Simulations made with coupled ocean-atmosphere models frequently "drift" away from the observed climate. The situation has improved in recent years (e.g., Boville et al., 1997), but the tuning required to minimize this drift remains a sensitive and unfortunate necessity. Identifying and understanding biases in the simulated air-sea exchange of heat, momentum and buoyancy is an important step towards resolving the climate drift in coupled ocean-atmosphere models.   AMIP offers the opportunity to examine in detail surface fluxes simulated with AGCMs constrained by realistic sea surface temperatures (SST) and sea-ice conditions. This serves as a useful test to examine the readiness of an AGCM to be successfully (minimal drift) coupled to an OGCM.

For many of the AMIP I simulations the implied ocean heat transport in the Southern Hemisphere was equatorward. These unrealistic transports resulted from deficiencies in the simulated cloud-radiative effects (Gleckler et al., 1995). Additionally, systematic biases in the simulated surface heat fluxes were identified by direct comparison with observationally-based estimates (Randall and Gleckler, 1996 and Gleckler and Weare, 1997).

Simulated ocean surface fluxes and their implied meridional transports need to be studied in further detail. The continuing primary objective of this subproject is to identify and explain systematic errors by making use of the best available observations.  The emphasis on seasonal climatologies (and not interannual variability) reflects that: 1) despite prescribed SSTs, AGCMs do not yet adequetly simulate the seasonal cycle of surface heat fluxes (Randall and Gleckler, 1996), and 2) observations are sparse, and consequently it is believed that available composite climatologies are more credible than estimates of interannual variability. However, several recent satellite products show promise in capturing interannual variability during part of the AMIP II period, and may deserve consideration for comparison with the simulations. Regional field studies (e.g., TOGA-COARE intensive flux array) may also prove useful.  The proposed research consists of the following:
 

1. Evaluation of simulated global ocean seasonal climatologies of  a) all terms of the surface energy balance b) surface wind stress c) buoyancy (freshwater flux)
2. Evaluation of the annual mean implied meridional transports of heat, momentum and density in the Atlantic, Indian, Pacific and global oceans
3. Attempt to identify the reasons for existing surface flux and transport biases (e.g., Gleckler et al., 1995).
 

Monthly mean model output needed:

Because of the large uncertainties associated with "state-of-the-art" observational estimates, it is likely that many of them will continue to be compared with models.  Randall and Gleckler (1996) found that despite the large differences between the observational estimates, collectively they have helped to identify several important model biases.  Thus it is expected that a diverse collection of observations will be used to study the AMIP simulations, including a variety of surface and satellite based estimates, reanalysis, and field studies.