Theoretical, observational, and modeling studies have suggested the existence of an "ocean tunnel"-the water constituting the equatorial undercurrent and therefore the upwelling water in the equatorial Pacific comes from the subtropical/extratropical region. The effect of subtropical surface cooling on ENSO through this "ocean tunnel" is investigated using a coupled model. The atmospheric component of this coupled model is an empirical atmospheric model. The ocean component of the coupled model is a primitive equation model (the NCAR Pacific Basin Model) and therefore explicitly calculates the heat budget of the entire equatorial upper ocean. The model simulates major characteristics of the observed ENSO. Cooling in the subtropical ocean is introduced by reducing the radiative convective equilibrium SST in that region. Except with a delay in the response of the ENSO, the cooling imposed outside of the equatorial region has the same effect as that from an enhanced equatorial heating on the amplitude of ENSO. An enhanced cooling in the subtropics is found to reduce the temperature of the water feeding the equatorial undercurrent. The colder temperature of the equatorial undercurrent then results in colder upwelling water. This further leads to a cooling in the central and eastern equatorial Pacific SST. When this cooling eventually coincides with an ongoing La Niņa, the equatorial zonal winds are enhanced which further depress the thermocline in the western Pacific. The equatorial ocean then becomes less stable and a stronger El Niņo event develops. The stronger El Niņo event cools the western Pacific and warms the eastern Pacific and thereby largely reverses the perturbation from the subsurface to the equatorial zonal SST contrast during the La Niņa phase. The change in the time-mean SST contrast between the eastern and the western equatorial Pacific is found to be small. In the enhanced subtropical cooling case, more heat is transported out of the equatorial Pacific to the higher latitudes and the transport becomes more episodic. These findings provide further support for the "heat-pump" hypothesis for ENSO.The hypothesis states that the magnitude of El Niņo warming is proportional to the meridional differential heating over the Pacific: either an enhanced surface heating over the equatorial region or an enhanced cooling over the subtropical/extratropical ocean may result in stronger El Niņos. Moreover, El Niņo may be a mechanism that prevents the time-mean state of the equatorial Pacific from becoming substantially unstable.