Gavin’s Falsifiable Science

Guest Post by Willis Eschenbach

Gavin Schmidt is a computer programmer with the Goddard Institute of Space Sciences (GISS) and a noted climate alarmist. He has a Ph.D. in applied mathematics. He’s put together a twitter thread containing what he sees as some important points of the “testable, falsifiable science that supports a human cause of recent trends in global mean temperature”. He says that the slight ongoing rise in temperature is due to the increase in carbon dioxide (CO2) and other so-called “greenhouse gases”. For simplicity, I’ll call this the “CO2 Roolz Temperature” theory of climate. We’ve discussed Dr. Schmidt’s ideas before here on WUWT.

Now, Gavin and I have a bit of history. We first started corresponding by way of a climate mailing list moderated by Timo Hameraanta back around the turn of the century, before Facebook and Twitter.

The interesting part of our interaction was what convinced me that he was a lousy programmer. I asked him about his program, the GISS Global Climate Model. I was interested in how his model made sure that energy was conserved. I asked what happened at the end of each model timestep to verify that energy was neither created nor destroyed.

He said what I knew from my own experience in writing iterative models, that there is always some slight imbalance in energy from the beginning to the end of the timestep. If nothing else, the discrete digital nature of each calculation assures that there with be slight roundoff errors. If these are left uncorrected they can easily accumulate and bring the model down.

He said the way that the GISS model handled that imbalance was to take the excess or the shortage of energy and sprinkle it evenly over the entire planet.

Now, that seemed perfectly reasonable for trivial amounts of imbalance coming from digitization. But what if it were larger, and it arose from some problem with their calculations? What then?

So I asked him how large that energy imbalance typically was … and to my astonishment, he said he didn’t know.

Amazed, I asked if he had some computer version of a “Murphy Gauge” on the excess energy. A “Murphy Gauge” (below) is a gauge that allows for Murphy’s Law by letting you set an alarm if the variable goes outside of the expected range … which of course it will—Murphy says so. On the computer, the equivalent would be something in his model that would warn him if the excess or shortage of energy exceeded some set amount.

Nope. Not only did he have no Murphy Gauge set, but he also had no idea how far the model went off the rails regarding the conservation of energy, either on average or in individual timesteps. He just pushed it back into balance with each timestep, turned his back, and kept rolling.

At that point, I concluded that Gavin was far from suspicious enough of his model. Me, I wrote my first computer program in 1963, about the time that Gavin was born. And I don’t trust computer models one bit. They’ll bite the hand that feeds them at the slightest opportunity even if you fence them in with Murphy Gauges … and Gavin didn’t understand that basic problem.

This is particularly an issue with what are called “iterative” models. These are models that go step by step through time, with the output of each timestep being the input to the next timestep. Errors tend to accumulate in such models, so it’s very easy for them to spiral out of control … and climate models are all iterative models. Here’s a large number of runs from an iterative climate model.

Figure 1. 2,017 climate model runs from climateprediction.net. 

Figure 1 (b), the lower of the two graphs, shows the change in temperature. Note how during the “control phase”, when there is no change in the inputs, even a small ongoing drop in temperature can lead to the model spiraling down to the “Snowball Earth” off the bottom of the graph, as shown in the control phase of the modeled temperature in Figure 1 (a). 

So I’m suspicious as can be of all of the modern iterative climate models. They are all tuned to hindcast the past … but the climate sensitivities in all of them are different. How can that be? Well … it can’t. It means that they’re just making it up. I discussed this problem here, it’s a big one.

Next, let me make a heel turn to set the stage with an overview of the recent changes in climate. Back in Medieval times, around 1000 or so, the surface temperature was as warm or perhaps even warmer than it is today. But then the earth cooled and went into what is called the “Little Ice Age”. This was a hard time for plants, animals, and us humanoids alike. Shorter growing seasons, frozen rivers and harbors, crop failures. No bueno.

Why were the Medieval times so warm? We don’t know. Why did the temperatures drop down to the cold Little Ice Age? We don’t know. Why did temperatures stop dropping around 1700 and not 1400 or 1800? We don’t know.

After that very cold time, temperatures started rising again. And since about the year 1700 or so, temperatures have been rising, in fits or starts, at about a half a degree per century for the last two-plus centuries.

Why didn’t the temperature continue to cool after the Little Ice Age and put us into a glaciation? We don’t know. Why did it start to warm at the end of the Little Ice Age, rather than simply staying cold? We don’t know. Why did it start to warm around 1700 or so, rather than in 1900? We don’t know. Why have we seen slow warming since the Little Ice Age? We don’t know.

As you can see, although we know a lot about the climate … we also don’t know a lot about the climate.

In any case, with that as prologue, here is Gavin’s “falsifiable science” tweet. Bear in mind that I’m not saying he’s wrong because he is a careless programmer. That’s a separate question. I’m saying he’s wrong because he’s conflating three very different theories and treating them as one. Here’s his tweet.

---

We develop theories.
1) Radiative-transfer (e.g. Manabe and Wetherald, 1967)
2) Energy-balance models (Budyko 1961 and many subsequent papers)
3) GCMs (Phillips 1956, Hansen et al 1983, CMIP etc.)

We make falsifiable predictions. Here are just a few:
1967: increasing CO2 will cause the stratosphere to cool
1981: increasing CO2 will cause warming at surface to be detectable by 1990s
1988: warming from increasing GHGs will lead to increases in ocean heat content

1991: Eruption of Pinatubo will lead to ~2-3 yrs of cooling
2001: Increases in GHGs will be detectable in space-based spectra
2004: Increases in GHGs will lead to continued warming at ~0.18ºC/decade.

We test the predictions:
Stratospheric cooling? ✅
Detectable warming? ✅
OHC increase?✅
Pinatubo-related cooling?✅
Space-based changes in IR absorption? ✅
post-2004 continued warming?✅

---

Let me start by saying that Gavin is badly conflating three very separate and distinct theories.

• Theory 1) Increasing CO2 increases atmospheric absorption, which affects the overall temperature of the various layers of the atmosphere, and increases downwelling so-called “greenhouse” radiation.
• Theory 2) In the short term, large changes in downwelling radiation change the surface temperature.
• Theory 3) In the long term, small continuing increases in downwelling radiation lead to corresponding small continuing increases in global surface temperature.

Here the spoiler alert: I think that the first two of these are true (with caveats), but we have virtually no evidence that the third one is either true or untrue.

So let’s go through his six lines of evidence, consider which theory he’s actually discussing, and see if they stands up to critical examination.

a) Increasing CO2 will cause the stratosphere to cool. This is obviously evidence in support of theory 1. Here’s the record of stratospheric temperatures, from the Microwave Sounding Units on a succession of satellites.

Figure 2. Global stratospheric temperatures measured from space.

As you can see, although the stratospheric temperature has indeed dropped, the drop has been quite complex. The two peaks in the record are from the volcanoes noted in the graph. After each one, the stratosphere has warmed for about five years. Each time it seems to have stabilized at a lower temperature. There has been a slight drop since the second eruption. It’s likely that this is from the changes noted in Theory 1, although that is far from clear.

b) Increasing CO2 will cause warming at surface to be detectable by 1990s. This is supposed to be evidence in support of Theory 3. However, while this is true, the temperature has been rising for a couple of hundred years. So unless you believe in Little Ice Age SUVs, this is not evidence in support of any part of the “CO2 Roolz Temperature” theory.

c) Warming from increasing GHGs will lead to increases in ocean heat content. Same as (b) immediately above, and the same objection. It’s supposed to be in support of Theory 3, but in a warming world, a warming ocean is expected and not probative of anything.

d) Eruption of Pinatubo will lead to ~2-3 yrs of cooling. This is evidence in support of Theory 2 … but then so is the surface warming up when the sun rises. We know that large transient changes in the amount of downwelling radiation (which is called “forcing” in climate science) will change the surface temperature.

However, the models didn’t do a very good job of predicting the size of the cooling. Here are some results which I discussed in a post ten years ago:

Figure 2. Comparison of annual predictions with annual observations. Upper panel is Figure 2(b) from the GISS prediction paper, lower is my emulation from digitized data. Note that prior to 1977 the modern version of the GISS temperature data diverges from the 1992 version of the temperature data. I have used an anomaly of 1990 = 0.35 for the modern GISS data in order to agree with the old GISS version at the start of the prediction period. All other data is as in the original GISS prediction. Pinatubo prediction (blue line) is an annual average of their Figure 3 monthly results.

Note that the Hansen/Schmidt GISS model predicted more than twice the drop from Pinatubo compared to the actual reality. It also predicted that the drop would last longer than what happened. I’ll return to this question in a bit, but for now, we’ll note that Theory 2 is true—short-term changes in forcing, whether daily, monthly, or from volcanoes, do change the temperature.

e) Increases in GHGs will be detectable in space-based spectra. With more greenhouse gases in the atmosphere, we expect to see more infrared absorbed by the atmosphere. We’ve measured this change in a variety of ways. This is evidence in support of Theory 1.

f) Increases in GHGs will lead to continued warming at ~0.18ºC/decade. This is put up in support of Theory 3. However, it’s been warming for two or more centuries now, and this prediction in 2004 is nothing but the continuation of the prior thirty years of warming. Once again, the fact that it is still warming is not proof of anything.

To summarize:

• Theories 1 and 2 are clearly true and are supported by a variety of evidence. Three of his six bullet points are evidence in support of those two theories.

• The other three pieces of evidence are saying that after more than two centuries of slow warming … the warming is continuing. This says exactly nothing about Theory 3.

This is the continuing problem with the “CO2 Roolz Temperature” theory … it’s really three very separate theories in one, and while two of the theories are clearly true, there is very little evidence in support of the third leg of the stool. And the stool will not stand up with only two legs.

Gavin closes out his tweet with the following:

We can also look at the testable, falsifiable, theories that were tested, and failed.

Solar forcing? Fails the strat cooling test.❌

Ocean circulation change? Fails the OHC increase test ❌

Orbital forcing? Fails on multiple levels ❌

If you have a theory that you don’t think has been falsified, or you think you can falsify the mainstream conclusions, that’s great! We can test that too! (But lots of people have tried this already so expect there to be an answer already).

PS. Actually, it’s even a bit harder. Not only would you need to find a theory that does as well as the current theory, but you’d also need to show why the current theory isn’t operative.

Now, for folks unfamiliar with my work, I do have a theory. I also have a heap of evidence in support of it. But I’m not a climate skeptic—I’m a climate heretic, someone who denies their basic claim that changes in the temperature are a simple linear function of the changes in forcing.

Folks are interested in why the temperature of the planet changes over time. That’s at the center of modern climate science. My theory, on the other hand, arose from my being interested in a totally different question about climate—why is the temperature so stable? For example, over the 20th Century, the temperature only varied by ± 0.3°C. In the giant heat engine that is the climate, which is constantly using solar energy to circulate the oceans and the atmosphere, this is a variation of 0.1% … as someone who has dealt with a variety of heat engines, I can tell you that this is amazing stability. The system is ruled by nothing more solid than waves, wind, and water. So my question wasn’t why the climate changes as it does.

My question was, why is the climate so stable?

And my answer is, there are a host of what are called “emergent phenomena” that arise when local temperatures go above some local threshold. They include the timing and strength of the daily emergence of the cumulus cloud field in the tropics; the development of thunderstorms; the emergence of dust devils when temperatures get hot; the action of the El Nino/La Nina pump moving warm water to the poles; and various “oscillations” like the Pacific Decadal Oscillation.

These emergent phenomena arise out of nowhere, last for some length of time, and then disappear completely. And acting together, they all work to prevent both the overcooling and the overheating of the planet. And as mentioned above, I say that these phenomena acted to reduce the length and the depth of the effect of the Pinatubo volcano. See my post called “When Eruptions Don’t” for another look at how the climate system responds to a decrease in incoming solar energy due to volcanic eruptions.

I originally published this theory in the journal Energy and Environment. I followed that up with a posting of the same ideas here at Watts Up With That in a post called The Thermostat Hypothesis.

I have continued this quest by writing a number of posts over the last 20 years that have added observational evidence to the theory and explored its ramifications. These included “Emergent Climate Phenomena“, describing what emergence is and why it is so important; “The Details Are In The Devil“, explaining why the “climate sensitivity” type of analysis doesn’t work in a thermostatically controlled system; “Watching Thunderstorms Chase The Heat“, about how thunderstorms operate to cool only the warm parts of the tropical oceans; and most recently “Drying The Sky“, discussing the evolution of different stages in the tropical thermal regulation system.

In all, I’ve written some 40 or so posts exploring this theory of how the climate works. There’s an index to a number of them here, divided up by subject which covers up to January 2018 … hmm, I need to update the index. More recent posts of mine, not separated by subject, are listed here in reverse chronological order. [2021 update—my updated index to my work is now here.]

Now, I fear that my theory is of little interest to the climate establishment because they’re looking for headlines about THERMAGEDDON! CLIMATE EMERGENCY! My theory doesn’t have any of that, in fact, the opposite. My theory says that future warming is likely to be slow and small. So mostly, as with all good heretics, I’m shunned by the powers that be.

Let me close by saying that I have absolutely no academic qualifications at all. I took Physics 101 and Chemistry 101 in college. That’s it. 

Since then, however, I have followed my education by teaching myself a host of subjects. For example, I taught myself and have made money writing programs in the following computer languages—Basic, VBA, Mathematica (2 of 3), Hypertalk, Vectorscript, Pascal, C/C++, and R. I taught myself refrigeration so I could take a job constructing and installing a blast freezer on a boat … in Fiji. As that post discusses, that was instrumental in understanding how thunderstorms operate in exactly the same manner as your household refrigerator.

And to return to the current discussion, I’ve spent thousands and thousands of hours researching and writing and learning about climate … all with zero certificates on my wall.

So please, don’t bother telling me that I’m an uneducated jerk or an ignorant fool. First, I already know that, and if I forget, my gorgeous ex-fiancee will gladly remind me … and second, that’s not the question. The question is absolutely not are my educational bona fides up to your high standards? That’s meaningless.

Nor is the question is Watts Up With That believable or not?  I say this because where something gets published is never the question. There are folks out there that truly seem to think that if E=MC^2 is written on the bathroom wall it’s not true because of where it was published.

The question, the only valid question in science, is are the claims true? Does my theory stand up to close inspection? Are my ideas backed, not by climate models, but by actual real-world observations? Can you find flaws in the logic, the data, the math, or any other part of what I’ve written?

I have great confidence in what I’ve written about my theory, for a simple reason. Watts Up With That is the premier spot on the web for public peer-review of scientific theories and ideas about climate. This doesn’t mean that it only publishes things known to be valid and true. Instead, it is a place to find out if what is published actually is valid and true. There are a lot of wicked-smart folks reading what I write, and plenty of them would love to find errors in my work.

So when those smart folks can’t find errors in what I’ve written, I know that I have a theory that at least stands a chance of becoming a mainstream view.

My best wishes to all,

w.

Post Scriptum: As is my custom, I politely ask that when you comment, you quote the exact words that you are referring to, so we can all be totally clear about both what and who you are discussing.