Behavioural responses in wild cod (Gadus morhua L.) exposed to fish holding water

Local fishermen in several areas of Norway assert that salmon farms have caused the wild migrating cod to change their migratory behaviour, so that they no longer enter their natural spawning grounds in the fjords. If the asserted changes in behaviour of wild cod populations can be linked to establishment of salmon farms, water-soluble odorants are then possible candidates to explain such a connection. Chemical stimulants are important to the individual fish's conception of the surrounding environment. High density stocks of fish in a farm are expected to release large amounts of waterborne information. The present laboratory experiments were conducted to test behavioural responses in Atlantic cod exposed to water containing metabolites and waste from farmed salmon. The trials were conducted on single fish in a multiple chamber preference system. The results show that migrating wild Atlantic cod chose to spend more time in chambers without addition of water from the salmon tank, regardless of their maturation status, and even at very low concentrations (0.2%). The avoidance is probably due to presence of chemical compounds with olfactory properties from salmon, since anosmic cod did not elicit such response. Farmed cod, on the other hand, does not avoid water from the salmon tank, and stationary wild cod caught nearby a fish farm had a less pronounced response as compared to wild migrating cod. The response seen is not species specific, as wild migrating cod responded similar to water from a tank holding farmed cod as to water from salmon. These results do not preclude fishermen's observations that cod change their behaviour in areas with fish farming activity. Such a change in behaviour could be a response to water-soluble odorants, but needs to be validated and detailed in further laboratory experiments as well as in nature before any conclusions can be made.     

Migratory behaviour of adult wild and escaped farmed Atlantic salmon, Salmo salar L., before, during and after spawning in a Norwegian river

The migratory behaviour of adult wild and escaped farmed Atlantic salmon, Salmo salar L., before, during after spawning in the River Namsen, Norway, was analysed using radio telemetry. The fish were caught, radio tagged and released into the fjord between 7 and 25 km from the river mouth. A significantly higher proportion of wild (74%) than farmed (43%) salmon was subsequently recorded in the river. Wild salmon (33%) were more frequently captured in the sea and in rivers than farmed salmon (14%). The migration speed from release to passing a data logger 11 km upstream from the river mouth was not significantly different between wild (20.6 km day^?1) and farmed (19.8 km day^?1) salmon. Wild salmon tagged when water flow in the river was increasing had a significantly higher migration speed than wild salmon tagged when water flow was decreasing. This was not true for farmed salmon. Farmed salmon were distributed significantly higher up the river than wild salmon during spawning, although both types of fish were found together in spawning areas. Thus, there was no geographical isolation to prevent spawning between wild and escaped farmed salmon. Farmed salmon had significantly more and longer up- and downstream movements than wild salmon during the spawning period. Unlike farmed salmon, the number of riverine movements by wild salmon increased significantly when variation in water flow increased. A smaller proportion of wild (9%) than farmed (77%) salmon survived through the winter after spawning.     

Movements and spatiotemporal distribution of escaped farmed and local wild Atlantic cod (Gadus morhua L.)

Commercial farming of Atlantic cod (Gadus morhua L.) is now being developed in several countries. The ecological consequences of cod culture are poorly understood, but recent research suggests that Atlantic cod are more prone to escape from net pens than Atlantic salmon. Here, we describe the movements and the spatiotemporal distribution of farmed cod after escape relative to wild cod, both during and outside the natural spawning season. The experimental design included simulating escape incidents of farmed cod tagged with acoustic transmitters and using an array of automatic listening stations to monitor their dispersal and distribution. For comparison, local wild cod were monitored using the same array of receivers. The farmed cod dispersed rapidly after a simulated escape, they randomly distributed over large areas and their distribution overlapped with local wild cod. Moreover, escaped farmed fish were found at local cod spawning areas during the spawning season. The study also indicated that the recapture rate of escaped farmed cod was high compared with that of escaped farmed salmon. Thus, while our results showed that there is a considerable potential for ecosystem effects caused by escaped farmed cod, mitigating actions such as an efficient recapture fishery for escapees may be possible.     

Tracking escaped Atlantic cod (Gadus morhua L.) aggregated around Norwegian farms: Considerations for management strategies

To examine site fidelity and potential for ecosystem impacts, 20 escapee cod (Gadus morhua L.) from fjord‐based cod farms were tagged with acoustic transmitters, released in the vicinity of the same farm, and their subsequent behaviour and survival monitored using a fixed array of acoustic receivers. Most individuals demonstrated high fidelity to their farm where tagged, and this increased during the polar night. Only five individuals (25%) were occasionally detected at another farm (4 km distance), and only three cod (15%) migrated out of the fjord during the 6‐month study period. Vertical movements were consistently greater in the hours around noon than at night, a behaviour inconsistent with that observed for wild fish. A high mortality rate (40%) was observed during the first 4 months post‐release, and actual mortality may have been even higher, limiting the potential for negative ecosystem effects from cod escapees. Findings from this study are used to discuss the development of recapture strategies for minimising potential impacts from cod escapees to ensure sustainable farming activity.     

Effect of smolt age on migratory behaviour of Baltic salmon, Salmo salar L., transplanted to the east Atlantic

Abstract. This paper describes tests which were made to determine if smolt age at release influences the migratory pattern of three different stocks of salmon, Salmo salar L. The fish were hatchery-reared and released in two different rivers, the River Imsa and the River Akerselv. Based on the tag returns we found that Baltic salmon from the River Neva, USSR differed in migratory pattern from two Norwegian stocks from the River Lone and the River Imsa. A large proportion of the 2+ River Neva smolts stayed in the fjord during the summer and autumn after release. On the other hand, 2+ smolts of the Norwegian stocks left the fjord and migrated to the feeding areas in the Norwegian Sea within a short time after release. The 1+ smolts of all stocks showed the same migratory pattern as the 2+ smolts of Norwegian origin. We propose that the observed differences in migratory pattern are influenced by the developmental rate of the smolts. The effect of developmental rate on the migration may differ among stocks. Our results show that it is possible to develop a fjord fishery by releases of 2+ smolts of Neva salmon. However, such releases must be carried out with the utmost caution, preferably in fjords with no salmon rivers, so that the possibility of gene flow between populations is minimized.     

Dispersal of large farmed Atlantic salmon, Salmo salar, from simulated escapes at fish farms in Norway and Scotland

Abstract  Individually tagged farmed Atlantic salmon, Salmo salar L., were released from fish farms in simulated escapes in Scotland ( n  =   678) and Norway ( n  =   597) to compare migratory behaviour and dispersal. Large fish (510–870 mm fork length) were released to simulate the escape of aquaculture growers. Fish were released in spring and all recaptures of tagged fish were reported during summer and autumn of the year of release. Recapture rates were respectively 7.0 and 0.6% of the salmon released in Norway and Scotland, indicating a higher mortality or a lower exploitation or reporting rate for Scottish fish. Recaptures of Norwegian fish were all from Norway and mostly within 150 km of the release site; 64% were taken by anglers in fresh water. By contrast, the three salmon recaptured from the release in Scotland were reported from Norway (Hardangerfjord and Lofoten) and western Sweden (River Göta); two detached tags were found on beaches in Scotland north of the release site. These findings establish the capacity for long distance dispersal among escapees from aquaculture facilities and suggest a net easterly bias in long distance dispersal of salmon escaping from Scottish fish farms.     

Salmon farming affects the fatty acid composition and taste of wild saithe Pollachius virens L.

There is an ongoing controversy on whether fish farming affects the quality of wild fish in fjords. In northern Norway, local people prefer not to eat saithe, Pollachius virens L., from areas in the vicinity of fish farms because they say the taste is inferior to saithe from other areas. To address this issue, saithe were collected in the vicinity of a salmon fish farm in a fjord in northern Norway and in two reference areas: one site 6 km away from the nearest fish farm in the same fjord, and the other in a fjord nearby with no fish farms. The objective of this study was to clarify whether the physiology or taste of saithe near fish farms differs from saithe in areas with no or limited influence from fish farms. The fish collected near farms were larger than those from control sites of the same age and had been eating pellets. Analysis of fatty acid composition of the fillet indicated that fish collected near the farms had fatty acid profiles that resembled that of pellets more than fish from reference sites. The results suggest that there were individual differences in the time spent near cages for saithe collected near the farm, and that saithe collected at the reference area in the same fjord had also been eating pellets, i.e. had visited at least one of the fish farms in the fjord. In a sensory test, the test panel found that saithe from the fjord without fish farms tasted better than saithe collected near the cages. The test panel found no clear differences in taste between saithe collected near the fish farm and saithe from the same fjord, but 6 km from the nearest fish farm.     

Prespawning migratory behaviour and spawning success of sea-ranched Atlantic salmon, Salmo salar L., in the River Gudenaa, Denmark

The migratory behaviour of sea-ranched Atlantic salmon, Salmo salar L., was analysed by radio-telemetry in the River Gudenaa, Denmark. The main objectives were to: (1) estimate mortality of returning adults through the fjord; (2) observe rate of progression and migratory pattern in the fjord and river; and (3) record whether spawning occurs in the river. Forty-two returning salmon (19 males and 23 females of total body length from 60–97 cm) reared and released as smolts, were caught and equipped with external radio transmitters in the outer estuary of the River Gudenaa in 1994 and 1995. Of the tagged salmon, 18 (43%) were caught in the estuary, four (10%) were not recorded after release and 20 (47%) entered the river. The mean rate of progression through the fjord was 7.6 km d−1 (range 1.4–18.2) in 1994 and 5.4 km  d−1 (range 1.6–17.1) in 1995. Eleven salmon were alive at the onset of the spawning period. Eight were retrieved dead from the river during or after the spawning period; four with empty gonads assumed to be successful spawners, and four with intact gonads. In 1994, unsuccessful spawners (found dead with intact gonads) entered the river earlier and had a longer total migration distance in the river compared to successful spawners. This suggests that spawning success of sea-ranched salmon is associated with time of river entry and river migration length.     

Survival of Atlantic salmon captured in and released from a commercial trap-net: Potential for selective harvesting of stocked salmon

The major wild Atlantic salmon stocks in the Baltic Sea began to recover in the late 1990s. This recovery has been partly due to strict regulations in the Gulf of Bothnia that effectively prevent salmon fisheries during the peak migration. About half of the migrating salmon, however, are reared fish that could be harvested. We simulated a limited trap-net fishery that selectively harvested reared salmon and released wild fish, and studied the survival and migration of the released salmon. We tagged and released 1970 salmon caught in the trap-nets along the coast in 2001 and 2002. The mean maximum capture and release induced mortality of salmon was 11%, ranging between 4% and 21% in different release groups by year, sea age and number of releases. The cumulative mortality for the total salmon population on their spawning migration in the Gulf of Bothnia was below 5%, and it would not increase considerably after the first capture and release events, provided fishing effort is not excessive and fish are handled properly. Survival of trap-net captured and released Baltic salmon appears high and their migration behavior is not altered due to this handling. Several preconditions, however, should be considered before selective fishing is introduced in the Gulf of Bothnia salmon fishery.     

Residency of adult Atlantic cod (Gadus morhua) in the western Gulf of Maine

A mark and recapture study of cod in the western Gulf of Maine was conducted to study the seasonal movements of fish, particularly as they related to areas closed to commercial fishing. A total of 27,772 cod were tagged, and 1334 (4.8%) were recaptured with sufficiently detailed recapture location and date to be included in the study. Results indicated that the group is resident to the area and sedentary. Although there were a small percentage of fish (2.5%) that traveled long distances (>100 km), most were recaptured near their release location. There was no linear relationship between fish length and linear distances traveled, and linear distance traveled was only weakly related to days-at-large. For all groups of cod tagged and released in particular areas and months, mean distances traveled were small (<65 km), rates of travel were slow (<0.2 km/day), and rates of group dispersion were <50 km²/day. There were no recognizable spatial or temporal patterns in the mean angles of travel for groups released in various month/area combinations, and the angular deviations associated with the mean angles were quite large. Although there seems to be little pattern in the movement of cod in our study area, temporal and spatial changes in abundance indicate that movements are occurring.

The general pattern was a concentration of large cod into one small, inshore area (Area 133) in both the spring and winter, and dispersion from this area in the ensuing months. Monthly percentages of ripe cod in Area 133 provide evidence for two spawning groups; a winter group that spawns from November through January, and a spring group that spawns from April through July. Thus it is likely that the observed spring and winter concentrations of fish in Area 133 were associated with spawning. Fish in the spring group were the largest encountered in the study. The study also provided some evidence of natal homing, i.e. a return to the same spawning grounds year after year, for the spring spawning group. Although there is some exchange of fish between most of the closed areas, the timing of the closures appears to protect the largest aggregations of cod.     

Comparative analysis of stomach contents of cultured and wild cod, Gadus morhua L.

In a cod enhancement project in a small fjord in western Norway, two groups of reared cod, Gadus morhua L., were released with 6 months’difference in time. After the last release a 3-week fishing survey was conducted. Stomachs from wild and recaptured reared cod of similar size were sampled and analysed. The last released cod had learned to catch the same prey types and in the same relative proportions as the wild cod the second week after release, but mean weight of stomach contents and mean numbers of prey items per stomach were smaller, and a larger proportion of the newly released fish had empty stomachs. Three weeks after release the proportion of empty stomachs was similar to that of wild cod and the first released reared cod, but the food intake was still smaller. The first released group of reared cod, which had spent 6 months in the fjord, seemed to feed as efficiently on wild prey as did wild cod.     

Quality of wild-captured saithe (Pollachius virens L.) fed formulated diets for 8 months

Fish farms may attract wild fish that feed on waste feed from the cages. Saithe, Pollachius virens L., are particularly numerous around salmon cages in northern Europe and may obtain a significant proportion of their diet from waste feed. It has been claimed that these fish are of inferior quality to saithe that feed on natural diets; differences are said to include soft muscle tissue and a different taste. In order to document such changes in quality we performed a feeding experiment. Young wild saithe were collected and fed either a lipid-rich salmon diet or a lean cod diet for 8 months. All fish were individually tagged and growth was monitored throughout the experiment. Parameters related to flesh quality were measured. Diet clearly influenced the growth rate of the fish, and many fish reached a very high hepatosomatic index when fed on a salmon diet. However, many fish had a low feed intake and thus a low rate of growth. There were some differences in skin and muscle colour, pH and in sensory parameters between wild-caught and artificially fed saithe at the end of the experiment. Those fed the cod diet were more similar to wild saithe than those fed the salmon diet.     

Marine migrations in anadromous brown trout (Salmo trutta). Fjord residency as a possible alternative in the continuum of migration to the open sea

Partial migration is a common phenomenon in many fish species. Trout (Salmo trutta) is a partially migratory species where some part of the population migrate to the marine environment, while another remains in freshwater. In the years 2008 and 2009, a total of 159 wild sea trout smolts were tagged with acoustic and PIT‐tags in the river Villestrup (Denmark) to study the initial postsmolt marine behaviour within a fjord system. We found that the strategies of the sea migrants vary: some stay in the fjord, while others migrate to the sea, suggesting that partial migration occurs even in the marine environments. Overall, a total of 53% of the tagged smolts migrated from the fjord to the sea, and 47% stayed (or potentially died) in the fjord. The ratios of fjord‐resident versus sea‐migrating postsmolts were consistent at the study times, and no differences between the early and late migration periods of the smolts were observed. The individual's size or body condition at the time of tagging did not affect survival or the migratory decisions in the fjord. High overall initial survival (74%) was found 30 days after the fjord entry. We suggest that within a continuum of migration to sea, there is a migratory decision point when sea trout postsmolts encounter a fjord system. At this point, postsmolts will assess the possibility of migration versus the alternative of fjord residency.     

Prespawning migratory behaviour of wild and farmed Atlantic salmon, Salmo salar L, in a north Norwegian river

The migration patterns of wild and released farmed Atlantic salmon, Salmo salar L., were studied by radio-telemetry during migration from entering a river to after spawning. The wild salmon were caught during return migration in bagnets and the farmed salmon were produced in a near by fish farm. Both groups were tagged and released at sea near the river mouth of the River Alta, North Norway. A pronounced individual variation in migration pattern was observed. No significant differences was found between wild and farmed salmon in the distance from entering the river mouth to the place of the first stop (recorded in the same pool for 5 days or longer) and the days from entering the river mouth to arrival at the first stop. For both wild and farmed salmon, mean migration speed was 2.6 km day¹, varying from 0.4 to 11.0 km day^-1. A larger proportion of farmed salmon distributed to the upper part of the river at spawning; mean distances from the river mouth were 30.1 and 19.1 km. respectively. Farmed salmon spent a significantly longer time from entering the river to reaching the area occupied during spawning.     

Caligus elongatus Nordmann genotypes on wild and farmed fish

Two mitochondrial genotypes have been described for Caligus elongatus Nordmann in Norway. This article reports on the distribution of C. elongatus mitochondrial cytochrome C oxidase 1 genotypes from wild fish hosts from the SE Norwegian coast. For comparison, lice from areas with fish farming were included in the study. The genotype distribution of 841 lice from wild coastal (n = 535), wild North Sea pelagic (n = 26), farmed (n = 160) and wild hosts in areas of fish farming (n = 89) is presented. The genotype frequencies of C. elongatus on wild coastal hosts varied significantly between spring and autumn. Lice from these fish show a large proportion of genotype 1 lice in March-June every year. Genotype 2 lice were found more frequently in autumn. Genotype 1 was clearly associated with the lumpfish, Cyclopterus lumpus L. The genotype frequency appeared to be different in areas with aquaculture. Caligus elongatus from farmed fish and wild fish caught close to Atlantic salmon fish farms in Norway were predominantly genotype 1 in autumn. Genotypes of C. elongatus on the SE coast of Norway vary according to season and fish species. Factors involved in the encounter between fish and lice are important for the establishment of lice on their hosts.     

Occurrence of Francisella piscicida in farmed and wild Atlantic cod, Gadus morhua L., in Norway

Francisellosis, caused by the bacterium Francisella piscicida , has become one of the most serious diseases in Atlantic cod production in Norway. The major aim of this study was to determine the distribution of F. piscicida in farmed and wild fish in areas with cod farming along the Norwegian coast, and its occurrence in cod from areas without cod farming. Two real-time PCR assays, targeting the 16S rRNA gene and the FopA gene of F. piscicida , were developed since sensitive and specific diagnostic tools are required for detecting asymptomatic carriers of the bacterium. A total of 422 wild cod from 13 sampling areas and 955 farmed cod from 10 areas along the coast of Norway were examined. Using the real-time polymerase chain reaction (PCR) assays, F. piscicida was detected in wild populations of cod from all counties examined south of Sogn og Fjordane in southern Norway (overall prevalence 13%, n  = 221). Wild cod north of Sogn og Fjordane were negative for the bacterium ( n  = 201). Farmed cod from most parts of Norway were F. piscicida positive. The apparent absence of the bacterium in wild populations of cod in the northern parts of Norway and its widespread occurrence in wild cod from southern parts of Norway is believed to relate to differences in seawater temperatures.     

The selectivity for cod (Gadus morhua L.) in two experimental trammel-nets and one gillnet

The present study analyses the selectivity for cod in two trammel-nets (three-wall nets) and one gillnet (single-wall net) used during an experimental cod fishery conducted monthly from october 1985 to December 1988 in a Norwegian fjord. It is part of a cod enhancement experiment conducted in the fjord. The three nets were grouped and utilized to sample the wild cod population. In addition, the gillnet and one trammel-net were utilized to sample a tagged group of cod with a known length composition. Relative selectivity curves were estimated indirectly for the two trammel-nets by comparing the catches of each length class within the length range 18–58 cm. Absolute selectivity curves were estimated directly for the gillnet and one trammel-net for the length range 15.0–22.0 cm. These were utilized to verify the indirectly estimated curves. The trammel-net selectivity curves obtained were used to provide an estimate for the true length-frequency distribution of cod in the fjord. Finally, it was tested whether a group of the three nets may be utilized as a unit effort to provide approximate random samples from the population.     

Sea lice infestation rates on wild and escaped farmed Atlantic salmon (Salmo salar L.) entering the Magaguadavic River, New Brunswick

A monitoring program for the prevalence and intensity of sea lice infestations of wild and escaped farmed salmon has been underway on the Magaguadavic River since 1992. Fish are screened in a fish ladder trap located in freshwater a short distance above the head of tide. No trends with time were evident in observed sea lice burdens, and in all years the majority of salmon, both wild and escapees, had no or low levels of infestation with sea lice. In the spring of 2002, 23 landlocked salmon moving to sea from the Magaguadavic River were acoustically tagged. Two fish returned to the river after a brief period of residence in Passamaquoddy Bay, with significant dermal damage from sea lice. These fish were tracked to areas close to commercial salmon farms.     

Migratory delay leads to reduced passage success of Atlantic salmon smolts at a hydroelectric dam

Passage of fish through hydropower dams is associated with mortality, delay, increased energy expenditure and migratory failure for migrating fish and the need for remedial measures for both upstream and downstream migration is widely recognised. A functional fish passage must ensure safe and timely passage routes that a substantial portion of migrating fish will use. Passage solutions must address not only the number or percentage of fish that successfully pass a barrier, but also the time it takes to pass. Here, we used radiotelemetry to study the functionality of a fish bypass for downstream‐migrating wild‐caught and hatchery‐released Atlantic salmon smolts. We used time‐to‐event analysis to model the influence of fish characteristics and environmental variables on the rates of a series of events associated with dam passage. Among the modelled events were approach rate to the bypass entry zone, retention rates in both the forebay and the entry zone and passage rates. Despite repeated attempts, only 65% of the tagged fish present in the forebay passed the dam. Fish passed via the bypass (33%), via spill (18%) and via turbines (15%). Discharge was positively related to approach, passage and retention rates. We did not detect any differences between wild and hatchery fish. Even though individual fish visited the forebay and the entry zone on multiple occasions, most fish passed during the first exposures to these zones. This study underscores the importance of timeliness to passage success and the usefulness of time‐to‐event analysis for understanding factors governing passage performance.     

Life‐history differences of juvenile Chinook salmon Oncorhynchus tshawytscha across rearing locations in the Shasta River,California

Salmon from different locations in a watershed can have different life histories. It is often unclear to what extent this variation is a response to the current environmental conditions an individual experiences as opposed to local‐scale genetic adaptation or the environment experienced early in development. We used a mark–recapture transplant experiment in the Shasta River, CA, to test whether life‐history traits of juvenile Chinook salmon Oncorhynchus tshawytscha varied among locations, and whether individuals could adopt a new life history upon encountering new habitat type. The Shasta River, a Klamath River tributary, has two Chinook salmon spawning and juvenile rearing areas, a lower basin canyon (river km 0–12) and upper basin spring complex (river km 40–56), characterised by dramatically different in‐stream habitats. In 2012 and 2013, we created three experimental groups: (i) fish caught, tagged and released in the upper basin; (ii) fish caught at the river mouth (confluence with the Klamath River, river km 0), tagged and released in the upper basin; and (iii) fish caught at the river mouth, tagged and released in the lower basin. Fish released in the upper basin outmigrated later and at a larger size than those released in the lower basin. The traits of fish transplanted to the upper basin were similar to fish originating in the upper basin. Chinook salmon juvenile life‐history traits reflected habitat conditions fish experienced rather than those where they originated, indicating that habitat modification or transportation to new habitats can rapidly alter the life‐history composition of populations.