FISHERY BULLETIN: VOL. 84, NO. 2 



able to calibrate the sonar systems so that the sensi- 

 tivity of all transducers in the array were equal. 

 Hence, the probability of detecting a given school 

 in one sector of the sonar display was not necessarily 

 the same as detecting it at an equal distance in 

 another sector. The inability to calibrate the trans- 

 ducers may have compromised our ability to detect 

 all schools directly below the transect. This is the 

 most important assumption of line transect surveys; 

 school abundance estimates will be biased downward 

 if it is violated. Intercalibration of the transducers 

 would also help establish a more accurate detection 

 function which would apply throughout the sonar's 

 range 



The limited lateral resolution of sector scanning 

 sonar hampers the accurate measurement of school 

 width, an important value for determining mean 

 school biomass. Each transducer in the fan-shaped 

 array acts as an independent echo sounder and if 

 any portion of a school enters the radiation pattern, 

 the entire width of the 9° -10° sector sampled by that 

 beam will be displayed as a reflective target (Fig. 12). 

 This results in an overestimation of school width and 

 a distortion of the school's size and location, yield- 

 ing overestimates of biomass and inaccurate 

 measures of distance from the transect plane The 

 detection function will be altered by these inac- 

 curacies and may modify estimates of school abun- 



dance depending on the magnitude and the direc- 

 tion of the errors. The distortion may be aggravated 

 by interference of side lobes in the directivity pat- 

 tern of individual transducer beams (Fig. 13). Even 

 these lower power lobes can produce echo signals if 

 very dense targets are encountered and may inter- 

 fere with the acoustic signals from adjacent trans- 

 ducers. 



Another weakness of sector scanning sonar in this 

 application is insufficient detection sensitivity. This 

 weakness became apparent during calculations of 

 the lengths of individual schools. Lengths were cal- 

 culated twice for each school, once based on echo 

 sounder data and again based on sonar data. The 

 theoretically correct method would employ the sonar 

 data because schools could be detected further to 

 each side of the vessel. The echo sounder could only 

 detect the portion of the school within the 10° -11° 

 beam directly below the vessel. Consequently, if a 

 large part of the school was outside the beam, its 

 length was underestimated. In practice however, the 

 length estimates based on sonar detections were 

 usually shorter than those based on echo sounder 

 data (Table 9) due to the lower sensitivity of the sonar 

 system. The sonar-based lengths were chosen, how- 

 ever, because they measured the dimensions of the 

 part of the school having densities above the thres- 

 hold required to trigger the sonar. This is probably 



Segment covered by 

 one element in sonar 

 transducer array 



True school center 



Apparent school 

 center 



Apparent location of school 

 as seen on the sonar display 



True location of school 



Figure 12— A facsimile of the sector scanning sonar output display exemplifying biases in apparent school loca- 

 tions resulting from the limited resolution of the instrument. 



306 



