vehicle Is continuously measured. f^rre and Ryan (1984) have developed 

 a method of combining these sources of Information Into a detailed con- 

 tour chart of bathymetry. These contours, when evaluated for depth on 

 an evenly spaced grid, comprise the data set used In this study. 



Figure 6-11 Illustrates a contour chart of the study area, which 

 encompasses the Carteret Canyon, continental slope, and upper continen- 

 tal rise. The smaller area outlined was used to produce the spectral 

 parameters Illustrated In Figure 6-12. The data grid uses a spacing of 

 only 5 meters and represents much higher resolution than the surface- 

 derived sonar data In the other two examples. Unfortunately, due to 

 round-off errors producing a whlte-nolse level of 0.3 meters In the data 

 (only whole meter depths were retained), most spectra were only above 

 noise to the 100 m wavelength band, similar to the resolution of SASS. 



Figure 6-12 Illustrates the results of the azlmuthal dependence of 

 spectra study. The slope parameter (b) shows a mean of -1.88, lower 

 than the other two study areas. The data also Indicate what might be an 

 example of azlmuthal dependence of b, such as that observed on the 

 Mendocino Fracture Zone and discussed In Chapter 4 and Appendix E. The 

 Intercept (a) parameters are u > .87 m, v > .23 m and relative azimuth 

 (6^) > 85°, which represent a level of roughness and anlsotropy Inter- 

 mediate to the previous two examples. Because the original survey was 

 collected at an azimuth of 140°, the true azimuth of anlsotropy Is 45". 



In conclusion, the simple theoretical model of the effect of linear 

 trends on frequency domain statistics, can adequately describe anlsot- 

 ropy in three test areas. The parameter b(6) appears to be Independent 

 of azimuth In at least two areas, although quite noisy. The sinusoidal 

 construction of the Intercept parameter as 



99 



