The SAW layer remains thick along a U-shaped region between the NE Icelandic shelf and the Jan-Mayen ridge (Fig. 2.2). The thickest layer is generally observed near the deep and steep trough where the Jan-Mayen Ridge intersects the Iceland-Faroe Ridge. The SAW layer thickness varies substantially between years, both within this U-shaped region and particularly in the southeastward ‘leakage’ into the Norwegian Sea and into the Faroe-Shetland Channel. Volumes of SAW were large from 1995 to 2002, but the leakage region thinned abruptly in 2003. Apart from smaller pulses (e.g. in 2012), the apparent eastward flow remained weak until 2018, after which a much thicker SAW layer was again observed north of the Faroe slope and along the Norwegian slope.

The SAW proxy record from the north Faroe slope (Section N, updated from Kristiansen et al., 2019) confirms that the eastward flow of SAW increased after 2017 (Fig. 2.3a). This increase was steep, and the SAW volume in the 2020s reached values matching the highest levels of the 1990s. The SAW proxy record from the Norwegian slope (Svinøy Section) closely follows the development at Section N (Fig. 2.3). The post-2017 increase was also dramatic in this more eastern location, and appears to have lagged behind Section N by about a year. The longer record available from the Svinøy section shows that the SAW volume in the southern Norwegian Sea was likely highest in the late 1970s, declined to very low values in 1987–1990, increased rapidly in the early 1990s, and remained high until 2002. It dropped, as evident in the spatial view (Fig. 2.2) in 2003 and remained low until the aforementioned rapid post-2017 increase.

Figure 2.3 SAW volume in the southern Norwegian Sea and oceanic zooplankton. a) Indicators of SAW volume at Section N at the north Faroe slope and the Svinøy Section (Norwegian slope, see Fig. 2.1). These are actually estimates of the area of SAW at each respective section; b) As in panel (a), complemented with the abundance of the subarctic calanoid copepods Calanus finmarchicus and Calanus hyperboreus at the subarctic (northern) flank of Section N, in May.

The post-2017 SAW increase, and especially the very large SAW volumes in 2021–2022, coincided with marked peaks in overwintering C. finmarchicus (in 2021) and C. hyperboreus (in 2022) at Section N, both exceeding the levels of the 1990s (Fig. 2.3b).

As proposed by Kristiansen et al. (2019), large volumes of SAW lead to increased biomasses of subarctic zooplankton in May, thus supporting this hypothesis. To recapitulate the findings from Kristiansen et al. (2019, 2022) and Skagseth et al. (2022), the high SAW volumes during the 1990s and early 2000s (a subarctic period) were associated with high abundances of these subarctic copepods, the Atlantic period (2004–2014) with low SAW volumes was characterized by low abundances. This was followed by the above-mentioned increase, which commenced in 2016 (before the SAW increase) and amplified in the early 2020s.

Figure 2.4 The Faroe cod stock and environmental drivers. a) Cod recruitment, the 0-group length index and the SAW volume, divided into subarctic periods (blue shades) and Atlantic period (red shades), based on a selected threshold. B) Cod biomass and the SAW volume. The timing of cod recruitment peaks is emphasized with vertical dashed lines, in both panels. The question mark in the 2020s refers to the discussed uncertainty.

Peaks in cod recruitment occurring at 6–10 years are clearly linked to peaks in the 0-group length index, which represent the biological production of the Faroe shelf. The recruitment peaks in 1973–1974, 1994, 2000, 2009, and 2017 align precisely, while the recruitment peak in 1982 precedes the 0-group peak by a year, and the small recruitment peak in 1988 follows the 0-group peak by a year (Fig. 2.4a). However, recruitment peaks magnitude varies much more than the peaks in the 0-group length index.

The high recruitment peak in 1982 coincided with large volumes of SAW, and the high peaks in 1994 and 2000 were preceded by peaks in SAW volume in the previous years (1993 and 1999, respectively). The three low peaks in 1988, 2009, and 2017 correspond with low SAW volumes (Table 1). Years with large SAW volumes and low 0-group length indices generally do not result in good recruitment, although the very high SAW peak in 1979 did actually coincide with a low recruitment peak. These findings support our hypothesis, except for the 2020s increase in SAW volume and recently elevated 0-group length index (2023–2024), which has not led to increased cod recruitment – yet.

Strong recruitment peaks lead to more rounded peaks in total cod biomass after a lag of 1–3 years (Fig. 2.4b). The biomass declines rapidly after these peaks, and the low recruitment peaks in 1988, 2009, and 2017 result in strong declines in cod biomass in the late 1980s/early 1990s and after 2004. Although direct correlations between cod biomass and SAW volume are not high, high SAW volumes (subarctic periods) are associated with large biomasses, while low SAW volumes (Atlantic periods) correspond to reduced biomasses. The only exception occurs in the 2020s, when the SAW volume has been very high, while the cod stock has remained very small.

Table 2.1 Assessment of the proposed hypothesis, that high recruitment requires both strong on-shelf production and a large SAW volume (proxy for off-shelf food abundance).

Figure 2.5 Attendance of guillemots in the Faroe Islands against: a) the Faroe cod biomass and b) the SAW volume.

The biomass of the Faroe cod stock and the attendance of guillemots have correlated during the last half century (Fig. 2.4a). A clear negative trend exists in both records, primarily caused by the marked cod decline during the late 1980s and early 2000s and a corresponding guillemot decline during these periods (Olsen et al., submitted). Rounded peaks in guillemot attendance roughly coincide with cod recruitment peaks, which lead the cod stock size by 1–3 years. The guillemot attendance closely follows the volume of SAW from the late 1970s to immediately before the post-2017 increase (Fig. 2.5b). The 1970s to 1990 and early 2000s declines, and the 1990s increase are evident in both records, and even shorter-term variability aligns; e.g. a ‘plateau’ during the early 1980s and peaks around 1993-1995, 1999–2000, and 2008-2009. However, there exists a strong contrast during the 2020s, when guillemot counts remained very low despite the SAW increase.

The cod stock on the Faroe Bank, located immediately west of the Faroe Plateau, is included to provide a broader spatial perspective. The cod stock size, here proxied by the CPUE of cod in both March and September show increases in the 2020s (Fig. 2.6). The March records shot up in 2020, while the September record soared in 2021 and 2023. The apparent link to the SAW volume record has also persisted back in time, with low and declining CPUE during the late 1980s, high CPUEs in the late 1990s and early 2000s, and a rapid drop after 2003.

Figure 2.6 The cod stock size on the Faroe Bank, proxied by the Catch Per Unit Effort (CPUE) in March and September, respectively, and both indicators of SAW volume.