The connection between global warming and the changes in ocean heat content has long been a subject of discussion in climate science. This was explicitly discussed in Hansen et al, 1997 where they predicted that over the last few decades of the 20th Century, there should have been a significant increase in ocean heat content (OHC). Note that at the time, there had not been any observational estimate of that change (the first was in 2000 (Levitus et al, 2000)), giving yet another example of a successful climate model prediction. At RC, we have tracked the issue multiple times e.g. 2005, 2008 and 2010. Over the last few months, though, there have been a number of new papers on this connection that provide some interesting perspective on the issue which will certainly continue as the CMIP5 models start to get analysed.
The most recent paper was a new study from NCAR out last week that looked into what happens to OHC in models when there are occasional 10 year periods with no trends in global surface temperatures (Meehl et al, 2011).
It is well-known (or at least it should be) that simulations for late 20th C and early 21st Century do not produce monotonically increasing temperatures at the annual or decadal time-scale. For the models used in AR4, the decadal trends expected under estimates of present day forcings are roughly N(0.2,0.14) (i.e. a Gaussian distribution centered on 0.2 ºC/decade with a standard deviation of ~0.14ºC/decade. This implies that one would expect an 8% probability of getting surface temperature trends less than zero in any one decade.
The Meehl et al study looked at the changes in ocean heat content during these occasional decades and compared that to the changes seen in other decades with positive surface trends. What they found was that decades with cooling surface temperatures consistently had higher-than-average increases in ocean heat content. This makes perfect sense if there is internal decadal variability in the fluxes that connect the deeper ocean to the surface ocean (which of course there is). An anomalous downward heat flux reduces the ocean surface temperature (and hence global surface temperature), which generates an anomalous heat flux into the ocean from the atmosphere (because the flux into the ocean is related to the difference between atmospheric and ocean temperature). And this of course increases total OHC.
A related study from the UK Met. Office looked at the relationship between the ocean heat content changes in the top 700m and the total ocean heat content change in models (Palmer et al, 2011). They found that (unsurprisingly) there is more variability in the top 700m than in the whole ocean. This is important to quantify because we have better estimates of the upper ocean OHC change than we do of changes in the whole ocean. Observational studies indicate that the below-700m increases are not negligible – but they are poorly characterised (von Schuckmann et al, 2009). The Palmer study indicates that the uncertainty on the decadal total OHC change is about 0.15 W/m2 if one only knows the OHC change for the top 700m.
So what can we infer about the real world from these tests? First, we can conclude that we are looking at the right quantities. Total OHC changes are a good measure of the overall radiative imbalance. Second, there is likely to be a systematic issue if we only look at the 0-700m change – this is a noisy estimate of the total OHC change. Third, if the forcings are close to what we expect, we should anticipate that the deeper ocean (below 700m) is taking up some of the slack. There are of course shorter term sources of variability that also impact these measures (OHC changes associated with ENSO, solar irradiance variability over the solar cycle) which complicate the situation.
Two further points have come in comment threads recently that are related to this. The first is whether the changes in deep ocean heat content have any direct impact other than damping the surface response to the ongoing radiative imbalance. The deep ocean is really massive and even for the large changes in OHC we are discussing the impact on the deep temperature is small (I would guess less than 0.1 deg C or so). This is unlikely to have much of a direct impact on the deep biosphere. Neither is this heat going to come back out from the deep ocean any time soon (the notion that this heat is the warming that is ‘in the pipeline’ is erroneous). Rather, these measures are important for what they tell us about the TOA radiative imbalance and it is that which is important for future warming.
The second point is related to a posting by Roger Pielke Sr last week, who claimed that the Meehl et al paper ‘torpedoed’ the use of the surface temperature anomaly as a useful metric of global warming. This is odd in a number of respects. First, the surface temperature records are the longest climate records we have from direct measurements and have been independently replicated by multiple independent groups. I’m not aware of anyone who has ever thought that surface temperatures tell us everything there is to know about climate change, but nonetheless in practical terms global warming has for years been defined as the rise in this metric. It is certainly useful to look at the total heat content anomaly (as best as it can be estimated), but the difficulties in assembling such a metric and extending it back in time more than a few decades preclude it from supplanting the surface temperatures in this respect.
Overall, I think these studies show how we can use climate models to their best advantage. By looking at relationships between key quantities – those that can be observed in the real world and those that are important for predictions – we can use the models to interpret what we are measuring in the real world. For these cases the inferences are not particularly surprising, but it is important that they be quantified. Note that the assumption here is akin to acknowledging that since the real world is more complicated than the (imperfect) models, inferences in the real world should at least be shown to work in the models before you confidently apply them to reality.
However, it is the case that none of these studies prove that these effects are happening in the real world – they are merely suggestive of what we might strongly expect.
- S. Levitus, J.I. Antonov, T.P. Boyer, and C. Stephens, "Warming of the World Ocean", Science, vol. 287, pp. 2225-2229, 2000. http://dx.doi.org/10.1126/science.287.5461.2225
- G.A. Meehl, J.M. Arblaster, J.T. Fasullo, A. Hu, and K.E. Trenberth, "Model-based evidence of deep-ocean heat uptake during surface-temperature hiatus periods", Nature Climate Change, vol. 1, pp. 360-364, 2011. http://dx.doi.org/10.1038/nclimate1229
- M.D. Palmer, D.J. McNeall, and N.J. Dunstone, "Importance of the deep ocean for estimating decadal changes in Earth's radiation balance", Geophysical Research Letters, vol. 38, pp. n/a-n/a, 2011. http://dx.doi.org/10.1029/2011GL047835
- K. von Schuckmann, F. Gaillard, and P. Le Traon, "Global hydrographic variability patterns during 2003–2008", Journal of Geophysical Research: Oceans, vol. 114, 2009. http://dx.doi.org/10.1029/2008JC005237