Ocean Thermal Expansion and Heat Uptake in Climate Change Experiments

Jonathan Gregory
Hadley Centre
jmgregory@meto.gov.uk

Some of the experiments submitted to the IPCC DDC have included or will include information about global-average thermal expansion. Since these experiments have been run with similar scenarios (historical CO2 to 1990, then IS92a or 1%), these results will be comparable, and that is useful. However, it is not sufficient to quantify the uncertainty in thermal expansion, because it doesn't separate the effects of differing climate sensitivity and heat uptake processes.

To get some information about this, I'd like to have a look at some CMIP data, since some of the models concerned are the same, and from the range of models one could get an idea about the systematic uncertainty. Analyses which would be useful are:

• From the 3D temperature and salinity fields, I would calculate the global heat uptake and the global average thermal expansion (assuming a common equation of state). This would show how consistent an indicator the thermal expansion is of the heat uptake. They are not necessarily proportional because of the effect of T and S on thermal expansivity. However, I suspect that in practice they may be related in a similar way for different models.
• Using the net heat flux into the ocean, which will probably be similar to the net heat flux at the TOA for 20-year means (since the heat capacity of land and atmosphere is small) and assuming a value for the radiative forcing of 2*CO2, I would compare the relative influences of climate feedback and heat uptake on determining the rate of climate change, as John and I did for HADCM2 in Geophys Res Lett 24 (1997) 1943-1946, where we compared the time- development of climate change in 1% experiments using HADCM2 with and without flux adjustment. We found then that
  • The heat flux into the ocean was roughly proportional to the temperature change, so could be expressed as a linear term, like the climate feedback. (This would only work for a linearly increasing forcing, I expect.)
  • Both this coupling and the climate feedback were different in the two models. Moreover, the differences were of equal importance in bringing about the different rates of climate change.

It is the final point which bears upon the question of how the rate of sea- level rise will be related to the rate of temperature rise, and it would be interesting to know whether a similar analysis works for other CMIP models.