This section deals with some of the processes and feedbacks in the coupled atmosphere-ocean system; those involving land and the cryosphere are dealt with in Sections 7.4 and 7.5, respectively. Increasingly, natural modes of variability of the atmosphere and the atmosphere-ocean system are also understood to play key roles in the climate system and how it changes, and thus the teleconnections which are believed to be most important for climate change are briefly discussed.

Fundamental issues in the global climate system are the relative amounts of heat and fresh water carried by the oceans and the atmosphere to balance the global heat and water budgets. Because the top-of-the-atmosphere radiation largely determines the distribution of net energy on the Earth, the fluid components of the climate system, the atmosphere and ocean, transport heat and energy from regions of surplus to regions of deficit, with the total heat transport constrained. In the SAR a major problem with atmospheric models was that their surface fluxes with the ocean and the implied oceanic heat transports did not agree with the quite uncertain observational estimates and the biggest problem was clearly with clouds. The result was that all coupled models of the current climate either had a bias in their simulation of the mean state or that a �flux adjustment� was employed to keep the mean state close to that observed. Although some recent coupled ocean-atmosphere models (Boville and Gent, 1998; Gordon et al., 2000) do not have adjustments to the modelled surface heat, water and momentum fluxes (see Chapter 8, Section 8.4.2), the drift in surface climate is small. This implies that the surface air-sea heat and water fluxes calculated by the model are reasonably consistent with the simulated ocean heat and fresh water transports. However, only comparison with flux values of known accuracy can verify the predicted fluxes, and thereby increase confidence in the modelled physical processes.

The role of the oceans in maintaining the global heat and water budgets may be assessed using bulk formulae, residual and direct (hydrographic) methods. Direct measurements of the sensible and latent heat fluxes have led to improved parametrizations (Fairall et al., 1996; Weller and Anderson, 1996; Godfrey et al., 1998), but large 20 to 30% uncertainties amounting to several tens of Wm^-2 remain (DeCosmo et al., 1996; Gleckler and Weare, 1997). Recent global air-sea flux climatologies based on ship data and bulk formulae (da Silva et al., 1994; Josey et al., 1999) exhibit an overall global imbalance; on average the ocean gains heat at a rate of about 30 Wm^-2. This was adjusted by globally scaling the flux estimates (da Silva et al., 1994), but spatially uniform corrections are not appropriate (Josey et al., 1999). While satellites give data over all ocean regions (e.g., Schulz and Jost, 1998; Curry et al., 1999), and precipitation estimates over the oceans are only viable using satellites (Xie and Arkin, 1997), their accuracy is still unknown. Overall, surface evaporation minus precipitation (E�P) estimates are probably accurate to no better than about 30%.

Table 7.1: Ocean heat transport estimates, positive northwards in PetaWatts (1 PetaWatt = 1 PW = 1015 W), from analyses of individual hydrographic sections from pre WOCE sections (Macdonald, 1998), indirect methods (Trenberth et al., 2001), and years 81 to 120 of the HadCM3 (UKMO) coupled model (Gordon et al., 2000) and from the CSM 1.0 (NCAR) (Boville and Gent, 1998). Typical error bars are ±0.3 PW. For the Atlantic the sections are: 55°N (Bacon, 1997), 24°N (Lavin et al., 1998), 14°N (Klein et al., 1995), 11°S (Speer et al., 1996), and at 45°S (Saunders and King,1995). For the Pacific: 47°N (Roemmich and McCallister, 1989), 24°N (Bryden et al., 1991), 10°N (Wijffels et al., 1996). Because of the Indonesian throughflow, South Pacific and Indian Ocean transports make most sense if combined.
Atlantic	Sections	Macdonald	Trenberth		HadCM3	CSM 1.0
55°N	0.28	-	0.29		-	0.63
48°N	-	0.65	0.41		0.54	0.81
24°N	1.27	1.07	1.15		1.14	1.31
14°N	1.22	-	1.18		-	1.27
11°N	-	1.39	1.15		1.12	1.21
11°S	0.60	0.89	0.63		-	0.65
23°S	-	0.33	0.51		0.67	0.61
45°S	0.53	-	0.62		0.64	-
Pacific
47°N	-0.09	-0.08	-0.06		0.17	0.11
24°N	0.76	0.45	0.73		0.50	0.69
10°N	0.70	0.44	0.85		0.68	0.87
Pacific + Indian
32°S	-	-1.34	-1.14		-1.19	-1.13

In the residual method the atmospheric energy transport is subtracted from the top-of-the-atmosphere radiation budget to derive the ocean heat transport. Until fairly recently, the atmospheric poleward transport estimates were too small and led to overestimates of the ocean transports. This has improved with local applications of global numerical atmospheric analyses based on four dimensional data assimilation (Trenberth and Solomon, 1994; Keith, 1995). With reanalysis of atmospheric observations, more reliable atmospheric transports are becoming available (Table 7.1). Sun and Trenberth (1998) show that large interannual changes in poleward ocean heat transports are inferred with El Niño events in the Pacific. The moisture budget estimates of E�P are generally superior to those computed directly from E and P estimates from the model used in the assimilation (e.g., Trenberth and Guillemot, 1998) and are probably to be preferred to bulk flux estimates, given the inaccuracy of bulk estimates of P.

Continues on next page

IPCC Homepage