tude and longitude. Figure E-6 shows a graphic representation of this 

 surface and its spectral parameters as a function of azimuth, in the 

 same format as the artificially generated surfaces shown in Figures E-1 

 through E-4. 



Compare the bathymetric surface shown in Figure E-6 to the artific- 

 ially generated surface shown in Figures E-3 and E-4. The overall mor- 

 phologies are quite similar, with a longer wavelength, higher amplitude 

 component in the north-south direction relative to the orthogonal trend. 

 This similarity in morphologies is also expressed as a similarity in the 

 azimuthally dependent roughness models, although the parameters gener- 

 ated from the bathymetric surface are somewhat noisier than the artific- 

 ially generated examples. In all cases, the slope parameter b is not 

 constant with azimuth, as was the case for the surfaces studied in Chap- 

 ter 6. Like the artificial surfaces of Figures E-3 and E-4, the spec- 

 tral slope is approximately b = -1.5 in the 0° or 180° azimuth and 

 approaches b = -1.0 for the 90° azimuth. The intercept parameter in 

 Figure E-6 reaches its maximum at e =90°, similar to the example in 

 Figure E-4. In the case of the Mendocino Fracture Zone, this parameter 

 is approximately doubled in the east-west direction over the north-south 

 direction. This indicates that for wavelengths near 1 km, the Fracture 

 Zone surface is twice as rough for e = 90° as for 6=0°. In longer 

 wavelengths, this relationship is reversed as evidenced by the much 

 higher total relief in the north-south direction. Such a reversal is 

 only possible for near orthogonal trends with different spectral slopes. 

 Sinusoidal regression lines, which comprise the basic model of Chapter 

 6, are included in Figure E-6 to emphasize how poorly this simple model 

 describes the two trend case. 



213 



