FISHERY BULLETIN: VOL. 74, NO. 2 



progress on a number of objectives established in 

 early 1968. In order to develop "sonar mapping" 

 as a stock assessment tool, it was decided that 

 such a system should be able to: 1) count the 

 number of schools per unit area in the upper 

 mixed layer from a ship proceeding at 12 knots, 2) 

 measure the horizontal size of each fish school, 3) 

 calculate the biomass of each school, 4) estimate 

 the size of individual fish within a school, and 5) 

 distinguish the northern anchovy from all other 

 schooling species. 



Smith (1970) developed a technique for "map- 

 ping" fish schools in the area where the northern 

 anchovy, Engraulis mordax, is abundant off the 

 coast of southern California. Sonar mapping dif- 

 fers from echo sounding; with sonar, estimates 

 can be made of the number offish schools per unit 

 area, of their horizontal dimensions, and of the 

 degree of aggregation of fish schools. We do not 

 routinely estimate depth of the school in the 

 water column, nor thickness of the school in the 

 vertical plane. Hull-mounted echo sounders pro- 

 vide estimates of the number of schools per line 

 transect deeper than 4 m, measures of chords 

 across the horizontal dimension of the school in 

 the plane of ship travel, depth in the water col- 

 umn, and thickness or vertical height of the fish 

 school. Experience indicates that the process of 

 "sonar mapping" encounters one or two orders 

 of magnitude more fish-school targets per unit of 

 ship time as compared to echo sounding from 

 the same vessel. It is important to emphasize that 

 this technique was developed because fish 

 schools are frequently found in the upper mixed 

 layer of the ocean where echo sounders are rela- 

 tively ineffectual at counting or measuring them. 



In the first report on this project. Smith (1970) 

 described a series of experiments designed to de- 

 termine the feasibility of the use of sonar to count 

 and measure the size of pelagic fish aggregations 

 (objectives 1 and 2). Optimum instrument set- 

 tings were determined for source level, receiver 

 gain, pulse length, transducer bearing, trans- 

 ducer directivity, and range. Methods were de- 

 veloped for correcting target width (dimension 

 measured on axis parallel to ship's track) for the 

 effect of the beam angle and for correcting target 

 count "edge biases." Since no target was counted 

 unless it lay entirely within a specified range, the 

 latter adjustment was made to compensate for the 

 narrowing possible interval of detection for larger 

 targets. 



Holliday (1972, 1974) investigated the fre- 



quency domain processing of fish school echoes 

 using experimental equipment brought aboard 

 the David Starr Jordan. By detecting and 

 measuring Doppler spread, Holliday was able to 

 calculate tail beat amplitudes of schooled fish 

 and, indirectly, their length (objective 4). 



Holliday also examined the resonance struc- 

 ture of pulse returns from fish schools and was 

 able to detect the presence or absence of a swim 

 bladder in the school constituents. This informa- 

 tion, when supplemented by observations on 

 school behavior and free vehicle camera drops, 

 may be used to distinguish anchovy from other 

 pelagic schooling organisms in a sample taken 

 randomly from targets encountered during a sur- 

 vey (objective 5). The statistical base thus ob- 

 tained would be applied to the entire survey. 



The California Department of Fish and Game 

 (CF«&G) has been engaged in sea surveys using 

 sonar methods since 1967 (Mais 1974). Its ap- 

 proach has been the collection of large amounts of 

 data and its interpretation, while the work at the 

 Southwest Fisheries Center (SWFC) has been in 

 the isolation of sampling errors and the develop- 

 ment of an automated hydroacoustic data acqui- 

 sition and processing system. As such, the two 

 groups complement each other with field experi- 

 ence and technological development. 



SOURCES OF 

 SAMPLING VARIABILITY 



We have made the assumption that quantita- 

 tive errors associated with system instrumenta- 

 tion are small in comparison to errors generated 

 by sampling an adult schooling population whose 

 behavior is little understood. For this reason, we 

 monitored our sonar system response when it was 

 operated in a variety of circumstances and 

 changed that system in answer to practical rather 

 than theoretical considerations. Using operating 

 techniques developed in 1968, school size fre- 

 quency distributions were generated and a lower 

 detectable size threshold defined; school target 

 strengths were calculated and compared with 

 similar work conducted by the Navy and the 

 CF&G; the relationship between the detected oc- 

 currence of pelagic fish schools and bottom topog- 

 raphy was investigated; and the variable range of 

 detection of schools due to internal waves was 

 studied (Smith^). 



^Smith, P. E. 1973. The effects of internal waves on fish school 



282 



