SECKEL and YONG: SEA-SURFACE TEMPERATURES AND SALINITIES 



change of temperature and salinity ( Murphy et al. 

 1960; Seckel 1960, 1962). Advection is the product 

 of the surface temperature (salinity) gradient and 

 the component of the current normal to the 

 isotherm or salinity isopleth. Generally near 

 Hawaii the temperature increases and the salinity 

 decreases equatorward. Consequently, with a 

 northward component of flow, advection would in- 

 crease the temperature and decrease the salinity 

 at Koko Head. 



The usual spring salinity decline at Koko Head 

 is best explained by advection. It is estimated that 

 at the latitude of Hawaii there is an excess of 

 evaporation over precipitation with the highest 

 excess occurring during spring and summer 

 (Seckel 1962). Thus, since the salinity is increas- 

 ing with depth, the only source of lower salinity 

 water lies south of the islands. 



On average the orientation of isotherms is 

 northwest-southeast and that of salinity isopleth 

 is zonal. In this case only the meridional com- 

 ponent of flow causes salt advection, but both 

 meridional and zonal components of flow cause 

 heat advection. Consequently, meridional com- 

 ponents of flow causing salinity variations do not 

 necessarily produce temperature variations. Co- 

 incident changes of salinity and temperature 

 that appear to be advection related, tend to occur 

 during late winter and early spring when the 

 North Equatorial Current is weak. For example, 

 between days 60 and 110 of 1973 (Appendix B), 

 decreasing and increasing temperatures corre- 

 sponded with increasing and decreasing salini- 

 ties. Pronounced coincident temperature and 

 salinity variations occurred during the first half 

 of 1959 and are most evident in the residual 

 curves, panel D of Figures 1 and 2. 



Coincident changes in temperature and salinity 

 during specific seasons are not necessarily associ- 

 ated in the longer term. From 1956 through 1959 

 when the long-term salinity variations were pro- 

 nounced, there was no long-term temperature 

 change (panel B of Figures 1, 2). Later, a strong 

 salinity decline lasting from 1966 to 1968 corre- 

 sponded with a temperature increase. Then, as the 

 salinity returned toward 35%o, the temperature 

 also returned to the pre-1965 values. The first 

 situation may mean that there were climatic 

 shifts in the general northwest-southeast direc- 

 tion, parallel to the isotherms, thus causing a 

 long-term change in the salinity but not in the 

 temperature. In the second situation the climatic 



shift was first northward and then southward, 

 affecting both temperature and salinity. 



White ( 1975) described secular changes in baro- 

 clinic transport and morphology of the North Pa- 

 cific subtropical gyre and indicated that during 

 the years of low maximum transport the south- 

 west portion of the gyre extended farther south 

 than during the years of large transport. Sim- 

 ilarly, it is possible that higher baroclinic flow 

 and tightening of the gyre near Hawaii will result 

 in lower salinity and a relaxation of flow will 

 result in higher salinity. The long-term changes 

 in the Koko Head salinity do not correspond with 

 the changes described by White and are only in 

 partial agreement with the supposition when 

 tested against Wyrtki's (1974) North Equatorial 

 Current index. The supposition, therefore, is in 

 error or, the local wind induced surface flow, 

 superimposed on the baroclinic flow, plays an 

 important part in the long-term salinity changes. 



At Christmas Island, in addition to the heat 

 exchange and advection, the effect of wind- 

 induced equatorial divergence is a process affect- 

 ing the sea-surface temperature. Unfortunately, 

 meteorological observations suitable for the cal- 

 culation of heat exchange across the sea surface 

 were not made on the island. Estimates made by 

 Wyrtki (1966) and Seckel (1970) indicate the net 

 heat exchange across the sea surface near Christ- 

 mas Island to lie in the range of about 100 to 

 300 cal cm" 2 day" 1 . Assuming that the heat is 

 distributed through a column of water 50 m deep, 

 this process can produce temperature changes 

 from about 0.6° to 1.8°C/mo. Temperature in- 

 creases within this range are observed (Fig- 

 ure 3A). 



An important term in the net heat exchange is 

 the radiation from sun and sky that is affected 

 in the equatorial region of the central Pacific by 

 large variations in cloudiness (Bjerknes et al. 

 1969). The effect of such variability is most pro- 

 nounced in late fall and early winter (Seckel 1970, 

 figure 6). For example, the average net heat ex- 

 change near Christmas Island for November 1963 

 to January 1964 was calculated to be 177 cal cm -2 

 day" 1 , and for the same months 1 yr later, 274 cal 

 cm" 2 day" 1 . The average calculated radiation 

 from sun and sky during the same periods was 

 372 cal cm 2 day" 1 and 440 cal cm " 2 day" 1 , respec- 

 tively, and accounted for 70^ of the interyear 

 difference in the net heat exchange. The Christ- 

 mas Island water temperature declined in the 



777 



