esls, amplitude spectra were generated for several profiles la the Gorda 

 Rise area which spanned multiple stationary provinces. The effect of 

 these extensions was to reduce the degree of statistical homogeneity of 

 each profile. The computed spectral slope (b) parameters for the 

 resulting amplitude spectra converge consistently to the value b - -1.0 

 as longer profiles are tested. Figure 5-10 Illustrates one such long 

 profile extending nearly 500 km from the Oregon coast. The profiles 

 analyzed by Bell (1975b) were often much longer. No formal statistical 

 argument for this convergence to b - -1.0 will be attempted here, how- 

 ever, It would appear that the concatenation of multiple profiles of 

 differing spectral characteristics results In a profile which resembles 

 a random walk. The need to define statistically homogeneous sample 

 spaces before generating statistics Is clearly demonstrated. 



Most slope values shown In Figure 5-7 cluster about -1.5 with the 

 exception of the smooth ridge axis segment south of 42''20'N. The two 

 large sedimentary provinces are represented by two values for spectral 

 slope. Figure 5-11 Illustrates a typical amplitude spectrum from these 

 sedimentary provinces. The spectrum clearly separates Into two distinct 

 straight-line segments of different slope. The average values of these 

 distinct slopes for all profiles Is given In the corresponding boxes 

 (see Figure 5-7). If the t^pothesls Is accepted that the characteristic 

 spectral slope of amplitude spectra of sea-floor topography represents a 

 dominant relief-forming process, these spectra should represent areas 

 where two processes are at work, affecting the relief In different spa- 

 tial frequency bands. It Is likely that the higher frequency band rang- 

 ing from X - 2.5 km and with b - -.6, represents the sedimentary regime 

 in these areas. The lower frequency process with b < - 1.4, Is less 



72 



