Reply by Ray Bates to the blog post by Peter Thorne of Maynooth University
A) General comments
1) Prof. Thorne states that my critique of SR1.5 was not peer-reviewed and should not be referred to as a paper.
His statement is incorrect. My critique was peer-reviewed. I wouldn’t list it in my CV as a journal article, but it is correct to call it a paper (see the Oxford Dictionary). That said, it matters little to me whether my publication is called a critique, a piece, or a paper.
2) Press freedom and right of reply
Prof. Thorne states, in relation to critical comments of his published in the Irish Times of 21/12/2018 regarding my SR1.5 critique: “To be crystal clear, a free press is an essential component of a healthy, vibrant democracy and it would be strange for the media to completely censure views.” I find it very comforting to hear Prof. Thorne express this viewpoint in such a clear manner. I would request him (Prof Thorne) to note, however, that I have not been accorded a corresponding right of reply in a number of instances involving my name in the Irish media. It takes the website of what Prof. Thorne describes as a ‘highly questionable think tank’ to provide me with the opportunity to point this out.
3) Dynamic meteorologists cannot be counted as climate scientists.
In reply to Prof. Thorne’s assertion that dynamic meteorologists cannot be counted as climate scientists, it will suffice to look at an example. The first assessment report addressed to policymakers warning of the risks associated with increasing carbon dioxide was “Carbon Dioxide and Climate: a Scientific Assessment” (US National Academy of Sciences, 1979). That report, known as the ‘Charney Report’, had nine authors. Five of these (including Charney, its chairman) were dynamic meteorologists. Is Prof. Thorne suggesting that the US National Academy of Sciences did not know what it was doing when it selected this committee?
B) Reply to comments by Prof. Thorne on the scientific content of my paper
1) Prof Thorne accuses me of “wilful misinterpretation of AR5 attribution findings”
Apart from its being invalid, I find the tone of this accusation totally unprofessional. Nevertheless, in the interest of setting the record straight, I will answer it.
The section of my paper referred to here by Prof. Thorne is Section 2, “Departure from the IPCC’s Fifth Assessment Report”, which contains the following paragraph:
“The central attribution statement of Working Group I in the Fifth Assessment was as follows:
It is extremely likely that more than half of the observed increase in global average surface temperature from 1951 to 2010 was caused by the anthropogenic increase in greenhouse gas concentrations and other anthropogenic forcings together.
This statement did not necessarily attribute all the observed post-1950 warming to anthropogenic effects, nor did it attribute the substantial early 20th century warming (1910–1945) to such effects. In contrast to this caution, SR1.5 portrays all the global warming observed since the late 19th century as being human-induced (see Figure 1). This major departure from the Fifth Assessment is presented without any rigorous justification.”
The indented sentence in the above paragraph is a quote from Section D.3 (page 17) of the AR5 SPM. In my paper, I omitted the sentence following it, which reads:
“The best estimate of the human-induced contribution to warming is similar to the observed warming over this period.”
Prof Thorne asserts that my omission of this second sentence was “wilful misrepresentation”.
In reply to the above, I first wish to point out that Prof. Thorne misrepresents me by failing to make any mention of my comment on the substantial early 20th century warming (1910–1945). In portraying a substantial part of this warming as human-induced, SR1.5 clearly in a major way went beyond the attribution statement in the IPCC’s Fifth Assessment1, without giving any rigorous justification for doing so.
With regard to the omitted sentence, one of my reasons for omitting it is that I regard it as being in conflict with the sentence preceding it (which I did quote): on the one hand, the observed post-1950 warming is extremely likely to have been more than half due to anthropogenic effects, on the other hand it is most likely to have been all due to anthropogenic effects!
A further reason to omit this sentence in the present context is that the best estimate of the anthropogenic warming provided in SR1.5 is based on a recent detection and attribution study (Haustein et al. 2017) whose conclusions are open to debate. The study states that it uses a classical, multiple regression based, method that scales modelled responses (here global mean temperature from a simple model) to different types of forcing to give a best match with observations. A GCM-based spatio-temporal version of this method was used for the AR5 attribution findings. Such studies will necessarily attribute almost all of the observed trend to a weighted sum of trends in their explanatory variables. In the study in question these are modelled responses to anthropogenic forcing and to natural forcing. The sum of natural solar and volcanic forcings had very little trend over the analysis period (AR5 WG1 Figures 10.1 and 10.4; Haustein et al. 2017, Figure 1). Internal variability was not treated as an explanatory variable. Hence, the study will necessarily find anthropogenic forcing to be the cause of essentially all the warming, even if some of it was actually caused by multi-decadal or multi-centennial internal variability.
Haustein, K. et al., 2017: A real-time Global Warming Index. Scientific Reports, 7(1), 15417, doi:10.1038/s41598-017-14828-5.
2) Appropriate use of temperature period and series
Prof. Thorne states: “Ray Bates goes on in his piece to imply nefarious intent behind the IPCC considering only the post-1960 series of GMST in their SPM figure. The choice is reasonable because the attribution statement in AR5 (and the prior figure) pertained to post-1950 changes. Furthermore, in the underlying chapter 1 Figure 1.2 the full series from 1850 is shown. The SR1.5 is hardly ignoring the early period as implied is it?”
In its Summary for Policymakers (SPM), SR1.5 did ignore pre-1960 temperature trends, as I stated. It is no defence of an omission in the SPM to point to a chart in the underlying report, the exclusion of which from the SPM is the whole matter at issue.
In Figure SPM.1, why was the obvious choice not made to show the post-1950 (corresponding to the period of the AR5 attribution statement) rather than the post-1960 changes? Prof. Thorne provides no explanation. Was the purpose to make it difficult for policymakers to see that cooling trends could occur while CO2 was increasing, as in the period 1950-1960?
3) The neglect of satellite temperature data in SR1.5
I pointed out in my paper that SR1.5 did not discuss satellite-observed temperature trends and expressed the view that this was a serious defect. Prof. Thorne replied that “satellite data are [or ‘aint’] no gold standard” and asked “What’s the issue?”.
If Prof. Thorne believes climate science should not use satellite temperature data because the temperature-measuring instrument had to be launched into space on a rocket, then he needs to advocate against a wide range of essential data products; for example satellite altimeter data on sea level rise, which was extensively used in AR5. But he seems not to mind all the other data series, disparaging only one in particular, which makes his reasoning inconsistent.
The issue at stake in relation to SR1.5’s neglect of satellite temperature measurements, a topic on which Prof Thorne states he is “uniquely qualified” to comment, is clearly shown in the figure below (courtesy of Dr. John Christy). This figure shows that in the period 1979-2017 the average of three satellite datasets for the tropical mid-tropospheric temperature agrees closely with both the average of three balloon datasets and the average of three reanalyses. This is sufficient evidence as far as I am concerned to show that the satellite datasets in question cannot be brushed aside. It is very noteworthy that all three of the observational datasets, while agreeing well with each other, disagree strongly over the past two decades with the average warming trends projected by 102 CMIP5 models.
It is important to note here that the moist adiabatic lapse rate demands that surface warming in the tropics be accompanied by larger warming in the troposphere above.
In view of the above considerations, I adhere to my assertion that the neglect of satellite temperature data in SR1.5 was a serious defect.
4) Ocean ‘red herrings’
The ‘cold ocean warm land’ (COWL) pattern is a well-known feature of the observed global warming and one whose nature and extent are a subject of ongoing research. One aspect of this phenomenon that is not understood is the question of why there is such a difference between the period 1900-1980, when the land and sea-surface temperatures (SSTs) rose and fell at the same rate over multi-decadal periods, and the period from 1980 onwards, when the pattern of unequal warming became pronounced, with the land temperatures rising much faster than the SSTs. In the section “Observations of ocean warming” of my paper I referred to this issue and pointed out that the average SST for the period 2000–2014 (before the onset of the recent El Niño) was only 0.36°C warmer than the average for the period 1936–1950. Prof. Thorne refers to the above considerations as being a “red herring”. I hold that the above matters are important and reject Prof. Thorne’s attempt to dismiss them.
Prof. Thorne then goes on to dismiss my presentation of results from the ECMWF CERA-20C reanalysis as being another “red herring”. This ECMWF reanalysis may be experimental, but it used vast quantities of ocean data that had not previously been assimilated and a data assimilation system that is superior to anything that went before it. It is reasonable to assume that the results it provides for the ocean heat content OHC (0-300m) from the middle of the last century onwards are the best indication available of the value of this quantity. The fact that the results indicate the OHC (0-300m) around the middle of the last century to have been greater than recent values cannot easily be dismissed. Furthermore, it is consistent with the findings of Soon et al. (2015; referred to in my paper) that land temperatures over rural areas in the middle of the last century were comparable to what they are now.
Prof. Thorne seems to put much more faith in climate model projections than in observations (see his discussion of Figure 12.11 from AR5), paying little heed to the limitations of these models that are now openly acknowledged by some modellers.
Continuing with his section headed “Ocean red herrings”, Prof. Thorne states: “Variations of the degree shown would lead to changes in sea level that simply are not present in the tide gauge records around the world.”
The ECMWF OHC results presented in my paper refer to the upper 300m. This is less than a tenth of the average ocean depth. Sea level rise due to thermal expansion involves the entire ocean depth and also involves salinity variations, which are not included in OHC (0-300m). There are four other contributors to sea level rise in addition to thermal expansion listed in the table below (taken from AR5 WG1, page 1151).
Table of contributions to sea level rise from different sources (mm per year)
Thermal expansion 1.1
Glaciers and ice caps 0.76
Greenland ice sheet 0.33
Antarctic ice sheet 0.27
Land water storage 0.38
Observed sea level rise 3.2
In view of the above, it is clear that no conclusions about consistency or inconsistency with observed sea level rise can be drawn from OHC (0-300m).
5) Climate model tuning
Prof. Thorne states that my section on Climate Model Tuning is very confused, but does not reply to the papers of Hourdin et al. (2017) and Voosen (2016) cited there. These state that GCM tuning has a major influence on GCM response to GHG increase and provide examples. Why not respond to these points and to the authors’ call for documentation of model tuning in the IPCC reports, thereby avoiding their charge of “lack of transparency”?
In addition to the above, Zhao et al (2016) have shown that by tuning the new GFDL GCM – varying the form of convective precipitation parameterisation – they could change its climate sensitivity by a factor of almost two, without any clear observational constraint that they could find favouring one version of the model over the others .
Zhao, M. et al. (2016). Uncertainty in Model Climate Sensitivity Traced to Representations of Cumulus Precipitation Microphysics. J. Climate, 29, 543-560. DOI: 10.1175/JCLI-D-15-0191.1
6) Equilibrium Climate Sensitivity
The subsection of my paper dealing with equilibrium climate sensitivity (ECS) was entitled “Some recent independent estimates of climate sensitivity”. It summarized the AR5 position on ECS in its second paragraph. Prof. Thorne describes this subsection as “to put it politely, selective”. Yes, the subsection was selective and was not put forward as being anything other.
It presents results from four papers that were not referred to in SR1.5; Lindzen and Choi (2011), Mauritsen and Stevens (2015), Bates (2016) and Park and Choi (2017). These papers discuss results on the Earth’s tropical radiative response coefficient (TRRC) from NASA’s CERES satellite instrument, which has been accurately measuring incoming and outgoing radiation for the past 20 years. These measurements show that CMIP5 model-simulated TRRCs are in serious disagreement with observation. When inserted into two-zone tropical/extratropical energy balance models the measurements provide an estimate of ECS that lies below the lower limit of the AR5 1.5 – 4.5°C range. These results provide an easily understood reason why the GCMs may be overestimating ECS.
Prof. Thorne’s comments do nothing to invalidate my critique of SR1.5. My conclusion stands that SR1.5 is not a scientifically rigorous document.
- AR5 concluded (Section 10.3.1.1.3) regarding the early 20th century warming: “It remains difficult to quantify the contribution to this warming from internal variability, natural forcing and anthropogenic forcing, due to forcing and response uncertainties and incomplete observational coverage.”