Biological production on the Faroe shelf is characterized by marked peaks with 6- to 10-years intervals: specifically in 1974, 1983 (weak), 1994, 2000, 2009, 2017 (Jacobsen et al., 2019), and a recent weak peak in 2023-2024 (Olsen et al., 2025). These peaks are reflected in on-shelf primary production, as proxied by a so-called primary production index (PPI) (Gaard et al., 1998), the size of juvenile fish (Eliasen et al., 2011) [represented by a 0-group length index, Jacobsen et al., 2019], recruitment to commercial fish stocks (ICES, 2024) and the breeding success and nest attendance of seabirds (Olsen et al., 2025). Although there is coherent synchrony in these parameters, the magnitude of peaks in the PPI and 0-group length index are not always reflected in recruitment to the adult Atlantic cod (Gadus morhua) (hereafter just cod) stock. In particular, production peaks after 2003 have not benefitted cod as they did before. While the Faroese fish stocks and seabird populations exhibit peaks following productive periods, they also show clear longer-term shifts and negative trends (Olsen et al., 2025), which cannot be explained by variation in the PPI or in the 0-group length index. This calls for consideration of additional environmental drivers, likely with a larger-scale ‘footprint’ than just the Faroe shelf itself. Acknowledging previously suggested links between ecological shifts in the NE Atlantic and indices like the North Atlantic Oscillation (NAO) index, and the Atlantic Multi-decadal Oscillation (AMO) (Hátún et al., 2009), we discuss potential links to the influence of nutrient- and zooplankton-rich subarctic waters (SAW) from two main sources: the subpolar gyre (SPG) and the East Icelandic Current.

Recent research has demonstrated linkages between SPG dynamics and both oceanic and shelf ecosystems in the subpolar North Atlantic (Hátún et al., 2022b). Strong winter convection and an associated large SPG increase the contribution of SAW to the mixed Atlantic waters surrounding the Faroe shelf, which in turn can increase biological production in these waters (Hátún et al., 2022a).

While including the SPG driver, this study focuses on a potentially similar influence of SAW from the East Icelandic Current, which has received less attention to date. Following the development of accurate oceanographic instruments (the Nansen bottle) in the late 1800s, a cold tongue of SAW from northeast of Iceland extending into the southern Norwegian Sea was among the first thoroughly studied oceanographic features (Figure 1). In their seminal book, ‘The Norwegian Sea’, Helland-Hansen and Nansen (1909) hypothesized that “increased amount of Arctic water carried in contact with Atlantic water in the region of the Faroe-Shetland Channel will improve the biological conditions of food-fishes during the same season in the North Sea and one year later near Lofoten”. The suggested reasons were that SAW possesses physical (e.g., temperature) and chemical (e.g., nutrients) properties which are of great importance for plankton development. For the open ocean, they observed large and nutritious calanoid copepods Calanus hyperboreus in this Icelandic tongue and along the axis of the southeastward flow of cold and heavy SAW into the southern Norwegian Sea, while Calanus finmarchicus were observed in the mixed water of Arctic and Atlantic origin. C. finmarchicus is a critical prey item, transferring energy from primary production to higher trophic levels in the Norwegian Sea (Melle et al., 2014), and on adjacent shelves (Sundby, 2000). C. hyperboreus plays a similar role north and northeast of Iceland during spring (Gislason and Silva, 2012). The biophysical relationships proposed by Helland-Hansen and Nansen (1909) received relatively little attention until recent studies by Kristiansen et al. (2016, 2019) demonstrated that the variable extent of this tongue into the waters north of the Faroe Islands regulates zooplankton communities there: a large Icelandic tongue results in large biomasses of C. hyperboreus, and large stages of C. finmarchicus can reach the Norwegian slope as well (Skagseth et al., 2022). The fact that these subarctic copepods largely vanished from the southern Norwegian Sea when the Icelandic tongue retracted westwards after 2003 (Kristiansen et al., 2019), likely had significant ecological implications (Hátún et al., 2022a).

Similar shifts have, however, taken place before. For the Norwegian shelf, Helland-Hansen and Nansen (1909) found that increased recruitment to the cod, haddock (Melanogrammus aeglefinus) and Norwegian spring spawning herring (Clupea harengus) stocks was proceeded by increased volume of SAW and decreased temperatures in the Atlantic water. The SAW volume was estimated as the sectional area of SAW in a section across the southern Norwegian Sea (Sognefjord section) in May. The SAW volume was large in 1901, and fish thrived. This changed abruptly in ‘the famous year 1903’, with reduced SAW volume, increased Atlantic water temperatures and poor cod conditions, reflected in reduced development of sexual organs, reduced quantity of cod-roe and cod-liver oil and exceptionally late arrival at the fishing banks near Lofoten. Haddock were also small (in size) in 1903, and “thousands of dead seabirds (common guillemots, Uria aalge) were said to be in completely emaciated conditions, and washed ashore”. The deteriorated conditions coincided with a sudden occurrence of tropical plankton forms, which had previously never been observed in the northern seas.

SAW north of the Faroes circulates clockwise around the northeastern tip of the Faroe Plateau and into the Faroe-Shetland Channel, and establishes the hydrographic foundation below the Atlantic inflows towards higher latitudes. Confluence of lower trophic level biomass along the interface between the SAW and the overlying Atlantic waters – the permanent thermo/pycnocline, feeds a large biomass of krill, mesopelagic fish, and their larger predators (e.g. blue whiting (Micromesistius poutassou), Greenland halibut (Reinhardtius hippoglossoides), and several other fish species). This “Wall of mouths” (Hamner et al., 1988) is clearly identifiable as a Deep Scattering Layer in acoustic data from these waters (Cisewski et al., 2021). Zooplankton that have ascended and other near-surface, passive, food items can potentially also be advected from the Iceland tongue towards the Faroe shelf break in a low-saline surface layer, which builds during the summer and fall (Perkins et al., 1998). If the link between the SAW and the Norwegian shelf ecosystem suggested by Helland-Hansen and Nansen (1909) holds, such a mechanism might be even more robust for the Faroe shelf ecosystem. A thorough discussion of the ocean-to-shelf advection is, however, beyond the scope of the present work, which is to establish synchrony between key large-scale oceanographic drivers and biological indicators on the Faroe shelf.

Cod has been one of the main commercial fish species, sustaining a considerable portion of the Faroese economy. This stock has therefore been closely monitored, providing extensive data material, including long and robust time series (ICES, 2024). Guillemots have historically constituted the largest and most exploited seabird population in the Faroe Islands (Nørrevang, 1977). A major decline in this species during the 1960s elicited concern (Olsen, 1990; Reinert, 1976) and initiated structured seabird counts in the Faroe Islands (Dyck and Meltofte, 1975; 1973). The common guillemot (hereafter just guillemot) record discussed here is the longest of its kind from the Faroe shelf (Olsen et al., 2025). Cod and guillemots share similar prey species, e.g., sandeel (Ammodytes spp.), and adult guillemots in the Barents Sea (Myksvoll et al., 2013), and likely also on the Faroe shelf (B. Olsen, personal communication; Figure S1) also prey directly on juvenile cod. This shared predation makes the breeding performance of this seabird a useful indicator of the status and recruitment of the cod stock. Motivated by this expected ecological coupling, economic, ecological and cultural significance, and large data availability, cod recruitment, total cod biomass and total attendance of guillemots are employed as indicator records for principal higher trophic level species on the Faroe shelf. Building on Helland-Hansen and Nansen (1909), we will use these indicator records, the now well-documented SPG-ecology linkages and ongoing ecological research on biological production on the Faroe shelf, to introduce the following new theoretical framework: the biomass of higher trophic levels on the Faroe shelf is driven by both on-shelf production and off-shelf food abundance. This forms basis upon which more refined hypotheses can be built and rigorously tested.

The data material is described in Section 2. The SPG influence on SAW contribution to the Atlantic waters west of the Faroes (proxied by a salinity record) is presented in Section 3.1, and results from spatial and temporal changes in SAW from the Icelandic tongue are illustrated in Section 3.2. Updated oceanic zooplankton records north of the Faroe slope are provided in Section 3.3, recruitment and biomass of cod is presented in 3.4, and guillemot attendance is presented in 3.5. A comparison with the Faroe Bank cod is provided in 3.6 and the results are discussed in Section 4.

Table S1 in the Supplementary Online Information (SOI) contains raw data for all indicator records, which are provided as annual values (excluding the salinity record and the gyre index).

Acknowledging that fishery and hunting impact Faroe cod stock size and Faroese guillemot abundance, respectively, we explore a bottom-up perspective in this study. Our aim is to introduce a new theoretical framework, that biomass of higher trophic levels on the Faroe shelf is influenced by both on-shelf production and the import of off-shelf food resources. The data presented support the validity of this perspective, although our aim is not to provide definite conclusions. We encourage the development of more specific, testable hypotheses, that build on this framework.

The off-shelf food resources are in this framework related to the volume of SAW in Faroese waters, derived from both the SPG and the Icelandic tongue. In this initial presentation, we mainly focused on SAW from the Icelandic tongue, while the 0-group length index serves as the on-shelf production indicator. Despite this oversimplification, the framework can qualitatively describe three key aspects of the Faroe ecosystem: variability in Faroe shelf cod stock recruitment, biomass development of both the Faroe shelf and Faroe Bank cod stocks, and guillemot attendance at Faroese seabird cliffs. We acknowledge, however, that a realistic representation of stock development for cod and guillemots must involve top-down pressures (including fishery and hunting) and likely several environmental drivers not included here, for example declining silicate levels (Hátún et al., 2017), increasing temperature and regional/local processes.

We show that a rapid ‘subarctification’ in the southern Norwegian Sea in the 2020s resulted in a dramatic increase in subarctic zooplankton abundance north of the Faroes in May, which supports the previous hypothesis (Kristiansen et al., 2019; Skagseth et al., 2022) that a variable influx from the Icelandic tongue regulates the zooplankton community in the southern Norwegian Sea. Our SAW volume indicator therefore likely represents fundamental aspects of the marine environment and ecosystems in these waters. It should be noted that the total zooplankton dry weights at the north of the Faroes (Section N) in May increased more gradually than the copepod counts (Figure 4 and Figure S2, in the SOI). Since dry weight is used as an indicator for the Norwegian Sea (https://www.ices.dk/community/groups/Pages/WGINOR.aspx), the causes of this discrepancy should be further investigated.

Our new theoretical framework can, in a mechanistic way, explain why seemingly favorable on-shelf conditions do not always translate to high adult cod stock numbers. The weaker-than-expected cod recruitment peaks (1987-1988, 2008-2009 and 2016-2017) coincided with Atlantic conditions, characterized by small SAW volume and, by inference, limited ocean-to-shelf food supply throughout the year.

The tight synchrony between the cod recruitment record (#1-year-olds) and the 0-group length index suggests that the feeding conditions of 1-year old cod are a bottleneck for recruitment to the adult stock. The 0-group length index is composed of data from the cod, haddock, Norway pout and sandeel stocks, where the sandeel biomass likely plays a critical trophic role. Predation from cod on the other species, including cannibalism, should however also be considered. The high recruitment peaks in 1994 and 2000 were preceded by both very high SAW volumes (Figure 5A), and by increased numbers of juvenile cod the year before (i.e., 1993 and 1999; see Olsen et al. (2025), their Figure 4). The high recruitment peak in 1983 was also preceded by high numbers of juveniles, while the SAW volume coincided; the SAW peak occurred in 1983 and was at near average values in 1982 (Figure 5A). Our knowledge of the first year of Faroe cod is limited by data scarcity, and the trophic pathways of how onto-shelf advection can impact cod’s living conditions through their first year are manifold, including hydrographic conditions, fluxes of nutrients, zooplankton and higher trophic level biomass. The presented results suggest that this first preconditioning year is associated with intensifying influence from the SPG (Figure 2) and/or the Icelandic tongue (Figure 5), and the following bottleneck year coincides with maximum influence from these SAW sources. Guillemots are known for their ability to dive deep (up to 150-200 meters) to catch prey, making them efficient marine hunters, but their prey size is generally limited to smaller fish species (e.g., sandeel, capelin (Mallotus villosus) and juvenile cod; Piatt and Nettleship, 1985). This seabird species feeds both on and off the Faroe shelf (Olsen et al., 2025), and thus also over regions directly influenced by the SAW. However, elaborating on this bottom-up trophic cascade is beyond the scope of the present work.

Although the SAW volume indicator and cod biomass are not strongly correlated, they share important characteristics; low SAW volumes and low recruitment peaks quickly resulted in major declines in the cod stock biomass in the late 1980s and after 2000. Excluding the 2020s peaks in the SAW volume, both series clearly show a negative trend, and it is evident that Subarctic periods have historically led to higher biomasses, while Atlantic periods result in reduced biomasses.

Furthermore, a correlation between cod biomass and guillemot attendance in the Faroes enables us to use the guillemot record as an assessment-independent indicator for the cod stock. A major decline in Faroe guillemot attendance, which started in the 1960s (Olsen et al., 2025), has continued since the early 1970s, with significant decreases in the late 1980s and early 2000s (Figure 6), and Olsen et al. (2025) discussed both drops in light of reduced winter convection in the Labrador-Irminger Seas, a weakened subpolar gyre and reduced silicate levels and krill abundance in large parts of the subpolar North Atlantic. Regarding the early 2000s drop. Olsen et al. (2025) furthermore mentioned the concurrent reduction in SAW volume from the Icelandic tongue. Our findings strengthen the proposition of a link between the guillemot numbers and the SAW volume, which even could help explain persistent guillemot decline from the 1970s to 1990 (Figure 6B).

Excessive fishing and bird hunting pressure have certainly contributed to the cod and guillemot declines, but these cannot account for the increased guillemot attendances during the 1990s and the peak in 2008-2009, nor the good cod recruitment during the decade 1995-2005. This suggests that our new bottom-up perspective should be considered as well, and that this could contribute to more rational management of these resources.

Counterintuitive to our framework is that the Faroe shelf cod stock and guillemot attendance have apparently not reacted to the 2020s increase in SAW volume and increased abundance of oceanic subarctic copepods. However, this 2020s discrepancy does not necessarily invalidate our new theoretical framework. Signs of Faroe Bank cod recovery were already evident by 2019, when groundfish surveys began detecting an increase in the stock biomass (ICES, 2024). This stock has since surpassed the previously observed peaks of the 1990s and early 2000s (ICES, 2024). Age structure analyses indicate that this recovery is largely driven by strong year classes from 2015–2018 (Salter et al., in prep.), though this cohort was not clearly reflected in survey estimates until after 2019, possibly due to underrepresentation of younger ages in survey. The rapid decline of cod biomass on the Faroe Bank in the early 2000s was likely caused by a combination of excessive fishing pressure and decreased SAW volume. And between 2008 and 2021, the Faroe Bank (< 200 m) was entirely closed to commercial fishing, with the exception of a limited longline fishery.

Relationships between the 0-group length index, the SAW volume and haddock recruitment are similar to the cod linkages back to the late 1970s (Figure S3). And the Faroe shelf haddock recruitment did indeed improve during the Faroe shelf production peak in 2016-2017, and the biomasses of both the Faroe shelf and the Faroe Bank haddock biomass increased after the 2016-2017 event and especially into the 2020s (Figures S3, S4). Furthermore, some seabird species have experienced improved conditions on the Faroe shelf after 2015 [e.g. Faroe puffins (Fratercula arctica) and kittiwakes (Rissa tridactyla); (Olsen et al., 2025)].

Based on the rapid intensification of the SPG after 2015 (Figure 2, and see Hátún and Chafik, 2018), and its potential to increase ocean-to-shelf fluxes of nutrients (Eliasen et al., 2017; Hátún et al., 2022a) and zooplankton (Jacobsen et al., 2025, 2022, 2019), Hátún et al. (2016) anticipated increased biological productivity on the shelves adjacent to the SPG after 2014. The biological signals observed on the Faroe Bank and Faroe shelf support this projection. Similar to the linkage between SPG dynamics and zooplankton biomass on the south Iceland shelf (Hátún et al., 2016), we suggest that the correlation between peaks in the gyre index and the 0-group length index (Figures 2, 5) could strengthen a causal understanding of the Faroe shelf production peaks.

The Faroe Bank is most directly impacted by the SPG, while the relative influence of SAW from the SPG and the Icelandic tongue, respectively, on the Faroe shelf is more difficult to disentangle. Our metrics of SAW from these two sources may not be completely independent. The SPG regulates the water properties in the Atlantic inflows, both those flowing past the Faroe Plateau [Figure 2 and Larsen et al. (2012)] and those flowing north through the Denmark Strait (Hátún et al., 2005). When diluted by variable winter convection and mixing between Atlantic and polar water masses north of Iceland (Malmberg and Valdimarsson, 2003), a lagged SPG-induced hydrographic signal could reach the Icelandic tongue and even become integrated into our SAW volume indicator (which is based on temperature and salinity thresholds, Section 2.1.2). However, our SAW volume indicator primarily represents the amount of Icelandic tongue water reaching the Faroe slope, which is regulated by regional processes, such as the Norwegian Sea gyre circulation (Hátún et al., 2021).

The amount of drift-ice exported south through the Fram Strait and transported in the East Greenland Current declined after 2003 (Sumata et al., 2023), and virtually disappeared after 2015, which led to an ecological regime shift in slope areas of southeastern Greenland (Heide-Jørgensen et al., 2023). A recent study, however, shows a marked increase in southward freshwater fluxes from the Fram Strait in 2020 (de Steur, personal communication), and the sea-ice cover farther south has increased thereafter. While Dickson et al. (1988) linked such ‘Great Salinity Anomalies’ (reduced salinity) to reduced biological production north of Iceland, we suggest that such anomalies can contribute to increased SAW volumes, and associated increased biological production, in Faroese waters.

So why have the Faroe shelf cod stock and the guillemot colonies not increased again in the 2020s? A first plausible explanation is that intense fishery has reduced this cod stock to levels too low for recovery, consequently impacting the guillemot population more severely than other seabird species (such as puffins and kittiwakes). The fishing ban on the Faroe Bank in 2008, and the Faroe Bank cod recovery after 2015, support this explanation. A second reason is that the total volume of SAW surrounding the Faroe shelf might be less than apparent from the SAW indicator from the Icelandic tongue alone, since the SPG weakened again after 2016 (Figure 2). A third cause could be found in other environmental drivers beyond the SAW volume from the Icelandic tongue and the SPG, for example the long-term decline in pre-bloom silicate concentrations in the Atlantic water passing the Faroe Plateau (Hátún et al., 2022a, 2017), may have impacted the cod-guillemot trophic pathway more severely than it did other trophic paths. Although marine scientists often, too one-sidedly, associate environmental drivers with seawater temperature (Edwards et al., 2021), we do not consider temperature as the sole driver of the discussed trends. While the unproductive Atlantic period (2004-2014) was associated with high temperatures, the highly productive period of the late 1950s-early 1960s experienced comparably high temperatures (Larsen et al., 2008). Correlations between the presented biological time series and temperature are, furthermore, not significant. A fourth possibility is that an anticipated increase in cod recruitment has indeed taken place around 2023-2024, but this remains not yet evident in the assessment records, given that the cod recruitment estimate is uncertain for the most recent 2-3 years (Beverton and Holt, 1993). Under this scenario, we should expect an increase in cod biomass and guillemot attendance during the coming year(s).

Similar conditions to the early 2020s occurred in the early 1990s; SAW volume reached a high level in 1993, followed by high on-shelf production (proxied by the 0-group length index) in 1994 (Figure 5). Cod, spawned in 1993, recruited very well in 1994, and cod biomass rose from 1991 levels, which were comparable to current very low estimates, to very high values in already 1995 and 1996 (Figure 5B). Notably, due to poor fishery in the late 1980s-early 1990s, the Faroe Islands went bankrupt and the fishing fleet was largely dismantled in the early 1990s, resulting in very low fishing pressure (ICES, 2024), which likely has helped this rapid recovery. The 1990s experienced major fisheries management challenges and deteriorated trust between fishing industry and the scientific community, which underscores the critically importance of closely monitoring the present ecological development in Faroese waters. Our framework could potentially also help understanding trends in neighboring shelf ecosystems.

Ecological changes on the Faroe shelf have previously been discussed in relation to primary production on this shelf. These earlier studies provided foundational insights into the ecosystem’s dynamics, and their explanation model primarily explains the marked production peaks, recurring in this system at 6- to 10-year intervals. By including oceanic food abundance, represented by the volume of subarctic water masses around the Faroe Plateau, our new theoretical framework enables the explanation of longer-term trends that also characterize this ecosystem. This expanded approach offers a more comprehensive understanding of the ecosystem’s complex temporal dynamics, and can support improved management of marine resources. Multiple hypothesis inference could be a strategy to get closer to a truly evidence-based understanding of the drivers of the Faroe shelf ecosystem. This would involve formulating statistical models that are related to different hypotheses, which could be evaluated, ranked, and some rejected. A priority area of research enabling this approach should be on how advection of nutrients and prey from the open ocean subsidize local production, both on the Faroe shelf and on the Faroe Bank. Key questions are: what is advected, where (both geographically and vertically in the water column), during which season, and how does this vary between years? What are the relative roles of seawater temperature per se, local primary production, and the onto-shelf nutrient and food advection, respectively?