are connected by Equatorial undercurrent regime that transports warm and saline west box water into cold and less saline east box.
The conservation equations for density in the west box and in the east box are expressed into simple nonlinear equations for the ratio of the salinity effect and temperature effect. Since Tw-TE >0, and Sw-SE>O, the ratio r=(Tw-TE)/(SW-SE) is always positive. In this case, it is shown that the nonlinear equation has only two stable solution (Cahalan and North, 1979 (18))
This is an application of "a slope stability theorem" into the equatorial thermohaline circulation, i.e., if the local slope of the steady-state strength of salinity flux forcing (or fresh water forcing) versus the salinity advection rate is positive, then the steady state solution is stable. It is proved that one stable solution strengthen the Equatorial undercurrent when temperature dominates the west box. This is a case when the salinity forcing in the west box is negligible. The other stable solution weaken the Equatorial undercurrent when salinity dominates the west box. This happens when salinity forcing in the west box becomes important due to eastward shift of barrier layer asociated with eastward shift of warm water(that is El Nino) and increase of salinity role due to decrease of isotherm depth in the west box. Note that the average barrier layer thickness decreases during El Nino years compared with normal ones, suggesting increase of salinity effect due to decrease of r= (Tw-TE)/(Sw-SE), or decrease of (Tw-TE) >0 in the western equatorial Pacific.
The strengthing and weakning of the Equatorial undercurrent in the connecting channel between the western and eastern boxes may be a good candidate of validation of the above conceptual model for an interannual tropical thermohaline circulation. It is also interesting to test the above conceptual model by numerical experiments using the isopycnal ocean 0CM. We feel that the observation of temperature and salinity is needed to define the boxes. Where is the boundary between the western and east eastern boxes to define the variables such as the ratio of the salinity and temperature for the validation of the model?
4.Interrelation of Pacific to Indian ocean.
In the previous section we proposed a hypothesis on the effect of salinity on the equatorial thermohaline circulation. However, we feel that the wind stress on global SST variation may also be an important factor in global scale climate variation.
Ward and Hoskins(1996)(19) reported that windstress in the tropical North Atlantic and at the tropical Indian ocean froml969 to 1988 intensified in compared with that froml949 to 1968(Fig.12). This wind increase has brought SST decrease at the North Atlantic, and SST increase in the South Atlantic and in the Indian ocean. They suggested the drought at Sahaelian region in the subsahara has related to the Atlantic and Indian ocean SST change. We noticed similar SST trend in the Indian and the western equatorial Pacific between the same periods and are encouraged to conduct numerical experiments to mimic possible climate drift in SST in these oceans.
