TM No. 377 



at its fixed location. However, the most desirable location for a wind 

 speed indicator would be either on the west side of the tower, or high 

 enough on the beacon column to have an unobstructed view of the wind field* 



When the portable system was placed adjacent to the BBELS system, a 

 consistent discrepancy showed up in the speed readouts of the two systems. 

 Wind speed values were recorded simultaneously on 8 and 9 June 19^5? and 

 the results are plotted in figure IV-14. The BBELS system was clearly biased 

 toward lower speeds of about 2.4 m sec"! as the correlation graph indicates. 

 The direction sensing records were comparable for the two systems.. 



The biasing of the BBELS anemometer was reported to the Coast Guard on 

 10 August 1965.> whereupon the unit was replaced with a new F420 system within 

 a week. This, however, did not solve the problem of the errors in wind 

 speed records already made during the wave measurements taken from August 1964 

 through June 1965* It was decided to attempt a rough correction on this 

 potentially biased wind data, based upon the assumption that the error in 

 the BBELS anemometer system increased linearly from the time of the system 

 checkout in August 1964. A straight line was drawn through the points 

 (see figure IV-14), assuming a linear relationship between the two systems « 

 The slope of the line is close to unity, but a clear ordinate intercept 

 is produced. Therefore, it was assumed that whatever was wrong with the 

 BBELS system was only affecting the threshold velocity— or, in effect, the 

 ordinate intercept. The linear time variability equation of conversion was 

 determined in the following manner. The line in the graph may be represented 

 by: 



Vn =V 8 flW J (IV-2) 



where V^j is the desired corrected wind speed, and V-g is the recorded wind 

 speed. l(t) is the ordinate intercept, which increased with time* Assuming' 

 that the ordinate intercept increased linearly from August 1964 to August 1965, 

 one may write the following equation for the time period of one year; 



Vn * v 6 + [-rr)' Ll j ( IV - 3 ) 



where N = the month (i.e., 1, 2, ...12) and I, is the biased wind speed if the 

 calibration bias changed linearly with time for one year (ii = 3-2) 



V W = V B ; 

 and for the tenth month (June 1965) 



V N = V B + 2.4 (m sec" 1 ) 



87 



