HEWITT ET AL.: DEVELOPMENT AND USE OF SONAR MAPPING 



Based on Smith's (1970) work, sonar mapping 

 cruises aboard the David Starr Jordan were con- 

 ducted with a 30-kHz sonar unit directed 90° to 

 starboard and 3° down. The sampled range band 

 was 200 to 450 m from the transducer. The re- 

 ceivers were rebuilt using solid state circuitry 

 with the remaining system as described by Smith 

 (SIMRAD 580-10 Scientific Sonar and Sounder).^ 



Target Size 



Frequency distributions of fish school sizes 

 were generated from data taken on several 

 cruises (April-May, November, December 1973; 

 and March-April 1974) using the maximum dif- 

 ference between the leading and trailing edge of 

 the echo envelope, corrected for pulse length, on 

 an axis perpendicular to the ship's track. The cal- 

 culation of target widths (measured on an axis 

 parallel to the ship's track) was discontinued due 

 to uncertainties in choosing the effective beam 

 width (see Smith 1970), fluctuations in the ship's 

 speed, and the inability to quantify other factors 

 which may affect apparent target width (i.e., 

 target strength). 



School size distributions (based on range differ- 

 ences) remained nearly constant during several 

 sampling periods and agreed well with a much 

 larger sample collected by the CF&G. A total of 

 4,355 sonar targets were counted and assigned to 

 size classes on three cruises approximately 6 mo 

 apart. Ten-meter class intervals were used and 

 frequencies were corrected for recording edge bias 

 employing the method described by Smith (1970). 

 This bias is encountered when one excludes 

 targets which do not entirely occur within the 

 observation band. Thus, frequencies of targets 

 other than point sources, are underestimated by 

 virtue of the fact that their physical size limits 

 the probability of their detection. To determine 

 unbiased relative proportions of target sizes, one 

 must correct observed target count (those targets 

 which lie entirely within the observation band) to 

 a count of targets whose centers lie within the 

 observation band.^ 



mapping. Presented at the ICES-ICNAF-FAO Symposium on 

 the Acoustic Methods in Fisheries Research, Bergen, Norway, 

 Contrib. No. 8, 13 p. 



"•Reference to trade names does not imply endorsement by the 

 National Marine Fisheries Service, NOAA. 



^It is assumed that range-dependent, size-specific target los- 

 ses are a minimum for the observation band sampled (Smith 

 1970). A similar study expanded to include the effects of target 

 strength on detection ranges would be of value. 



In developing a correction for recording edge 

 bias, a diagram may be useful. In Figure 1 a 

 school of diameter d is shown at the maximum 

 and minimum ranges of detection for an observer 

 on a ship sampling an observation band of k 

 units. The difference between the maximum and 

 minimum range of detection isk - d units. 



Let A represent the event that a school of d 

 diameter has its center within an observation 

 band of ^ units. LetB represent the event that a 

 school of d diameter is not intersected by either 

 edge of the observation band. Then the probabil- 

 ity of event B occurring given that event A has 

 occurred may be expressed: 



P[BIA] =^^. 



Further, let A^^ represent the count of targets of 

 diameter d who lie entirely within the observa- 

 tion band. Let N'^ represent the count of targets 

 of diameter d whose centers lie within the obser- 

 vation band. Since A^'^ represents both edge inter- 

 sected and non-edge intersected targets of diame- 

 ter d, the portion of non-edge intersected targets 

 may be estimated by: 



Nd = N'aP[BIA] =N'a 



d 



5 



o 



Ship 



k-d 



© 



e 



Observation band 



Figure l. — Plan view of sonar mapping technique showing 

 maximum and minimum ranges of detection for a target of 

 diameter d within an observation band of k units. 



283 



