TM No. 3^12 



defined crests, visually estimated at I.5 second periods. Wind speed was 

 about 9 m sec-1, wave height about 50-75 cm, and water depth below the 

 instrument about 7 meters. The assumption was that the waves were essentially 

 surface (ice., deep water) waves. 



The co-variance U' ¥* bet^reen the II88 pairs of data was fo\ind to be 

 -7 08 cm^ sec"^, which is an estimate of the magnitude and sign of the Reynolds 

 stress (down -'jard momentum flux) . The Reynolds stress is here defined as 

 2^= /? U' ¥' . /5 is the density of water, which is assumed constant. The 

 linear correlation coefficient between U' and ¥' was -0.21. 



Spectral analysis was made of the time series data utilizing the methods 

 described by Tukey3 , Auto co variance spectra were made of the time series 

 data of the two velocity components and are shown in the upper curves of 

 figure 2. The spectra show peaks at wave periods of about I.5 sec« which 

 was approximately the observed wave period. Both of the velocity components 

 exhibit similarly shaped peaks. The co-spectra component of the covariance 

 spectra exhibits a negative peak at the dominant wave frequency. This indi- 

 cates that the downward momentum flux, due to the negative correlation of the 

 velocity components, occurred at frequencies equi-valent to those of the waves. 



The interpretation given to dynamics of wave measurements from this 

 statistical analysis is critically dependent upon the sensing character of 

 the ducted meter system. A continuing study is being conducted of the 

 response characteristics of the meters, including a series of laboratory 

 and field tests. (The OMDUM II system was 'developed and was geometrically 

 similar to OMDUM I. This new system utilized similar magnets mounted on the 

 impellers, and pickup coils in lieu of the oscillator and amplifying circuits 

 which proved unreliable in OMDUM I.) 



The OMDUM II was first calibrated for steady-state flow parallel to the 

 axis of each of the cylinders. Then the axes of the cylinders were set at 

 arbitrary angles between and 90° from the direction of steady flow to 

 obtain the variation in instrument response to the "off angle flow". 

 Finally, the impeller response to accelerative flow was measured. 



The results indicate that the meters have essentially identical cali- 

 brations for steady flow parallel to the respective cylinder axis. The 

 flow response as the meter axis is rotated with respect to the flow direction 

 follows closely the cosine law. In other words, the component of flow varies 

 as the cosine of the angle subtended by the cylinder axis and the mean flow 

 direction, Fiirther tests are being conducted for p-urposes of assessing the 

 exact angle-response relationships » 



To assess the response time of the impellers, the OMDUM II was sinus- 

 oidally oscillated vertically in a test tank^ and the fluctuating response 

 of the meter was recorded. These velocity variations, obtained from the 

 calibration curve for the steady flow, were compared directly with the 



3 



