ICES Journal of Marine Science, 56: 422–432. 1999 Article No. jmsc.1999.0453, available online at http://www.idealibrary.com on Evaluation of an alternative strategy to enhance salmon populations: Cage rearing wild smolts from Conne River, Newfoundland J. B. Dempson, V. A. Pepper, G. Furey, M. Bloom, T. Nicholls, and G. Hoskins Dempson, J. B., Pepper, V. A., Furey, G., Bloom, M., Nicholls, T., and Hoskins, G. 1999. Evaluation of an alternative strategy to enhance salmon populations: Cage rearing wild smolts from Conne River, Newfoundland. – ICES Journal of Marine Science, 56: 422–432. Five-thousand wild Atlantic salmon (Salmo salar L.) smolts from Conne River, Newfoundland, were captured during their downstream migration in May 1995, and transferred to an estuarine aquaculture rearing site at Roti Bay, 23 km away. Survival was monitored throughout the experiment. The greatest mortality occurred in July, approximately 6–8 weeks following transfer. Survival of smolts to one-sea-winter salmon was 18.5%, over four times higher than the average survival of wild salmon to Conne River during the past 6 years. Growth was monitored at monthly intervals until November 1995, with additional sampling in the spring and early summer of 1996. Survivors were split into two groups and released directly into the Bay d’Espoir fjord; one group was released 27–28 June 1996, at a site approximately 7 km from the mouth of Conne River. The second group was retained at Roti Bay and released 23 July 1996. Lotek radio transmitter tags were used in evaluating the success of the experiment by tracking migration timing and subsequent distribution of cage-reared salmon throughout the Conne River system. Approximately 80% returned to Conne River and 20% strayed to other streams. Less than 50% of the surviving fish were later accounted for in local Bay d’Espoir rivers. Results are discussed in relation to the utility of this technique to enhance salmon populations. 1999 International Council for the Exploration of the Sea Key words: Atlantic salmon, enhancement, growth, survival, migration, wild smolts. Received 27 March 1998; accepted 25 September 1998. J. B. Dempson, V. A. Pepper, G. Furey, M. Bloom, T. Nicholls: Science Branch, Department of Fisheries and Oceans, PO Box 5667, St John’s, Newfoundland, Canada A1C 5X1; G. Hoskins: Fisheries Division, Miawpukek Mi’kamawey Mawi’omi, Conne River, Newfoundland, Canada, A0H 1J0. Correspondence to J. B. Dempson: tel: +1 709 772 4475; fax: +1 709 772 3578; e-mail: [email protected] Introduction Enhancement of Atlantic salmon (Salmo salar L.) populations is a common practice in many areas where the species occurs naturally. The idea to enhance salmon stocks has arisen as a consequence of a combination of effects including poor marine survival, over exploitation of the resource, loss and destruction of habitat, and the introduction of foreign organisms (Jonsson and Fleming, 1993). For example, production of Atlantic salmon has been severely reduced in many rivers in Nova Scotia, Canada, as a result of acidification (Lacroix, 1992; Korman et al., 1994), while salmon have now been eradicated from a number of rivers in Norway 1054–3139/99/040422+11 $30.00/0 (Jonsson and Fleming, 1993) due to the unintentional introduction of the parasite Gyrodactylus salaris. Stocking of various salmon life stages (fry, parr, smolt) has been the most common means to try to offset declines in natural production (Jonsson and Fleming, 1993). Hatchery propagation, however, has not been without its problem and critics (Hilborn, 1992; Meffe, 1992; Brannon, 1993; Hilborn and Winton, 1993; Riddell, 1993; Washington and Koziol, 1993). Many of the concerns regarding hatchery propagation of salmonids for release to natural habitats relate to the loss of genetic diversity in a population and reduced fitness of individuals (Jonsson and Fleming, 1993; Riddell, 1993; Verspoor, 1997). Additional concerns 1999 International Council for the Exploration of the Sea Evaluation of an alternative strategy to enhance salmon populations 423 N illtown) 0 15 30 km N.W. Brook (Tailrace) S.E. Bro ok (M 48°00' N 47°55' g in nt n Co ) ill s on ok Roti Bay Bay d' Espo ir 47°50' ro i eR ttl Li 55°50' n lli Co 55°45' k oo r sB 2 Bois Island 55°55' r ve S. 0 47°45' (C B E. Conne River Village em n fe u Co Ri ite e nc Swanger' s Cove Brook r ve ne 55°40' 4 6 km 8 10 55°35' W Figure 1. Location of Conne River, Bay d’Espoir, Newfoundland, and the Roti Bay aquaculture grow-out site, along with other place names mentioned in the text (Scale 1: 250 000). Insert illustrates the general location of Conne River in Newfoundland. Release sites of tagged salmon: , cage release site; , river release site. include increased competition with local wild populations, impacts of accidental straying, and the introduction of diseases (Saunders, 1991; Hilborn, 1992; Jonsson, 1997; McVicar, 1997). In contrast, others believe and provide evidence that hatcheries are useful and have been successful in various places (Bielak and Davidson, 1993; Incerpi, 1996). The difference in opinions regarding the merits of enhancement of wild fish populations likely is due to a historic failure to recognize the inherent dichotomy between fish culture techniques applied to wild stock enhancement and those appropriate to aquaculture initiatives (Pepper and Crim, 1996). Nevertheless, a variety of enhancement techniques are now commonly applied to Atlantic salmon (Bielak and Davidson, 1993; Davidson and Bielak, 1993). One strategy not discussed by Davidson and Bielak (1993) but advocated by Thomas (1996) consists of capturing out-migrating smolts from a river, transporting them to a salt-water rearing facility, rearing them to maturity, and subsequently returning them to their native stream. A modification of this was applied to Atlantic salmon from the Big Salmon River, New Brunswick (Bruce, 1995). Here, hatchery-reared smolts were transported to a sea cage and reared for about 16 months to maturity. Mature fish were trucked back to the river and released. Such strategies lend themselves to situations in which local stocks are perceived to require rebuilding assistance, where logistics and infrastructure are readily available, and where future reliance on natural spawning requires minimizing the potential for domestication selection. This situation exists in Newfoundland on the Conne River. Conne River flows into Bay d’Espoir on the south coast of Newfoundland (Figure 1). It is a one-sea-winter (1SW) salmon river with a modal smolt age of 3 years (Dempson and Stansbury, 1991). Total returns of adult Atlantic salmon, obtained from enumerating salmon at a fish counting fence, have ranged from a high of 10 000 fish in 1987, to a low of 1600 fish in 1994 (Beacham and Dempson, 1998; Dempson et al., 1998). Over this interval, marine survival from smolts to 1SW salmon declined from 7–10% to 2.6%, and the river was closed to recreational fishing from 1994 to 1996. The decrease 424 J. B. Dempson et al. Table 1. Summary of the numbers of smolts transferred from the fish counting fence trap at Conne River to the sea cage at Roti Bay (n), along with the mortalities (M) and the numbers of fish surviving (S) in the sea cage at various dates. Date 1995 13 May 16 May 17 May 18 May 18 May 20 May 21 May 23 May 25 May 26–31 May 1–30 Jun 1–31 Jul 1–31 Aug 1 Sep–31 Dec 1996 1 Jan–15 May Total n 237 325 472 594 752 771 782 633 434 5000 M S 1 77 3 1 9 107 50 524 2543 230 168 237 562 1034 1627 2302 3070 3851 4475 4802 4752 4228 1685 1455 1287 361 4074 926 (18.5%) in adult salmon returns continued even though smolt production was relatively stable. In an effort to compensate for precipitous declines in marine survival of Conne River Atlantic salmon smolts, an experiment similar to that advocated by Thomas (1996) was carried out. Wild Atlantic salmon smolts were captured as they migrated out of Conne River, transported to a sea cage rearing facility, and reared for 13 months. Surviving salmon were released directly into the Bay d’Espoir fjord which extends inland from the coast about 45 km. This paper evaluates the utility of this technqiue as an alternative strategy to enhance salmon populations by comparing survival and growth of cage-reared fish with wild salmon returns. Information on homing response and return rates of cage-reared salmon to Conne River is provided along with the results of radio-telemetry tracking investigations in the context of discussing modifications that may be required before the method can be widely applied. Materials and methods Smolt capture, transfer, and rearing Five-thousand Atlantic salmon smolts, or 8% of the total estimated run in 1995, were captured at a fish counting fence (Dempson and Stansbury, 1991) located in the lower section of Conne River about 0.75 km above the head of tide (Figure 1). Smolts were caught over a 13 day period from 13 to 25 May 1995 (Table 1), during which time approximately 67% of the total smolt run emigrated from Conne River. Smolts were dipped from the fish trap into a transfer box with recirculated water. There was no selection for smolts of any particular size. Smolts were then transported by boat 23 km to Roti Bay. Generally, the trip was completed in 4–5 h. At the sea cage site, smolts were dipped from the transfer box into a sea cage that had a circumference of 70 m and depth of 6 m. All equipment used in the transfer process was disinfected before and after use with an iodine-based solution (Wescodine). During spring and summer, Roti Bay is characterized by a large gradient in both temperature and salinity (Sutterlin and Stevens, 1992), and the smolts could select the salinity they desired. The feeding regime varied among seasons and is summarized in Table 2. Fish sampling Dead smolts and postsmolts were removed each day and counted. At approximately 1 month intervals, samples of at least 25 specimens were measured for length and weight. Monthly sampling was limited to this smallsample size to minimize handling stress. Fish were anaesthetized with 2-phenoxy-ethanol, weighed on an electronic balance (0.1 g), and fork length was measured to the nearest millimetre. Specimens were allowed to regain their equilibrium in a recovery bath before being returned to their respective cages. Sampling continued until mid-November, 1995, when a complete inventory of remaining fish was carried out. Subsampling for length and weight resumed in May 1996, when another inventory of surviving fish was done. Release of surviving adults and tagging operations Prior to releasing the salmon, a fish health evaluation was undertaken on 15 May 1996, by collecting 60 wild smolt specimens from the Conne River fish counting fence, concurrent with sampling of 55 of the cage-reared wild adult salmon from Roti Bay that were part of this experiment. All samples were taken to Halifax, Nova Scotia, and processed as per procedures of the Canadian Fish Health Protection Regulations by personnel of the Department of Fisheries and Oceans fish health laboratory. All specimens were negative for known bacterial and viral fish pathogens. Remaining salmon were separated into two different cages. Group 1 fish (n=434) were tagged with Floy ‘‘T-bar’’ anchor tags inserted immediately anterior to the dorsal fin. On 7 June 1996, fish in this cage were towed approximately 16 km to an estuarine location adjacent to the village of Conne River (Figure 1), about 7 km from the mouth of the Conne River. On 19 June, radio tags were inserted into the stomachs of 37 fish that ranged in fork length from 430 to 500 mm (Table 3). Salmon were anesthetized with Benzocaine in Evaluation of an alternative strategy to enhance salmon populations Table 2. Feeding regime used for the cage-reared wild Atlantic salmon project, Roti Bay, 1995–1996 (Eff: Ewos fry feed; SSED: Suregain smolt Excel diet). Date 1995 13 May–25 May 26 May–30 Jun 1 Jul–20 Sep 21 Sep–28 Oct 29 Oct–12 Dec 1996 13 Dec–7 Apr 8 Apr–14 May nfeeding d 1 Feed type Pellet size 6 8 5 4 3 Eff Eff SSED SSED SSED 1.5–3.5 1.5–3.5 3.5–5.0 3.5–5.0 3.5–5.0 1/2 3 SSED SSED 3.5–5.0 5.0–8.0 preparation for implant. Group 1 fish were released 27–28 June 1996. The second group of salmon (n=433; 4 fish lost in transfer) was held in a separate sea cage at Roti Bay. These fish were also tagged with Floy ‘‘T-bar’’ anchor tags but with the tag insertion immediately posterior to the dorsal fin. This was done to assist in the identification of fish from each release group as they passed through and were recorded on the video camera system used at the Conne River fish counting fence. Group 2 fish, held for another month as part of a comparative study on growth, were released 23 July. This group was sampled for length and weight prior to release. At the time of release 16 fish had lost their tags leaving 417 tagged individuals. In addition, 22 wild Conne River salmon caught at the adult salmon fish counting fence trap were also provided with radio transmitter tags on 19 June (n=19) and 27 June (n=3). These fish, which ranged in fork length from 475 to 555 mm (Table 3), were immediately released above the fish counting fence. Radio-telemetry monitoring Radio tags used were Lotek’s (114 Cabot Street, St John’s, Newfoundland, Canada A1C 1Z8) digitally encoded transmitter model MCFRT-3BM. This tag, which has a rated battery life of 238 days, is 1143 mm in size and weighs 3.7 g in water. The digitally encoded tag provides a unique numerical identification for each fish (channel; code). Tags transmitted at a frequency ranging from 149.320 to 149.800 MHz. A radio-telemetry receiving station was set up at the Conne River fish counting fence (Figure 1) on 27 June 1996. The station consisted of Lotek’s SRX–400 telemetry receiver/datalogger connected to two directionally positioned 4-element Yagi antennae. The telemetry data were downloaded from the receiver to a portable computer for processing. Downloaded information consisted of the first and last dates a tag was detected, times of 425 detection, and the number of occurrences (events) over a pre-set interval that a signal was obtained. The telemetry receiver was operated until 31 August 1996. The Conne River fish counting fence was in operation until 23 September 1996. On 4 September 1996, the lower portions of Southeast Brook (Milltown), Southeast Brook (Conne Mill), and Northwest Brook (Tailrace) were surveyed on foot with the Lotek receiver and a hand-held antenna. A section of Little River, where a fish counting fence was also in operation, was surveyed below the main falls. Ground surveys were followed by three helicopter overflights: 2 October 1996; 7 November 1996; and 10 January 1997. Watersheds covered during all overflights included: Little River, Southeast Brook (Conne Mill), Southeast Brook (Milltown), Northwest Brook (Tailrace), and Conne River. In addition, Collins Brook and Swanger’s Cove Brook were added to the second and third overflights, respectively. During the overflights, the telemetry receiver was set to scan sequentially through the six frequencies at a rate of one frequency every 6 s. When a radio-tag signal was detected, scanning was stopped and the receiver fixed on the particular frequency in question. The approximate position of the radio-tag in the river was determined by adjusting the altitude and air speed of the helicopter to obtain the maximum power signal being emitted from the tag. At that point, latitude and longitude coordinates were recorded from a Global Positioning System (GPS) aboard the helicopter. Scanning then resume and the overflight continued. Statistical analyses Size comparisons among groups of fish were addressed by analysis of covariance of log transformed weight– length relationships, unless sample sizes were too small. In this case, analysis of variance was used to compare lengths and weights of 1SW salmon. Comparisons of the recovery rates of released salmon among Bay d’Espoir rivers, and of the various tagging experiments were conducted using a test of independence (log-likelihood ratio testG-test). Results Survival and growth in cages There were few mortalities resulting from the transfer of smolts to the Roti Bay cage rearing site, with 95% of the fish alive at the end of May 1995 (Table 1). Mortalities increased during the month of June and peaked in July leaving 1685 survivors. Subsequently, mortalities levelled off and at the inventory carried out in midNovember, there were 1297 postsmolts (25.9%) still surviving. Mortality remained low over the winter. 426 J. B. Dempson et al. Table 3. Summary of length and weight data for various groups of both cage-reared and wild Conne River Atlantic salmon (C, cage; W, wild; M, mortality; L, live; R, radio tag). Lifestage Smolt Post-smolt 1SW Category Status C C W C C C C C C C C C C W C C W M L L L M L L L L L L L R R L L L Date May 1995 May 1995 May 1995 Jun 1995 Jul 1995 Jul 1995 Aug 1995 Sep 1995 Oct 1995 Nov 1995 May 1996 Jun 1996 Jun 1996 Jun 1996 Jul 1996 Oct 1996* Summer 1996 n 114 25 249 25 53 25 25 25 25 31 36 35 37 22 35 22 72 Fork length (mm) Mean s.d. Range 138 145 143 173 153 238 358 342 374 374 413 454 463 518 452 465 518 17 14 15 20 26 26 18 39 42 59 55 44 22 19 36 24 22 105–192 121–173 103–179 126–204 119–254 184–273 309–388 228–382 240–436 218–438 265–484 317–574 430–500 475–555 324–518 415–500 475–573 Whole weight (g) Mean s.d. Range 24 28 29 49 26 168 565 577 780 762 840 1182 1281 1542 1209 8 9 8 24 36 56 73 179 236 300 328 332 223 201 278 13–49 14–51 10–51 14–91 9–205 51–260 401–715 150–878 166–1196 136–1194 190–1450 399–2275 980–1860 1160–1920 441–1946 1523** 219 1160–1920 *Caught at the fish counting fence and held in river until October. **Weight data based on 19 fish. Another inventory was carried out 16 May 1996, after the sample of 55 fish had been removed for disease analysis. At this time 871 fish remained. Thus, overall survival, with the inclusion of the 55 salmon sacrificed for disease analysis was 18.5%. A comparison of the transferred smolts that died showed that they were smaller than the average size of wild smolts sampled as they emigrated from Conne River in 1995 (Table 3; Figure 2). Analysis of covariance of weight–length relationships indicated significant differences in slopes (r2 =0.913, Fslopes =5.09, p=0.025) with the length-specific increase in weight less in those fish that died. Of the 53 dead smolts sampled in July, 49 had a whole weight that was less than the mean of the group transferred in May (Figure 2). Apparently, these fish had not taken to feeding and exhibited characteristics of failed smolt syndrome (McCarthy et al., 1996). This is undoubtedly the cause of the increased mortality during this period. Rapid increase in size occurred from July to August, and by October 1995 caged-reared wild salmon averaged 374 mm in length and 780 g in weight (Table 3; Figure 3), 82 and 66% of their subsequent size achieved by June 1996, respectively. Little growth occurred over the period October 1995 to May 1996. Two samples of surviving adults were obtained from June 1996. One sample (n=37) was measured 19 June 1996 (Table 3). These had been taken off feed in mid-May, then transferred to the Conne River estuary and were subsequently implanted with radio transmitter tags. An additional sample of 1SW salmon was retained at Roti Bay and maintained on feed. An analysis of covariance indicated that neither slopes (Fslopes =0.15, p=0.699) nor intercepts (Fintercept =2.09, p=0.153) were statistically different between these two groups. In contrast, wild 1SW Atlantic salmon returning to Conne River in 1996 were significantly larger in length (F=190.27, p<0.001) and weight (F=19.30, p<0.001) than 1SW caged-reared salmon. Distribution of tag returns among rivers Cage-reared wild salmon from each of the two groups released in Bay d’Espoir subsequently returned to Conne River and Little River (Table 4), both monitored by fish counting fences, as well as to Northwest Brook (Tailrace) and Southeast Brook (Milltown). In total, 38.5 and 45.8% of the salmon from release groups 1 and 2, respectively, were accounted for. Based upon known recoveries, there was a strong affinity for fish from each release group to return to Conne River (80%). The distribution of salmon among rivers, after pooling information from Northwest Brook and Southeast Brook (Milltown) due to low expected cell frequencies, was independent of release group in the likelihood ratio test (G=2.264, p=0.322). That is, the distribution and straying of fish among these Bay d’Espoir Rivers was the same for each of the two release groups. Similarly, the recovery rates of salmon in Conne River or in Little River were independent of the early vs. late release groups (Conne River, G=2.964, p=0.085; Little River, G=2.508, p=0.113). Evaluation of an alternative strategy to enhance salmon populations 60 8 (a) 427 (c) 7 50 Frequency 6 40 5 4 30 3 20 2 10 0 35 1 0 5 10 15 20 25 30 35 40 45 50 0 0 50 (b) 30 25 50 75 100 125 150 175 200 225 250 275 (d) 40 Frequency 25 20 30 15 20 10 10 5 0 0 5 10 15 20 25 30 35 40 45 50 Whole weight interval (g) 0 0 25 50 75 100 125 150 175 200 225 250 275 Whole weight interval (g) Figure 2. Weight–frequency distributions of: (a) wild Conne River salmon smolts, May 1995 (n=249, mean=28.6 g); (b) smolt mortalities in the sea cage, May 1995 (n=114, mean=24.2 g); (c) surviving smolts in the sea cage, July 1995 (n=25, mean=167.7 g); and (d) smolt mortalities in the sea cage, July 1995 (n=53, mean=25.5 g). Run timing of cage-reared wild salmon 2400 2200 2000 1800 1600 1400 1200 1000 800 600 400 200 0 Radio-telemetry monitoring M ay 1 Ju 995 n 19 Ju 95 l1 A 995 ug 1 Se 995 p 19 9 O ct 5 19 N 9 ov 5 19 M ay 95 1 Ju 996 n 19 Ju 96 l1 99 6 Whole weight (g) Cage-reared wild salmon from group 1 first entered Conne River 2 days following their release on 27–28 June, but most of the fish delayed their return for several weeks. The median date of return was 13 July (Group 1) with the last fish recorded at the fish counting fence on 29 July. Run timing of wild salmon returning to Conne River in 1996 was the earliest since 1989, with a median return date of 27 June and 90% of the run completed by 12 July. The last wild fish was recorded on 23 September. In contrast, cage-reared salmon from the second group released at Roti Bay 23 July, had a median return date of 31 July with the first record of these fish at Conne River on 26 July. The last salmon from this release group was recorded at Conne River on 21 August. Return of these salmon was more rapid than those from the first release even though they had to migrate a greater distance to return to Conne River. Month and year Figure 3. Summary of the change in weight (g) of Conne River Atlantic salmon smolts reared in a sea cage to the one-seawinter (1SW) adult stage. Vertical line represents the minimum and maximum values, the rectangle the 95% confidence intervals of the mean, and the marker within the rectangle is the mean. Of the original 37 cage-reared radio tagged fish released in Bay d’Espoir, 15 were identified as having returned to Conne River. The first of these was detected at the fixed receiving station on 27 June, the last on 20 July. There was no difference in the recovery rate of radio tagged fish vs. the Floy tagged fish at Conne River (G=0.889, p=0.346). The ground survey conducted 4 September, found a single radio tag in Little River. No signals were detected in the lower sections of the other rivers. However, a radio tagged salmon was angled in 428 J. B. Dempson et al. Table 4. Distribution of Floy tagged (1 and 2: cage-reared) and radio tagged (WC, cage reared; W, caught and released directly in Conne River) Atlantic salmon among Bay d’Espoir rivers by release group (CR, Conne River). Group Number released Release site Recovered Conne River Little River Northwest Brook (Tailrace) Southeast Brook (Milltown) Collins Brook Total Percentage recovered 1 434 CR Village n (%) 2 417 Roti Bay n (%) 134 21 6 6 80.2 12.6 3.6 3.6 152 31 6 2 79.6 16.2 3.1 1.0 167 100.0 38.5 191 100.0 45.8 the lower section of Southeast Brook (Milltown) on 7 July 1996. The 2 October aerial survey of Conne River detected 14 of the cage-reared radio tagged salmon; seven in the main stem, five in the Bernard Brook tributary and two in Twillick Brook (Figure 4). There was also one radio tag detected at the fixed receiving station that was not found on the first overflight. Thus, 15 of the 37 cagereared radio tagged salmon were accounted for in Conne River. In addition, aerial overflights detected one other radio tag in Little River. No signals were detected at any of the other rivers surveyed at this time. Sixteen of the radio tagged wild fish released above the fish counting fence were found during the October overflight; eight in the main stem, two in the Bernard Brook tributary, and six in Twillick Brook (Figure 4). Two tags had been recovered earlier above the fence where fish had regurgitated them. Thus, 18 of the 22 river-released fish were accounted for. The distribution of radio tagged salmon within Conne River, after pooling information from the two tributaries because of low expected cell frequencies, was independent of release group (G=0.016, p=0.898). Subsequent aerial surveys carried out on 7 November 1996, and 10 January 1997, provided similar results at Conne River and Little River, with three additional radio tagged river-released salmon located. Some of the cage-released and wild salmon detected in Conne River had moved downstream or into ponds. At Collins Brook, surveyed for the first time during the November overflight, two radio tags were detected approximately 10 km upstream from the river mouth. Based on three overflights, 21 of 22 (95.5%) radio tagged wild salmon released in Conne River were accounted for (including two regurgitated tags found). With respect to the radio tagged cage-reared salmon (Table 4), 20 of 37 (54%) were accounted for [2, Little River; 2, Collins Brook; 1, Southeast Brook (Milltown); 15, Conne River]; 75% were found in Conne River. WC 37 CR Village n (%) 15 2 15.0 10.0 1 2 20 5.0 10.0 100.0 54.1 W 22 CR n (%) 21 100.0 21 100.0 95.5 Discussion Wild Atlantic salmon smolts were successfully cagereared at sea to 1SW adults with a resulting survival rate four times greater than the average survival of wild salmon returning to Conne River during the period 1991–1996 (4.2%; Dempson et al., 1998). Results of the radio-telemetry investigations confirmed that cagereared wild salmon that returned to Conne River were able to navigate and utilize virtually the entire watershed, and had a distribution in the river similar to their natural counterparts. Thus, our results demonstrate that this technique could be used as a strategy in stock rebuilding. The technique was similar in some respects to the delayed release approaches carried out in the Baltic (Eriksson, 1991; Eriksson and Eriksson, 1991) where survival was similarly improved. To our knowledge, this was the first time that a project of this nature had been attempted using wild smolts and releasing adults into an ocean–estuary environment rather than directly into the river. The 18.5% survival to 1SW salmon realized is comparable to 20% obtained in the Big Salmon River experiment (Bruce, 1995). Studies that have examined the interaction of wild and escaped farmed salmon occasionally report that farmed fish enter the rivers later (Gausen and Moen, 1991), tend to spawn more in the lower reaches of a river, and later in the year (Webb et al., 1991) than wild salmon although this is not consistent among all studies (Okland et al., 1995). In our experiment, the first group was released to coincide with the timing of the natural salmon return to Conne River, but the cage-reared wild fish generally delayed their return to the river for several weeks. In contrast, the second group which was released later returned to Conne River more rapidly. There was no evidence from the radio-telemetry investigations that the cage-reared wild salmon remained in the lower sections of the Conne River. The strategy to use cage-reared wild salmon to enhance salmon populations is novel by comparison Evaluation of an alternative strategy to enhance salmon populations 429 48°15' N Conne Pond 48°10' Be rn ar 48°00' dB Tw ill ick Br oo k 48°05' roo N em St Counting fence site 55°40' em St k n ai M 47°55' n ai M 0 55°35' 2 4 6 km 55°30' 8 10 55°25' W Figure 4. Distribution of cage-reared ( ) and wild ( ) radio tagged Atlantic salmon throughout the Conne River watershed as determined from an aerial overflight survey conducted 2 October 1996. Scale 1: 250 000. with conventional fry- or smolt-stocking programmes. However, there are several aspects that would require attention before this method could be widely applied. These can be partitioned into three stages: capture and transfer; cage rearing; and release. Although initial smolt mortality following transfer to the estuarine pen was low relative to the mortality that occurred 6–8 weeks later, survival might be enhanced by further optimizing catch and transfer procedures to reduce handling. Stress associated with transporting smolts has been found to reduce survival (Hansen and Jonsson, 1988). Secondly, there is a strong need for improvements to cage design and feeding strategy to encourage a greater incidence of feeding among smolts on introduction to the estuarine cage. This could include use of a locally-formulated moist diet, the addition of aromatic fish oils, and more frequent feeding during the day. In the experiment with enhancement-oriented aquaculture of wild Conne River Atlantic salmon smolts, greater experimental flexibility would be required to allow for various treatments and replicates of experimental protocols which was not feasible using the 70-m circumference production cage. Initial feeding of smolts on introduction to sea cages has been problematical in aquaculture operations, even with hatchery-reared smolts that are predisposed to commercial dry diets. In light of these observations, the performance of wild Conne River salmon smolts in the estuarine cage is remarkable given that most of the smolt had been feeding naturally in the river for at least 3 years. While the observed feeding response among some of the smolts in our experiment confirms that commercial diets are acceptable to some wild smolts, feeding disposition would have to be improved greatly before protected rearing of wild smolts in sea pens could be advocated as a viable stock enhancement tool. The decision to release surviving cage-reared wild 1SW salmon to the ocean–estuary environment of Bay d’Espoir provided the opportunity to examine the local home-water homing response in fish that originated in, and subsequently were intercepted as they emigrated from Conne River as wild smolt. The results obtained have implications for the release phase where additional improvements might be possible to ensure maximum contribution to the spawning populations and to prevent straying. This is because only 42% of the survivors that 430 J. B. Dempson et al. were assumed to have retained their external tags were subsequently accounted for, of which 80% returned to the Conne River and 20% strayed into other Bay d’Espoir rivers. Overall, 5.7% of the 5000 smolts transferred returned to Conne River, which is almost identical to the return rate of wild smolts that returned as 1SW spawners in 1996 (Dempson et al., 1998). The fate of the remaining survivors of the cage-reared wild fish is unknown, but there was no evidence of any returning as two-sea-winter fish to either the Conne River or Little River fish counting facilities in 1997. One strategy to alleviate the straying problem and thus ensure that cage-reared wild salmon survivors contribute to the spawning escapement would be to simply release surviving fish directly into their natal stream as was done in the Big Salmon River experiment in New Brunswick. Transferring smolts from the river in an open cage that allowed continued contact and exchange of the river and estuarine water is another option that could decrease the rate of straying. Salmon have a remarkable ability to home to their natal river (Harden Jones, 1968; Stabell, 1984; Hansen et al., 1993). Recent studies have shown that farmed (Heggberget et al., 1993) or hatchery reared salmon (Jonsson et al., 1991; Hvidsten et al., 1994) stray more than wild fish. According to the ‘‘sequential imprinting’’ hypothesis (Harden Jones, 1968; Hansen et al., 1993; Dittman and Quinn, 1996), juvenile salmon learn sequentially from various cues associated with their home river which enables them to orient back to their natal stream. The most sensitive period for learning appears to occur during the parr–smolt transformation stage (Hasler and Scholz, 1983; Dittman and Quinn, 1996). Quinn (1993) states that disruptions in the sequence of odours associated with their natal freshwater streams and subsequent seaward migration could result in straying. Also, smolts may gain essential experience during their outward migration allowing them to navigate with precision to their home stream (Hansen et al., 1993; Hansen and Jonsson, 1994). In our experiment, smolts were captured in the lower part of the river about 0.75 km above head of tide. Given the largely fresh-to-brackish characteristics of the Conne River estuary, transferred smolts could have been prevented from fully ‘‘learning’’ and thus imprinting to their natal stream and estuarine characteristics. This may have contributed to a higher straying rate than wild salmon typically have (Stabell, 1984; Quinn and Tallman, 1987; Thorpe, 1994) and was more comparable with rates observed in hatchery-reared (Jonsson et al., 1991; Hvidsten et al., 1994), or smolt transportation experiments (Mills, 1994). Alternatively, maintaining smolts in a grow-out sea cage for 13 months could have also acted to alter the normal sequence of cues that fish key on to ensure precise navigation. There are a number of advantages in using cagereared wild smolts as an alternative enhancement strategy to assist in rebuilding natural salmon populations. First, the naturally present genetic strains are utilized and maintained. In addition, nature has already contributed to part of the selection process, in our case, throughout the first 3–4 years of freshwater life. Past enhancement strategies that stocked salmon fry into a river system required 4 years following stocking before any adult production materialized. Few if any of these projects were thoroughly evaluated. In contrast, adult salmon were produced here in just 1 year. Given the infrastructure at Conne River, it was rather straightforward to monitor and subsequently evaluate the success of the project. On the basis of results obtained from this initial experiment, the approach appears to offer considerable potential as a stock rebuilding tool. In situations where the infrastructure already exists in support of aquaculture operations, or in which wild stock declines are sufficiently critical to warrant immediate implementation of recovery strategies, rearing of wild smolts in estuarine pens appears to have more immediate benefits than traditional enhancement techniques. Given either of these conditions, and good quality freshwater habitat for both natural reproduction and early survival, broodstock enhancement is one means of minimizing domestication effects on salmon populations during the interval it takes to increase wild salmon stocks to sustainable levels of conservation egg deposition requirements. Acknowledgements The authors thank Anne-Margaret MacKinnon of the DFO Moncton Fish Health Laboratory for coordinating the fish health analyses. A research team from the University of Waterloo (Dr Scott McKinley, Dr Rick Booth, and Mr Eric Bombardier) was instrumental in applying the radio tags and helping to set up the monitoring equipment. Technical assistance and the telemetry receiver were provided by Dave Scruton, DFO St John’s. The authors also wish to acknowledge the technical and logistical support provided by Lotek Marine Technology staff, in particular Keith Stoodley (St John’s, Newfoundland) and Mitch Sisak (Aurora, Ontario), and the cooperation of the Canada Coast Guard helicopter pilots, Casey Buckles and Gordon Simmons, for their patience and careful attention to our demands during the aerial overflight surveys. Drs Bror Jonsson, John Browne, and Niels Daan, kindly reviewed the manuscript and provided editorial advice. Funding support for this research was provided by the CanadaNewfoundland Cooperation Agreement on Salmonid Enhancement/Conservation and by the Aboriginal Fisheries Strategy (Government of Canada). Evaluation of an alternative strategy to enhance salmon populations References Beacham, T. D., and Dempson, J. B. 1998. Population structure of Atlantic salmon from the Conne River, Newfoundland as determined from microsatellite DNA. Journal of Fish Biology, 52: 665–676. Bielak, A. T., and Davidson, D. 1993. New enhancement strategies – an overview. In Salmon in the sea and new enhancement strategies, pp. 267–298. Ed. by D. Mills. Fishing News Books, Oxford. Brannon, E. L. 1993. The perpetual oversight of hatchery programs. Fisheries Research, 18: 19–27. Bruce, H. 1995. Big doings on the Big Salmon: volunteers spark special stocking. Atlantic Salmon Federation Newsletter, 14(1), February/March. Davidson, K., and Bielak, A. T. 1993. New enhancement strategies in action. In Salmon in the sea and new enhancement strategies, pp. 299–320. Ed. by D. Mills. Fishing News Books, Oxford. Dempson, J. B., and Stansbury, D. E. 1991. Using partial counting fences and a two-sample stratified design for markrecapture estimation of an Atlantic salmon smolt population. North American Journal of Fisheries Management, 11: 27–37. Dempson, J. B., Furey, G., and Bloom, M. 1998. Status of Atlantic salmon in Conne River, SFA 11, Newfoundland, 1997. Department of Fisheries and Oceans, Canadian Stock Assessment Secretariat Research Document 98/28. 43 p. Dittman, A. H., and Quinn, T. P. 1996. Homing in Pacific salmon: mechanisms and ecological basis. Journal of Experimental Biology, 199: 83–91. Eriksson, T. 1991. Sea releases of Baltic salmon: increased survival with a delayed-release technique. American Fisheries Society Symposium, 10: 562–566. Eriksson, T., and Eriksson, L.-O. 1991. Spawning migratory behaviour of coastal-released Baltic salmon (Salmo salar). Effects on straying frequency and time of river ascent. Aquaculture, 98: 79–87. Gausen, D., and Moen, V. 1991. Large-scale escapes of farmed Atlantic salmon (Salmo salar) into Norwegian rivers threaten natural populations. Canadian Journal of Fisheries and Aquatic Sciences, 48: 426–428. Hansen, L. P., and Jonsson, B. 1988. Salmon ranching experiments in the River Imsa: effects of dip-netting, transport and chlorobutanol anaesthesia on survival. Aquaculture, 74: 301–305. Hansen, L. P., and Jonsson, B. 1994. Homing of Atlantic salmon: effects of juvenile learning on transplanted postspawners. Animal Behaviour, 47: 220–222. Hansen, L. P., Jonsson, N., and Jonsson, B. 1993. Oceanic migration in homing Atlantic salmon. Animal Behaviour, 45: 927–941. Harden Jones, F. R. 1968. Fish migration. Edward Arnold, London. 325 p. Hasler, A. D., and Scholz, A. T. 1983. Olfactory imprinting and homing in salmon. Springer-Verlag, Berlin. 134 p. Heggberget, T. G., Okland, F., and Ugedal, O. 1993. Distribution and migratory behaviour of adult wild and farmed Atlantic salmon (Salmo salar) during return migration. Aquaculture, 118: 73–83. Hilborn, R. 1992. Hatcheries and the future of salmon in the Northwest. Fisheries, 17: 5–8. Hilborn, R., and Winton, J. 1993. Learning to enhance salmon production: lessons from the Salmonid Enhancement Program. Canadian Journal of Fisheries and Aquatic Sciences, 50: 2043–2056. Hvidsten, N. A., Heggberget, T. G., and Hansen, L. P. 1994. Homing and straying of hatchery-reared Atlantic salmon, 431 Salmo salar L., released in three rivers in Norway. Aquaculture and Fisheries Management, 25 (Suppl. 2): 9–16. Incerpi, A. 1996. Hatchery-bashing: a useless pastime. Fisheries, 21(5): 28. Jonsson, B. 1997. A review of ecological and behavioural interactions between cultured and wild Atlantic salmon. ICES Journal of Marine Science, 54: 1031–1039. Jonsson, B., and Fleming, I. A. 1993. Enhancement of wild salmon populations. In Human Impacts on Self-recruiting Populations, pp. 209–238. Ed. by G. Sundnes. An international symposium, Kongsvoll, Norway, 7–11 June 1993. Tapir, Trondheim Norway. Jonsson, B., Jonsson, N., and Hansen, L. P. 1991. Differences in life history and migratory behaviour between wild and hatchery-reared Atlantic salmon in nature. Aquaculture, 98: 69–78. Korman, J., Marmorek, D. R., Lacroix, G. L., Amiro, P. G., Ritter, J. A., Watt, W. D., Cutting, R. E., and Robinson, D. C. E. 1994. Development and evaluation of a biological model to assess regional-scale effects of acidification on Atlantic salmon (Salmo salar). Canadian Journal of Fisheries and Aquatic Sciences, 51: 662–680. Lacroix, G. L. 1992. Mitigation of low stream pH and its effects on salmonids. Environmental Pollution, 78: 157–164. McCarthy, I. D., Carter, C. G., Houlihan, D. F., Johnstone, R., and Mitchell, A. I. 1996. The performance of allfemale diploid and triploid Atlantic salmon smolts on transfer together to sea water. Journal of Fish Biology, 48: 545–548. McVicar, A. H. 1997. Disease and parasite implications of the coexistence of wild and cultured Atlantic salmon populations. ICES Journal of Marine Science, 54: 1093–1103. Meffe, G. K. 1992. Techno-arrogance and halfway technologies: salmon hatcheries on the Pacific coast of North America. Conservation Biology, 6: 350–354. Mills, D. 1994. Evidence of straying from wild Atlantic salmon, Salmo salar L., smolt transportation experiments in northern Scotland. Aquaculture and Fisheries Management, 25 (Suppl. 2): 3–8. Okland, F., Heggberget, T. G., and Jonsson, B. 1995. Migratory behaviour of wild and farmed Atlantic salmon (Salmo salar) during spawning. Journal of Fish Biology, 46: 1–7. Pepper, V. A., and Crim, L. 1996. Broodstock management. In Principles of salmonid aquaculture, pp. 231–289. Ed. by W. Pennell, and B. A. Barton. Elsevier, Amsterdam, The Netherlands. Quinn, T. P. 1993. A review of homing and straying of wild and hatchery-produced salmon. Fisheries Research, 18: 29–44. Quinn, T. P., and Tallman, R. F. 1987. Seasonal environmental predictability and homing in riverine fishes. Environmental Biology of Fishes, 18: 155–159. Riddell, B. E. 1993. Salmonid enhancement: lessons from the past and a role for the future. In Salmon in the sea and new enhancement strategies, pp. 338–355. Ed. by D. Mills. Fishing News Books, Oxford. Saunders, R. L. 1991. Potential interaction between cultured and wild Atlantic salmon. Aquaculture, 98: 51–60. Stabell, O. B. 1984. Homing and olfaction in salmonids: a critical review with special reference to the Atlantic salmon. Biological Reviews, 59: 333–388. Sutterlin, A. M., and Stevens, E. D. 1992. Thermal behaviour of rainbow trout and Arctic char in cages moored in stratified water. Aquaculture, 102: 65–75. Thomas, R. L. 1996. Enhancing threatened salmonid populations: a better way. Fisheries, 21: 12–14. Thorpe, J. E. 1994. Significance of straying in salmonids and implications for ranching. Aquaculture and Fisheries Management, 25 (Suppl. 2): 183–190. 432 J. B. Dempson et al. Verspoor, E. 1997. Genetic diversity among Atlantic salmon (Salmo salar L.) populations. ICES Journal of Marine Science, 54: 965–973. Washington, P. M., and Koziol, A. M. 1993. Overview of the interactions and environmental impacts of hatchery practices on natural and artificial stocks on salmonids. Fisheries Research, 18: 105–122. Webb, J. H., Hay, D. W., Cunningham, P. D., and Youngson, A. F. 1991. The spawning behaviour of escaped farmed and wild Atlantic salmon (Salmo salar L.) in a northern Scottish river. Aquaculture, 98: 97–110.
© Copyright 2026 Paperzz