slope parameter Is different la two areas within a single shading pat- 

 tern (bottom center of chart). Many roughness province boundaries coin- 

 cide with obvious physiographic province boundaries. In these cases, 

 the bathymetry was used to trace the province boundaries between sample 

 spectra. In many other cases, no boundary was obvious In the bathymetry 

 and such tracing was not possible; that Is, what appears as an abrupt 

 boundary may represent simply an arbitrary contour of a continuous gra- 

 dient. Notice also that the Illustrated bathymetry from Chase et al. 

 (1981) was based on a totally different data set than that used for 

 spectral model generation, and therefore some provinces appear In the 

 model which are apparently unsupported Independently In the bathymetry. 



The gross distribution of the roughness statistic a corresponds 

 fairly well with what one would expect Intuitively. The roughest areas 

 (a > 2.5 ra) are located at the Gorda Rise crest, Blanco Fracture Zone 

 and over most seamounts. In particular the President Jackson Seamounts. 

 The Ijftast rough areas correspond to the sedimentary provinces of the 

 Tufts Abyssal Plain, Cascadla Basin, the Astoria deep-sea fan, and the 

 continental shelf. The Cascadla Channel appears as an Intermediate 

 roughness province which can be traced very easily through the Blanco 

 Fracture Zone and onto the Tufts Abyssal Plain. The Astoria Channel, 

 located to the east of the Cascadla Channel, Is too narrow (<8 km) for 

 this analysis and thus does not appear as a separate province. 



Within this gross distribution of roughness, some more subtle pat- 

 terns can be Identified. The continental margin between the shelf and 

 abyssal plain appears to be banded with the topography becoming gener- 

 ally rougher down slope. This particular continental slope represents a 

 slowly converging margin between the North American Plate and the Juan 



65 



