LORD ET AL.: ACOUSTIC ASSESSMENT OF MIGRATING SALMON 



acoustic assessment system that can be used 

 alone or in conjunction with the purse seine. 

 Some of the anticipated advantages of such a 

 system were that it could obtain abundance 

 estimates and swimming depths with around- 

 the-clock operation in a wide range of weather 

 conditions. 



PRELIMINARY CONSIDERATIONS 



The final configuration of the system was 

 determined, to a large extent, by consideration of 

 problems related to obtaining adequate numbers 

 of representative samples. The preliminary indi- 

 cations were that most of the fish of interest were 

 concentrated near the surface at a depth of 10 m 

 or less. This conclusion was based on the results 

 of experiments in which longline and gill net gear 

 were fished at various depths (Manzer 1964; 

 Machidori 1966; French et al. 1967) and also by 

 direct visual observation of salmon in the purse 

 seines. This concentration offish near the surface 

 precluded the use of a hull-mounted device since 

 such a system would necessarily exclude the top 3 

 or 4 m of the water column. This led to considera- 

 tion of a transducer suspended in some manner 

 below the main body of fish. 



A transducer mounted on a towed platform 

 with a coaxial cable to the towing vessel was con- 

 sidered initially but was abandoned because of 

 the anticipated difficulty of developing a platform 

 that could maintain depth and attitude stability 

 while maintaining position to the side of the ves- 

 sel. Since the extreme water depths precluded an 

 anchored system the approach eventually 

 adopted was to suspend the transducers from 

 free-floating surface buoys with self-contained 

 electronics. Consideration of the anticipated 

 sampling statistics indicated the need for a 

 multibuoy system which in turn suggested radio 

 telemetry of the data from the buoys to a central 

 shipboard receiver and recorder This is the type 

 of system that was eventually constructed. 



The sampling statistics of particular relevance 

 to the design and operation of the buoy system 

 concerned: a) the level of effort required to obtain 

 a specified precision in the estimation of the 

 density of the fish and b) the purse seine effort 

 required to obtain comparable precision in the 

 estimation of the species composition of the 

 population. 



Extensive purse seining over a period of several 

 years indicated that the salmon were relatively 



sparse and probably did not school or otherwise 

 interact to a significant degree. Under these con- 

 ditions the echo counts may be assumed to have a 

 Poisson distribution with a parameter, /jl, that is 

 proportional to the number density^ of the fish. 

 Thus we have. 



M = PoVs 



(1) 



where V, is the sampling volume of a single 

 counter and po is the average fish density defined 

 so that fjL is the expected number of echoes per 

 acoustic pulse. If we assume large sample theory, 

 the minimum number of acoustic pulses required 

 to be 100a% confident that the relative error of 

 the estimate of po does not exceed e is given by, 



M„ 



e'PoVs 



(2) 



where c?^, is the lOOa^f point (two-sided) of ^^(0,1). 

 The crucial feature of Equation (2) is that the 

 sampling effort must be increased as either po or 

 Vg decrease. Preliminary estimates of po based on 

 purse seine data, while quite crude, indicated 

 that V, should be as large as possible subject only 

 to the tradeoffs necessary to obtain an adequate 

 signal to noise ratio. Also, the need for multiple 

 buoys sampling mutually disjoint volumes was 

 indicated. 



The high-seas salmon population generally 

 consists of a mixture of species so that it is neces- 

 sary to determine species composition by some 

 means. In the area south of the central Aleutians 

 significant numbers of chum salmon, O. keta, 

 occur mixed with immature sockeye salmon dur- 

 ing the sampling period, and occasionally pink, 

 O. gorbuscha; coho, O. kisutch; and chinook 

 salmon, O. tshawytscha, are present in small 

 numbers. The only nonsalmonid species gener- 

 ally found in this area, at the depths being sam- 

 pled, is the Atka mackerel, Pleurogrammus 

 monopterygius. This species generally occurs in 

 small numbers relative to the salmon so that, if 

 counted, it will not seriously affect the estimates 

 of the density of the salmon. Further, this species 

 does not have a swim bladder so that by the 

 proper choice of the detection threshold level 

 these fish will not be detected by the sonar. 



^For echo counting the number density is the quantity of 

 interest. If acoustic echo integration is utilized, the density on 

 a mass basis is appropriate. 



105 



