The flux of water from the mixed layer into the thermocline/intermediate la
yers of the Pacific Ocean is quantified using chlorofluorocarbon (CFC) and
hydrographic data. The total ventilation flux of at least 123 Sv for the So
uth Pacific (SP) only slightly exceeds that of at least 111 Sv for the Nort
h Pacific (NP). Although the overall ventilation flux (to 27.3 sigma (theta
)) is similar in the NP and SP, the partitioning amongst the water masses i
s markedly different. In the NP the partitioning is equal between the wind-
driven (less than or equal to 26.5 sigma (theta)) and thermohaline (> 26.5-
27.3 sigma (theta)) layers. While in the SP the ventilation flux of the the
rmohaline layers exceeds by nearly 2:1 the wind-driven layers. The wind-dri
ven subtropical gyre thermocline ventilation flux for the NP (41 Sv) exceed
s the SP (25 Sv), and both agree well with literature estimates of Sverdrup
transports. The ventilated volumes and ages are related to the wind stress
curl and surface buoyancy fluxes. In the thermocline ventilation of Shallo
w Salinity Minimum Water (22 m yr(-1) in the NP, 15 m yr(-1) in the SP) and
Subtropical Mode Water is more effective in the NP than in the SP. In cont
rast, in the thermohaline layers direct air-sea exchange during convective
formation of Subantarctic Mode and 'Antarctic Intermediate Water is more ef
fective in ventilating the SP than processes in the NP. These same differen
ces are also used to explain the larger volume of the shadow zone in the NP
. In the subpolar regions the ventilation fluxes can be used to infer forma
tion rates of 8 Sv for the NP Intermediate Water and 9 Sv for the Subantarc
tic Mode Water. Into the tropical Pacific there is a substantial flux of 35
Sv of extratropical water for the wind-driven layers and 36 Sv for the the
rmohaline layers. The relatively young (5-20 years increasing with increasi
ng density) CFC-derived ages show that a climate anomaly introduced into th
e subtropical thermocline could be transported into the tropics relatively
quickly.