Debaters Behaving Badly, Part 3 - Confusing Local and Global Temperatures

In my first two posts in this series, I described the bad behavior of cherry-picking short-term trends and choosing the wrong scale to hide the incline of global temperatures. Here I want to consider the trick of using local (or regional) temperatures in place of global temperatures. As best I can tell, there are at least two reasons why people might want to do this. Some people want to say that global warming is good, the world is starved of CO2, and we need to add more CO2 to the atmosphere. Unsurprisingly, two of the most prominent proponents of this tactic are former geologists for the fossil fuel industry. They want to show that societies thrive when global temperatures are warmer. Others using this tactic seem to want to say that global temperatures don't correlate with CO2. And what better way to make that point than to use local temperatures instead of global temperatures? Prepare for some snarkiness. 

Using Local Temperatures to Say Global Warming Is Good

Imagine you have a weather station in your back yard that is perfectly sited and produces a temperature record to rival the stations in the best networks in the world. Do you think your weather station would help you understand the weather and climate in Greenland? How about the climate 500 m below the surface of the Makassar Straight in the Indonesian Throughway? Would you imagine that your back yard temperatures cause the rise of great philosophers or affect the rise and fall of empires? When your back yard is cold do you feel an overwhelming urge to burn witches? If you answered "no" to these questions, you might be able to see why these debaters are participating in some bad behavior. If you answered "yes," you might be Andy May.

Back in 2016, Andy May posted a blogpost attempting to show that warm periods in Greenland correspond to wonderful and exciting events all over the world. In this penetrating historical analysis, Andy May misplotted the temperature proxy record from the GISP2 ice core (he used 2000 instead of 1950 for "present") and then picked what he considered good events to assign to the warm periods and bad events to assign to the colder periods. When Greenland was cold in a.d. 784, there was "starvation, extreme cold" and "over a third of the European population died." At another time when Greenland was cold in a.d. 1600, "witch hunts reach maximum" and "50,000 witches [were] burned." Who knew that historical analysis could be so insightful? It's just a matter of picking a local temperature proxy (Central Greenland is as good as your back yard) and pretending that the entire globe conforms to the local proxy.

There was only one problem, though. His proxy was 50 years off. He thought "present" was a.d. 2000 instead of 1950. That had to be fixed. Oh wait. There was a second problem. Another proxy from the same GISP2 ice core used different isotopes, and Kobashi's reconstruction had some significant differences. You know that cold period in Greenland that caused a third of England to die? Kobashi's proxy showed the Greenland Summit to be warm at that time. So much for that. Except May tried to re-explain these cherry-picked events with a different temperature proxy for the Greenland summit that didn't always agree with Alley's.

Kobashi's GISP2 reconstruction differs from Alley's

May seems to be determined to make the globe conform to a local temperature proxy. If not the GISP2 ice core, maybe ocean temperatures will do the trick? In 2022, May decided to use another temperature proxy for temperatures 500 m below the Makassar Straight. May seems to think  this proxy is able to explain the rise of Plato and the peak of the Roman Empire. And of course, temperatures in the Indonesian throughway also seemingly compel people to burn witches.

Andy May's antics are probably the most humorous I've seen, but there is no shortage of graphs and "arguments" attempting to make claims about the entire globe on the basis of a local temperatures proxy. The GISP2 proxy at the Greenland Summit from R.B. Alley appears to be the favorite local proxy to confuse with global temperatures. Except R.B. Alley had something to say about that:

...although the central Greenland ice-core records may provide the best paleoclimatic temperature records available, multiple parameters confirm the strong temperature signal, and multiple cores confirm the widespread nature of the signal, the data still contain a lot of noise over short times (snowdrifts are real, among other things). An isotopic record from one site is not purely a temperature record at that site, so care is required to interpret the signal and not the noise. An extensive scientific literature exists on this topic, and I believe we are pretty good in the community at properly qualifying our statements to accord with the underlying scientific literature; the blogospheric misuses of the GISP2 isotopic data that I have seen are not doing so, and are making errors of interpretation as a result.

So, what do we get from GISP2? Alone, not an immense amount. With the other Greenland ice cores… and compared to additional records from elsewhere, an immense amount… Using GISP2 data to argue against global warming is, well, stupid, or misguided, or misled, or something, but surely not scientifically sensible.

Andy May wants to show that both increasing CO2 and global warming is good. That's the motivation for cherry picking events to fit into the warmer and colder portions a local proxy. Even combining these two proxies, you can say little to nothing about GMST from them. The GISP2 proxy isn't even corrected for elevation changes - as the ice sheet increases in elevation, some of the cooling is from the fact that the summit is at a higher elevation, not because of global climate changes. By itself, it's not even that good at representing Greenland, let alone the globe. But think tanks like the CO2 Coalition want to support this notion that fossil fuel emissions are good. That's what is driving his line of reasoning. So May quotes Happer of the CO2 Coalition saying,

There is no scientific evidence that global greenhouse gas emissions will have a harmful effect on climate. Quite the contrary, there is very good evidence that the modest increase in atmospheric CO2 since the start of the Industrial age has already been good for the Earth and that more will be better.

Gregory Wrightstone of the CO2 Coalition agrees. In an article titled Global Warming Isn't the Problem, It's the Solution, he claims:

A review of human and climate history reveal a strong link between the rise and fall of temperature and the rise and fall of civilization, and it is just opposite of what the climate doomsayers are telling you. Past warming periods were much warmer than our modern temperatures and were associated with times of great prosperity. The intervening cold eras had names like Greek Dark Ages, the Dark Ages and Little Ice Age and were linked to crop failure, pestilence and mass depopulation.

The only way this case can be made using a temperature reconstruction is to attempt to force the globe to conform to the temperature variability of a local proxy. If you use a global multiproxy reconstruction, the variability in global temperatures across the Holocene through the end of the Little Ice Age all but disappears. Global temperatures stabilized about 8 thousand years ago, and the the only significant warming period globally occurs at about the same time we began emitting carbon through our fossil fuel emissions. Locally, temperatures may fluctuate a lot, but largely heat was getting redistributed around the globe without shifting GMST that much.

Osman's Global Temperature Reconstruction

Global, multiproxy reconstructions won't do, but local temperature proxies, which show more variability than GMST, can do the trick. So that's the trick they use. Global reconstructions get criticized for statistical flaws that don't stand up to serious scrutiny or they get dismissed as products of a political narrative, but none of these critics have been able to produce a global multiproxy reconstruction that has a different shape.

Using Local Proxies to Say CO2 does not Correlate with Global Temperatures

Proxy evidence for CO2 levels in the preindustrial Holocene show concentrations to be remarkably stable. At the beginning of the Holocene they were ~265 ppm, though they decreased to ~260 ppm about 7000 years ago. Since then, they slowly increased to ~280 ppm in 1750, at the start of the Industrial Revolution. The newest reconstructions of global temperatures conform to this remarkably well. Global temperatures were roughly stable, with slight warming until the Little Ice Age. It's hard to make a case that CO2 and GMST do not correlate if you use the current evidence for global temperatures. So contrarians use the GISP2 ice core instead.

The distribution of warmth across the Holocene was not static. Our understanding of orbital cycles lead us to believe that the Arctic was much warmer during what has been called the Holocene Thermal Maximum (HTM). The Arctic then likely began to cool in sync with orbital cycles. This is one reason why the GISP2 ice core shows an overall cooling trend for the last several thousand years (it's also a source of seasonal bias affecting the Marcott 2013 reconstruction). So if you use a local proxy in the Arctic, you can exploit this long term variability of a local proxy to your advantage. Here's an example.

CO2 Doesn't Correlate Well with a Local GISP2 Proxy

What point can we make from this? Simply that local temperatures conform to the expectations of climate scientists' understanding of orbital cycles. But what contrarians seem to want to do is to claim that a local temperature proxy proves that CO2 and global temperatures don't correlate. Of course, if they used global temperatures, they couldn't make that point. So you get the above graph.

And they don't just use proxy reconstructions to make this point. They also use local portions of the instrumental record. Above, someone decided to use US temperatures. Below, someone decided to use Central England temperatures. 

These examples also use the deceptive scale trick I described in part 2. I suppose I could take these data sets and plot the correlation between them and CO2, but that would defeat the purpose of this post. Climate science has shown through multiple lines of evidence that increasing atmospheric CO2 pushes GMST towards warming, and this is observable in empirical data. Local areas may or may not correlate as well as GMST, since the distribution of heat around the globe is affected by many different factors. If you use the instrumental record for global temperatures, CO2 and GMST do correlate with an r^2 of 0.88. So those that want to challenge this can't use the long term instrumental record for GMST. They must cherry pick local temperatures and then perhaps use one of the other tricks I described in previous posts to make their case. This is yet another clear example of debaters behaving badly.