APPENDIX A 



DESCRIPTION OF THE DATATRON 20^ PROGRAM 



The five main functions of this program are: 



1, Determination of the depth at which speed of sound is maximum. This 

 level is defined as the layer depth. Sound velocity is computed from the 

 following polynomial: 



Sovmd velocity = a + T^ [b + T^C-c + d T^)] + eZ^ 



where a = 47^2.2473 

 b = 15.316241 

 c = .18545565 

 d = .00103379 

 e = .05938308 



Ti = Temperature (OQ) at ^± 

 Zi = Water depth in meters 



If the maximum computed sound velocity is at the lowest depth on the BT 

 trace; this depth is recorded as the layer depth with an X overpunch in 

 column 58. 



2. To search the digitized BT trace between the surface and 80 feet for 

 the surface effect . The surface effect must have a nearly constant slope 

 with an absolute value greater than 1°f/100 ft between its lower bound 



and the surface. A leeway factor of 0.5°f/100 ft is allowed between slopes • 

 If this slope extends below a depth of 60 feet, the surface effect is 

 recorded as 60 feet with an X overpunch in column 76 of the master card for 

 card output and is preceded by a negative sign on print out. In all other 

 cases the depth of the surface effect is recorded as the upper bound of the 

 first slope interval that fails to meet the stipulated req.uirements. The 

 mean gradient of the surface effect in °F/100 ft is also recorded. If 

 the search indicates no surface effect, zero is entered in all appropriate 

 columns of the master card. 



3. To determine the maximum gradient , each depth interval below the 

 recorded surface effect is examined.* The BT depth intervals are always 

 considered to be 20 feet. Thus all 40-foot intervals (360 to 880 feet) 

 are linearly interpolated into two 20-foot intervals. The upper bound 

 of the first interval having a slope equal to or less than -2PP/100 ft 

 is recorded (in the computer program -3°F, etc., are less than -2°F) . 



*When this program was written, temperature data were assumed to be 

 continuous. Occasionally a temperature value was omitted from the punch 

 card at an intermediate depth. The card was punched 000, indicating no 

 data. The Datatron program recognized this depth as the last recorded 

 interval and accordingly computed the gradients only to this depth. Expe- 

 rience shows that a linear interpolation over the adjacent depths should 

 be made. Gradients computed by use of the interpolated values are more 

 realistic. 



15 



