Sensitivity of Ocean Circulation and Tracer Distributions to Changes in Surface Heat and Freshwater Fluxes in an OGCM

We vary the sea-surface restoring temperature and salinity in an ocean general circulation model (OGCM) to obtain a suite of unique ocean circulation states. We find that the differences in implied surface heat and freshwater fluxes between model runs are relatively small (within the error of previous estimates for the modern ocean), but observe significant differences between runs in the large-scale ocean circulation. In particular, the relative fractions of North Atlantic Deep Water (NADW) and Antarctic Bottom Water (AABW) in the ocean interior vary widely between runs. The fraction of the interior ocean ventilated from the North Atlantic is observed to vary between 12% and 61% throughout the entire range of sensitivity runs, while the corresponding fraction ventilated from the Southern Ocean varies between 68% and 15%. The sensitivity of northern and southern end-member water mass fractions to small changes in surface heat and feshwater fluxes has important implications for interpreting paleoceanographic records. Records of past ocean tracer concentrations are an important source of information in reconstructing the past ocean circulation. For example, it has been argued that increases in the top-to-bottom radiocarbon age in the North Atlantic at the Last Glacial Maximum (LGM) resulted from a weaker North Atlantic overturning circulation, in turn supporting reduced northward heat transport. In contrast, we show that top-to-bottom radiocarbon ages in the North Atlantic can increase by 400-500 years with little change in the strength of the overturning circulation or in the ocean northward heat transport. This is achieved without changing the air-sea equilibration timescale for radiocarbon. Our results highlight the difficulty of estimating past ocean transports from paleoceanographic tracer data. Furthermore, if the real ocean is as sensitive to changes in surface heat and freshwater fluxes as our model suggests, we could expect significant temporal variability in tracer concentrations without drastic changes in the energy or hydrological cycles.