tain the monthly heat of evaporation--these val- 

 ues are equivalent to QiEg) of the above listing. 

 The results of the three methods of computation 

 are presented in figure 3. 



In method 1, if for each day the meteorologi- 

 cal observations were uniformly distributed in 

 the 5° square, Q(E) would represent the mean 

 monthly evaporation for the area. In methods 2 

 and 3, Q(E|) and Q(E2) represent the results for 

 the mean location of observations. Since the 

 heat exchange processes in table B are based 

 on the interpolated, mean meteorological prop- 

 erties, they, too, represent values at the loca- 

 tions listed rather than for the 5° square. 



For the 24 months, the differences between 

 Q(E) and Q(E|) range from 2 to 32 cal. cm."^ 

 day"' and average 17 cal. cmT^ day"'. The dif- 

 ferences between Q(E|) and Q(E2) range from 

 -56 to 69 cal. cmT^ day"' and average 35 cal. 

 cmT^day"'. The average difference between 

 Q(E) and Q(E|) is less than 7 percent, about as 

 reported by Malkus (1962), and reflects the ef- 

 fects of variability in the data. The differences 

 between Q(E|) and QCEg) are about the same as 

 those for the computations at Johnston Island 

 and Weather Station November, listed above. 



Figures shows that in all months QiEj) is the 

 lowest value except for December 1964, when it 

 was 10 cal. cmr^day"' above Q(E|), and Feb- 

 ruary 1965, when it was 56 cal. cmr^day"' above 

 Q(E|). The large deviation in Q(E|) - QCEg) 

 during February 1965 from the common occur- 

 rence illustrates that fluctuations in the evap- 

 oration are not only due to the variable wind 

 factor, G(W), but also due to changes in the sea- 

 air vapor pressure difference. Daily values 

 showed that meteorological conditions departed 

 from those commonly observed in this 5° 

 square. During February 1965, a relatively 

 high frequency of winds above 10 m. sec."' cor- 

 responded with low values of the sea-air vapor 

 pressure difference. The winds below 10 m. 

 sec."' corresponded with higher sea-air vapor 

 pressure differences. In consequence, the mean 

 monthly meteorological properties yielded a 

 higher evaporation rate than the mean daily 

 evaporation rates. The departure in meteoro- 

 logical conditions for this month is apparent 

 from the data summaries in table A (lat. 25°- 

 29° N., long. 140°-144° W.). Relatively strong 

 southeasterly winds with a lower sea-air vapor 

 pressure difference contrasted with northeast- 



300 



a 



1963 



1964 



1965 



Figure 3. — Monthly heat of evaporation, cal. cm. day , lat. 

 25° to 29° N., long. 140° to 144° W. , July 1963 to June 

 1965, Q(E), mean heat of evaporation based on computations 

 made for each set of meteorological observations, Q(Ei), 

 mean heat of evaporation based on daily mean meteorological 

 properties, and QCEa), mean heat of evaporation based on 

 monthly mean meteorological properties. 



12 



