Berliand and Berliand (1952) has been used in this 

 study: 



^^ = -[sa(0.s)'<O-39-O.O5OV^)(l-AC=) 



+4va0i(ei-0fl)] 



where i = 0.97 (the ratio of the radiation of the 

 sea surface to a black body); 

 O5 = absolute sea surface temperature (°K): 

 d(, = absolute air temperature (°K); 

 a= 1.175 X 10"" (the Stefan-Boltzmann 

 constant); 

 ea = vapor pressure (mb); 

 k = a function of latitude (values given in 



Johnson et al., 1965); 

 C = cloudiness in tenths of celestial dome 

 covered. 



Additional energy enters or leaves the sea surface as 

 evaporation lQ^,> and sensible heat iQs)- Evaporation 

 depends upon 1 ) the velocity of the wind and the vapor 

 pressure difference between the sea surface and air 

 above it. and 2) a coefficient of proportionality. The 

 coefficient of proportionality used in our computations 

 is given by Tabata (1958): 



Qc = - 4.70 (f,v - eaW 



where cy = saturation vapor pressure at tempera- 

 ture of sea surface (mb); 

 i'li = vapor pressure of air (mb); and 

 W = wind speed (m/sec). 



Bowen (1926) established the relation between 

 evaporation and the heat conduction at a water sur- 

 face. The equation used here for sensible heat loss by 

 the ocean is derived from the relation found by 

 Bowen: 



Qs = - MTs - Ta)W 



where T^ = 



Ta = 



W = 



sea temperature (X); 

 air temperature (°C); and 

 wind speed (m/sec). 



Preparation of the Heat Exchange Charts 



Preparation of the charts presented in this report 

 began by transferring the heat exchange values from 

 computer printouts to a map of the eastern North 

 Pacific and then contouring them by hand. In drawing 

 the isolines. subjective smoothing was used, and, 

 therefore, the lines do not always conform to the num- 

 bers as printed. This technique was used to eliminate 

 the influence of values that could be an order of mag- 

 nitude different from surrounding ones, a problem that 

 is usually due to observation errors. 



Caution should be exercised in interpreting the indi- 

 vidual energy exchange values in regions having limited 

 observational coverage (see Fig. la, lb). Small errors in 

 observation and transmission can cause large errors in 

 some of the computations. In quadrangles having few 



observations, considerable bias can be introduced by \ 

 the relative positions of the reporting ships and their 

 timing with respect to the calendar month. All compu- 

 tations presented assume the data centroid to be at the 

 center of each respective quadrangle and for the middle 

 of the month. Energy exchange calculations were not 

 made for 5° quadrangles having fewer than five obser- 

 vations per month. 



CHARTS PREPARED FROM THE HEAT 

 EXCHANGE COMPUTATIONS 



Monthly Average and Anomaly Charts 



Monthly averages and anomalies of the total (net) 

 heat exchange, 2^ across the air-sea interface are 

 shown for each month of the ll-yr period, 1961-71, in 

 Part 1 of the chart section. A detailed description of 

 these charts will not be given since readers can draw 

 their own conclusions depending upon the intended 

 use of the charts. However, some remarks will be 

 made concerning spatial characteristics and mag- 

 nitudes of the Qi anomaly patterns. 



The anomaly patterns of total heat exchange Qf 

 vary widely over the chart for a particular month and 

 from month-to-month during the ll-yr period. Mag- 

 nitudes of the Qj anomalies range from less than \% of 

 the monthly average values to over 200%, with the 

 largest values occurring in summer months. In addi- 

 tion, there appears to be very little month-to-month 

 persistence in the Qt anomaly patterns. This result is 

 not surprising, since there is also very little persistence 

 in the anomaly patterns of the four heat exchange 

 terms that determine Qj. 



Since anomalies of Q; and Q^y are usually small 

 compared to those of ^^ and Q^, spatial and temporal 

 variations in Qt anomalies are primarily due to fluctua- 

 tions of evaporative and sensible heat flux. Anomaly 

 patterns of Qg and Qs tend to be fairly large in geo- 

 graphical scope and coherence; at times, more than 

 50% of the eastern North Pacific is covered by an 

 anomaly pattern of the same sign and magnitude. In 

 addition, magnitudes of the anomalies can vary widely 

 from month-to-month, ranging from less than 1% to 

 over 100% of the monthly average. 



Long-Term Mean or Normal Charts 



In order to facilitate description of the heat ex- 

 change normals, the first two harmonics (first har- 

 monic has a period equal to 1 yr or 12 mo) of the 

 Fourier Series for each heat exchange term were com- 

 puted for each of the 93 5° quadrangles on the chart. In 

 addition to the two Fourier coefficients, the per- 

 centage of series variance accounted for by each of the 

 harmonics was also computed. This type of analysis 

 was useful in interpreting seasonal cycles of the data, 

 since the two harmonics usually accounted for over 

 95% of the variance in each quadrangle; in fact, over 

 most areas of the chart the first harmonic or yearly 

 cycle accounted for over 90% of the variance. By de- 



