mately 60% of the data grid (Fig. 7) and imply less oceanic heat loss 

 for the flux estimates computed from monthly mean variables. Higher 

 heat loss would be expected over a large area adjacent to the Pacific 

 Northwest. The high linear correlations between heat exchange com- 

 ponents estimated from monthly mean atmospheric properties and 

 those computed from synoptic reports imply that the principal spatial 

 and temporal patterns evident in the charts in Appendix I are indepen- 

 dent of the averaging methods used to compute the monthly mean 

 fluxes. 



Effects of Variable Exchange Coefficients 



The empirical formulae which have been applied to studies of 

 large-scale turbulent latent and sensible heat transfer vary primarily 

 in the form of the exchange coefficients which have been used. In 

 the derivations of the bulk aerodynamic formulae, it is usually 

 assumed that the coefficients C E and C w are constants, approxi- 

 mately equal to the drag coefficient, C , for momentum transfer 

 (Roll 1965). For example, Kraus (1972) suggested the value 

 C E =C H = C o =0.0013. In recent years, however, a great deal of 

 effort has gone into the determination of these bulk exchange coef- 

 ficients and their dependence on windspeed, atmospheric stability, 

 and the aerodynamic roughness of the sea surface as a function of 

 the spectral shape of the ocean wave field (Davidson 1974). Until 

 these coefficients have been accurately determined for all combina- 

 tions of stability, windspeed, and sea conditions, the bulk exchange 

 formulae are best regarded as dimensionally correct parameteriza- 

 tions which need further experimental verification (Pond 1975). 



Busch (1977) discussed the results of recent research and described 

 the work of Friehe and Schmitt (1976), who compiled data from sev- 

 eral sources, including their own, and found that the exchange coeffi- 

 cient for the turbulent flux of latent heat was well described by 

 C £ = 1.32 xlO' ±0.07x10'. The coefficient for sensible heat transfer 

 could be approximated by C„= 1.41 x 10"' ±0.02 x 10° over wider 

 ranges of windspeed and sea-air temperature differences. Their analy- 

 ses suggested that the coefficient. C t . was constant and larger than that 

 for sensible heat transfer. C H , for equivalent windspeeds, and demon- 

 strated a dependence of C H on amiospheric stability and windspeed. An 

 extensive review of comparable data for momentum transfer (Garratt 

 1977) substantiates the windspeed dependence for the drag coefficient. 

 C D . The dependence of the transfer coefficients on atmospheric stabil- 

 ity has been predicted by Deardorff (1968). Based on theoretical con- 

 siderations and a review of experimental data. Coantic (footnote 3) 

 recommended the values. C £ = C„= 1.3x10"' for windspeeds <10m/s 

 and C E = C„= (1 .0 + 0.05 1/ 10 ) x 10 -' for the range, < U w < 20 m/s. 

 Eddy flux measurements discussed by Anderson and Smith (1981) 

 indicated a positive windspeed dependence for the neutral evaporation 

 coefficient. C c = (0.55 + 0.083 UJx 10-\ for windspeeds of 5 to 11 

 m/s. No dependence of C E on atmospheric stability was observed, 

 although this result may have been influenced by the lack of numerous 

 direct measurements in stable conditions. The corresponding tempera- 

 ture flux data showed a 40% increase in the coefficient for sensible heat 

 transfer from C H = 0. 82x10' in a stable boundary layer to 

 C H = 1 . 1 2x 1 ' for unstable conditions. Because there is large scatter in 

 open ocean determinations of Q and C H . lack of agreement among 

 individual observers, and a tendency for uncertainties in the flux esti- 

 mates introduced by errors in the observed atmospheric properties to be 

 larger than errors caused by variations in the exchange coefficients, we 

 assumed identical and constant values for C E and C H in this study. 



To investigate the possible effects of variable exchange coeffi- 

 cients on the turbulent transfer processes, the monthly mean fields 

 of latent and sensible heat flux have been recomputed, and the con- 



stant values C E and C H in Equations (4) and (5) were replaced by 

 coefficients which varied with windspeed and atmospheric stabil- 

 ity. The linear relationship proposed by Coantic (footnote 3) was 

 used to define the windspeed dependence of C E and C H , and Dear- 

 dorff 's (1968) empirical expressions, which are functions of the 

 bulk Richardson number (Ri) g , were adopted to parameterize the 

 dependence of the exchange coefficients on stability. Neutral stabil- 

 ity was assumed when the absolute value of the air-sea temperature 

 difference was < 1 °C and the available data were further reduced 

 by restricting the calculations to the range -0.2<(^?/) B <0.1. In the 

 following discussion we consider only the calculations for latent 

 heat flux. Complete details will be reported elsewhere. 



The combined effect of windspeed and atmospheric stability was 

 computed as a percentage increase (decrease) in the magnitude of 

 the monthly mean latent heat flux above (below) the corresponding 

 value computed with a constant coefficient. Values corresponding 

 to the percentage differences for June and December, respectively, 

 are displayed in Figures 8 and 9. In June, the average percentage 

 increase is between 10 and 15%, although in the region of the wind 

 stress maximum between Point Conception and Cape Blanco (Nel- 

 son 1977). the computed increase may be larger than 25%. The cor- 

 responding evaporative heat loss increases from approximately 45 

 to 56 W/m 2 . The winter distribution (Fig. 9) shows an alongshore 

 gradient. South of Point Conception the average percentage 

 increase is approximately 15%, while in the region adjacent to the 

 Pacific Northwest differences >40% might be expected. 



The principal difference between estimates of latent heat flux 

 computed with a variable as opposed to a constant exchange coeffi- 

 cient, C E , is caused by the windspeed dependence of C E . Additional 

 analyses were performed to identify the relative effects of atmo- 

 spheric stability and windspeed. The results for June suggest a 10 to 

 20% increase in latent heat transfer due to the dependence of C E on 

 windspeed. Stability effects account for <5% of the overall 

 change. Higher windspeeds and unstable conditions associated 

 with transient winter storms produce larger differences. Stability 

 effects account for a 10 to 15% increase in latent heat flux during 

 winter, and the windspeed dependence of C F contributes between 5 

 and 25% of the difference. The relative differences between esti- 

 mates of latent heat flux computed with a constant exchange coeffi- 

 cient as opposed to a coefficient which depends on windspeed and 

 stability, are larger than values of <5% reported by Dorman et al. 

 ( 1974) at OWS-N, but more consistent with increases of 6 to 15% 

 discussed by Husby and Seckel (1975) for OSW-V and with results 

 for the North Pacific trade wind zone described by Seckel (1970). 



SEA LEVEL OCEAN-ATMOSPHERE 

 PROPERTIES AND THEIR 

 SEASONAL VARIABILITY 



The predominant factors influencing the seasonal variations in the 

 air-sea interaction processes of the California Current region are the 

 seasonal movements and changes in intensity of the subtropical high 

 pressure system and the continental low pressure system and variations 

 in the properties of the waters of the California Current. The variations 

 in the atmospheric pressure centers modify the surface wind field and 

 the distribution of cloud cover, and, through the action of the surface 

 wind stress, induce upwelling at the coast during summer. Upwelling 

 has a marked effect on the climate along the west coast of North Amer- 

 ica (Smith 1968). 



Figure 10 shows the average monthly sea level atmospheric pressure 

 distributions over the eastern North Pacific Ocean and the west coast of 



14 



