Comparison of ocean heat content estimated using two eddy-resolving hindcast simulations based on OFES1 and OFES2

Abstract

In this study, we have compared the ocean heat content (OHC), estimated using two eddy-resolving hindcast simulations based on Ocean General Circulation Model for the Earth Simulator version 1 (OFES1) and version 2 (OFES2). Results from a global objective analysis of subsurface temperature (EN4) were taken as a reference. Both EN4 and OFES1 suggest that OHC has increased in most regions of the top 2000m during 1960-2016, which is mainly associated with the deepening of neutral density surfaces and variations along the neutral density surfaces of regional importance. Upon comparing the results obtained from the two OFES hindcasts, we found substantial differences in the temporal and spatial distributions of the OHC, especially in the Atlantic Ocean. A basin-wide heat budget analysis showed that there was less surface heating for the major basins in OFES2. The horizontal heat advection was mostly similar; however, OFES2 had a significantly stronger meridional heat advection associated with the Indonesian Throughflow (ITF) above 300m. Additionally, large discrepancies in the vertical heat advection were also evinced when the two OFES results were compared, especially at a depth of 300m in the Indian Ocean. We inferred that there are large discrepancies in the vertical heat diffusion (those that cannot be directly evaluated in this study due to data unavailability), which, along with the different magnitudes of sea surface heat flux and vertical heat advection, were the major factors responsible for the examined differences in OHC. This work suggests that OFES1 provides a reasonable multi-decadal estimate of global and basin-integrated warming trends above 700m, except for the top 300m for the Pacific Ocean and between 300-700m for the Indian Ocean. Although the estimates of the global OHC during 1960-2016 are consistent with observations between 700-2000m, caution is warranted while examining the basin-wide multi-decadal OHC variations using OFES1. The seemingly suboptimal OHC estimate based on OFES2 suggests that any conclusions on long-term climate variations derived from OFES2 might suffer from large drifts, necessitating audits.