The derivation of the functional form of this 9 dependence would 

 best be investigated and verified in areas where complete two- 

 dimensional measures of depth are available. Areas surveyed with multi- 

 beam sonar systems such as SASS and SEABEAM are ideal for this purpose. 

 If a functional form is determined to be consistent, discrete samples 

 from randomly oriented tracklines could be fitted with this functional 

 model and F(6,s) estimated. Such studies will be presented in Chapter 6 

 and Appendix E. 



One complication that could arise due to anisotropic roughness is 

 in the use of the province picker for the delineation of stationary 

 provinces. In a highly lineated area, the power level of particular 

 narrow frequency bands would be dlrectionally dependent. The province 

 picker, however, measures total energy (integrated power) rather than 

 discrete power, and as mentioned previously, the peak shift due to 

 oblique sampling of a linear feature does not affect the total energy 

 measured. Unless a major spectral peak is shifted beyond the low cut- 

 off frequency of the band pass used in the province picker, the results 

 should not be adversely affected by directional sampling problems. Even 

 this problem seems unlikely to arise as thus far in all spectra sampled, 

 none have shown unusual low spatial frequency peaks such as are seen by 

 Hayes and Conolly (1972), and which are probably unique to a few tec- 

 tonic provinces such as the Antarctic-Australian Discordance. 



41 



