Computational Efficiency and Accuracy of Methods for Asynchronously Coupling Atmosphere–Ocean Climate Models. Part I: Testing with a Mean Annual Model

View More View Less
  • 1 National Center for Atmospheric Research, Boulder, CO 80307
© Get Permissions
Full access

Abstract

Using the mean annual, globally-averaged, coupled atmosphere–ocean energy balance model of Harvey and Schneider, the effect on the transient climate response to a step function solar constant increase using a variety of asynchronous coupling methods is investigated. In asychronous coupling, the atmosphere is integrated with fixed ocean temperature for a period τa, and the ocean is integrated for a longer period τ0 with one of several possible atmospheric assumptions. The process is repeated until equilibrium occurs. The assumption of fixed atmospheric temperature during ocean integrations, used in some atmosphere–ocean GCMs, leads to a large energy conservation error and a very slow transient response compared to the synchronously coupled response. Assuming fixed heat fluxes or a fixed atmosphere–ocean temperature difference greatly improves the asychronous transient response, but large errors still remain. Assuming fixed turbulent fluxes subject to a time lag, or extrapolating the trend in the atmosphere–ocean temperature difference with a variable τ0 gives very good globally-averaged results, but the former method is of questionable utility in a nonglobally-averaged model as it involves large compensating errors in individual heat flux terms.

Abstract

Using the mean annual, globally-averaged, coupled atmosphere–ocean energy balance model of Harvey and Schneider, the effect on the transient climate response to a step function solar constant increase using a variety of asynchronous coupling methods is investigated. In asychronous coupling, the atmosphere is integrated with fixed ocean temperature for a period τa, and the ocean is integrated for a longer period τ0 with one of several possible atmospheric assumptions. The process is repeated until equilibrium occurs. The assumption of fixed atmospheric temperature during ocean integrations, used in some atmosphere–ocean GCMs, leads to a large energy conservation error and a very slow transient response compared to the synchronously coupled response. Assuming fixed heat fluxes or a fixed atmosphere–ocean temperature difference greatly improves the asychronous transient response, but large errors still remain. Assuming fixed turbulent fluxes subject to a time lag, or extrapolating the trend in the atmosphere–ocean temperature difference with a variable τ0 gives very good globally-averaged results, but the former method is of questionable utility in a nonglobally-averaged model as it involves large compensating errors in individual heat flux terms.

Save