FISHERY BULLETIN: VOL. 74, NO. 1 



near the surface can be obtained by increasing 

 the depth resolution to 0.5 m from the present 

 1 m. This increased depth resolution would 

 necessitate the elimination of the first 14 m of the 

 46-m water column above the transducer since 

 only six bits are available for each range word. 

 This is a desirable tradeoff, however, in view of 

 the concentration of the fish near the surface. 



Figure 6 is a series of plots of the computed 

 areal densities of the salmon obtained by inte- 

 grating the depth distribution histograms over 

 the depth. Also plotted are the purse seine 

 catches which were obtained in reasonable time 

 and space proximity to the buoys. The data are 

 reasonably consistent although significant 

 departures occasionally occur. There are several 

 possible sources for the observed discrepancies: 

 a) set-to-set variations in seine hauls, b) similar 

 variations in the sonar counts, c) the inability 

 of the purse seine and the acoustic buoys to 

 sample precisely the same volumes of water, 

 and d) possible attraction or avoidance of the 

 acoustic gear by the fish. The variations within 

 gear types can be explained by the "patchiness" 

 of the salmon. The digital printouts tend to show 

 small groups offish, rarely giving more than three 

 echo counts, occurring with widely varying inter- 

 arrival times. This observation indicates the exis- 

 tence of relatively large areas that are nearly 

 devoid of fish thus explaining the occasional 

 twofold variations in successive seine hauls 

 made at the same station. 



Sonar gear avoidance or attraction by the fish 

 is a potentially serious problem, the magnitude of 



30 5 



JUNE 



10 15 20 2530 5 10 I 



JULY JUNE JULY 



20 25 



Slolion 3 



Stotion 4 



Figure 6. — Plots of computed relative areal densities ( ) 



and purse seine catches ( — ) by date and station for 1974. 



which is not yet known. Occasional sea lions have 

 been observed around the buoys but they usually 

 departed after several minutes. Also, there is lit- 

 tle evidence to indicate that the fish are attracted 

 to the sonar gear since none of the observed 

 targets remains in the ensonified region for more 

 than a few pulses. Sonar gear avoidance is a more 

 likely prospect. The Stellar sea lion is common in 

 the area being sampled and it is a known pred- 

 ator of salmon. The sonar buoy is similar in size 

 to that of a sea lion so that avoidance is a distinct 

 possibility. Secchi disc readings of 15 m are typi- 

 cal so that the buoy or cable may be detected at 

 significant distances by the fish. Day versus night 

 data differ slightly but as yet there are too few 

 data on which to base a conclusion concerning 

 gear avoidance. 



All of the acoustic data from which Figure 6 

 was obtained were pooled and the sample correla- 

 tion coefficient for the buoy-purse seine was com- 

 puted. A value of 0.547 was obtained which, 

 under the assumption of normality, is significant 

 at approximately the 0.5% level {t distributed 

 with n - 2 = 19 df). The results indicate that the 

 acoustic buoys can obtain statistically significant 

 population information as well as such ancillary 

 information as depth distribution and density 

 during both day and night. Additionally, indirect 

 information on schooling is available by observ- 

 ing the interarrival tirnes of the fish although 

 this has not been investigated in detail. 



The design of the acoustic buoy system is essen- 

 tially fixed although modifications for use in 

 other situations are possible. For example, a bot- 

 tom anchored version for use in water depths of 

 about 100 m has been designed but fabrication 

 has not begun. Another possible design change is 

 in the radio telemetry system. The present sys- 

 tem, while reliable, is inefficient. An improved 

 system has been designed and will be fabricated 

 upon allocation of a suitable frequency by the 

 Federal Communications Commission. 



The current approximate unit cost per buoy, in- 

 cluding the radio transmitter, is $6,000. The 

 shipboard receiver-decoder cost is approximately 

 $2,000. The tape recorder currently used is a 

 Kennedy Model 1400 digital incremental which 

 records at 556 bits/inch on Va-inch magnetic tape 

 on 10-inch reels. It is "off the shelf but is inter- 

 faced to the receiver-decoder. The interfacing cost 

 is approximately $1,000 which should remain 

 constant for interfacing to any digital incremen- 

 tal recorder. 



110 



