Models are failed hypotheses!

John Christy has written a report for the Global Warming Policy Foundation called climate models have been predicting too much warming. The basic conclusion of the report is that climate models predict far more warming in the tropical troposphere than is observed; essentially it’s suggesting that there is a missing tropical tropospheric hot spot.

One thing that the report doesn’t make clear is that amplified tropical tropospheric warming is not simply a signature of greenhouse gas-driven warming; it’s expected for any kind of warming. I think it’s also worth reading this Climate Dialogue discussion, in particular that by Steve Sherwood, who pointed out that

[w]eaker upper-tropospheric warming and hence weaker water-vapour feedback actually implies, on average, slightly stronger overall positive feedback due to lapse rate and water vapour combined

There are also papers that do indicate amplified warming in the tropical troposphere. I disussed one in this post and another is discussed in this article.

However, what I wanted to highlight was the end of the report, which discusses what one might conclude from the comparison between the observations and the models and which suggests that one option is that

  • [t]he models are failed hypotheses.

The report then finishes with

I predict that the ‘failed hypothesis’ option will not be chosen. Unfortunately, that’s exactly what you should do when you follow the scientific method.

The reason that it would not be chosen is because it doesn’t really make sense. Unless a model is so simple that it really only incorporates one bit of physics, it isn’t a hypothesis. Models are typically a combination of multiple bits of physics that, together, allow one to investigate a complex system. Typically, the underlying physics is so well tested that even if the model doesn’t match the observations, one wouldn’t conclude that some of the underlying physics had been falsified.

Of course, a model can be wrong, but that’s not the same as it being a falsified hypothesis. The problem, though, is that all models are wrong, so how does one decide if it’s so wrong that it has no use whatsoever? Even though it might appear that the models are predicting more tropospheric warming than is observed, there are many areas where models have been shown to be skillful.

Determining tropospheric temperature trends is also very tricky, and these are observations that have been corrected on a number of occasions. One really can’t rule out that the mismatch is a problem with the observations, rather than with the models.

As George Box said all models are wrong, but some are useful. Climate models are very useful for trying to infer what might happen if we continue to dump greenhouse gases into the atmosphere. This is especially true if we are interested in trying to understand how it will likely respond on various time, and spatial, scales. If we throw out complex climate models, we’d simply become less well informed. This is not, however, to suggest that we shouldn’t be open about their limitations.

However, our basic understanding of climate change is not really dependent on complex climate models. Our understanding of how our climate responds to radiative perturbations is based on many different lines of evidence. We have models of a variety of complexities. We have recent observations of surface warming, ocean warming, sea level rise, ice sheet mass loss, and many other indicators. We can study changes to our climate that have occured in the past to also understand how it responds to external perturbations. It’s bizarre to suggest that a potential mismatch between complex climate models and observations of one region of the climate system indicates that these models are failed hypotheses.

Even though I’ve been commenting on this kind of thing for a reasonable amount of time now, I still find myself surprised that supposedly serious people will present such simplistic arguments. It’s almost as if they’re looking for something that suits their preferred narrartive, rather than actually trying to help improve our understanding of how our climate is likely to respond to increasing atmospheric CO2 concentrations. Surely not, though?

The missing tropical hot spot – Climate Dialogue discussion about the supposedly missing tropical hot spot.
Tropospheric hot spot – a post of mine about a paper that finds amplified warming in the tropical troposphere.
One satellite data set is underestimating global warming – John Abraham article about another paper that finds amplified warming in the tropical troposphere.
Climate change is real and important – article a group of us wrote that also discusses why climate models are actually skillful.
More errors identified in contrarian climate scientists’ temperature estimates – John Abraham article highlighting the occasions on which John Christy’s satellite temperature dataset has had to be adjusted.
Climate model projections compared to observations – Realclimate post showing various model-observation comparisons.

This entry was posted in Climate change, Climate sensitivity, ClimateBall, Global warming, Greenhouse effect and tagged , , , , , , , , . Bookmark the permalink.

61 Responses to Models are failed hypotheses!

  1. David B. Benson says:

    Why are we paying any attention to anything that John Christy writes?

  2. David,
    I did wonder that myself. However, I thought it was a useful way to highlight why it’s ridiculous to suggest something like “models are failed hypotheses”.

  3. dikranmarsupial says:

    There is no lower bound on what GWPF is prepared to publish. Models are a means for testing hypotheses, however hypotheses can only be falsified to the degree that both parties agree the model is reliable/meaningful.

    In the climate dialogue discussion, my alter-ego (Gavin Cawley) tried to explain to him why the statistical test used in Douglass et al was wrong, but he flatly refused to engage with a simple thought experiment that illustrates the error, and he demonstrated a very naive “statistics cookbook” approach to statistics (that ignored the meaning of the statistics) that suggests he should be listening to statisticians, rather than telling them how statistics should be done.

  4. lerpo says:

    Risk is only increased if models can’t accurately predict the response of any particular region of the climate system. It makes planning for adaptation that much more difficult and costly.

  5. Joshua says:

    Surely not, though?

  6. Dave_Geologist says:

    UAH dataset temperature model shows that [t]he models are failed hypotheses.

    Oh well, rubbish-in, rubbish-out.

  7. JCH says:

    So, according to the great John Christy, ECS is above 3 ℃?

  8. David writes:
    “Why are we paying any attention to anything that John Christy writes?”

    With Gell-Mann’s death yesterday, it’s useful to to recall the so-called Gell-Mann amnesia effect”

    The tropical hot-spot is reminiscent of this effect in the sense that the skeptics have to believe that the the fundamental model behind the warming should be given credence in the first place. In other words, the skeptics have enough confidence in the scientists’ expertise to consider the fundamental effect is important and then go “Hah! told you so!” when some aspect of the model does not pan out quite as expected.

    Thus the amnesia that they are showing is a fundamental belief in the basic model that they think they are attacking. This happens all the time with skeptics. It invariably happens when some sophisticated computer-savvy skeptic complains about the inability of models to do anything useful, while they are oblivious to the fact that the integrated circuitry on the computer chip they are writing their screeds on — “I don’t believe in models!” — is the result of CAD model verification costing millions of dollars.

  9. Mitch says:

    Focusing on one problem spot in a complex model as a test is a standard confusianist technique, a focus on the tree rather than the forest. I find it interesting that the newer models that get ENSO and eddies right tend to be running warmer, around 5°C for an ECS.

  10. Dave_Geologist says:

    Oooh, the delicious irony. The Michael Crichton who coined the phrase wouldn’t be Michael Crichton, the medical doctor and show-biz icon? The one who wrote a novel about climate science being fake, and was given at least one ill-advised non-fiction award for it (AAPG jumped the shark on that one, but it worked out in the end because they encountered so much anger from members, and so much ridicule in the scientific community, that they moved from denial to lukewarmism shortly afterwards)? Who jumped onto the hockey-stick brouhaha and testified to Congress on the denier side? Who believed Mann et al. was wrong, in defiance of the conclusions of actual climate scientists and physicists of the National Academies and other actual authorities on the actual subject? Based on his expertise, decades earlier, as a medical doctor not a climate scientist, physicist, statistician or tree biologist? I presume he’s recanted now that the hockey stick has been validated and replicated dozens of times? Where can I find his public apology to Mann and his team? I hope he asked Congress to correct the record. It would be embarrassing for the guy who invented the phrase to have such an egregious example of the sin sullying his own reputation.

    What, it is the same Michael Crichton, and he can’t see the plank in his own eye? Words fail me!

  11. Dave,
    I’m not getting the context of your comment. What phrase did Michael Crichton coin?

  12. Joshua says:

    Speaking of Gell-Mann:

  13. paulski0 says:

    I’ve noticed recently among much of the “we accept the basic science” crowd that there is a major misconception about how the greenhouse effect works, with a belief that it’s about direct warming of the mid-troposphere due to local CO2 IR absorption which then warms the surface (presumably by convection would be the idea). Although not new as a misconception this appears to have become popular because it ties CO2 influence directly with the mid-tropospheric hotspot. And therefore allows people to say “no hotspot, no (or little) CO2 influence”.

    Of course, this is wrong and doesn’t make any physical sense. And yet there are lots of people who claim to “accept the greenhouse effect” who accept this version of the greenhouse effect rather than real one. I can’t say whether Christy has fallen prey to this misconception, but claiming the hotspot as ‘effectively a diagnostic signal of greenhouse warming’ is suggestive.

    I do believe we are now at a point where we can say that the models are probably getting something significant wrong with regards mid-upper tropospheric warming, although that’s perhaps not much of a statement (“all models are wrong” etc.). For example, I was looking at spatial tropical tropospheric amplification due to ENSO variations in models and observations and found some curious and substantial discrepancies. Specifically I was interested in how the equatorial ENSO signal spread over the tropical Pacific, and made a difference plot between a tropical Nino3.4 area (20S-20N, 195-240E) and standard Nino3.4 in TLT records (12-month running means, RSS4 is similar to RSS3.3). Interestingly, I found that El Nino warming over the full tropical Nino3.4 region was stronger than that directly over the equatorial region, and seemed to be highly non-linear with the 1983 and 1998 super El Ninos clearly standing above everything. Also, that effect was much stronger in UAHv6, which I understand has a sampling shifted towards higher altitudes.

    When I checked how this compared with models I got this result. This is from an AMIP simulation, which has prescribed SSTs from observations. In the models the greatest El Nino warming is conversely directly above the equatorial Nino3.4 region and the ENSO signal amplitude shows much more of a linear transposition of Nino3.4 SSTs.

    It seems possible to me that models have underestimated the influence of a negative ENSO trend (as has happened since 1979) on mid-tropospheric temperatures.

  14. Dave_Geologist says:

    Crichton coined the Gell-Mann amnesia effect, according to Wiki.

    Although I confess I did realise after I’d written it that it wasn’t a perfect match. The arrogance and hypocrisy of Crichton thinking he was right and Mann was wrong is consistent with the attitude ascribed to sufferers by Paul. Although Paul is extending it beyond the Wiki definition to include more than just newspapers, as in “why do you distrust some numerical models but stake your life or career on others”. I’ve used that one myself with denialist industry colleagues. A reservoir simulation model is conceptually and numerically very similar to a climate model and makes use of sometimes-poorly-constrained parameterisation just like climate models. And the geological properties were probably kriged, like Cowtan & Way. But I bet Crichton got his take on the hockey stick from publications whose medical or showbiz coverage he could rip to shreds based on his inside knowledge. So I’ll claim it as a meta-Gell-Mann Effect 😉 .

  15. Dave,
    I missed that, thanks. It does seem rather ironic that Crichton would have coined that term.

  16. Paul,
    Thanks, that’s very interesting. What do you mean by a negative ENSO trend on mid-tropospheric temperatures?

  17. JCH says:

    Paul S. – to you, is there an obvious cause of the negative trend?

  18. russellseitz says:

    Jim Hansen’s 1988 congressional testimony reported that GISS had modeled three hypothetical future forcing trajectories and got a spread of temperature resultsi. But the one that came closest to the observed secular trend was not the focus of popular communication,as many representatives and Senators who heard him testtify went on to publically emphasize the high end results.

    This aroused some to respond just as hyperbolically, witness what the gatekeepers of the NASA satellite temperature record, Christy and Spencer wrote in Science in 1990 :

    “But how good is the evidence, and how likely is substantial global warming? When might it happen? Applying the customary standards of scientific inquiry, one must conclude that there has been more hype than solid facts … Modeling of global climate is largely concentrated on examining effects of doubling the atmospheric content of greenhouse gases. As might he expected, the answers they get are functions of the models they employ. The spread is from 1. 5′ to 5’C; that is, there is great uncertainty. If one examines the subject, one finds virtually unanimous agreement that the models are deficient….What have been the warming effects, if any, of anthropogenic gases? The typical answer is 0.5’C.

    But the answer depends on what time interval is chosen. There was substantial increase in temperature from 1880 to 1940. However, from 1940 until the 1960s, temperatures dropped so much as to lead to predictions of a coming ice age. New, precise satellite data raise further questions about warming. From 1979 to 1988 large temperature variability was recorded, but no obvious temperature trend was noted during the 10-year period.’ …A fashionable estimate of the time when doubling of atmospheric CO, will occur is the middle of the next century. But past predictions of energy usage have been notoriously inaccurate.. What should he the national response to the above uncertainties? … Whatever we do should he based on well-thought-out long-range goals. It should not result from a half-baked political response. ‘

    –R.W. Spencer and J. R. Christy, “Precise Monitoring of Global Temperature Trends from Satellites , “Science 247 (March 30, 1990): 1558

    Years passed before it was realized that, while Hansen’s midrange model had mildly overemphasised the rate of warming , Christy and Spencer had gotten the sign wrong in translating their micrwave radiometer data into a temperature trend.

  19. Dave said

    “Based on his expertise, decades earlier, as a medical doctor not a climate scientist, physicist, statistician or tree biologist? I presume he’s recanted now that the hockey stick has been validated and replicated dozens of times? “

    Crichton may have recanted by now … if he hadn’t died like 10 years ago.

    “Now don’t be scared. I’m a doctor” The Andromeda Strain scared the heck out of me as a kid.

  20. paulski0 says:


    Well, the ENSO trend since 1979 is negative (by MEI, detrended Nino3.4, SOI etc.) which indicates a negative internal variability influence and lower expected tropospheric trends, particularly in the tropics, from 1979 to present. That’s well accepted I think, and AMIP models run from 1979 produce substantially lower tropical tropospheric warming than free-running CMIP models.

    However, if models are underestimating (or not properly capturing non-linearity of) the ENSO influence on tropical tropospheric temperatures then that could cause them still to produce slightly too great a tropical tropospheric warming trend since 1979.

  21. Paul,
    Okay, I see, thanks.

  22. “Thanks, that’s very interesting. What do you mean by a negative ENSO trend on mid-tropospheric temperatures?”

    I looked at his last chart and am curious about the anti-correlation shown.

    The first thing I thought of is that since ENSO is a standing wave dipole, if one makes a measurement or does a simulation at the wrong spatial location, you will invert the the sign of the excursions. For example, SOI is anti-correlated with NINO34 due to the defined way the indices are measured.

    Here is an animation of the ENSO dipole I created a few days ago — you can the anti-correlation at different locations:

  23. Mitch says:

    ” I find it interesting that the newer models that get ENSO and eddies right tend to be running warmer, around 5°C for an ECS.

    From that article, the author writes: “Its rendering of the El Niño cycle, the periodic warming of the equatorial Pacific Ocean, looks “dead on,” says Michael Winton, a GFDL oceanographer who helped lead the model’s development.”

    That’s a strong claim if the author quoted properly. Getting ENSO “dead on” is a significant finding all by itself.

  24. paulski0 says:

    Paul Pukite,

    I should perhaps repeat that both curves are difference plots of a tropical Nino3.4 area average (defined as 20S-20N, 195-240E) minus the standard equatorial Nino3.4 area average. The GFDL model does produce a large amount of tropical warming in response to El Nino, it’s just that the distribution is more concentrated towards the equator rather than amplified at higher latitudes within the tropics as in the observations.

    The model is a high resolution global GCM running in AMIP mode – using prescribed SSTs (HadISST I think).

  25. bless, John as been predicting cooling since 2001, how is that for a failed hypothesis

  26. Steven Mosher says:

    funny. i read his whole report .. and tweeted that line.. models are failed hypothesis .. as the stupiest thing i could recall from his piece.

  27. Dana Nuccitelli says:

    It never ceases to amaze me how these guys who are supposed to be reasonably good scientists say such scientifically stupid things. The psychology of it is pretty intriguing – what makes these guys believe such garbage? Is ideology really that strong?

  28. russellseitz says:

    It’s theology, Dana, not ideology.

    Those in the orbit of The Cornwall Alliance hail C&S as Dominionist culture heroes for standing up for the principle that climate change can’t lead to SLR , because Noah.

    Christy does missionary work in Africa, and Spencer has recieved the Evangelical Climate Scientist of the Year Award.

  29. izen says:

    Hypothesis is a collective noun.
    Like ‘herd’ or ‘society’.

    It is formed of tree of an increasing number of subsets of units that may also be called hypotheses down to the level of the description of the quark.

    The unity reified metaphor of a ‘hypothesis’ is the emergent property of its constituent subsystems which are themselves hypotheses. It always has a nested structure of interacting ‘smaller’ parts.

    A hypothesis about aspects of the material reality within which we can assume we exist can be said to fail when the emergent property of that system of smaller hypothesis does not replicate in its emergent properties the behaviour we observe in reality.
    (Although the problems with such simplistic falsification should be well known.)

    I would conjecture that Christy is ether intentionally, or from ignorance, is confusing the differences from observations in the emergent properties of climate models, which can be regarded as mathematical instantiations of many other hypotheses about the physical and chemical processes involved, to try and falsify the whole of the underlying set of hypotheses in physics and chemistry of which it is formed. Or at the very least to imply that the emergent property of climate from what we do ‘know’ (or hypothesize) is intrinsically false.

    As others have pointed out, the fact that his own work relies on many of the same hypotheses about the physics of the atmosphere being accurate. Given his own record of suspect observations, there is an element of irony in his eagerness to dismiss the output of climate models based on those systems as false. It plays a to a binary dichotomy that one aspect of an emergent system that is not correct, falsifies the whole hypothesis.
    But to paraphrase, all maps are wrong, but many are useful.

  30. Dave_Geologist says:

    I’ve never really got that argument russell, although I’m aware it exists and there’s a YouTube video of a Congressman making it, to eyerolls from the audience behind.

    Surely God didn’t say there wouldn’t be any more floods? Just that He wouldn’t send another Flood. And AFAIK he didn’t promise to protect us from floods induced by human folly. And from my reading, the Old Testament God is a tricksy God, and you’re wise to read the small print and parse His exact language carefully. A flood which covered Mount Ararat clearly required more water than is available on the entire planet, which had to be magicked into existence then magicked out of existence forty days later. Surely steric expansion of seawater and melting of ice, with no added water inventory, doesn’t contradict the “no Flood” promise?

  31. Chubbs says:

    There is no error analysis – so can’t test any hypothesis.

  32. David B. Benson says:

    Dave _ Geologist, the story of the Flood is clearly one that grew with the telling. There is a popular account of the flooding of the Black Sea, when the current passageway to the Aegean opened about 8000 years ago. Likely Noah and many others had to keep moving to keep ahead of the sea level rise.

    Possibly now a parable for our times.

  33. Eli Rabett says:

    Russ, it was pretty clear early on that the Hanson 1988 model had the climate sensitivity too high, like about 4+, and he even said so soon after. However over 10-15 years that makes close to no difference.

    The important thing that Christy and most people miss is that the models do a pretty good job on atmospheric circulation, which is the big heat engine and more attention should be paid to that than the single parameter global temperature anomaly.

  34. Dave_Geologist says:

    Yes, I’ve heard the Black Sea story. And of course the expansion of humans into Europe and growth of settlements coincided with post-glacial SLR. I bet the former inhabitants of Doggerland had their own Flood story. Especially with the tsunami caused by the Storegga Slide, which reached up to 80km inland and 20m above sea level. At a time when NW Europe was forested and most populations were probably living near coasts and rivers, it would have been devastating even if the water later receded.

  35. russellseitz says:

    Dave_Geologist says:
    ” AFAIK he didn’t promise to protect us from floods induced by human folly. And from my reading, the Old Testament God is a tricksy God, and you’re wise to read the small print and parse His exact language ”

    Preoccupation with my own humble superstition, physics, as kept me from checking on progress in exegesis since Hobbes and Barth, but it think it being “wise to read the small print ” should be one of the Ten Commandments of Climateball, especially as it applies to Materials and Methods, Executive Summaries, and Manifestos

  36. Mal Adapted says:

    “Cornball Alliance” – LOL, Russell! Indeed, the Cornbwall Alliance promulgated the 2009 Evangelical Declaration on Global Warming, which our host discussed some years ago on his previous blog. Roy Spencer is a signer. The referenced blog post concludes:

    If Roy Spencer has indeed signed this declaration then he has essentially made up his mind about global warming. His research is not to consider if global warming is natural, it’s to show that global warming is natural. This completely changes how one should assess his research and anything he writes about global warming and climate change. Many accuse climate scientists of being biased but this seems like a classic example of explicit bias. Essentially it seems that Roy Spencer’s research is aimed at confirming his view that global warming and climate change are simply a consequence of some natural process and are not anthropogenic. I think everyone should bear this in mind when considering Roy Spencer’s views on global warming and climate change.

    John Christy appears not to have signed the Declaration, but as you point out, is centrally involved in the CA. If “The first rule of science is not to fool yourself,” then open association with the CA announces one’s willingness to be fooled. IMHO Christy has forfeited all scientific credibility, just as Spencer has.

  37. JCH says:

    RC Wills, since 1979 Links to relevant papers are in the article, and Zhou’s low clouds are discussed. As Issac Held said to Judith Curry’s group at the physics society, it all goes through the eastern Pacific.

  38. David B. Benson says:

    Dave _ Geologist — In addition, Eden was like what is now the Persian Gulf.

  39. David B. Benson says:

    Dave_Geologist — In addition, Eden was likely what is now the Persian Gulf.

  40. KiwiGriff. says:

    Dr Roy Spencer Ph,D has not only sighed up to the CA.
    He is both on their Advisory Board and is a Cornwall Alliance Scholar Senior Fellow.

    As to awards.
    Dr Roy Spencer Ph.D proudly accepted last year .
    The Frederick Seitz Memorial Award,
    Past recipients of this august award have included Dr’s Soon, Legate and Happer .
    The same organizations ceremony gave Viscount Monckton of Brenchly a Dauntless Purveyor of Climate Truth Award in 2014.
    After that interesting internet research.
    I need a long shower and some mind bleach.

  41. russellseitz says:

    The salient feature of is that The Heartland Institute is a full service diploma miil– these awards are for sale to all comers at $4,000 a pop

    While the living can decline them, the dead have no recourse against their names being attached to such awards.

  42. David B. Benson says:

    For comments regarding Energy or Climate Change, consider the
    BNC Discussion Forum

    Post facto moderated.

  43. russellseitz says:

    David, registration is disabled at the BNC forum.

    If your’e aboard you might bring this to their attention-
    What could possibly go wrong with adding floating glass to floating plastic in the oceans?

  44. David B. Benson says:

    Russell, I passed your complaint to the moderator.

  45. When writing for the general public in his non-peer-reviewed political GWPF “report”, Christy claims the “scientific method” requires explaining differences by saying the models are in error:

    “Will the next IPCC report discuss this longrunning mismatch? There are three possible ways they could handle the problem:
    • The observations are wrong, the models are right.
    • The forcings used in the models were wrong.
    • The models are failed hypotheses.
    I predict that the ‘failed hypothesis’ option will not be chosen. Unfortunately, that’s exactly what you should do when you follow the scientific method [page 8].”

    Click to access JohnChristy-Parliament.pdf

    But when writing for scientists in the peer-reviewed literature, Christy admits he cannot discount alternative explanations besides model error:

    “Satellite bulk tropospheric temperatures as a metric for climate sensitivity
    As noted, we cannot totally discount that natural variability or errors in forcing might also account for the discrepancy between modeled and observed TTCR. However, given the facts that the processes controlling the uptake of energy by oceans and the transfer of heat in the tropical atmosphere are largely parameterized, it is not scientifically justified to dismiss model error, possibly substantial, as one source of the discrepancy [page 517].”

    Click to access christymcnider2017.pdf

    Of course, Christy doesn’t bother to address the published evidence in support of the “forcings error” explanation and against his “model error” explanation. This includes research that looks at near surface warming trends; forcings-error for near surface warming trends would manifest in the bulk troposphere as well, since the near-surface air and bulk troposphere are radiatively and convectively coupled:

    “Causes of differences in model and satellite tropospheric warming rates”
    “The ‘pause’ in global warming in historical context: (II). Comparing models to observations”
    “Robust comparison of climate models with observations using blended land air and ocean sea surface temperatures”
    “Reconciling warming trends”
    “Natural variability, radiative forcing and climate response in the recent hiatus reconciled”
    “Reconciling controversies about the ‘global warming hiatus’”

    Figure 7: “Volcanic radiative forcing from 1979 to 2015”

    And how about the other times in which differences were not due to model error, but another error, such as Christy messing up on observational analyses? Examples include:

    1) Christy claiming in the 1990s that the troposphere didn’t warm; he didn’t adequately correct for orbital decay of satellites in his analysis.
    2) Christy claiming the lower troposphere was warming too slowly; his diurnal drift correction had the wrong sign.
    3) Christy claiming the mid-to-upper troposphere wasn’t warming much; he didn’t correct for stratospheric cooling contaminating his satellite-based mid-to-upper troposphere analysis.
    4) Christy claiming the tropical lower stratosphere was at 150mb and had warmed overall since the 1950s or 1970s; contradicted by his own satellite-based analysis and other sources [the tropical troposphere doesn’t start until 70mb].

    By Christy’s poor reasoning, in each of those cases the “scientific method” would have required saying the models were wrong, even though the actual explanation was that Christy was wrong (whether intentionally or otherwise).

    At this point, I’m no longer amazed that Christy gets away with the sort of stuff when writing in politically-motivated non-peer-reviewed documents, since he’s been doing it for decades.

  46. russellseitz says:

    Christy & Spencer’s high-profile 1990 Science editorial , ( vide supra) which was predicated on their erroneous results, left policy analysts in a daze until their belated publication of a eight years later corrigendum.

  47. izen says:

    I know the scientific accuracy with which the climate change issue is discussed is important, so Christy’s dubious characterisation of models is fair game.
    But it does seem to pale into insignificance compared to the political spin that the US government is applying.
    Recent moves to exclude the term ‘climate change’ from official documents and the elimination of climate predictions any further than 20 years ahead are bad enough.
    But this…!?!? Freedom Gas !
    ““Increasing export capacity from the Freeport LNG project is critical to spreading freedom gas throughout the world by giving America’s allies a diverse and affordable source of clean energy.””

  48. russellseitz says:

    Horizontal integration of the petroleum and potato industries promises great new meme opportunities.

    A new Freedom Oil feedstock could be created by recycling rancid corn oil from cooking Freedom Fries as a drilling fluid additive in petroleum production. Over to you, CanMan & Dave_Geologist.

  49. Dave_Geologist says:

    We already use vegetable oils in drilling fluid russell, and have done for decades. My first exposure was in 1994. Much easier to get permits than with diesel, especially offshore. Yep, the drilling industry has been using biodegradable fluids since last century. Not as an additive, generally about a 50:50 mix with brine. The brine is weighted up with heavy salts to compensate for the lighter oil. It brings lubrication benefits, and is also beneficial with water-sensitive shales. It’s an invert emulsion, so looks to the rocks like oil. Some of the emulsifiers in the early days were a bit nasty. Strip the oils out of your skin. I remember one rig where the crew complained it was giving them dermatitis. Of course they should have been wearing gloves…

    Shales usually have water as the mobile phase, so you want an immiscible drilling fluid. For the reservoir, you usually prefer water-based drilling and completion fluids, since oil is the mobile phase. However gas is so mobile you get invasion anyway, so we often use oil-based muds there, with solids designed to build a mud-cake on the wellbore wall and minimise invasion. What invasion there is should ideally just be the oil phase. The solids include sand grains which won’t penetrate the rock pores, and mud with pores too small to pass the brine micelles. One of the lessons from that 1994 well was to make it Bubba-proof. The lab found a sweet spot which didn’t damage the reservoir. Too much or too little did. Unfortunately they dumped it in by the sackload and hoped it would mix on the way down. It didn’t.

    Corn starch was widely used in fraccing, right back to the early days, as a viscosifier. Cheap, and already permitted for food. Most of the shale oil and gas plays use slickwater fracs though.

  50. russellseitz says:

    I see, Dave, but where do you stand on extra virgin crude oil a salad dressing ingredient ?

  51. The Very Reverend Jebediah Hypotenuse says:

    Increasing export capacity from the Freeport LNG project is critical to spreading freedom gas throughout the world by giving America’s allies a diverse and affordable source of clean energy.

    Throughout the world?
    America has but five allies left… Israel, Saudi Arabia, the UK, Australia, and Canada.
    Neither the Saudis nor the Aussies need any US freedom gas.
    And Canada is just being polite.

    The only things America is currently ‘giving’ its allies and spreading throughout the world is what comes out of Trump’s lie-hole, and the ensuing paroxysms of bemused distraction.
    That is most definitely a renewable.

  52. Dave_Geologist says:

    If by extra-virgin you mean clean and pure russell, we’re basically talking liquid paraffin. So only if the fibre alone is not enough, and you need a laxative.

  53. russellseitz says:

    What finer condiment could vegans ask for than pure Texas Crudite Oil?

    If Dave can persuade API to lobby for the use of balasmic vinegar as a completion fluid, ,by the end of the Antropocene, salad dressing will gush from the nation’s wells.

  54. Dave_Geologist says:

    Stimulation fluid, russell, not completion fluid. That’s usually KCl brine. Acetic acid is indeed used in acid washes and acid fracs. In one of my last jobs it worked a treat in dissolving stress-caging material. But somewhat more concentrated than vinegar, so not advised on salads.

  55. mt says:

    Santer gave a talk at Texas A&M on work in progress a few years ago which I attended. It was frustrating for me to be asked not to write about (a sort of informal embargo).

    But here’s the publication:

    “When the impact of lower-stratospheric cooling on TMT is accounted for, and when the most recent versions of satellite datasets are used, the previously claimed ratio of three between simulated and observed near-global TMT trends is reduced to approximately 1.7. Next, the validity of the statement that satellite data show no significant tropospheric warming over the last 18 years is assessed. This claim is not supported by the current analysis: in five out of six corrected satellite TMT records, significant global-scale tropospheric warming has occurred within the last 18 years. Finally, long-standing concerns are examined regarding discrepancies in modeled and observed vertical profiles of warming in the tropical atmosphere. It is shown that amplification of tropical warming between the lower and mid-to-upper troposphere is now in close agreement in the average of 37 climate models and in one updated satellite record.”

    Santer, despite his reputation, is committed to understatement. He did not use the words “Christy’s dataset is bullshit” in his talk, nor in the paper. But he emphasized in the talk that Christy’s dataset was incompatible with ALL other evidence, and my understanding is that but for Christy’s dataset, the problem has been resolved to within the expected uncertainties of modeling and observation.

    Maybe Santer is wrong about all this, but maybe he is right. Given that this position is at least being advanced by a credible scientist, I think it is necessary to regard this posturing

    “Will the next IPCC report discuss this longrunning mismatch? There are three possible ways they could handle the problem:
    • The observations are wrong, the models are right.
    • The forcings used in the models were wrong.
    • The models are failed hypotheses.
    I predict that the ‘failed hypothesis’ option will not be chosen. Unfortunately, that’s exactly what you should do when you follow the scientific method [page 8].”

    as self-serving and misleading. While the possibility is open that all of the observations *except Christy’s* are in agreement with models, Christy doesn’t mention that.

  56. John Ridgway says:

    “Of course, a model can be wrong, but that’s not the same as it being a falsified hypothesis.”

    Yes, but Christy did not refer to ‘falsified hypothesis’ but to ‘failed hypothesis’. I see a subtle but important distinction here.

    You speak as if hypotheses can only apply to a proposition regarding the nature of a physical law. I don’t think this is a necessary restriction. It can apply to anything taken to be true for the purposes of an argument or investigation. An hypothesis is therefore just an assumption. A model is a structured collection of such assumptions, and whether or not the collection is accurate and sufficient to make the relevant argument is the matter in hand. For example, when the hypothesis happens to be that the model captures the dynamics of the system, with sufficient accuracy as to ensure that its projections will fit with future data, it is natural to evaluate such potential for fidelity by reference to a model’s track record. It is in that sense, I presume, that Christy refers to models as ‘failed hypotheses’. This doesn’t mean that anything has been falsified, but it does mean that they are not, in his view, providing a strong enough premise for some of the conclusions that are being drawn. In saying so, I presume Christy understands that adherence to physical law (as much as it is currently understood) serves to constrain the model; the laws are not themselves hypotheses to be falsified by the model. As to whether or not he is over-reacting to specific failures of specific models, I am not qualified to comment.

    I am not wishing to take sides in the argument regarding tropospheric warming, but I am suggesting it may be that objections to the manner in which Christy has chosen to express his misgivings may be the basis for a phony war and that some of the extreme disdain hitherto articulated on this thread may be unwarranted.

  57. John,
    But there is a difference between wrong and a failed hypothesis. Also, all models are wrong, so the key things is whether or not they are useful. Even if a model turns out to be completely wrong about something, it can still provide useful information.

    In this particular context (as some have pointed out above) it’s not even clear that the mismatch between the model and the observations is quite what John Christy suggests. It’s maybe a pity that he didn’t at least acknowledge this possibility.

    Also, the tropospheric hot spot is based on some pretty basic physics. Water vapour increases as the air warms. As it rises the atmosphere can become saturated. As the water vapour condenses it then heats that part of the atmosphere. We already know that this must be broadly correct, because the atmospheric lapse rate matches what we’d expect for a moist atmosphere.

    If, however, this is not manifesting itself as we warm, then it could mean that water vapour is increasing by less than we expect. Given that lapse rate feedback is negative, this actually implies a slightly higher, rather than lower, climate sensitivity. Alternatively (as this Isaac Held post points out), it could imply that the atmosphere in the tropics is destabilising through enhanced convection. The observations, however, suggest that this isn’t the case.

    So, one should be careful of assuming that the problem is with the models, rather than with the observations, especially when dealing with observations that have – in the past – undergone numerous adjustments and when some observations do indeed suggest enhance tropical tropospheric warming that is in line with what is expected.

  58. John Ridgway says:

    “Also, all models are wrong, so the key things is whether or not they are useful.”

    Indeed, you are right, all models are ‘wrong’ (that’s what makes them models) but some are still more useful than others. That is why I think the distinction between a failed hypothesis and a falsified hypothesis is important. Wrong models are still successful if they have predictive skill in at least one important regard, but they may nevertheless be deemed a collective failure if the overarching argument to be made is not adequately supported by them. This may be for a variety of reasons – even in the absence of hypothesis falsification. I think it is important to appreciate that we are talking about evidence based decision-making here and the question as to whether the evidence is strong enough is ultimately a political one, since it depends upon the nature of the decision to be made.

    I thank you for the scientific insight regarding Christy’s specific views, but I still cannot be drawn into that debate since I have no competence to comment, other than to agree that data quality is an important consideration when seeking to evaluate models.

  59. angech says:

    Models are failed hypotheses! or hypotheses are failing models.
    “One thing that the report doesn’t make clear is that amplified tropical tropospheric warming is not simply a signature of greenhouse gas-driven warming; it’s expected for any kind of warming”
    It is good to know that any kind of warming should have a tropical hot spot.
    Ups the ante on the hypothesis even more.
    Hence it is not good enough to ignore the lack of evidence for such a hot spot or to claim that it is their but being missed.
    Searching desperately for straws in this regard is entertaining.
    More pertinent perhaps is asking why reality* is out of step with observations.
    *Reality = hypothesis.

  60. I wanted to offer a follow-up to my comment before, to further expand on the problems with what Christy said. There are at least four possible explanations for differences between model-based projections vs. observational analyses:

    A1) Observational uncertainty / heterogeneities in the observational analyses
    A2) Internal variability / chance fluctuations
    A3) Error in inputted forcings
    A4) Model error (ex: models over-estimate climate sensitivity)

    These explanations are not mutually exclusive. Christy’s problem is that for decades he’s prematurely leaped to explanation A4-model-error, when the primary explanation was actually something else. For instance, in my previous comment I went over examples where the correct explanation was not A4-model-error, but instead A1-observational-uncertainty, in the form of Christy screwing up on diurnal drift, orbital decay, etc. in his satellite-based UAH analysis. That makes it rather ironic when he falsely claims that the “scientific method” requires leaping to explanation A4-model-error.

    And in my previous comment I also discussed some evidence in support of explanation A3-forcings-error. But what about some evidence-based arguments against Christy’s version of A4-model-error, where he claims that models over-estimate climate sensitivity? Well, one problem with Christy’s claim is that the mid-to-upper tropospheric warming in question represents the lapse rate feedback. Relatively greater mid-to-upper tropospheric warming would indicate a more negative lapse rate feedback, which would act as a negative feedback limiting climate sensitivity. So it makes no sense to claim, as Christy does, that the models under-estimating this tropospheric warming reflects the model’s over-estimating climate sensitivity.

    For some further arguments, let’s take the following recent paper from Santer et al. that I cited in my previous comment:

    “Causes of differences in model and satellite tropospheric warming rates”

    Pages 482 and 483 of the paper offer the following arguments against Christy’s A4-model-error explanation in terms of models over-estimating climate sensitivity:

    1) If the models are much too sensitive to CO2, then there should be a specific discrepancy between observed climate responses to volcanic eruptions vs. the models’ predicted response to said eruptions. But this discrepancy does not appear. Santer et al. cite a bulk tropospheric analysis to show this, but near-surface analyses show this as well:

    “Volcanic contribution to decadal changes in tropospheric temperature”
    Figure 3: “The equilibrium sensitivity of the Earth’s temperature to radiation changes”
    “Effect of climate sensitivity on the response to volcanic forcing”
    “Climate response to volcanic forcing: Validation of climate sensitivity of a coupled atmosphere‐ocean general circulation model”

    Wigley et al.: “Comment on ‘‘Climate forcing by the volcanic eruption of Mount Pinatubo’’ by David H. Douglass and Robert S. Knox” [DOI: 10.1029/2005GL023312]
    Robock: “Comment on ‘‘Climate forcing by the volcanic eruption of Mount Pinatubo’’ by David H. Douglass and Robert S. Knox” [DOI: 10.1029/2005GL023287]
    “Response to the eruption of Mount Pinatubo in relation to climate sensitivity in the CMIP3 models”
    “Global cooling after the eruption of Mount Pinatubo: A test of climate feedback by water vapor”

    2) If the over-sensitivity accounts for most of the post-1999 model-observations discrepancy, then models should exaggerate pre-1999 CO2-induced warming as well. So there should be a similar pre-1998 model-observations discrepancy with respect to tropospheric warming. Yet this pre-1998 discrepancy is not evident.

    3) A statistical, model-based test using a proxy for each model’s climate sensitivity argues against Christy’s model-sensitivity explanation.

    So that’s four arguments against Christy’s model-error explanation. Maybe one day he’ll come up with a cogent response to them, instead of writing non-peer-reviewed articles for politically-motivated organizations, where he acts as if the scientific method involves ignoring these arguments and just leaping to model-error as an explanation anyway.

  61. Pingback: 2019: A year in review | …and Then There's Physics

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.