Some profile editing proved necessary to remove redundant sampling in those 

 instances where repeated observations were made at the source location and time. For 

 example, at the Strait of Juan de Fuca site there were two periods of 24 to 48 hours where 

 Nansen casts had been taken each hour. These were replaced by "typical" profiles. This 

 removed possible bias for statistical analyses. Also, in the Korea Strait data set, a large 

 number of casts were taken in 1943. Those were edited out as a separate set. An occasional 

 obviously bad profile was also removed. 



Detailed discussion of the selected shallow water environments and details of the 

 computerized methods of the statistical approach to classification and tabulation of pro- 

 files and profile gradients are given in the Appendix. 



2.2.2 Characteristic Profile Selection 



Extreme complexity and variabihty of sound speed profile shapes were observed in 

 the shallow water data samples, particularly in those from the spring and fall seasons. 

 Procedures established to select a representative sound speed profile for deep ocean 

 environments did not give meaningful results for the shallow water data. Because the 

 gradient of the vertical sound speed structure is important to acoustical propagation, 

 analytical procedures were developed to segregate profile types based on the vertical 

 gradient. From these subsets, representative profiles can be selected and acoustic results 

 evaluated on the probabihty of occurrence of selected profile types. For this study, pro- 

 files were objectively grouped as positive or non-positive gradient profiles with gradient 

 values and percentages of occurrence tabulated. 



To classify single profiles as surface duct (positive gradient) or downward refracting 

 (non-positive gradient) a new definition was required. This definition should be based on 

 the expected effect of the overall profile shape on the primary mode of acoustic propaga- 

 tion and must remain relatively simple to apply to all shallow water environments. The 

 basic definition chosen for this analysis will classify a profile as positive if the depth of 

 the absolute maximum observed sound speed value (MAX Z) is greater than 10 meters and 

 negative (non-positive) if the absolute maximum is observed within 10 m of the surface. 

 Thus a large number of possible profile shapes can be grouped together in either category. 

 As a first order profile-type separator for an assumed near surface sound source, this 

 sorting approach works well. 



Examples of possible positive gradient and non-positive (negative) gradient profile 

 types are shown in Figure 2.3. Profiles a through d are positive gradient profiles while 

 e through h are negative (non-positive) gradient profile types. Throughout the remainder 

 of this report the two classes of profiles will be referred to as positive and non-positive 

 gradient profiles. Also, positive gradient profiles are regarded as surface ducted profiles. 



Two methods of computerized statistical tabulation for the vertical sound speed 

 gradients were used. Gradients were computed in units of meters/second per meter (or 

 inverse seconds). The Appendix gives examples with tabulated results. The first approach 

 produced a tabular hsting, by season, of the number of positive and non-positive gradients 

 observed in the data set from the surface to each successive standard depth and provided 

 statistics on gradient strengths. The tables indicated if the overall gradients were positive 

 or non-positive and some measure of correlation between gradient and depth of observation. 



14 



