bruhn et al., 2021). Cases of HAB have been noticeably increasing northward in recent years, with many reports from the Pacific Arctic in particular (Anderson et al., 2021; Anderson et al., 2022; Einarsson et al., 2022; Fachon et al., 2024), with important impacts on top predators, including humans (Hendrix et al., 2021; Lefebvre et al., 2022; Van Hemert et al., 2022).

Changes in the environment and in primary productivity (species composition, spatial and temporal patterns) are propagating through the food web, driving the borealization of the pelagic compartments. Indeed, environmental changes trigger contrasting responses of the boreal and Arctic pelagic communities, due to their strong differences in biogeography, life history strategies and functions. For example, many zooplankton communities dominate successively across seasons and regions, varying in their ecological niches and functional traits (Questel et al., 2013; Balazy et al., 2018; Kimmel et al., 2018; Kimmel et al., 2023; Mańko et al., 2020; Mańko et al., 2022; Hop et al., 2021; Darnis et al., 2022; Mazanowski et al., 2023; Van Engeland et al., 2023; Wold et al., 2023). Large, lipid-rich Arctic copepod species such as C. glacialis and C. hyperboreus can utilise allochthonous production, and survive low productivity periods owing to their large lipid reserves (Hunt et al., 2014; Hirche et al., 2024). Arctic species thus tend to dominate the biomass in late summer/autumn during years with low advection and late sea ice retreat (Kimmel et al., 2018; Spear et al., 2020). However, Arctic zooplankton have a slower increase in respiration rate with rising temperatures compared to their Atlantic counterparts, which implies that, although they might be resilient to warming, Atlantic zooplankton may become more competitive under such conditions (Kaiser et al., 2022). In contrast, boreal species typically consist of smaller zooplankton with less reserves. Along the continental slope of the Atlantic Water inflow, high concentrations of C. finmarchicus prevail (Bluhm et al., 2020). Differing traits of Arctic and boreal zooplankton might be adaptative of their respective environment (Renaud et al., 2018), and a shift from lipid-rich Arctic to more boreal communities might not have strong impacts on plankton-feeders as boreal zooplankton occur in larger numbers, over a longer period of time, and have a faster life cycle. However, reduced abundances of lipid-rich C. glacialis on the Bering Sea shelf have been associated with poor energetic condition and reduced survival of young walleye pollock (Heintz et al., 2013).

Warm water advection and sea-ice retreat are major determinant of zooplankton community dominance across the Arctic (Ershova et al., 2015; Kimmel et al., 2018; Kimmel et al., 2023; Spear et al., 2019; Abe et al., 2020; Dalpadado et al., 2020; Drits et al., 2023; Ramondenc et al., 2023). Krill, gelatinous zooplankton, ichthyoplankton and young-of-the-year pelagic fish also display different life history strategies and habitat preferences and respond strongly to varying advection regimes (Eriksen and Dalpadado, 2011; Eriksen et al., 2015; Eriksen et al., 2020; Oviatt et al., 2015; Benkort et al., 2019; Deary et al., 2021; Skjoldal et al., 2022). Adult pelagic fish are often less directly affected by variations in environmental conditions. Boreal mobile pelagic fish can relocate to follow the production and expand into newly suitable habitats, thus propagating the borealization of the lower trophic levels up the food chain, although other important factors such as recruitment and availability of suitable spawning habitat also have a strong impact on population spatial-temporal dynamics (Hollowed et al., 2013b; Eriksen et al., 2017). In contrast, some Arctic species such as polar cod are very dependent on ice to complete their life cycle, and loss of sea-ice thus drives a loss of essential habitats for those species (David et al., 2016; Huserbråten et al., 2019; Gjøsæter et al., 2020; Marsh and Mueter, 2020; Deary et al., 2021; Geoffroy et al., 2023).

Long-term borealization has thus been reported on for many pelagic compartments. Shifts from lipid-rich Arctic zooplankton to communities of smaller individuals are of concern in western Greenland (Møller and Nielsen, 2020), the Bering sea (Kimmel et al., 2018; Kimmel et al., 2023), the Chukchi Sea (Abe et al., 2020; Spear et al., 2020; Ershova et al., 2021; Mueter et al., 2021a), Labrador, eastern Greenland and the Barents Sea (Trudnowska et al., 2012; Weydmann et al., 2014; Eriksen, 2016; Eriksen et al., 2017; Freer et al., 2022), and the Fram Strait, main gateway of Atlantic inflow into the Arctic (Gluchowska et al., 2017). Subtle changes in species composition may also be underway within neritic species complexes (Ershova et al., 2015). Krill is also shifting in species composition from Arctic to temperate in the Barents Sea (Eriksen and Dalpadado, 2011; Orlova et al., 2015; Eriksen, 2016), Svalbard, and Bering Sea (Hunt et al., 2016). Lack of long time series hampers our ability to detect changes in zooplankton community composition on the other shelves. Similarly, boreal gelatinous species are thought to also extend northward on the inflow shelves (Brodeur et al., 2002; Brodeur et al., 2008; Eisner et al., 2014; Eriksen et al., 2017; Geoffroy et al., 2018). But drivers of their abundance may be complex (Decker et al., 2023) and historical data and from other shelves are limited, making temporal and regional comparisons difficult.

Further up in the food web, the geographical retreat of polar cod due to the loss of sea-ice is often concomitant with the northward expansion of more southern species, such as capelin, saffron cod or walleye pollock in the Atlantic or Pacific Arctic (Hop and Gjøsæter, 2013; Stige et al., 2019; Wildes et al., 2022; Levine et al., 2023; von Biela et al., 2023). Increased abundance of more southern fish species, such as capelin and Pacific sand lance, are also observed at interior and outflow shelves: in the Canadian Arctic Archipelago (Falardeau et al., 2017), in the Hudson Bay (Gaston et al., 2003), and northern Baffin Bay (Ulrich and Tallman, 2021). While mesopelagic fish have been suggested to be limited in their northward expansion because of the light conditions (Ljungström et al., 2021; Langbehn et al., 2022) or a combination of both light and temperature (Chawarski et al., 2022), studies based on observations rather than modelling have partially challenged the light hypothesis by reporting mesopelagic sound scattering layers partly formed of boreal fish species (e.g., glacier lanternfish Benthosema glaciale and juvenile redfish) in the high Arctic (Knutsen et al., 2017; Geoffroy et al., 2019; Snoeijs-Leijonmalm et al., 2022; Ingvaldsen et al., 2023). Moreover, studies looking at both temperature and latitude (a proxy for irradiance) have concluded that temperature fronts might also contribute to hindering the northward range expansion of mesopelagic fish (Chawarski et al., 2022). Data are more limited for the inner shelves of the Kara, Laptev and East Siberian seas, but those ecosystems are often relatively poor in pelagic fish (Hirche et al., 2006; Schmid et al., 2006).

Pelagic borealization is affecting populations and ecosystem functioning. Changes in zooplankton development and community succession along the seasons have been linked to pelagic fish diet and biomass trends (Orlova et al., 2010; Dalpadado et al., 2024), propagating climate effects to upper trophic levels. They also affect first year juvenile fish survival, mainly walleye pollock and polar cod, as lipid-rich species associated with autumn blooms during cold years are a critical source of reserve for the juvenile fish before the winter (Kimmel et al., 2018; Kimmel et al., 2023; Bouchard and Fortier, 2020; Dupont et al., 2020), and as prey size impacts that of its visual predators, as shown by bioenergetic and behavioural modelling studies (Ljungström et al., 2020). In contrast, the sustained advection of krill in the Barents, Chukchi and western Beaufort seas is altering trophic controls and stabilising predator populations (Berline et al., 2008; Eriksen and Dalpadado, 2011; Orlova et al., 2013; Ashjian et al., 2021). Increasing co-occurrence of polar cod with other species such as capelin or Pacific sand lance would lead to more competition for food at inflow and interior shelves, with negative consequences for both species (Orlova et al., 2010; Falardeau et al., 2014; McNicholl et al., 2016; McNicholl et al., 2018; Pedro et al., 2020). Changes in dominant forage fish and zooplankton prey have altered marine mammals’ and seabirds’ migration patterns and diet across the pan-Arctic, sometimes leading to lower growth rate for chicks or altered demography (Provencher et al., 2012; Chambellant et al., 2013; Gaston and Elliott, 2014; Falardeau et al., 2017; Gall et al., 2017; Balazy et al., 2018; Kuletz et al., 2020; de la Vega et al., 2021; Merkel et al., 2021; Ulrich and Tallman, 2021; Descamps et al., 2022; Moore et al., 2022).

Many studies have reported changes in benthic (mostly invertebrates) and demersal (mostly fish) ecosystems across Arctic and sub-Arctic shelves. Signals of borealization are stronger on inflow shelves, where shifts in demersal fish communities have been widely reported, such as in the Barents Sea (Fossheim et al., 2015; Bergstad et al., 2018; Spotowitz et al., 2022), in western Greenland (Post et al., 2021), as well as the northern Bering Sea and, to a lesser extent, the Chukchi Sea (Mueter et al., 2021a; Zhang et al., 2022). An emerging pattern from recent studies is that borealization of inflow regions seems to impact neighbouring areas by spillover effect. In eastern Greenland (Emblemsvåg et al., 2022a; Emblemsvåg et al., 2022b), demersal communities across depths have rapidly reorganised in response to climate change, in particular between 350 and 1000m depth. Specifically, studies have found that Atlantic cod, beaked redfish and deep-sea shrimps (Pandalus borealis) observed on the northeast Greenland shelf originated from the Barents Sea populations (Christiansen et al., 2016; Andrews et al., 2019). In addition, across the Siberian Arctic (Kara Sea, Laptev Sea, East Siberian Sea), North Atlantic species have expanded eastward while North Pacific species have expanded westward (Orlov and Volvenko, 2024).

Borealization in inflow shelves has also been documented for benthic communities, although responses are lagged with regard to climatic signals (Grebmeier, 2012; Jørgensen et al., 2019; Waga et al., 2020; Zakharov et al., 2020; Calvet et al., 2024). Interestingly, the borealization of benthic communities does not necessarily lead to an increase in productivity or biomass (Zakharov et al., 2020 for the Barents Sea; but see Krause-Jensen et al., 2020, for pan-Arctic increases in seaweeds). In contrast, unchanged community biogeography could suggest resilient communities in the Laptev Sea (Kokarev et al., 2017), and to some degree in the Bering and Chukchi Seas, over several decades (Grebmeier et al., 2015a but see Grebmeier et al., 2006; Huntington et al., 2020, for biomass declines). The lack of repeated surveys hinders our ability to evaluate the degree of borealization of benthic communities, especially on interior and outflow shelves, but bycatch data from commercial or scientific bottom trawls can serve as a baseline for future assessments (Jørgensen et al., 2022), and the Distributed Biological Observatory in the Pacific Arctic provides time series for macro- and megabenthic fauna (Grebmeier et al., 2015b; Grebmeier et al., 2018).

At smaller scales, some studies have revealed the borealization of some coastal zones across the Arctic, as suggested by an increase in boreal fish species in lagoons on the Beaufort Sea (von Biela et al., 2023), and by increased macroalgal cover (Krause-Jensen et al., 2020), abundance of boreal species and coincident shifts in benthic invertebrate communities in Svalbard fjords and near Franz Josef Land (Kortsch et al., 2012; Węsławski et al., 2018; Al-Habahbeh et al., 2020; Dvoretsky and Dvoretsky, 2024). However, the degree of borealization of benthic communities in Atlantic Arctic fjords might depend on the bathymetric features of the coastline, such as fjord sills and the interannual strength of the Atlantic Water inflow (Beuchel et al., 2006; Renaud et al., 2007; Jordà-Molina et al., 2023). A more important driver of major changes in Arctic coastal ecosystems and landscapes could instead be the melting marine terminating glaciers (Lydersen et al., 2014; Meire et al., 2017; Williams et al., 2021; Kavan and Strzelecki, 2023). Some sub-Arctic coastal regions do not display any particular response to climate change. For example, western Greenland intertidal ecosystems do not vary with latitude, which would suggest resilience to climate change, based on space-for-time substitution (Thyrring et al., 2021), and in the Baydaratskaya Bay, Kara Sea, benthic communities were unchanged until 2013, and only seemingly impacted by other local anthropogenic impacts since then (Azovsky and Kokarev, 2019).

As in the pelagic systems, demersal and benthic borealization appears to be mainly driven by changes in suitable habitats and prey distribution induced by climate change. Rates of response to environmental changes are often lagged and species-specific (Alabia et al., 2018). At the inflow shelves, the main drivers are the fluctuating extent of colder and warmer water masses and timing of ice retreat. This has been shown in the northern Bering Sea and Chukchi Sea, for both demersal and benthic communities (Grebmeier, 2012; Nishio et al., 2020; Zhang et al., 2022), and in the Barents Sea (Ingvaldsen et al., 2021; Nascimento et al., 2023). In addition to expanding suitable warmer habitats, advected waters transport demersal and benthic early life stages. Variations in water mass extent across interannual to decadal scales have been linked to the varying abundance of advected ichthyoplankton species, leading to variation in dominance regime of fish larvae in the Eastern Chukchi Sea (Randall et al., 2019; Axler et al., 2023), and in the Barents Sea (Eriksen et al., 2015; Eriksen et al., 2017). These fluctuations in water mass inflows are linked to large-scale circulation patterns, such as the sub-polar gyre in the northern Atlantic in western Greenland (Post et al., 2021), or the Arctic dipole in other inflow shelves (Polyakov et al., 2023). While warm conditions are maintained at inflow shelves and neighbouring regions are gradually warming due to local changes in the environment, mobile populations such as fish are likely able to further expand beyond inflow shelves (Andrews et al., 2019; Orlov and Volvenko, 2024), but demersal Arctic fish species living on the edge of the shelves cannot move north as they run out of shelf (Wassmann et al., 2006b; Fossheim et al., 2015). The northward expansion of less mobile boreal benthos is likely more dependent on the dispersion of the larval phase (Ershova et al., 2019; Meyer-Kaiser et al., 2022), although those will may not settle successfully (Wassmann et al., 2015; Ershova et al., 2019; Descôteaux et al., 2022).

Shifts in species distribution can drive changes in both pelagic and benthic-demersal life cycles, oceanic seascapes, population structures and interspecific interactions. In the Chukchi Sea, increasing number of boreal pre-spawning, spawning and post-spawning walleye pollock have been observed in recent years (Emelin et al., 2022). This could indicate an expansion of their spawning location or range but does not necessarily correlate with juvenile survival. Survival of the early life stages depends on species capacity to take advantage of available food and to resist harsh winter conditions. Polar cod, adapted to Arctic conditions, have a unique fat storage strategy, from diatoms and Calanus origin, that enables it to thrive in cold ice-covered region (Copeman et al., 2022). In contrast, juvenile walleye pollock have better condition in the southern than in the northern Chukchi Sea (Copeman et al., 2022), which suggests that they are not yet adapted to colder northern conditions. In the Barents Sea, climate induced changes in temperature and food availability also drive trends in 0-group fish lengths (Skjoldal et al., 2022) and distribution (Eriksen et al., 2017). Apart from their lipid storage strategy, differences in spawning timing between species is likely a key driver of their successful survival during the Polar Night. Indeed, winter surveys around Svalbard show that the spring spawning strategy of boreal fish, which contrasts with winter-spawning Arctic and arcto-boreal species, might lead to a temporal mismatch between their planktonophagous larvae and zooplanktonic blooms (Berge et al., 2015; Geoffroy and Priou, 2020). Large incoming boreal fish are often generalists which are thus able to feed on boreal as well as Arctic prey. This was shown by the diet composition of walleye pollock in the northern Chukchi Sea, where Arctic copepods, euphausids and polar cod were observed in their stomach content (Benzik et al., 2022). Walleye pollock also consume other incoming species, such as Pacific sand lance, whose abundance has been increasing with the decline of sea-ice (Baker et al., 2022). Similarly, in the Barents Sea, ctenophore species, which might benefit from warmer conditions in the region, and snow crab (Chionoectes opilio), which overlaps increasingly with Northeast Arctic cod as these expand north-eastward, increased in frequency in Northeast Arctic cod diets (Eriksen et al., 2018; Holt et al., 2021). However, there is no sign that polar cod are increasing in Northeast Arctic cod’s diets (Holt et al., 2019).

Local changes in environmental conditions, driven by alterations in large scale circulation patterns and fluid (water, air) masses properties, are reported across the pan-Arctic. The strongest impacts seem to affect the inflow shelves, where high connectivity to very different systems result in a high interannual variability and sensitivity to external forcings. Borealization has thus progressed more rapidly on these shelves, affecting all trophic levels (Csapó et al., 2021). Ecosystems inflow shelves are currently transitioning towards a more boreal functioning but retain some Arctic characteristics. The potential degree of borealization in the future is discussed below. A new emerging signal, as revealed by many recent studies reviewed here, is that interior and outflow shelves are also experiencing borealization through climate-induced modifications of local environmental conditions (sea-ice cover, river runoff, temperature) that render habitat suitable for species that are transported or actively relocate there (Figure 6).

However, the extent of this “spillover” effect of borealization across space, time and functional groups is uncertain, due to several knowledge and data gaps. Indeed, a recent biodiversity assessment for the whole Arctic domain has highlighted the lack of data for the Siberian seas (Kara, Laptev, East Siberian seas) and the Canadian Arctic (Canadian Arctic Archipelago, Beaufort Sea, Hudson and Baffin Bay, Alfaro-Lucas et al., 2023). In addition, for all shelves and for the Arctic basins, data are largely restricted to the summer and autumn seasons, with recent but limited knowledge on winter and polar night processes (Berge et al., 2015; Berge et al., 2020; Gerland et al., 2019), and transition periods between ice-free and ice-covered seasons (Shogren et al., 2020). More seasonal and spatial coverage, including land-sea connectivity at more extensive spatial scales than currently conducted, are important for understanding physical processes such as sea ice loss, water freshening, nutrient and carbon cycles (Brown et al., 2020; Shogren et al., 2020; Timmermans and Marshall, 2020), as well as biological processes (Berge et al., 2015; Geoffroy and Priou, 2020). Deployment of dedicated autonomous technologies are often suggested to fill this data gap. In addition, approaches such as participatory modelling tools have a strong potential to fill the knowledge gaps around e.g., winter processes, using local ecological knowledge (Pedro et al., 2023). A more comprehensive overview and monitoring of Arctic ecosystem borealization would also necessitate the allocation of research resources to characterise and better understand processes of relatively understudied compartments such as microbial communities, epibenthos and infauna (CAFF, 2017).

Predicting the extent to which Arctic borealization will continue in the future across ecosystem compartments is a priority that has received much attention from the scientific community. Climatic scenarios from the coupled model intercomparison project (CMIP) suggest various trajectories for future Arctic climatic and oceanographic conditions, but all scenarios display similar trajectories until the 2040s. The Artic will likely continue to warm at a faster rate than mid-latitude systems. Arctic amplification could continue under a low emission scenario, where ice albedo feedback would maintain the phenomenon, but it would be less prominent in a high-emission scenario (Ono et al., 2022). The increase in heat transport from the Atlantic and the Pacific will likely continue and progress poleward through the Bering Strait and the Barents Sea (Drinkwater et al., 2021; Dörr et al., 2024). All shelves will become ice free in the summer, irrespective of the climatic scenario, and the Barents Sea could become ice free in winter by the end of the century (Årthun et al., 2021). Expanding open water areas and increasing precipitation will contribute to large and rapid changes in hydrological cycles (McCrystall et al., 2021). Increased river runoffs should increase stratification and subsurface warming in the central Arctic Ocean (Nummelin et al., 2016).

It is challenging to assess the degree of borealization of biogeochemical processes and microbial communities at pan-Arctic scales. Primary production is arguably the microbial process that has received most attention in the literature, and numerous studies suggest an evolution towards a more boreal functioning in the future (Ardyna and Arrigo, 2020; Noh et al., 2024). The Pacific and Atlantic inflow shelves will likely experience the strongest increase in primary productivity, although this is more uncertain for the western Barents Sea, where oceanographic models project variable directions and magnitudes of change depending on how stratification processes are represented (Mueter et al., 2021b; Mousing et al., 2023). However, increasing cloud cover (Bélanger et al., 2013) and strong seasonality in light conditions at higher latitudes are likely to limit the transition towards a completely boreal phenology, and primary production should retain some characteristics of an Arctic system, which would lead to arctic-boreal hybrid functioning. Yool et al. (2015) projected that, by 2,100, despite an increase in productivity driven by the loss of sea ice, the Arctic will not be more productive than the Atlantic, with both systems entering a low nutrient regime. In contrast, the Pacific, richer in nutrients, might experience a longer rise in productivity (Yool et al., 2015). However, the trajectory of nutrient limitation remains a significant knowledge gap, leaving predictions of Arctic productivity in the future uncertain (Slagstad et al., 2015; Mueter et al., 2021b; Mousing et al., 2023). While observational data on other microbial processes across the pan-Arctic remain limited, experimental studies offer valuable insights into potential shifts in crucial microbial functions. For example, temperature variations might influence bacterial metabolic rates, phytoplankton community structure and chlorophyll a concentrations and, consequently, carbon cycling (Kritzberg et al., 2010).

The climate driven environmental changes will continue to trigger distributional shifts and changes in biodiversity across functional groups and Arctic regions. These shifts are quantitatively explored based on species distribution models fitted to species’ current thermal preferences. Ensemble modelling predicts an increase in phytoplankton species richness all over the globe under the highest warming scenario, by the end of the century, except in the Arctic, where no trend is expected (Benedetti et al., 2021). Other projections with similar time horizons suggest that poleward shifts in zooplankton distribution could continue at a pace of 8.7 km/decade, accompanied by a change in seasonal occurrence peak up to 13 days earlier, a high species turnover, and a northward shift of many boreal gelatinous zooplankton, sometimes only into sub-polar regions (Villarino et al., 2015; Benedetti et al., 2021; Pantiukhin et al., 2023; Pantiukhin et al., 2024; Cheung et al., 2015) also predicted a northward range expansion for Pacific pelagic fish species of 30.1 ± 2.34 (S.E.) km decade⁻¹. However, regional predictions for Canadian waters predict a decline in biomass in both the Pacific and Atlantic sides (Bryndum-Buchholz et al., 2020). In contrast, projections under a high emission scenario reveal an increase in total forage fish biomass in Hudson Bay, but with decreasing fish size and a shift from lipid rich polar cod to smaller boreal fish (Florko et al., 2021). Overall, responses to climate change are likely region-specific, as well as species specific and are thus challenging to predict (Mueter and Litzow, 2008; Alabia et al., 2018; Husson et al., 2020; Calvet et al., 2024). The impact of climate change on benthic habitats may depend on the spatial scale of the projection. In the Bering and Chukchi seas, the habitat of many assemblages of epibenthic invertebrates is predicted to be reduced by 50% by 2050 and almost 100% by 2,100, with the exception of eurythermal species such as the basket star Gorgonocephalus cf. arcticus (Logerwell et al., 2022). At the pan-Arctic scale, however, habitat change for benthic species might be limited in the future (Renaud et al., 2019). Habitat loss should mainly affect currently cold and aragonite-oversaturated areas such as northern Barents Sea and east Greenland Sea (Renaud et al., 2019), and cold-stenothermal species such as gastropods and bivalve mussels Musculus spp. in the Bering and Chukchi Sea, leading to large losses of biodiversity locally (Logerwell et al., 2022). At the ecosystem-level, long-term projections using species distribution ensemble models suggest an increase in species richness and functional redundancy poleward with an increase in predatory taxa that will threaten Arctic species and decrease the modularity of Arctic food webs in the Bering and Chukchi seas (Alabia et al., 2020). It is predicted that the co-occurrence of boreal and arctic fish species in borealized regions will vary from 1 year to the other, with more Arctic species in cold years and more boreal species in warm years (Geoffroy et al., 2023), as seen in the Labrador and Bering seas (Marsh and Mueter, 2020).

The poleward progression of communities could lead to many changes in seascapes, with shifts in essential habitat localisation and altered connectivity, leading to declining stocks and a potential increase in population mixing and hybridization by the end of the century. Habitat forming species such as macroalgae are likely to continue their colonisation of coastal Arctic regions, with more of the Canada, Greenland and Svalbard coasts becoming suitable habitats due to less ice cover for three important macroalgae species by 2,100 (Jueterbock et al., 2013). In contrast, in the soft-sediment dominated Beaufort Sea, kelp production has been shown to be limited by wind speed that resuspends river and coastal sediments and limits light penetration to the seabed (Bonsell and Dunton, 2018). This is likely to affect nursery and feeding habitats, although their variations are only partially understood and might be affected by multiple factors. Along the Norwegian coast, Northeast Arctic cod spawning locations have been fluctuating with a northward trend, seemingly under the effect of population demography, while the effect of climatic factors is uncertain (Höffle et al., 2014; Opdal and Jørgensen, 2015). Similarly, spawning locations of Northeast Arctic haddock (Melanogrammus aeglefinus) are likely driven by density dependence, rather than temperature (Langangen et al., 2018). Finally, predation by boreal species has been shown to influence capelin’s spawning area (Olsen et al., 2024). It is thus challenging to model future locations of spawning sites. Nevertheless, modelling studies based on thermal preferences and bottlenecks of species early life stages can give a first estimate. Capelin is anticipated to shift spawning north-eastward in the Barents Sea (Huse and Ellingsen, 2008). For Northeast Arctic cod, further northward shifts in spawning sites could affect juveniles’ survival and increase their sensitivity to spawning timing (Endo et al., 2024).

Population mixing will likely increase and would probably occur via the Northeast and Northwest Passages (Vermeij and Roopnarine, 2008; Kelly et al., 2010; Wisz et al., 2015). This has likely already started, as suggested by the first observations of interoceanic dispersals (e.g., a Pacific diatom in the Atlantic waters, Reid et al., 2007). The Greenland halibut Pacific and Atlantic populations, with so far limited genetic mixing (Orlova et al., 2019), are now increasingly interacting along the Eurasian Arctic, suggesting that a potential reunification of those populations is possible in the future (Orlov and Volvenko, 2024).

To summarize, all climatic scenarios point to a profound transformation of the Arctic oceanic functioning. Under those conditions, many species will shift, but not all, and not all the way into the Arctic Basin, where conditions will remain harsh. Ecosystems’ primary productivity, biodiversity, functioning and food webs will be altered. All this suggests that Arctic marine ecosystems might not entirely transform into a boreal functioning, but rather stabilize in a hybrid state between Arctic and boreal ecosystem at a latitudinal limit north of the current one, with transient borealization during warmer periods.

Large uncertainties and bias surround projected trajectories for the Arctic and sub-Arctic ecosystems. Arctic Ocean circulation is complex, with numerous feedback and potentially tipping processes among and between the cryosphere, hydrosphere and atmosphere, and several possible future evolutions depending on climate change intensity (Lannuzel et al., 2020; Timmermans and Marshall, 2020). The most recent version of the Coupled Model Intercomparison Project (CMIP6) has persistent bias on Arctic water masses representations, inherited from the previous version, which necessitate major improvements, notably by using higher model resolutions (Khosravi et al., 2022). For example, other studies showed that heat transport from lower latitudes might be underestimated by model with too coarse spatial resolution (Xu et al., 2024).

Projections of community shifts in the future are also associated with considerable uncertainties, although hard to estimate and thus rarely mentioned in the studies. The potential for further borealization of the ecosystem depends on species habitat preferences, the location and connectivity of their essential habitats (spawning, nursery, feeding grounds), and their capacity to acclimatise and then adapt to new abiotic and biotic conditions. Projecting species future habitats using species distribution models is a first approach to anticipating species movement. However, those models are often based on strong assumptions, such as environmental niche stability, and are built around limited number of environmental covariables, ignoring the impact of population dynamics and interspecific interactions on species distributions. Among environmental variables, factors that do not vary over time are not always included. Indeed, while temperature, salinity or oxygen, which will be affected by future climate change, can be used as time-varying predictors of most fish species’ suitable habitats, poleward displacement might be limited by non-dynamic factors such as depth, light regimes, or seabed type. Considering a species’ traits and life cycle can help estimate the likelihood of that species to shift northward. Using this approach, it is assumed that cod and haddock have reached their northern limit at the shelf break of the Barents Sea and minke whale or harp seals will likely follow the ice edge, while other species, such as capelin, redfish or snow crab will likely be able to expand further into the Arctic (Hollowed et al., 2013b; Hollowed et al., 2013a; Haug et al., 2017; Mullowney et al., 2023).

When conditions in the Arctic are limiting, species that will not be able to shift their distribution will have to acclimate, then adapt, or die (Hoffmann and Parsons, 1990), and this will impact the pace and extent of borealization across regions and functional groups. Adaptation to environmental stress manifests itself mainly via two mechanisms: the species plasticity (capacity of a genotype to express different phenotypes under varying environmental stress), and the selection, on evolutionary scales, of certain phenotypes. Species phenotypic plasticity includes, among others, behavioural flexibility and physiological tolerance, both of which are challenging to observe in natural environments (Beever et al., 2017) and manifest differently across a species’ range (Donelson et al., 2019). Plasticity can be a way to cope temporarily with environmental stress, but can also hinder adaptation in some cases (Fox et al., 2019). It is thus important to assess species adaptive capacity to understand species responses to climate change and support conservation and management efforts (Hoffmann and Sgrò, 2011). This can be done by considering the species fundamental adaptive capacity, characterised via its functional traits such as longevity, mobility, and dispersal capacity (Beever et al., 2016). Such approaches, along with experiments on species thermal adaptivity, are often suggested as potential improvements to species distribution models (Drost et al., 2014; Drost et al., 2016; Logerwell et al., 2022). Evolutionary models could integrate that information and incorporate interspecific interactions to further explore possible trajectories of co-evolution for Arctic marine systems.

Finally, lasting borealization, emerging from species completing their life cycle, implies that environmental conditions are suitable for all life stages (Wassmann et al., 2015), of which early life stages and spawning adults are often the bottlenecks (Dahlke et al., 2020). Several species have been suggested to be able to complete their life cycle in areas where it was previously not observed, for example, the boreal capelin in Hudson Bay (Gaston and Elliott, 2014), the copepod C. finmarchicus and the Atlantic hyperiid amphipod Themisto compressa at the Fram Strait, with the latter two species much poorer in lipid than their Arctic counterparts (Kraft et al., 2013; Tarling et al., 2022). Connectivity between the habitats and along dispersal routes is thus a key consideration when projecting population shifts and species range shifts. Knowledge on thermal preferences of early life stages enabled modellers to estimate changes in spawning sites for cod and polar cod under various gas emission scenarios (Dahlke et al., 2020; Sanders et al., 2022). Results of these studies all suggest a deterioration of the current spawning sites for both species under high emission scenarios, and more limited impacts for low emission scenarios. In addition, some local conditions such as depth may act as barriers to dispersal of fish species across the pan-Arctic (Bouchard et al., 2018).

Despite some uncertainties and knowledge gaps, there is a clear consensus on the ongoing borealization of Arctic ecosystems, and its likely continuation in the future. Arctic endemic communities and ecosystems are most at risk, if they are unable to adapt to the multiple environmental stressors or to compete with new incoming species (Kędra et al., 2015). Sympagic species or species with a strong link to the ice, through their diet or life cycle, are particularly at risk (Kohlbach et al., 2016), while some may survive in the pelagic zone (Kunisch et al., 2020). The ice association is well documented for example, for polar cod. In the Chukchi Sea, its body condition appears to be lower during warm years, suggesting that the region is warming beyond the species’ thermal limits (Copeman et al., 2022). With sea ice loss, polar cod will lose its spawning habitat (Huserbråten et al., 2019), and new incoming species could outcompete it (Mueter et al., 2016). However, some studies suggest that competition for food might be limited at least for early life stages, through niche partitioning (Falardeau et al., 2014; Bouchard et al., 2022). Moreover, in the Canadian Arctic Archipelago and Central Arctic Ocean, the relaxation of the most extreme conditions with climate change could benefit polar cod, at least in the short term (Bouchard et al., 2017; Steiner et al., 2019; Geoffroy et al., 2023). There, larval survival of polar cod increases in years with earlier ice-breakup and warmer sea surface temperatures (Bouchard et al., 2017).

With sea-ice loss and increased open water area, carbon sinking and ocean acidification are likely to increase (Lannuzel et al., 2020; Terhaar et al., 2020; Qi et al., 2022), adding to the effect of warming to affect species habitats and distributions. The extent of the impacts of ocean acidification on marine populations is still not well understood. Experimental studies show that acidification affect coastal phytoplankton species growth, colony length and chain length, although the effect vary across species (Thoisen et al., 2015). In addition, increased acidification could lower phytoplankton growth rates and amplify the decrease in primary production with warming (Coello-Camba et al., 2014), and alter microbial community composition (Brussaard et al., 2013; Sugie et al., 2020). Finally, acidification can lead to the dissolution of Arctic benthic invertebrate shells (Comeau et al., 2009; Niemi et al., 2021) and, in combination with warming temperatures, increased consumption by Arctic brittle stars (Wood et al., 2011). However, some Arctic bivalves are more resilient to decreasing pH (Goethel et al., 2017).

In addition to those climate-driven pressures, Arctic ecosystems will experience increasing pressures from many other anthropogenic sources. Microplastic contamination is already affecting benthic organisms in the Chukchi Sea and other areas of the Arctic through local sources and transport (Fang et al., 2018; Bergmann et al., 2022). New ice-free conditions along the shelves and seasonally in the Central Arctic Ocean would allow Arctic shipping year-round starting in the 2070s (Min et al., 2022), leading to increased concentrations in pollutants (Svavarsson et al., 2021). This increase in shipping would also be prompted by industrial opportunities newly accessible in a future seasonally ice-free Arctic, such as mineral, oil and gas extraction, of fishing. Indeed, large global models project a high potential for Blue Economy in the Arctic, with potential for new fisheries, owing to increased local productivity and incoming new stocks (Cheung et al., 2015). However, the latest FAO report stresses the high risk associated with potential fishing in the Arctic due to very high uncertainties around the future biomass projections associated with the two main Earth system models used to force the simulations (Tittensor et al., 2021; Blanchard and Novaglio, 2024). In addition, climate is not the only factor affecting Arctic fisheries, and complex cross-scale, cross-domain (ecology, society, economy) processes can greatly affect the future of pan-Arctic fisheries (Niiranen et al., 2018). The potential growth of biomass of commercial stocks thus call for new management approaches (Christiansen et al., 2014; McBride et al., 2014) and careful consideration of potential new fishing opportunities. For example, a study investigating the potential for a fishery on walleye pollock that are expanding into the western Chukchi Sea highlight strong uncertainties around the amount of biomass entering the region, the state of the stock, and the possibility of harvesting considering the harsh local fishing conditions (Maznikova et al., 2023). Such complex systems are challenging to apprehend, and models are often a key tool to anticipate future changes. To date, few models explore regional ecosystem response to potential effects of climate change scenarios on ecosystem services and human wellbeing, with or without considering other human pressures such as fisheries. The few projections that have been completed anticipate a decrease in fish biomass or high uncertainties in biomass trajectories, thus challenging the hypothesis of a future Blue Growth in the Arctic (Bryndum-Buchholz et al., 2020; Whitehouse et al., 2021).

In the meantime, some traditional local fisheries are threatened by ecosystem changes. For example, the winter ice fishery on Greenland Halibut (Reinhardtius hippoglossoides) is now being threatened by the progressive loss of sea ice (Hussey et al., 2017). Fishermen in Greenland are shifting their traditional fishing practices in response to changes in ice cover, and are observing new species (Schiøtt et al., 2022). Local ecological knowledge, including knowledge from Indigenous communities whose livelihoods and cultures are tightly linked to Arctic ecosystems, is increasingly highlighted as a most valuable tool to face future challenges and apply an adaptive strategy to the fast changing Arctic (Riedlinger and Berkes, 2001; Schwoerer et al., 2021). Local fishermen and hunters have detailed knowledge on seasonal and spatial scales that are rarely covered by any scientific survey, and are highly useful to combine with scientific knowledge and data do detect changes in Arctic ecosystems and threats to traditional livelihoods (Chila et al., 2022; Falardeau et al., 2022; Bouchard et al., 2023; Dunmall et al., 2024). Better inclusion of this rich source of information could be done through community-based monitoring and co-construction of scenarios to directly inform management (Nilsson et al., 2021; Falardeau et al., 2022). Indeed, balancing ecosystem and livelihood conservation and the increasing economic interests in the Arctic necessitate consideration of the whole socio-ecological system, but a holistic, interdisciplinary and systematic approach is currently lacking to properly picture the complexity of the Arctic socio-ecological system and the pressures that affect it (Falardeau and Bennett, 2020; Eerkes-Medrano and Huntington, 2021). Identified gaps include for example: the heterogenous knowledge base across the Arctic, limited research on cumulative impacts of climate change with other anthropogenic pressures, a lack of knowledge on some feedbacks between ecosystem services and human well-being, and a lack of projection into the future for human systems (Falardeau and Bennett, 2020). Considering the local and global importance of Arctic climatic and biological processes, urgent mitigation and adaptive actions are needed (Overland et al., 2019).