The issue with local weather fashions – watts with them?

“We know that there are known acquaintances. There are things that we know we know. We also know that there are unknown unknowns; That said, we know there are some things that we don’t know. But there are also unknown unknowns – those we do not know, who we do not know ”- Donald Rumsfeld

Ed Zuiderwijk, PhD

An observation

There is something strange about climate models: they don’t converge. What I mean by that, I’ll explain based on historical determinations of what we now refer to as “Equilibrium Climate Sensitivity” (ECS), also known as “Charney Sensitivity” (Ref. 1), defined as the temperature rise at the bottom of the Earth’s atmosphere when the CO2 content is doubled (after all feedback has prevailed). The early models by Plass (2), Manabe & Arbeiter (3) and Rowntree & Walker (4) in the 1950s, 60s and 70s gave ECS values ​​of 2 degrees Celsius to over 4 ° C. In recent decades, developed these models into a collection of more than 30 climate models, which are summarized in the CMIP6 ensemble and form the basis for the upcoming AR6 (“6th assessment report”) of the IPCC. However, the ECS values ​​still cover the interval from 1.8 ° C to 5.6 ° C, a factor of 3 difference in results. After around four decades of development, climate models have still not converged into a “standard climate model” with a clear ECS value. Rather, the opposite is true.

What this means becomes clear when we consider what it would mean if, for example, the astrophysicists were to find themselves in a similar situation with their star models, for example. Years ago, the analytical polytropic description of the early 20th century gave way to complex numerical models that made it possible to study stellar evolution – caused by the change in internal composition and the associated changes in energy production and opacity – and that also in 1970, when I mean The first steps in this area offered a reasonable explanation, for example the Hertzsprung-Russell diagram of star populations in star clusters (5). Although these models are always in need of improvement, you can say that they have converged into something called a canonical star model. The different computer codes for calculating stellar evolution developed by groups in different countries give the same results for the same evolutionary phases, which is also in good agreement with the observations. Such convergence is indicative of the advancement of the knowledge on which the models are based, by improving understanding of the underlying physics and testing against reality, and manifests itself in many of the sciences and techniques in which they are used.

For example, if the astrophysicists were in the same situation as the climate model makers, they would still be working with a number of solar models that predict, give or take a value of X for the surface temperature. Or that in a technical application a new aircraft design should have a wing area of ​​Y, but it could also be 3 years. You don’t have to be a genius to understand that such models are not believable.

A thesis

So much for my observation. Well what does it mean. I will present my analysis here in the form of a thesis and defend it with an appeal to elementary probability theory and a little story:

“The fact that the CMIP6 climate models show no signs of convergence means that, on the one hand, none of these models are likely to represent reality well and, on the other hand, the true ECS value is likely outside the interval 1.8-5.6 degrees. “

Suppose I have N models that all predict a different ECS value. Mother Nature is difficult, but she is not evil: there is only one “real” value of ECS in the real world; If this were not the case, any attempt at a model from the start would be pointless. Therefore, at best, only one of these models can be correct. Then what is the probability that none of these models is correct? We know right away that N-1 models are incorrect and that the remaining model may or may not be correct. So we can say that the a priori probability that a model is wrong is [(N-1+0.5)/N] = 1–0.5 / N. This gives a probability that none of the models of (1-0.5 / N) ^ N is correct, about 0.6 for N> 3. So this is a probability of 3 to 2 that all models are wrong. This 0.6 is also the probability that the real ECS value is outside the interval of 1.8 ° C to 5.6 ° C.

I can already hear the objections. For example, what does it mean that a model is “wrong”? Algorithmic and coding errors aside, this means that the model may be incomplete, it may not contain elements that should be included, or that it is cluttered with aspects that are not (a bug that is often overlooked). In addition, these models have an intrinsic variation in their result and often contain the same elements, so these results are correlated. In fact, the ECS results completely tile the interval from 1.8 ° C to 5.6 ° C, and for each ECS value between the specified limits, models can be found that can provide this result. In such a case, consider the effective number of independent models M represented by CMIP6. When M = 1, it means that all models are essentially the same and the 1.8C-5.6C is an indication of the intrinsic error. Such a model would be useless. More realistic is an M ~ 5 through 9, and then return to the considerations above.

What rubs climatologists the most is my claim that the true ECS is outside the 1.8 ° C to 5.6 ° C interval. There are very good observational arguments that 5.6C is a gross overestimation, so I actually argue that the real ECS is likely below 1.8C. Many climatologists believe that this is instead a lower bound. Such a conclusion is based on an error, namely the premise that there are no “known unknowns” and in particular no “unknown unknowns”, that is, that the underlying physics is fully understood. And as mentioned earlier, the lack of convergence in the models shows that this is precisely what is not the case.

A short story

Imagine a parallel universe (theorists are not averse to this these days) with an alternate earth. There are two continents, each with a team of climatologists and their models. The A-Team on the Laputo landmass has 16 models that predict an ECS interval of 3.0 ° C to 5.6 ° C. If this is correct, it has significant consequences for the stability of the atmosphere. The B-Team in Oceania has 15 models that predict an ECS interval of 1.8 ° C to 3.2 ° C. The two teams may be unaware of the other’s existence due to political circumstances and believe that their models set hard limits to the true value of the ECS.

The fact that the models of both teams deliver so different results is because those of the A-Team contain ingredients that are not found in those of the B-Team, and vice versa. In fact, the climatologists on both teams are not even aware of the possible existence of such missing aspects. After a thorough analysis, both teams write a paper on their findings and at the same time send it to a magazine in Albion, a small island nation known for the strong sense of independence of its people. The editor sees the connection between the two works and decides to bring the authors into contact with one another.

A culture shock follows. The lower gods begin a screaming match. The members of the A-team call the members of the B-team “deniers”, who in turn call “chickens”. But the more mature teams of both realize that they had a massive blind spot on things that the other team knew but they themselves didn’t. That these “strangers” had bitten both teams hard in the buttocks. And the smartest realize that the combined 31 models are now a new A-Team, to which the above applies even more: that a new B-Team could emerge somewhere with models with ECS values ​​outside the range of 1.8 ° Predict C to 5.6 ° C.