Documentation of Unaccounted-for Losses of Chinook Salmon from Saltwater Cages

-"Shrinkage," or unaccounted-for losses of fish during cage rearing, is widely acknowledged among fish culturists who rear salmonid fishes, but the extent of such losses has not been documented . Over 60,000 Chinook salmon (Oncorhynchus tshawytscha) were reared in floating cages at four locations in Puget Sound, Washington, and were hand-counted at the beginning and end of 200+ d of rearing. Average unexplained losses were 8-38% for individual locations (three to four cages each) and 2.5-46.5% for individual cages . When cage netting remains intact, unexplained losses are probably the result of decomposition of carcasses (particularly during disease outbreaks) ; scavenging by birds, mammals, and fishes ; and to a lesser extent, escapes . The economic consequences of losing fish can be important in terms of lost harvest and increased food costs, resulting in true food conversions 10-68% higher than anticipated . It is widely known among commercial fish culturists that when fishes are held within nets in a body of water, a certain portion of fish assumed to be in cages disappears . Though this unexplained loss of fishes has been recognized for decades (Institute of Fisheries Economics 1971 ; Kennedy 1975 ; Secretan 1979), it has never been accurately documented . Even today, commercial fish culturists continue to lose important numbers of salmonid fishes from cages in salt water, estimated to range from 10% to as much as 30% (Lindsay 1980; Coche 1983; Leet et al . 1986 ; Hansen et al . 1987 ; Mills 1988; C. Mahnken, National Marine Fisheries Service, personal communication). Such losses are generally not discovered until a cage is emptied at the time of harvest or grading. Accurate documentation of the extent of such losses and the relationship of such losses to feed conversion (weight of feed fed/weight gain by fish) and rearing densities has not been made. Fish disappear even when there are no tears in the netting, the cages are covered, and daily inspections of cages are made . In a public or commercial operation, this loss can have economic importance-not only because of lost fish ("shrinkage") but also because food provided for these "phantom" fish often falls through the bottom netting and is wasted, such that assumed feeding rates and food conversions thus are both inflated . I conducted a 2-year study of the rearing of chinook salmon (Oncorhynchus tshawytscha) in floating cages to address two questions: How important are unaccounted-for losses of cultured stock? and, How do actual food conversions compare with predicted conversions based on the actual fish fed and the fish assumed to be present and needing food? Although the study was conducted over a decade ago, no documentation of this type has been provided in the interim and this information has direct application to current salmonid cage operations . Chinook salmon (Finch Creek stock) obtained from Hoodsport (Washington) State Fish Hatchery were held in cages at the Big Beef Creek Research Station and later transferred to floating cages at four sites in Puget Sound (three in 1971 and three in 1972): Clam Bay (central Sound), Squaxin Island (southern Sound), Big Beef Creek (a brackish-water pond near Hood Canal), and Friday Harbor (San Juan Islands) . Fish were reared in either three or four cages at each site in each year (total pens : 12 in 1971-1972 and 11 in 1972-1973) . Cages were cubical, 2.4 m per side, with an effective rearing volume of 12.7 m3/cage . Netting was 6 .4-mm-square mesh, and cage shape was maintained with square 2.4-m frames of weighted polyvinyl choride . A net was placed over each cage to control bird predation . Cages used today in commercial operations are similar to those used in the study except that commercial cages are somewhat larger and have netting extending higher above the water line (0 .7 m) to decrease predation by otters . Chinook salmon were individually counted into cages at the beginning of the culture period (1120 July 1971 and 22 May-8 June 1972) ; fish averaged 12 .3-12.8 g in 1971 and 4.8-5.8 g in 1972 . Fish were normally fed Oregon Moist Pellets four times a day at 1-4.5% of assumed fish biomass in the cages . Fish growth was monitored by biweekly sampling in spring through fall and monthly sampling in winter . Based on the samples, changes in feeding rates were made for all cages at all sites on the same days . Cages were inspected daily, and dead fish were removed . At termination of the study, remaining chinook salmon were again handcounted . Chinook salmon disappeared from every cage during the 214-260 d of rearing, even though there were no tears in the netting (Table 1) . As might be expected, the proportion of unaccounted-for losses varied widely but nearly all losses were important . Average percentage losses by site were : Clam Bay, 17 .5 in 1971-1972 and 37 .9 in 19721973 ; Squaxin Island, 18 .0 in 1971-1972 and 19.7 in 1972-1973 ; Big Beef Creek, 8.4 in 1971-1972 ; Friday Harbor, 28.7 in 1972-1973 . The initial number of chinook salmon stocked, minus known deaths, known escapement during biweekly or monthly sampling, and intentional removals was used to calculate the feeding rate for each cage . Because this net number included unexplained fish loss, actual food conversion was always higher than anticipated (Table 2) . Food conversions were higher than those encountered in most hatchery and commercial situations, possibly due to our experimental design . Feeding rates were not adjusted for environmental conditions at each site but were adjusted at all sites by date in order to eliminate location as a variable ; most fish culturists adjust feeding rates according to local conditions (e .g ., water temperature) . Similarly, the assumed rearing densities were always overestimated ; actual densities were 1 .5-7.6 kg/m3 lower . When fish disappear from a cage, escapement either through a hole in the netting or over the top of the cage is suspected . No holes were found in the netting (knotless 10.9-kg-test Ace netting, 6-mm mesh), and each cage extended out of the water 0.4 m and was covered with netting ; thus, there was little opportunity for chinook salmon to escape . Decomposition of dead chinook salmon is a likely contributing factor to unexplained losses . Cages were examined daily for dead fish, but decomposition sometimes is rapid . Carcasses can disappear from nets in less than a day, and some fish predators, such as spiny dogfish (Squalus acanthias), are known to prey on dead salmon through cage netting (C . Mahnken, National Marine Fisheries Service, personal communication) . Such decomposition and predation can be high during outbreaks of vibriosis, and there is a relation between known losses to Vibrio outbreaks and unaccounted-for losses during rearing (r = 0 .52 ; average losses to Vibrio at the sites I monitored were 1-3% in 1971-1972 and 11-19% in 1972-