Table 2. — Heat exchange processes for each month, cal. cm.~^day~^, July 1963 to June 1965, at Christmas 

 Island, lat. 2° N. , long. 157° W. ; Johnston Island, lat. 16° N. , long. 169° W. ; and Ocean Weather 

 Station November, lat. 30° N. , long. 140° W. Subscripts 1, exchange processes computed daily; sub- 

 scripts 2, exchange processes computed from mean monthly meteorological properties — Continued 



ruary 1964, at lat. 30° N., long. 140° W., Q(E|) = 

 266 cal. cmT^day"' and QCEg) - 192 cal. cmr^ 

 day"'. The evaporation computed from mean 

 monthly meteorological properties was 28 per- 

 cent below that computed daily. The mean wind 

 speed for this month was 6.4 m. secT' giving 

 G(6.4) = 26 cal. cm7^ day"'mb.~' The average 

 daily value of G(W) was 38.3 cal. cm7^ day~'mbr' 

 Here G(6,4) was 32 percent below G(W). Be- 

 cause of the nonlinearity of G(W), a changing 

 frequency distribution of wind speeds from 

 month to month would contribute to the chang- 

 ing differences between Q(E|) and Q(E2) in the 

 above listing for the three locations. 



The differences between Q(E|) and Q(E2) at 

 Christmas Island are small because winds are 

 generally below 6 m. sec."', so that they fall in 

 the approximately linear portionof G(W) (fig. 2). 



The discussion concerning the effect of tem- 

 poral changes in wind speed on the mean value 

 of G(W) applies also to spatial changes in wind 

 speed. It can be shown, however, that the effect 



of wind gradients of the magnitude occurring in 

 the trade wind zone on the area mean of G(W) is 

 small. 



The dependence of the computed heat of evap- 

 oration on the method of processing can be dem- 

 onstrated further for the 5° square lat. 25° to 

 29° N., long. 140° to 144° W., where the number 

 of observations were 164 to 225 per month. 

 Three methods of computations were used: (1) 

 the heat of evaporation was computed for each 

 set of meteorological data and the daily mean 

 value of the heat of evaporation was then used 

 to obtain the monthly average, which is desig- 

 nated Q(E); (2) mean daily meteorological prop- 

 erties for the 5° square were used to compute 

 the daily heat of evaporation from which the 

 monthly means were then obtained--these values 

 are equivalent to those in the above listing of 

 the three locations which were designated Q(E|); 

 (3) from the daily mean values for the 5° square 

 the monthly mean meteorological properties 

 were computed which, in turn, were used to ob- 



II 



