Lukas and Lindstrom(lS) have reported the presence of a fresh water lens which leads to form a barrier layer. Ando and McPhaden,1996~16~(Fig.9) reported that an average thickness of barrier layer is thicker in the normal years than in the El Nino years at the western equatorial Pacific. This barrier layer may keep SST at higher temperature due to hindering heat transmission downward. What are the mechanism controling a barrier layer thickness and equatorial Pacific interannual variabilities represented by El Nino?
She and Nakamoto(17) proposed a notion of an "information weight" derived from a 3-dimensional(x,y,z) EOF analysis of XBI data from surface to 500m depth. They obtained a patch-like first EOF mode having ca.2000 km of diametre.(Fig.1O). The area of large value of the information weight indicates the area of large variabilities of different modes. That region corresponds to a large current shear where an intensive diapycnal mixing may occur and thus the high salinity water from the southeastern Pacific accumulates at the depth l00m to 300m in the western equatorial warm water box and thus form a domain to support "negative fresh waterflux". This "negative fresh water flux" in the western equatorial box plays an important role in a conceptual model of the equatorial Pacific thermohaline circulation. The conceptual model will be discussed in the next section.
3. The effect of salinity forcing on the western equatorial Pacific
It is customarily assumed that buoyancy variations play little or only a passive role in the dynamics of ocean circulation experiments. Such an assumption has been useful and informative in understanding the role of wind driven ocean circulation. For a climate application, the spatial and temporal variations of temperature and salinity distribution may provide an important information in controling global scale ocean general circulation such as effects of clouds on shortwave forcing, biochemical distribution on downward shortwave penetration in the water, rate of production of turbulent kinetic energy on stress transfer rate, and advection on heat and salinity distribution. In this section, we proposed a simple conceptual model for salinity effect on an equatorial thermohaline circulation system.
From Levitus climatology map of sea surface salinity, we define the warm saline box in the western equatorial Pacific into which South equatorial current transports high salinity water. We assume that fresh water lenses do not mixed into this box. This assumption may be justified in El Nino time scale because a halocline depth does not change much compared with thermocline depth (Ando and McPhaden, 1996(16)). Since fresh water lenses in the warmpool region are found to be isolated to form a sharp pycnocline, temperature(TW) and salinity(SW) are characterized by warmer and more saline in the west box from 160E to 170W and from the equator to 105 . The east box from 170W to 140W,and from l0N to the equator is characterized by relatively lower temperature(TE) and lower salinity(SE)(Fig.11) due to fresh water flux in the Intertropical Convergence Zone (ITCZ) and the western Pacific water with fresher water transported by North equatorial countercurrent. These water mass boxes in the southwestern equatorial Pacific and the northeastern equatorial Pacific
