the assumed value of 42.00 feet. Random error in h is 

 7.5 feet; random error in H is 2.6 feet. Relative 

 vertical fish oscillations of this magnitude could be 

 expected, as can be seen in Figure 9, and because of 

 the sensitivity of fish depth to tow speed. However, a 

 random error in H of almost 3 feet is surprising, since 

 there is no hydrodynamic reason to expect this. At 

 any rate, the computed systematic and random errors 

 in h and H explain the pronounced spread in y^. An 

 example calculation in Appendix A shows that an 

 0.5-foot error in each of the four parameters, R^, R2, 

 H, and h yields an error of 3 feet in y^ for a 100-foot 

 depth. 



Table 3 gives the results of the one-time sweep 

 of the Carpinteria rock outcroppings. A single sweep 

 was made of a stretch of outcroppings approximately 

 4,000 feet long, running roughly parallel to the 

 beach. No attempt was made to repeat this sweep 

 because of the high probability of grounding one or 

 both fish. The values of y^ (Table 3) were obtained in 

 the same way as for Table 1. Figure 10 is a section ol 

 the stereo-sonar imagery obtained at the Carpinteri.i 

 site. 



There is a smaller spread in the y^ values in 

 Table 3 than one would expect from using the small 

 elevation angle, i3j. Since nominal water depth along 

 the towing track was about 25 feet, 80% of the y^ 

 values of Table 3 are possible. As shown in Figure 1 1, 

 a least-squares linear fit to the y^-versus-Rj data 

 yields a line having a slope of 3.9 degrees, which is 

 not more than 2 degrees greater than the actual 

 seafloor slope. The intercept (Figure 11) is 17.7 feet, 

 which, again, is consistent with the actual depth of 25 

 feet, since the two fish were about 5 feet under the 

 surface. 



The imagery of the Carpinteria seafloor is 

 highly photographic for the fish nearest the 

 target (upper chart, Figure 10). However, the 

 imagery from fish no. 2 is degraded, and it is 

 impossible to fuse the two charts into a three- 

 dimensional illusion.*^ The degradation is 

 apparently caused by the interference of the 

 near fish (no. 1) with the beam of the far fish 

 (no. 2). Staggering the two fish was not 

 possible by . the technique used in the 100-foot 

 scannings. Towing the two fish at different 

 distances would have increased the likelihood of 

 grounding. 



Distance, Rj 



A 



Time 



- '"m^ 



^™'^^"*'™™'™™"'"''™™"" 



mm 



V .; , 







-".«.^ " 



• 



">— 



^:^'~^^r——y^-^-^ 



;__.-_~,,,,-^^-_--_,;._-„->.--:-i^".^ 





^ 



^•^- 



Distance between horizontal lines =15 meters 

 Time between vertical lines = 2 minutes 



Figure 10. Two-fish sonar-chart readout of rock 

 outcroppings. 



Stereoscopic viewing would require that one of the charts be observed via a mirror, or made into a 

 transparency and viewed from behind. 



10 



