HEWITT ET AL.: DEVELOPMENT AND USE OF SONAR MAPPING 



Table 4. — Observed frequencies, edge corrected frequencies, and horizontal school 

 area contributions for size classes (metric tons, mt) up to a maximum of 160 m school 

 diameter. 



tively. Identification of the fish is not yet possible 

 on a routine basis. However, it is assumed that 

 the majority of schoohng fish in the Los Angeles 

 Bight are northern anchovy (Smith 1972; Squire 

 1972; Mais 1974). 



DISCUSSION 



It is our impression that the ultimate value of 

 sonar mapping is its potential to reconstruct 

 geographic patterns of school distributions at a 

 moderate cost of time both in data collection and 

 data reduction. However, before this potential 

 can be fully realized, several problems must be 

 recognized, investigated, and placed in proper 

 perspective. 



With regard to counting and sizing targets: 



1. An edge bias has been described which will 

 be present with any sonar system designed 



3. 



to count and size schools. The determination 

 of effective detection ranges establishes a 

 finite observation band. Larger schools tend 

 to be undersampled relative to smaller 

 schools; in terms of school area the bias may 

 be significant. 



Increasing the observation band would tend 

 to reduce the effect of edge bias. However, 

 the effects of target size and target strength 

 on maximum ranges of detection should be 

 investigated before defining the observation 

 band. Undersampling small schools may be 

 acceptable when considering their area con- 

 tribution. 



Effective detection ranges may also be lim- 

 ited by inhomogeneities in the medium 

 caused by short-period internal waves. 

 Smith (see footnote 3) investigated this 

 phenomenon and suggested the only prac- 

 tical solution is a statistical approach 



295 



