Migration routes and behaviour of Atlantic salmon and sea trout around operational wind turbine arrays in the Moray Firth, Scotland

3. Results

Atlantic salmon smolts

Timing and duration of river emigration

The proportion of tagged smolts that out-migrated from the River Wick was similar in Y1 and Y2 with 88% and 90% reaching Wick harbour, equating to a freshwater loss rate of 1.1% km^-1 and 0.9% in Y1 and Y2, respectively. Individuals took a median of 5 days 23h 24 minutes (range 11 h 17 minutes – 15 days 11 h 46 minutes) (Y1) and 4 days 20 h 24 minutes (range 6 h 29 minutes – 25 days 23 h 17 minutes) (Y2) after release to reach Wick harbour.

Temperature and depth (pressure) data (available for Y2 only) suggested that two salmon were preyed on by mammals within Wick harbour. Recorded temperatures rose rapidly from approximately 10°C to in excess of 30°C, concurrent with depth profiles suggesting rapid diving movements (Figure 2).

A total of 94 post-smolts in Y1 (87%) and 80 post-smolts in Y2 (80%) were assumed to have successfully entered the Moray Firth, having either been detected on the outer line of receivers in Wick Bay (173 individuals) or detected in the marine array after passing through the outer line without detection (1 individual). Most post-smolts passed through Wick Bay without delay, with 95% residing for 24 h or less (Figure 3). Outmigration occurred almost exclusively during darkness with movements predominantly associated with the period from 2100 to 0100 (Figure 4). In Y1, 39 individuals exited on an ebb tide and 55 on a flood tide compared to 37 exiting on an ebb and 42 exiting on a flood tide in Y2 (excluding the one individual for which we do not know the precise time of departure).

Marine migration characteristics

A total of 37 emigrating salmon post-smolts in Y1 (39%) and 39 in Y2 (49%) were subsequently detected by the marine receiver array. In both years, most of the first offshore detections were on the North Line of receivers: 51 % in Y1 and 49 % in Y2 compared to 24 % and 10% on the South Line in Y1 and Y2, respectively (Figure 5). The time between emigration into the Moray Firth and first detection in the marine array ranged from 7 h 56 minutes to 5 days 1 h 55 minutes (median 26 h 28 minutes). Neither current speed or bearing at Wick Bay entrance, nor release year, had a significant effect on the location of first detection after leaving Wick Bay.

The time between the first and last detections in the marine receiver array ranged from 1 s to 7 days 14 h 53 minutes (median 1 h 7 minutes) in Y1 and from 5 minutes 17 s to 6 days 17 h (median 7 h 48 minutes) in Y2. In both years, receivers in the north line detected the highest number of individuals (Figure 6).

Association with offshore wind development sites

Just under half of the post-smolts detected within the marine array were detected within the OWDs; a total of 17 post-smolts in Y1 (46%) and 16 in Y2 (41%), which in both years represented 16% of all smolts tagged. Sixty one percent of those 33 post-smolts were not intercepted by the North Line or South Line, i.e. their first marine detection was within a OWD boundary, mostly within Beatrice. By contrast, most of their last detections were in Moray East (Figure 7).

Post-smolts resided within the OWDs for a minimum total duration of between 1 s and 4 h 44 minutes (median 35 minutes 31 s), with no significant difference between years (W = 128, p = 0.79, n = 33). The spatial pattern of residency in the turbine areas indicated longest durations within the Beatrice OWD (Figure 8). Individual residence events lasted from less than 1 s to 2 h 22 minutes, with a median duration of 3 minutes 30 s.

Residency durations determined by summing the duration of individual residence events denotes certain but minimum time spent by fish in the OWDs because there was not complete receiver coverage of the turbine arrays, hence fish may have been in the vicinity of the turbines but undetected. The temporal pattern of detections suggests most post-smolts were present within the OWDs for one or two discrete events (Figure 9, Figure 10 ). The temporal spread of detections was generally longer in Y1 than Y2, spanning a period of 17 days in Y1 and 7 days in Y2.

Movement patterns within the OWDs, including the direction and total distance travelled, were inferred using assumed straight-line paths for post-smolts detected at two or more receivers within discrete periods that were temporally separated by less than 24 h (14 fish in Y1 and 12 fish in Y2). The bearings of assumed straight line trajectories between successive detection locations (i.e. steps) showed, in both years, a clear tendency for south and south-easterly movements (Figure 11). In Y1 there was also a notable trend towards easterly and northerly swim paths; 23% of steps were on a northerly or north-easterly bearing.

The total distance travelled by individual post-smolts within the OWDs ranged from 0.51 to 13.91 km (median 6.82 km) in Y1 and from 1.14 to 45.29 km (median 9.33 km) in Y2 (Figure 12). Individual trajectories were highly variable (Appendix 1).

Temperature and swimming depth

Of the 30 smolts tagged with transmitters with temperature and pressure sensors (Y2 only), temperature data were recorded from 14 individuals (298 detections) and depth data from 15 individuals (1211 detections) while within the marine receiver array. Recorded temperatures ranged from 7.85 to 10.67°C (mean 8.52 ± 0.43 S.D.) and recorded depths were consistently within the surface layer (range 0–3.31 m) with 98% of values 2.05 m or shallower. The median swimming depth recorded during the day (sunrise to sunset) was 0.30 m (range from 0.0001 to 3.31 m) (681 values from 15 individuals). The median swimming depth recorded during the night (sunset to sunrise) was 0.0001 m (range from 0.0001 to 0.31 m). These data provide no evidence of predation on post-smolts by mammalian predators while within the marine array. Predation by ectothermic predators such as piscivorous fish cannot be ruled out, but depth profiles did not show any sudden rapid changes in depth or use of deeper waters which can be indicative of such predation.

Adult sea trout

Detections from all of the 17 adult sea trout released in Y1, and from 45 of the 52 sea trout released in Y2 were retained within analyses. The overall time difference between the first and last marine detections of individuals ranged from 25 minutes to 17 days 6 h (median 31 h 19 minutes 48 s) in Y1 and from 37 minutes 10 s to 129 days 1 h 12 s (median 2 days 21 h 22 minutes) in Y2. Fish dispersed from their release points and were frequently detected within other parts of the array; 18% and 20% of the last marine detections were in the Moray East OWD in Y1 and Y2, respectively (Figure 13 and Figure 14).

Residency within the offshore wind developments

The total duration of residency events per fish within the OWDs ranged from 25 minutes 8 s to 16 h 36 minutes 36 s (median 3 h 6 minutes) in Y1 and from 10 minutes 41 s to 5 days 14 minutes 24 s (median 5 h) in Y2. The minimum total distance covered during these residencies (track length) ranged from 3.50 km to 201.32 km (median 27.47 km) and from 1.20 km to 384.60 km (median 21.76 km) in Y1 and Y2, respectively.

The longest total durations (all fish combined) were associated with receivers near the release sites in the Beatrice OWD, although there were total durations up to 2h (Y1) and 18 h (Y2) spread throughout the Moray East OWD (Figure 15). The maximum total duration per receiver was 7 h 13 minutes 12 s in Y1 and 49 h 13 minutes 48 s in Y2.

Comparing residency and behaviour in Beatrice OWD and control area

A minimum of 12 sea trout individuals were detected, beyond the 30 minute acclimation cut-off, in each of the three comparable ‘test areas’ (i.e. identical hexagons of receivers in Beatrice North, Beatrice South and “Control Area”, see Figure 2) per year (Table 3).

Residence duration in each area per fish per day ranged from less than one second to 8 h 10 minutes 48 s in Y1 and from less than one second to 14 h 40 minutes 48 s in Y2, with no difference between areas (test area explained 0.36% of residual deviance in GLMMs).

A Roaming Index, the average number of stations within an area visited by a fish per hour, used to infer exploratory behaviour, did not vary between test areas for either year. Sea trout in Y2 visited a median of 2.24 and 2.40 stations h^-1 in Beatrice North and Beatrice South, respectively, compared to a median of 2.90 stations h^-1 in the control area. Test area was not a significant predictor of Roaming Index, explaining just 0.2% of residual deviance in GLMMs.

Depth use

Swimming depth data which were available for all 17 of the sea trout released in Y1 and 34 of the sea trout released in Y2 showed greatest use of the top 10% and bottom 20% of the water column with fish tending not to dwell in the central zones (Figure 16,Figure 17). In Y2, sea trout were most frequently detected at in water depths of 40-55 m (Figure 17).

Comparison of swimming depth values recorded in both the OWDs (Beatrice and Moray East) with those from the Control Area revealed a similar bimodal distribution (Figure 18). In the OWDs, 75% of recorded depths were at 75% of seafloor depth or deeper compared to 54% of values in the Control Area, suggesting slightly less use of the deepest water depths in the Control Area.

Due to depth data being collected only at the timestamps when fish were in the detection zone of a receiver, robust reconstruction of individual depth profiles was limited by temporal gaps in the detection record. However, data from a small number of fish with extended periods of near continuous detections provided more detailed insight into the range of behaviours exhibited by sea trout in the current study. These included, for example, repeated dives from the surface to depths of 25 to 30 m, which occurred exclusively during the daytime, with the fish dwelling near the surface overnight (Figure 19). Among several other fish, there was a pattern of fish spending time near the surface (hours or days) before diving to great depth (> 40 m) and remaining there for extended periods i.e. multiple days (Figure 20, Figure 21). Some fish were only detected in surface waters, with the majority of their recorded swim depths within the top 10 m of the water column (Figure 22).

Temperature data

Temperature data recorded by the marine receiver array fell within a narrow range in both Y1 (11.30 – 15.38°C) and Y2 (10.67 – 13.96°C) (Figure 23) and therefore provided no evidence of predation of sea trout by mammals in either year.