Clark and Yarnall 



expense of digitally computing more accurate spectral estimates were 

 prohibitive. On-line computers are being incorporated into the MIMI 

 system and future experiments should yield more reliable spectra. 



An attempt has been made to see if the stability frequency, 

 N(b) , allows the distribution of frequencies noted in Figure 18 when 

 its calculation is based on local conditions. Since the required 

 knowledge of the vertical distribution of oceanographic parameters in 

 the sound path is not available for the period of LCT-1, recourse 

 was again made to hydrographic data obtained in 1961. For a station 

 located in the vicinity of the H3 hydrophone site, measurements of 

 temperature and salinity at various depths were used to calculate 

 density and sound speed. These (discontinuous) data were approximated 

 by "least squares fit" fourth order polynomials in order to obtain 

 equations for density, density gradient, and sound speed as functions 

 of depth. The formula for the stability frequency calculations was 

 obtained from Eckart,-'-' 



The data were inadequate in several respects for this 

 calculation. One immediate problem was the fact that hydrographic 

 data was available for only a portion of the full water depth at the 

 measurement site. This effectively chopped off the lower frequency 

 end of the calculated N(h) curve, permitting only an experimental 

 comparison with a range of frequencies which included the maximum 

 stability frequency. 



The solutions predict a stability frequency maximum of 

 0.0195 Ha (period of 5,35 min) occurring at a depth of 50 m for the 

 September (1962) data, and a maximum of 0.0190 Ha (5.50 min) at 70 m 

 for the October data. The solutions are illustrated in Figure 17. 

 These solutions suffer from the additional deficiency that the maxi- 

 mum for N(a) is quite sensitive to the value of the measured para- 

 meters near the depth of this maximum. The measured parameters were, 

 in the first place, discrete and, in addition, were effectively 

 smoothed by the polynomial approximation. On this basis low calcula- 

 ted values for the N(e) maximum might be expected. An improved re- 

 presentation of the density gradient would probably lead to a closer 

 match between the predicted and the observed maximum frequency. If 

 the density stratification in the Straits is not radically altered 

 from year to year, the distributions of Figure 18 may be represen- 

 tative for this region of the Straits during the fall and winter 

 months. (Solutions for data obtained in November, December and 

 January of the 1962-3 study yielded curves similar to the two illus- 

 trated,) 



It would be of interest to compare the spectral description 

 provided by the acoustic data with a "point" measurement obtained from 

 an array of environmental sensors specifically designed to measure 

 short period internal waves. 



The acoustical data evidently contains spectral signatures 



3Z4 



