The global ocean heat content anomaly is a particular effective climate metric to assess the climate system radiative imbalance as discussed in
Pielke Sr., R.A., 2003: Heat storage within the Earth system. Bull. Amer. Meteor. Soc., 84, 331-335.
Even Real Climate agrees with this viewpoint as they report in a May 3 2005 weblog Planetary energy imbalance? This weblog from Real Climate states that their key conclusions with respect to this issue are that
“i) model simulations with 20th century forcings are able to match the surface air temperature record, ii) they also match the measured changes of ocean heat content over the last decade, iii) the implied planetary imbalance (the amount of excess energy the Earth is currently absorbing) which is roughly equal to the ocean heat uptake, is significant and growing, and iv) this implies both that there is significant heating’ in the pipeline’, and that there is an important lag in the climate’s full response to changes in the forcing.”
Thus there is agreement on the value of using ocean heat content changes to diagnose global warming over recent decades and the need to predict the change in heat content in coming years. There is the question, however, if the close relation between ocean heat content changes and the global radiative imbalance occurs for yearly time scales, particularly when large events such as an El Niño or La Niña occurs.
Also, in the context of the Real Climate key conclusions, with respect to the questions posed by John Tierney of the New York Times and Roger Pielke Jr. in Prometheus on what tests should be used to test the skill of climate predictions, Real Climate states that the “…implied planetary imbalance”the amount of excess energy the Earth is currently absorbing) which is roughly equal to the ocean heat uptake, is significant and growing..”. Is the ocean heat uptake growing
The specific questions that Bryan is asking are:
- Does ENSO significantly affect the annual OHCA?
- Is the ocean heat content over annual periods too noisy to adequately diagnose any important system processes (ie equilibrated radiative forcings+feedbacks)?
- Are large net heat flux anomalies to space correlated with ENSO?
- Do AOGCMs replicate the magnitude of annual system heat content changes that are being observed? Do they need to do so in order to skillfully predict multi-decadal warming?
As
Bryan subsequently followed up with the insightful comments the following
I have been thinking about a possible way to address Gavin Schmidt’s apparent suggestion that ENSO might significantly alter the global heat budget on an annual to decadal time scale…… To probe such a question further, I would like to ascertain some basic information. Firstly, what is the net radiative flux (in W/m2, then converted to Joules) needed to raise the temperature of the troposphere (entire global integral) 0.2 degrees C ( or whatever the real number is, attributed to the 1998 Super El Nino) in a one year span? Over this same period, when considering the upper 750 m of the ocean, how does the amount of heat needed to warm the atmosphere 0.2 C (due to the 1998 El Nino) compare to the change in the ocean heat content observed for this same time period (measured in Joules) in Lyman et al. (2006)? From this data, are we then able to ascertain an estimate of the fraction of the observed change in the global heat budget that was directly attributable to the El Nino event (since the heat content anomaly in the atmosphere was certainly correlated to the large 1998 El Nino)? If this fraction can be estimated, it seems we could better speculate on the validity of the argument by Peixoto and Oort (1992) that even over annual periods of time, the change in ocean heat content is a proxy for the top of atmosphere radiative imbalance (which you point out is a snapshot in time of the sum of the non-equilibrium forcings+feedbacks).Is my question well posed, and would this be a sound methodology to address this question? (i.e., I understand the error bars on ocean heat content limit the precision of this comparison), but it seems important to understand if the numbers are even of the same magnitude)”
The resultant exchange on Real Climate between Gavin and Bryan, reproduced below, is a very constructive examination of this issue.
Comments on Real Climate
“Gavin, As a follow up to our discussion earlier this week, I have been considering your suggestion that ENSO significantly alters the global heat budget on annual or longer time scales. The magnitude of natural tropical variability (specifically ENSO), and its effect on the earth’s heat budget over annual to decadal periods seems an important issue in climate science. To probe such a question further, I would like to ascertain some basic information. Firstly, what is the net radiative flux (in W/m2, then converted to Joules) needed to raise the temperature of the troposphere (entire global integral) 0.2 degrees C(or whatever is the best number attributed to the 1998 Super El Nino) over the relevant one year span? Over this same period, when considering the upper 750 m of the ocean, how does this quantity of atmospheric heating compare to the change in the ocean heat content observed for this same time period (measured in Joules or converted to W/m2) in Lyman et al. 2006 (Figure 1)? From this data, are we then able to estimate the fraction of the observed change in the global heat budget that was directly attributable to the El Nino event (since the heat content anomaly in the atmosphere was certainly correlated to the large 1998 El Nino)? Once this fraction is determined, it seems we could better speculate on the validity of the argument by Ellis et al. (1978) that even over annual periods of time, the change in ocean heat content is a proxy for the top of atmosphere radiative imbalance, which Roger Pielke Sr. asserts approximately equals the mean non-equilibrium radiative forcing+feedbacks). I have e-mailed Roger, and also asked him to comment on this issue, which he indicated he would do on his weblog next week. I hope my question is well posed, and this would be a sound methodology to address this question? (i.e., I understand the error bars on ocean heat content limit the precision of this comparison), but it seems important to understand if the numbers are even of the same order of magnitude)
[Response: Update. Both ‘Energy into Ground’ and ‘Net radiation TOA’ are +ve down. Time series (including zonal means) are found here: “http://data.giss.nasa.gov/modelE/transient/Rc_jt.2.01.html“>I did a preliminary analysis from the sources I linked to. From early 1997 to mid 1998, the simulated atmosphere accumulated anomalous heat at about 0.2 W/m2 on average due to the El Nino event. This corresponded to an increase of surface air temperature (in these simulations) of around 0.3 to 0.4 deg C. Over that same time, the model ocean was losing heat at approximately 0.7 W/m2. The heat out at the TOA was thus ~0.5 W/m2. Now whether this is a good approximation to reality is unclear - my sense is that it is going to be sensitive to the details of the cloud feedbacks in the tropics, but assuming for the time being that this is reasonable, it implies that the interannual variability in OHC is certainly on the same order as the long term TOA imbalance, and that the factor between TOA imbalance and OHC might vary between 0.8 and 1.2 on interannual timescales. One would therefore expect to see a plateau or even a dip in OHC growth through this period in the real world. These kinds of model simulation do have a couple of disadvantages though that might be apropos. In mid-latitudes where SST variability is driven mainly by the atmosphere, starting off with the SST makes the fluxes go the wrong way. You could alleviate this by doing the analysis in the tropics alone - might even be easier to compare with data on clouds or rainfall too. But this is probably ok for a first cut. Comment by gavin - 19 January 2008 @ 9:36 PM”
The answers to the excellent questions posed by Bryan are
1. Does ENSO significantly affect the annual OHCA?
Yes. This, of course, should be expected as the ocean to the atmosphere fluxes of heat (both sensible and latent) are enhanced when the tropical ocean sea surface temperatures are above average over large regions. The result is both higher tropospheric temperatures and absolute humidity.
2. Is the ocean heat content over annual periods too noisy to adequately diagnose any important system processes (ie equalibriated radiative forcings+feedbacks)?
No. With the introduction of the Argo monitoring network, in particularly, as well as global monitoring of sea level and sea surface temperatures, quite accurate documentation of the upper OHCA is sufficiently accurate to diagnose the global average radiative imbalance on this time scale.
3. Are large net heat flux anomalies to space correlated with ENSO?
This is a really interesting (and important) research question. The rapid cooling after the 1997-98 El Niño (e.g., “see the RSS LT MSU data in their Figure 7″) suggests a large loss of heat to space (unless it was transfered deeper into the ocean). Gavin’s preliminary analysis adds to our understanding of this issue.
4. Do AOGCMs replicate the magnitude of annual system heat content changes that are being observed?
Excellent question. This should be a very visible result of all of the models. Moreover, the AOGCMS should make predictions for the coming years, in the manner suggested on Prometheus (see the several separate weblogs posted on this subject in January 2008) and the New York Times by John Tierney (see).
5. Do they need to do so in order to skillfully predict multi-decadal warming.
Certainly! The multi-decadal AOGCM predictions must predict the changes of Joules within the climate system if they are going to be accepted as skillful prediction tools. Since ENSOs are clearly involved with global average heat changes, they must be accurately simulated. As repeatedly emphasized on Climate Science, the assessment of AOGCM prediction skill of climate system heat changes should focus on the accumulation or loss of Joules of heat, not a global average surface temperature trend which has a multitude of serious issues (see; this paper will also be further weblogged on in the coming two weeks). The OHCA anomalies provide the best approach to diagnosis the actual climate system heat changes.