IV. DATA PROCESSING AND ANALYSIS 



Briefly, the data analysis technique consisted of subjectively 

 selecting cases from the temperature records wherein identifiable 

 "signatures," i.e., distinctive temperature variations related in form 

 (but not necessarily amplitude), could be ascertained with a high degree 

 of certainty to have occurred at all three measuring stations in a 

 single horizontal plane. The times of passage at each known station 

 were then used to calculate the speed and direction of the signatures. 

 A total of 95 cases were selected during the fall period (28 September - 

 15 October 1959) and 201 cases from the spring period (13 May -4 June 

 I960). 



In an attempt to preserve some measure of statistical significance, 

 all records were examined during each data period and every case 

 was recorded where the signature relative to background "noise" 

 was distinctive enough to almost completely assure nonambiguous 

 occurrence at each station. In cases where a "train" of several 

 signatures occurred, the most distinctive one was selected for analysis 

 or several were averaged. The magnitude of temperature variation was 

 not considered; distinctive features and uniform shape of the signature 

 observed at each station was responsible for selection. The middle of 

 the temperature peak or crest (internal wave trough) was normally 

 used to index the time of passage at each station. Signature speed 

 and direction were computed on an LGP-30 digital computer. 



No attempt was made to ascertain the effect of horizontal water 

 motion on signature speeds since accurate ocean current observations 

 were normally not taken concurrently. Appendix I indicates the 

 importance of accounting for the effect of field motion on individual 

 signature velocities. The lack of simultaneous current observations 

 leaves only the alternative of computing average speeds in each 

 quadrant for the single two-week period of reasonably good data. 

 This problem is further discussed in the following sections. 



V. RESULTS AND DISCUSSION 



A. SIGNATURE CHARACTERISTICS 



Figures 5 through 12 are typical sections of temperature records 

 selected to represent a cross section of observed activity. Figures 

 5 and 6 are samples of temperatures recorded in a single plane (or 

 level) at all three stations during 1.5-hour time spans. Note that 



