A presentation such as this, going as straight to the point as this one does, will necessarily rely on a fairly extensive ‘back catalogue’ of supporting graphs, figures and clarifications, most of which were generated in response to the likely questions and objections that might arise along the way, in the course of the argument being set up. Seeing how addressing these at every turn and having to substantiate all choices made at each single step would bog down and/or sidetrack the presentation to such an extent that the overall message would ultimately become lost in the noise, they will however have to remain in the background for now, rather kept ready at hand for the next round, so to speak. My point is this: Let the argument, as it stands, be presented first, to its completion – and then bring on the critique.
INTRODUCTION – THE THEORY BEHIND
The ‘AGW (CO2) warming hypothesis’ (really just another name for ‘the general idea of an «enhanced greenhouse effect» causing global warming’) says that, as the total content of CO2 in the atmosphere rises over time, so will global temperatures – in short: «Temps should go up». The scientific method demands that any scientific hypothesis should be able to make predictions like this, statements or claims about the world that can be tested, thus allowing us to either strengthen or weaken our trust in the explanatory power of our hypothesis. However, if there is to be any point in performing such a test, the prediction being tested needs to be relevant, i.e. it should be more or less unique to our particular hypothesis. So is «Temps should go up» a relevant prediction? No. It’s a prediction, but it’s not a relevant one. Because it isn’t specific enough. It isn’t unique to the ‘CO2 warming hypothesis’. It cannot separate between one proposed cause and another. For example, ‘more solar heat being absorbed by the Earth system over time’ would be an alternative explanation of multidecadal global warming to the «enhanced-greenhouse-effect» proposition. Both would predict the world to get warmer. So how do you choose one over the other? You hone in on an observation that would be unique to your favoured explanation. And now you’ve got yourself a relevant prediction to be tested …!
We, after all, want to find the cause behind the observed effect (‘global warming’), not the effect itself – that has already been found. That’s merely our starting point.
So what would be the relevant – the core – prediction of the ‘CO2 warming hypothesis’? In order to find that out, we must look at what ‘the idea of an «enhanced greenhouse effect» causing global warming’ really says. What is the proposed ‘greenhouse warming mechanism’, how does it work, and what is its identifying feature or physical fingerprint, distinguishing it from all other warming mechanisms?
Figure 1: From Held and Soden (2000).
«Adding more greenhouse gas [like CO2] to the atmosphere makes higher, more tenuous, formerly transparent portions of the atmosphere opaque to IR and thus increases the difference between the ground temperature [Ts] and the radiating temperature [Te]. The result, once the system comes into equilibrium, is surface warming.»
Note well what the caption of Fig.1 says: «(…) the effective emission temperature (Te) remains unchanged.» As Earth’s effective emission level (Ze) is forced up into generally higher (and thus colder) air layers, Earth’s average emission to space is reduced (lower temp = lower emission flux), which will have to be compensated for by warming at all altitude-specific layers from surface to tropopause, in order for the outgoing flux to be maintained at the same level as the incoming flux from the Sun. So, across each incremental step of the process, Te ends up at the same value as before (cooling first, then warming back up), while Ts (average global surface temp) and Ttropo (average global tropospheric temp) both end up at a higher value than before (Fig.1). And that is the gist of how the ‘greenhouse warming mechanism’ is supposed to work.
THE RELEVANT ‘CO2’ PREDICTION
Now, the question is: How can we determine whether or not such a process, of gradually (incrementally) raising the planet’s effective emission level as the atmosphere becomes more opaque to outgoing IR, is in fact at work in the Earth system? Enter the one relevant prediction of the ‘CO2 warming hypothesis’, which unambiguously points out what we’re theoretically meant to see, if this process really is the cause of ‘modern global warming’:
- The surface/troposphere temperature (Ts/Ttropo) should be observed to rise consistently over time, while Earth’s effective emission temperature (Te) stays at substantially the same level; ideally like this:
Which would give a ‘difference curve’ (Ts/Ttropo minus Te) looking like this:
A pretty clear signal of systematic divergence over time. So let’s turn to the climate models. They are the instruments – basing their output directly on the ‘greenhouse warming hypothesis’ outlined above – used by ‘Mainstream Climate Science’ to prognosticate about the future, and purportedly to explain the past. So let’s see how these models explain ‘global warming’ over the last few decades, using the so-called ‘CMIP5 model mean’ as our guide, by letting them compare our two variables: Ts/Ttropo, and Te*. (My data source (KNMI Climate Explorer) only provides temp data for the surface (Ts), not the troposphere (Ttropo), which would be more useful, considering how at least 85% of Earth’s final emission flux to space is drawn from the troposphere, and only about 10% from the surface. The two are, however, supposed to rise at equal rates (Fig.1).)
|* Te is found by converting the average OLR (‘Outgoing Long-wave Radiation’) flux anomalies at the ToA, designated ‘rlut’ in Climate Explorer, to the corresponding emission temperature anomalies by appropriately applying the Stefan-Boltzmann equation, in effect multiplying the radiation data by 0.266 (cf. Fig.4 below).|
Here are the results (01/1985 – 04/2020):
Which more or less exactly match the theoretical templates in Figs. 2 and 3 above:
The only significant discrepancy between the model output (Figs. 4 and 5) and the ideal cases (Fig.6) arises from the (cooling) volcanic influence in the former derived from two massive eruptions in 1982 (El Chichón) and 1991 (Pinatubo) respectively, pulling the model curves down somewhat prior to about 2001.
And with that we have reached the point where we are finally ready to put the ‘CO2 warming hypothesis’ to the test, by its one unique prediction.
Let’s first have our basic premise spelled out by the IPCC themselves, from the time, in their infancy, when they still bothered upholding a certain semblance of adherence to the scientific method, to ordinary logic and rational reasoning, in short, to an honest philosophy of always searching for the truth; from their 1990 First Assessment Report (FAR WG1, Section 8.1.2):
«The word ‘detection’ has been used to refer to the identification of a significant change in climate (such as an upward trend in global mean temperature). However, identifying a change in climate is not enough for us to claim that we have detected the enhanced greenhouse effect, even if statistical methods suggest that the change is statistically significant (i.e., extremely unlikely to have occurred by chance). To claim detection in a useful and practical way, we must not only identify a climatic change, but we must attribute at least part of such a change to the enhanced greenhouse effect. It is in this stricter sense that the word ‘detection’ is used here. Detection requires that the observed changes in climate are in accord with detailed model predictions of the enhanced greenhouse effect, demonstrating that we understand the cause or causes of the changes.»
They quickly moved away from this well-advised approach, of course, adopting instead the novel and – it would seem – near-magical detection standard of ‘unprecedented changes’. Apparently it is sufficient, after all, to simply look at the effect (like «an upward trend in global mean temperature») and then determine its cause directly from that, as long as the effect is significant enough, i.e. somehow ‘lacking historical precedent’. The only prediction you need from that point on is: «Temps should go up».
Problem is that this rather convenient change of mind revealed their mission to no longer be a scientific one, because employing a methodology where an effect is basically allowed to explain itself enables you to insert whatever cause or ‘driving mechanism’ that you wish to promote in the conclusion, from the beginning, effectively creating your own circular argument.
Which is why we will rather go back to 1990 to do a proper – real scientific – test of the ‘CO2 warming hypothesis’. And the test is simply this:
How well, if at all, does our theoretical/modelled prediction, highlighted in Figs. 2 & 3 (theoretical) and Figs. 4 & 5 (modelled) above, stack up against reality?
We have consistent, high-quality ToA radiation flux satellite observations from the real Earth system going back to 1985. In other words, we have a total of just over 35 years of reliable data, which should be more than sufficient to lend high confidence to a test like this. So here are the observation-based results – Ttropo (represented by UAHv6 TMT anomaly data) versus Te (represented by the combined ERBS Ed4 + CERES EBAF Ed4.1 series IR flux anomaly data, converted into corresponding emission temp anomalies), 01/1985 – 04/2020:
Bear in mind that this is the equivalent relationship on display, between our two relevant climatic variables, to the one in Fig.2 (theoretical) and the one in Fig.4 (modelled) above.
And here’s the resulting ‘difference curve’ between the two (Ttropo – Te):
What do we see? The radiative temperature variable (Te) appears simply to be tracking the altitude-specific temperature variable (Ttropo) over time (Fig.7). Which is fully in line with the fundamental physical knowledge that increasing temp → increasing emission, but obviously in direct contradiction to what the ‘CO2 warming hypothesis’ – the theory itself and the models based on it – all the same says we should observe as the world gets warmer.
Overall, the difference curve (Fig.8) lies pretty flat (along the 0 axis) from 1985 to 2020; even though there are of course quite significant ups and downs along the way, this only points to the fact that we are after all observing the progression in time of a complex, dynamic natural system, also known as the global climate. Taking this into account, I would say that the fairly tight overall correlation between the two curves in Fig.7, the clear tendency for them to converge over time rather than diverge, is rather impressive to behold. The keen eye might spot how the difference between Ttropo and Te (Fig.8) results in a pattern curiously reminiscent of the ENSO sequence (represented by the NINO3.4 SSTa data). In particular, the major ENSO events between 1985 and 2020 seem to leave their distinct mark on the Ttropo – Te residual curve, strongly suggesting that their well-established influence on the global climate in some way affects the direct relationship between the two variables being compared.
To show how significant this connection really is, I’ve superimposed the Ttropo – Te difference curve on the down-scaled ENSO signal (NINO3.4 SSTa), after delaying the latter by six months in order to visually optimize the fit:
I have added a quartic polynomial to both datasets to emphasize their variation tendencies over time; the graph spans a total of 35 years and 4 months.
(Another natural phenomenon strongly affecting the coincidence between the two curves in Fig.7 is the aforementioned large volcanic eruptions, such as that of Mt. Pinatubo in 1991. Note, I have intentionally removed the disproportionately – and thus rather distractingly – large impact of that incident by deleting both the Ttropo and Te data between mid ’91 and late ’92.)
SUMMARY – CONCLUSION
So how did the central prediction of the ‘CO2 warming hypothesis’ stack up against reality?
Short answer: Not at all!
It doesn’t take much to realise this is the simple truth of the matter. Let’s simply juxtapose below all the relevant diagrams presented above in order to carry the message home:
a) . . . . . . . . . . . . . . . . . . . . . . . . . b) . . . . . . . . . . . . . . . . . . . . . . . . c)
d) . . . . . . . . . . . . . . . . . . . . . . . . . e) . . . . . . . . . . . . . . . . . . . . . . . . f)
Figure 10 a), b), c), d), e) & f).
Total divergence from 1985 to 2020 (35+ years) between the Ts/Ttropo curve and the Te curve:
- MODELS (Fig.b) & e)): +0.7 K (+0.2 K/decade)
- OBSERVATIONS (Fig.c) & f)): ~ 0.0 K (0.0 K/decade)
Below I provide links to four other posts on this blog. Please read them for relevant background information and supporting material to this one.
In these links you will also find the references to the various data sources used …
adjustments AGW hypothesis AMO ASR atmosphere atmospheric mass CERES climate models clouds CMIP5 model mean co2 conduction convection data Earth's energy budget East Pacific energy transfer ENSO First Law of Thermodynamics GISS GISTEMP LOTI global temperatures global warming greenhouse effect hadcrut hadcrut3 HadCRUt4 heat heat transfer insulation internal energy lower troposphere nino3.4 OHC OLR OLR at the ToA radiation radiative forcing radiative heat transfer SSTa Steel Greenhouse step change temperature UAH v6 work