tlons of returned pulses from hull-mounted sonars were used to estimate 

 RMS roughness of micro topography (<0.2 km). Histograms of spectral 

 estimates (periodograms) were generated by hand and shown to map con- 

 sistently in an area southwest of Spain. 



Bell (1975b), also analyzing data from the eastern North Pacific 

 Ocean, used power spectral techniques to generate composite estimates 

 from several sources. He discovered a functional relationship for power 

 versus spatial frequency of the form ax^ (power law foinn) to be consis- 

 tent over a large range of observation scales. In a later paper (Bell, 

 1978), this relationship was shown to hold for an enlarged data base, 

 which is also confined to the same geographic area. The importance of 

 anisotropy was recognized, and an initial look at the aspect ratios of 

 submarine features was presented. 



Berkson (1975) generated spectra from a variety of bottom types and 

 attempted to fit these with a power law form. Although a wide variety 

 of coefficients were calculated, the power law form seemed to be consis- 

 tent over many types of topography. Akal and Hovem *(1978) generated 

 two-dimensional spectra of sea-floor roughness from two sets of stereo- 

 pair bottom photographs and a contoured bathymetric chart. They noted a 

 remarkable consistency of form in all three spectra. Matthews (1980) 

 developed a deterministic approach to describe bottom roughness. The 

 North Atlantic and North Pacific Oceans were divided into 30 x 30 nau- 

 tical mile squares and the maximum relief calculated. These cells were 

 then grouped by range of relief (0-1100 m, 1100-1900 m, >1900 m) and the 

 results mapped. Recently, Naudin and Prud'homme (1980) quantitatively 

 described bottom morphologies from several areas based on multibeam 



