Pacific Ocean Heat Transport at 24°N in a High-Resolution Global Model

Abstract

Meridional heat transport in the North Pacific Ocean in a seasonally forced high-resolution global ocean general circulation model is compared to observations. At 24°N, annual mean heat transport in the model of 0.37?1011W is half the most recent direct estimate of 0.76±0.3?1015W from hydrographic data. The model value is low because the model ocean loses too little heat in the region of the Kuroshio Current Extension. The water ventilated in this region returns southward across 24°N at depth between 200 m and 500 m approximately 2°?4°C too warm. If the model surface temperature were relaxed to a temperature adjusted for the influence of persistent atmospheric cooling in this region, rather than relaxed to climatological sea surface temperature, the model heat transport would improve. Assumptions inherent in estimating meridional heat transport from hydrographic sections are tested by examining the model. Rather than the abyssal circulation being steady, the model's deep western boundary currents vary seasonally to balance the seasonal cycle of Ekman transport, producing a larger seasonal variation in heat transport than is generally supposed for direct heat flux calculations. But the variability is such that there is no net contribution to the mean beat transport through a seasonal correlation between winds and surface temperature. The use of surface temperature observed during a single hydrographic section can seasonally bias an estimate of the wind-driven component of the beat transport, so a modification is proposed to the procedure by which compensation is made for seasonal variability in direct beat transport calculations. The most recent direct estimate was based on a springtime section, for which the model beat transport would be underestimated by about 0.05?1015W. Interannual timescale correlations in the transport and temperature of the Kuroshio Current contribute a net southward transport of some 0.07?1015W. The role or simulated mesoscale eddies is minor. Given the comparable order of the southward interannual heat transport and the northward seasonal bias, this present study does not suggest any significant revision to the latest direct heat transport estimate for 24°N in the Padfic. Other features of the model general circulation are noted, including a Kuroshio Current transport that is stronger than observed and the persistence of a branch of the Kuroshio that does not separate at 35°N but continues close to the coast forming unrealistically deep mixed layers through intense surface cooling.