Chapter 1 focuses on large-scale oceanographic processes in the North Atlantic Ocean. It is noted that the stock size of several demersal fish stocks and several seabird species have been declining during the last 2–3 decades. A critical question is whether this is due to anthropo­genic influence, natural environ­mental cycles or maybe (and more likely) a combination of these. A bottom-up perspective focuses on the most plausible physical drivers underlying ecosystem changes on the south Iceland, Faroe and Norwegian shelves – especially through the trophic pathway from large zooplankton (copepods) via forage fish (e.g. sandeel and Norway pout) to commercial fish stocks and seabirds. Historical temporal changes in the SPNA are here categorized into:

Future studies could focus on a review of recent ecosystem changes on the Iceland, Faroe and Norwegian shelves, that could be used to test and maybe adjust/improve previously proposed hypotheses on bio-physical linkages.

Seabirds associated with the Norwegian Shelf (Chapter 2) are in many respects good indicators of marine ecosystem dynamics, especially with regards to advection of nutrients and zooplankton onto the shelves and the growth and retention time of fish larvae and forage fish. Seabirds, like kittiwakes and puffins feed far offshore in their search for prey to sustain their offspring, although kittiwakes may also search for food in near-shore waters. These two seabird species thus reflect two different foraging niches in the pelagic ecosystem offshore, and the foraging ranges of both species in the breeding season cover the entire extent of the shelf area around Røst in Northern Norway. Productivity, e.g., chick fledging success for puffins or chicks per pair for kittiwakes, is the far best parameter linking seabird demography with trophic interactions in the breeding area. The productivity of Atlantic puffins in Røst, Norhern Norway has been monitored since 1964 and for kittiwakes in Røst since 1979. The results show the existence of high and low productive periods. Low productive periods have been observed during the 1970s and in the period from 2007 to 2015. This can be attributed to the availability of age 0 herring produced by the Norwegian spring-spawning stock. Further north, age 0 cod is shown to be an important diet component of adult common guillemots in the southwestern Barents Sea, which also affects the breeding success there. Crustaceans, such as krill, may also be important food for seabirds when forage fish are lacking. Physical drivers, although not well understood, may be of importance to seabird production since sea temperature and salinity within the NCC off the Lofoten Islands in March, i.e. two months prior to egg laying, explained more of the variation in puffin breeding success on Røst than the abundance or quality (size) of their main prey, age 0 herring.

Seabirds, i.e. puffins, in Vestmannaeyjar, Iceland, are also good marine ecosystem indicators (Chapter 3). A virtual production index, P(H.) is based on average age composition of harvest of birds ringed as chicks. Harvest is 75% composed of immature birds 2–4 years old. The P(H.) calculation sums up each cohort relative size, accounting for survival to age using estimated 87% annual survival in Vestmannaeyjar (Helgason 2012). This index reflects cohort sizes relative to a reference year of maximum cohort size (1882) and is the longest time series on seabirds in existence. It is important to keep in mind that this production index is essentially a product of breeding success (fledglings/pair) and breeding population size in a given year. However, by using fixed age composition, biases small cohorts upwards, and large cohorts downwards in numbers, thus reduces cohort real size variation. To gauge this filtering, another production index P(town) was created based on number of fledglings, either ringed (1971–2002) or rescued by the public annually in the town of Vestmannaeyjar (2003–2023). Each series is standardized by dividing by their respective maximum values (Figure 3.1). The number fledglings in town and direct production estimates are highly correlated. The length of the data series is impressive, as it covers the 1878–2023 period. Results show that there was a low-productive period from 1930 to 1962 and also from 2004 to 2015 while a low-productive period probably also was observed in the 1970s. These fluctuations are strongly correlated negatively with sea surface temperature in Vestmannaeyjar (1878–2005), likely via effects on sandeel egg developmental time (thus hatching time), winter survival, summer growth, and fertility by direct effect on basal metabolism. Also, warming of the Atlantic in 1995 coincided with a 50% reduction the abundance of northern krill in Icelandic puffins wintering area over the Atlantic ridge. Northern krill is likely key puffin prey in spring and might affect survival of puffin and many other seabird species wintering in the hotspot. In 2005 puffin production declined substantially faster than in the beginning of the last warm phase of the Atlantic Multidecadal Oscillation (1930–1965) despite similar SST increase as the current warm phase starting in 1995, suggesting another causative factor/s than SST alone. Furthermore, a large (1.5x SD, or 18 days) delay in Westman puffin breeding phenology (mean fledging time) was observed during this period (after 2005) in comparison to mean (mean fledgling time 1937–2023. This phenological delay is of great interest as it seems indicative of general trophic mismatch in Selvogsbanki, affecting not only production of puffin and their sandeel prey, but also many important commercial fisheries.

Seabirds at the Faroes confirm the above described patterns (Chapter 4). There is a strong positive correlation between Atlantic puffin reproductive success on Mykines and the mass of zooplankton in the Norwegian sea in May (g/m²) from 2011 to 2020 (r² = 0.74) maybe indicating that there is a common underlying driver capable of causing production peaks on both the Faroe and Norwegian shelves. A low-production periods is observed from 2004 to 2013. Local effects are also important since the 2009 and 2017 peak in tern reproductive success as well as mean number of terns throughout the country closely follows the kittiwake reproductive success these years and it is therefore also likely that these peaks can be explained at least partly by the local 0-group fish larvae index that is mainly driven by the local primary production. In addition to the abovementioned factors, seabird hunting may have contributed to the overall decline in seabirds (68% for guillemots 1972–2014, 60% for kittiwakes 1987–2014) at the Faroes.

Future research should focus on 1) the effects of large- and mesoscale oceanographic processes on the Norwegian shelf ecosystem, 2) top-down effects of mackerel as a predator on seabird prey and lower trophic levels, and 3) spatial and temporal dynamics of seabird foraging habitat use. Also, 4) time lags in productivity of seabirds among the Nordic shelves should be investigated as well as 5) getting more data on what adult seabird feed on and where the feeding area are located.

The sections above have shown that large-scale hydrographic conditions, e.g. the dynamics of the Subpolar gyre or the amount of zooplankton in the Norwegian Sea, have a large effect on the productivity of seabirds via the effects on zooplankton and/​or forage fish. Importantly, some low-productive periods were identified, such as the last one from around 2003 to around 2014 that seemed to affect both the Icelandic shelf, Faroe shelf and the Norwegian shelf.

The great question is whether something can be done to mitigate the effects of the low-productive periods on the shelf ecosystems. The short answer is yes. A novel approach is used to explain annual variations of demersal fish larvae in the pelagic phase that takes into account 1) ecosystem productivity (as shown above), 2) predation on fish larvae by pelagic fish (herring and mackerel) and 3) mitigation of the predation effect by demersal fish species (cod, haddock, ling, saithe).

The motivation for this approach is that pelagic fish larvae of demersal fish, e.g. sandeels, seem to be the crucial link between lower and higher trophic levels and that predation mortality is of paramount importance to the functioning of the shelf ecosystems. Although fish larvae may have many predators, it was focused on herring and mackerel that are present at the same time (March–April) and place as the fish larvae. It was, however, noted that the amount of herring/​mackerel were negatively correlated with the amount of adult demersal fish, such as cod, haddock, ling and saithe where the mechanism was interpreted as predation avoidance since these demersal fish are observed to prey on herring and mackerel in Faroese waters.

A modelling approach on the Faroe Plateau showed that the survival of fish larvae was a factor of 13 higher in demersal large-stock years compared with low-stock years. Environmental factors were also of great importance since the survival of fish larvae varied by a factor of 26 between years with the highest primary production compared with the lowest. Simpler modelling approaches for the Icelandic shelf, Faroe Bank and the Norwegian shelf seemed to confirm these results, i.e., that there was a positive relationship between the survival of fish larvae and the amount of demersal fish. This means that reducing the fishing mortality on demersal fish can mitigate the mortality on fish larvae that is caused by pelagic fish, although the extent of the effect was not evaluated.

A modelling approach was also performed for the Faroe Plateau where adult Norway pout were regarded as predators on fish larvae and were cod preyed on, and suppressed the biomass of, Norway pout. Results indicated that a low fishing mortality on cod was able to increase the amount of sandeels as well as giving higher yield per recruit for cod.

Future research should focus on the extent that pelagic fish (herring, mackerel, Norway pout) eat fish larvae. It should also be investigated whether the negative relationship between the amount of pelagic fish and demersal fish is real and, if so, whether demersal fish just scare pelagic fish away or they actually eat sufficient numbers of them to create the observed pattern.

It is worth noting that during this report a quite simple approach is used. Although we have used some models to evaluate the effect of e.g. primary production and the amount of demersal fish, via a scaring effect on predatory pelagic fish, on the survival of demersal fish larvae, such models only give a hint on whether the proposed mechanisms are valid or not. In order to do this work properly we, at least, need more information about the distribution of pelagic fish during the year and how much fish larvae they consume and whether adult demersal fish really have a scaring or predatory effect on pelagic fish. As such, we have only scratched the surface of a much bigger research topic. Also, when evaluating low- and high-productive periods of seabirds in relation to oceanographic forcing this is actually only to scratch the surface of a much bigger field of research. What is needed in the future is a much greater research effort that combines environmental factors as well as biological factors over a large area (NE Atlantic) that also needs to incorporate human effects, e.g. fishing on pelagic fish as well as demersal fish. This could certainly contribute to improved ecosystem-based management of marine sources in the NE Atlantic and probably elsewhere.