roughness. The artificial surface shown In Figure E-3, as well as the 

 Mendocino Fracture Zone shown in Chapter 4, are analogous to this hypo- 

 thetical example. 



The computed spectral parameters shown below in Figure E-3 also 

 reflect the complexities of this surface. The a(9) parameters appear to 

 vary regularly with azimuth, although not following the cosine curve 

 derived by simple regression, the variation occurs in spite of the fact 

 that the input signals have Identical a's. Only at exactly 90° azimuth, 

 does the a parameter jump to the input value. The b(9) values also fol- 

 low a complex pattern which is not described by the illustrated cosine 

 curve. Extensive analytical geometry would be necessary to reach an 

 understanding of these variations. 



A final artificial surface is presented as Figure E-4. In this 

 case, both the spectral slopes (b) and Intercepts (a) of the input sig- 

 nals are d^/ferent The construction is Identical to that shown in Fig- 

 ure E-3, with the exception that the intercept (a) parameter in the 

 east-west direction has been increased to 2.0. Between the input values 

 at 0°, 90°, and 180°, the variation of the spectral parameters appears 

 even more complicated than the results from Figure E-3. The combination 

 of signals at oblique azimuths, or the combination of more than two sig- 

 nals, would result in an even more complex pattern. 



With the insight gained by examining these artificially generated 

 surfaces, a more complete analysis of the anisotropy of the Mendocino 

 Fracture Zone can now be conducted. Figure E-5 presents a contour chart 

 of the surface used in this analysis, which represents a subarea of the 

 base chart shown in Figure 4-7. The digital data, collected by the SASS 

 multlbeara sonar system, is gridded at a spacing of 0.05 minutes of lati- 



209 



