CHIME Home Page

Alex Megann, Bablu Sinha and Adrian New, SOC

The majority of coupled climate models developed to date use an ocean component with a cartesian vertical coordinate; that is, the model fields at each water column are defined at a set of constant depth levels. This allows the ocean to have arbitrarily high resolution near the surface, but generally leads to excessive diapycnal mixing, especially in sill overflows of dense bottom water. This may have adverse consequences in long-term climate simulations, in particular on the stability of the meridional overturning circulation. Isopycnic models, which use potential density as their vertical coordinate, preserve the structure of water masses faithfully over long times but have poor vertical resolution in weakly stratified regions. The recently introduced hybrid-coordinate models, in which interior isopycnal layers transition to constant-depth levels in near-surface waters, should in principle combine the advantages of both model types without the weaknesses of either.

The Coupled Hadley-Isopycnic Model Experiment (CHIME) comprises a new coupled climate model, developed at Southampton Oceanography Centre. It consists of the atmospheric model used in the Hadley Centre's HadCM3 climate model, coupled to a hybrid coordinate ocean model ( HYCOM) via the OASIS coupler. The ocean model grid is identical over most of the globe to that used in HadCM3, allowing the effect of the different vertical representation of the ocean to be easily assessed. A further innovative feature of CHIME's ocean component is its use of a bipolar grid over the Arctic (see Figure 1). This avoids the problem of convergence of meridians toward the North Pole by replacing the latitude-longitude grid north of a given latitude circle (here 55¡N) with a bipolar grid. This matches the spherical grid perfectly at this latitude, but its two poles may be placed harmlessly in large land masses.

Development of the CHIME model was carried out partly under the NERC COAPEC programme. The analysis of the model has two main strands:

The preliminary 120-year integration of CHIME is now complete. This run shows that the climate drifts in CHIME are acceptably small, and that it is a useful research tool for investigating the natural variability of the climate system. The sea-surface temperature produced by CHIME is also realistic (see above). Furthermore, CHIME is showing intriguing differences with the HadCM3 model (mainly related to different positions of the major ocean current systems), which may lead to a better understanding of, and improvements in, both models.

Figure 2 shows the sea surface elevation in the North Atlantic in March of year 100 of the model run. The North Atlantic Current (NAC) can be clearly seen, approximately following the 90cm height contour from the Gulf Stream separation off the eastern seaboard of the US, across the north Atlantic, and finally passing into the Norwegian Sea between Iceland and Scotland. In CHIME the path of the NAC is remarkably stable, and in fact is more realistic than in HadCM3, where the NAC drifts southward during the first decade of the run.

Figure 3 shows the mixed layer depth, again in March of year 100. Wintertime deep mixing can be seen in the Labrador Sea and south of Iceland, which contributes to the formation of North Atlantic Bottom Water. Further south, a tongue of mixing down to 300-400 metres is visible across the Subtropical Gyre; this produces the characteristic mode waters in the gyre.

Comparisons with HadCM3

CHIME current status

We are currently porting CHIME to the Bull supercomputing cluster at NOCS, and the ocean component is being upgraded to the latest version of HYCOM, V2.1.34, which is more flexible and reliable.

Future plans

The control run will be repeated with the new model, and will be extended to at least 200 years.

We have obtained funding under the Rapid UK THC Intercomparison Project to carry out greenhouse sensitivity experiments with CHIME, and Adam Blaker will be employed in this capacity from 5 June 2006.

Last Updated: Alex Megann 25 May 2006

Top

Prev Next