Topic: A link between eddies & climate drift?
Info associated with a Journal of Climate paper (Delworth et al., 2011, in press) and a poster presented at the October 2011 WCRP Open Science Conference (Dixon et al, 2011).
To some degree, all free-running global coupled climate models experience what is commonly referred to as ?climate drift?. Stouffer and Dixon (1998) defined climate drift as an unforced trend away from some initial state, with the trend not being part of normally occurring variability about a constant mean state. We find that the ocean temperature drift simulated early in each of the CM2.1, CM2.5, and CM2.6 control experiments follows a similar pattern. The global mean ocean drift in all three models is characterized by a cool bias appearing in the upper 200m and a warm bias developing between depths of 500 and 900m. The sub-surface warming maxima occur in the subtropical gyres. However, the rates of drift are not the same in all three models. Both the surface cooling and subsurface warming are greater in CM2.5 than in either CM2.1 or CM2.6.
A hypothesis we are exploring in relation to the climate drift exhibited in these models is that, after initialization, wind-driven subduction in the subtropical gyres deepens the thermocline, leading to subsurface warming and enhanced horizontal temperature gradients at depth. The Ekman pumping-induced warming continues until other processes are strong enough to balance it. We suspect that lateral heat transport by meso-scale eddies is a key part of this balance (see schematic below). Subduction-enhanced horizontal temperature gradients around the deepened gyres should enhance meso-scale activity in models whether eddies are explicitly resolved or their effects are parameterized. However, if a model lacks sufficient lateral eddy heat transport, it follows that the thermocline would continue to deepen, implying a prolonged movement of heat from the near-surface to the interior. Such a process is consistent with the climate drift pattern of surface cooling, and sub surface warming with maxima occurring under the subtropical gyres.
This hypothesis also is consistent with the drift being largest in CM2.5 – a model that does not fully resolve eddies and which has no meso-scale parameterization. Less drift is seen in the eddy-resolving CM2.6 model and in CM2.1 (which uses a variant of the G-M parameterization of eddy effects [Gent & McWilliams, 1990; Gent, 2011]). TO investigate further, we have run an additional CM2.1 experiment without G-M and ofund that it exhibits more than twice the rate of drift of the standard CM2.1 run – also consistent with the hypothesis. More tests will explore the extent to which this mechanism leads to climate drift.
- Gent, P.R., and J.C. McWilliams, 1990: Isopycnal mixing in ocean circulation models. J. Phys. Oceanogr., 20 , pp. 150?155. [LINK]
- Gent, P.R., 2011: The Gent-McWilliams parameterization: 20/20 hindsight. Ocean Modelling, 39, pp 2-9. [LINK]
- Stouffer, R.J., Dixon, K.W., 1998. Initialization of coupled models for use in climate studies: a review. In Research Activities in
Atmospheric and Oceanic Modelling, Report No. 27, WMO/TD-No. 865. World Meteorological Organization, Geneva, Switzerland, pp. I.1?I.8.