FISHERY BULLETIN: VOL. 74, NO. 1 



Species and age discrimination by acoustic means 

 is not currently possible so that it is necessary to 

 obtain samples by purse seine in order to deter- 

 mine the species and age composition of the popu- 

 lation. 



Since sockeye salmon is the only species of in- 

 terest, we may treat the purse seine samples as 

 binomial events in which the parameter of in- 

 terest is the proportion, p, of sockeye salmon 

 present in the population. If asymptotic normal- 

 ity is again assumed, it is found that the sample 

 size required to be 100a% confident that the 

 relative error of the estimate of p will not exceed 

 8 is given by, 



(1 



n = 



p8 



(3) 



where d^ is that defined for Equation (2). 



Equations (2) and (3), as indicated, provide 

 information on the sampling necessary to achieve 

 prescribed levels of precision in the population 

 estimates. A direct but somewhat crude compari- 

 son of the purse seine and the acoustic buoys may 

 be made on an area basis. The purse seine has a 

 nominal length of 400 fathoms or about 732 m. 

 The area swept out in a round haul^ is about 

 42,600 m^. For a transducer having a beam width 

 of 28° to the 3 dB points and suspended 46 m 

 below the surface, the area ensonified is approxi- 

 mately 390 m^. Thus the purse seine sweeps out 

 an area about 110 times as great as the area 

 ensonified by a single acoustic pulse. The pulse 

 interval is approximately 10 s. However, it has 

 been found that the individual fish remain in the 

 pattern for longer periods of time, typically about 

 30 s, although a precise estimate is not available 

 at this time. Thus, a single buoy would have to 

 operate for at least 1 h to obtain coverage equiva- 

 lent to a single round haul. The additional cover- 

 age obtained using 30-min tow hauls is not 

 known precisely but the limited data available 

 indicate a factor of two or three over the round 

 hauls. Thus, to provide coverage comparable to 

 that of the purse seine a single buoy would have 

 to operate for a minimum of 3 h. A comparable 

 sampling time is obtained using Equation (2) 

 with poV,. estimated using a typical seine haul of 

 150 fish. The seine hauls may vary from zero to 



well over 1,000 fish from which it follows that the 

 time required for adequate acoustic samples may 

 vary, inversely, by corresponding amounts. The 

 sampling considerations just outlined played a 

 significant part in the choice of the hardware 

 configuration and the decision to utilize multiple 

 buoys. 



SYSTEM DESIGN 

 AND CHARACTERISTICS 



Figure 1 is a schematic illustration of the 

 high-seas assessment system showing only a 

 single buoy. In operation up to 10 buoys can be 

 deployed, each sending information to the ship- 

 board decoding and recording system. A four- 

 buoy system has been used at Adak to help assess 

 the migrating salmon population. A simplified 

 block diagram of the buoy system is given in Fig- 

 ure 2, and a photograph of the buoy is shown in 

 Figure 3. The buoy and shipboard system are dis- 

 cussed below. 



The buoy contains an acoustic system which 

 gathers fish count and depth distribution data, a 

 logic system which processes and provides tempo- 

 rary storage for these data, and a telemetry sys- 

 tem which sends data to the monitoring ship. The 

 acoustic system operates at 120 kHz and samples 

 the population every 10 s. Sample rates can be 

 changed to 5-s or 2.5-s intervals if desired. The 

 system transmits a 200-/xs pulse (24 cycles at 120 

 kHz) at a source level of + 106 dB. Target returns 

 must be greater than a preset threshold (approx- 

 imately 2V) for at least 100 fjLS before they are 

 validated. This technique, and an adequate 

 source level to give a worst case'' signal-to-noise 

 ratio of 10 dB, minimizes false target counts. 



Pulse elongation and amplitude testing tech- 

 niques are used to automatically adjust "end-of- 

 sample" so that surface returns and near surface 

 bubbles are not counted. 



Measurements at the University of Washing- 

 ton and at Adak during summer operation have 

 shown that the "average" target size of the mi- 

 grating salmon is about - 30 dB within the aspect 

 angles encountered in the sample volume. 



A typical plot of signal return versus aspect 

 angle from a single fish is shown in Figure 4. This 

 polar diagram shows target strength from the 



*Purse seining is normally done in a standard manner using 

 tow hauls in which the seine is held open in a semicircle for 

 30 min before closing and pursing. In a round haul the seine 

 is set in a circle and pursed immediately after closing the circle. 



'The worst-case condition exists for minimum target strength 

 ( -45 dB) at maximum range (46 m) at the -3 dB point in the 

 transducer beam pattern. 



106 



