is based on the formulas 



Tx = pC^W^ sin 6 and Ty = pCpW^ cos 6, 



respectively. 6 is the direction of the wind 

 vector; the zonal component is positive east- 

 ward, and the meridional component is positive 

 northward. 



The zonal (W^) and meridional (Wy) compo- 

 nents of the wind vectors were calculated dur- 

 ing the initial sorting of the meteorological data 

 so that recalculation of the trigonometric func- 

 tions is unnecessary. By substitution of W^/W 

 = sin 6 and Wy/W = cos 6 , the above formulas 

 become 



Tx = pCjjWWjj and 



Ty = pCjjWWy. 



The density of the air has been assumed to be 

 constant at p = 1.22 X 10"^ g. cm."^ . The stress 

 is expressed in dynes cm."^when the wind speed 

 is given in cm. sec."' . 



EVALUATION OF RESULTS 



The uncertainties in the estimation of large- 

 scale, sea-air heat exchange processes (dis- 

 cussed by Seckel, 1970) are due to (1) inad- 

 equacies in the distribution and quality of the 

 data, (2) the methods of processing that must be 

 used because of inadequate spatial and temporal 

 distribution of observations, and (3) uncertain- 

 ties in the empirical formulas. These factors 

 also affect the sea-air exchange process--the 

 wind stress — which is the subject of this paper. 



The derivation of formulas for the computa- 

 tion of wind stress is the subject of active re- 

 search and is thoroughly treated in the litera- 

 ture cited here and elsewhere. The method of 

 processing to be used must be chosen with care 

 because the wind stress is a vector whose mag- 

 nitude is a product of the square of the wind 

 speed and a drag coefficient that varies with 

 wind speed. Malkus (1962) examined the methods 

 of computation to be used in the light of these 

 factors. Here the evaluation will be concerned 

 with the data inadequacies, comparisons with 

 other results in the North Pacific, and with in- 

 terseason and interyear comparisons. Because 

 the wind stress is of principal interest in the 

 TWZO Pilot Study, only this property is com- 

 pared. The results of the comparisons, how- 

 ever, also apply qualitatively to the components 

 of the wind vectors. 



Inadequacy in the Distribution 

 and Quality of Data 



South of lat. 15° N. the number of observations 

 per month was small and in some 5° squares ob- 

 servations were lacking (table A). To complete 

 contouring of the smoothed charts, liberties 

 were taken when interpolation or occasionally 

 extrapolation was necessary. For these lati- 

 tudes the results presented in table B must 

 therefore be regarded with caution. 



North of lat. 15° N. sampling was more ade- 

 quate, particularly in the latitudes from Hawaii 

 northward, which contain the shipping lanes 

 from North America to the Far East and to 

 Hawaii. Reliable results in table B are those 

 from lat. 32° N., long. 142° W., which, in addi- 

 tion to a large number of observations from 

 merchant vessels, contain those from the U.S. 

 Coast Guard Weather Station November. At 

 Weather Station November, meteorological ob- 

 servations are made by trained observers from 

 the U.S. Weather Bureau. 



Results of table B at lat. 2° N., long. 157° W. 

 and at lat. 17° N., long. 167° W. should also be 

 reliable because they include the daily observa- 

 tions from Christmas and Johnston Islands. 

 There is evidence, however, that wind speeds 

 measured on atolls maybe below those over the 

 adjacent ocean. Ramage (MS.') made reference 

 to wind measurements on Palmyra Island (lat. 

 5°52' N., long. 162°6' W.) an atoll with a maxi- 

 mum elevation of 3 m. above sea level. He showed 

 that wind speeds depended on location of mea- 

 surement, and at some sites were as much as 

 38 percent below those measured at the exposed 

 windward side of the atoll. 



The magnitude of the island effect on the wind 

 speeds measured at Christmas and Johnstonls- 

 lands is not known. The interpolated values of 

 table B also depend on the observations in adja- 

 cent 5° squares that would tend to reduce the 

 island effect. At lat. 2° N., long. 157° W. (table 

 B) the 2-year mean of the zonal component of 

 the wind velocity is 4.1 m. sec."' . At the same 

 latitude but long. 152° and 162° W. the 2-year 

 means are 4.6 m. sec."' and 4.9 m. sec."' , re- 

 spectively. The lower value at lat. 2° N., long. 

 157° W. may be due to a Christmas Island bias. 

 At lat. 17° N., long. 167° W., no systematic bias 

 is apparent. 



'Ramage, C. MS. Monsoon meteorology. Hawaii 

 Institute of Geophysics, University of Hawaii, 

 Honolulu, Hawaii 96822. 



