Scottish Marine and Freshwater Science Volume 5 Number 18: Depth use and movements of homing Atlantic salmon (Salmo salar) in Scottish coastal waters in relation to marine renewable energy development

Discussion

The results presented here confirm and build on the findings of Godfrey et al. (2014), giving a more detailed picture of the swimming depths of Atlantic salmon principally on their first return migration to Scottish coastal waters. The sample is limited to larger (mainly 2 SW) salmon, and the swimming behaviour of grilse (one-sea-winter salmon) remains as yet unexplored. Data from Scottish coastal waters can be compared with the reported depth use by adult salmon in other situations. The proportion of time spent in the upper 5 m of the water column by salmon tagged as kelts was 60-90% on the high seas (Reddin et al., 2011; Lacroix, 2013), and 94-99% when migrating away from rivers in fjords and estuaries (Halttunen et al., 2009; Hedger et al., 2009). Salmon migrating home through a fjord were also surface-orientated (mean swimming depths 0.5-2.5 m; Davidsen et al., 2013). Thus, the general pattern of depth use by salmon returning through Scotland's northern coastal waters (median 80.4% of time at 0‑5 m depths, mean 75.0%) is more akin to that of kelts that had returned to sea to feed, and less surface-orientated than that of homing salmon migrating through fjordic or estuarine environments towards rivers. The range of maximum diving depths (13-256 m) of adult salmon in the coastal waters around northern Scotland was similar to the depths available in the study area (the depth of water column at the pop-off locations ranged from 0.5-176 m), while dives of over 600 m by salmon have been recorded in the Labrador Sea (Lacroix, 2013). These data are therefore consistent with most salmon using the full extent of the available water column. Thus salmon on their homeward migration in the coastal zone were predominantly surface-dwelling, but most passed regularly through the water column, and some spent extensive time at depth.

Figure 2 Map of Scotland showing satellite tag pop-up locations for salmon tagged and released at Armadale (shown by a cross) in 2013 (upper) and 2014 (lower). Time at sea varied: triangles (1-2 d); circles (2-5 d); squares (5-10 d); and stars (10-20 d). Locations were excluded if associated with an error of >25 km (n = 46) or if tags had been ingested by predators (n = 4).

There was considerable inter-individual variation in swimming depths in this study: for example, the individual proportion of time spent in the upper 5 m ranged from 8-99%. Sturlaugsson (1995) has previously noted different patterns of diving behaviour in Atlantic salmon, and Godfrey et al. (2014) identified two clusters of differing swimming depth behaviour in the 2013 data, but the reasons for differing uses of the water column are unclear. Diving behaviours may be associated with predator avoidance (Hastie et al., 2006, Lacroix, 2014), behavioural thermoregulation (Tanaka et al., 2000; Reddin et al., 2004), feeding (Reddin et al., 2011; Lacroix, 2013) and sampling the water column in search of olfactory cues from the home river (Westerberg, 1982; Døving et al., 1985). The behaviour of fish is likely to depend on the relative importance of different processes at any given time, which may vary with the geographic location (Lacroix, 2013) and the phase of migration (Middlemas et al., 2009) of the fish.

The swimming depths of salmon reported here have some general implications in relation to the development of MRE, while their similarity between years raises the confidence that these data can be used to extrapolate to future scenarios. Salmon used the entire water column and so could come into contact with any installed devices. The large majority of time was spent in the upper water column <20 m depth, but some individuals were more likely to use deep and mid-water zones. The preferential use of near-surface waters (median 80.4% of time at <5 m) implies that homing salmon are particularly likely to interact with surface-oriented installations such as those exploiting wave energy. With specific reference to the MRE development in the Pentland Firth region, the turbines currently licensed for this area have a minimum clearance of 8 m between the turbine blades and the water surface (Anon., 2012). All the tagged salmon in our study were recorded at depths below 8 m and therefore had the potential to interact with the proposed turbines, should they pass through the development area (detailed depth use data is shown in Appendix 1). Individuals spent a median of 15.2% of their time below 8 m depth ( IQR 8.5-26.6%). However, the scatter in pop-up locations (Figure 2) suggests that salmon may not have well-defined routes in the region. Furthermore, while some salmon in the sample appeared to undertake direct migrations, the movements of others were more peripatetic, implying the potential for these latter individuals to traverse individual development sites on more than one occasion. However, it is not currently possible to predict what proportion of the salmon in areas such as the Pentland Firth pass across specific zones.