- JoNova - https://www.joannenova.com.au -

New Science 19: The invisible nameless model that controls the whole field of climate science

Don’t underestimate the importance of the nameless basic model. It sounds small, but in the culture and philosophy of climate science it’s bigger and carries more weight than the massive hairy GCMs. Like an invisible gossamer web, it’s overarching. It spans and defines all the other models. When they produce “dumb” answers, the basic model holds them in, for thou shalt not stray too far from the climate sensitivity defined by the basic model. It defines what “dumb” is. (It’s just “basic physics” after all.) One model to bind them all. What could possibly go wrong?

A lot, apparently. The physics might be right, but the equations are calculating imaginary conditions. The answers might be arithmetically correct but useless at the same time. They miss the real route that energy flows through to space.

Basic climate model, climate change, David Evans, research, cartoon.

By definition, as long as the basic model is wrong, the GCM models can never  get it right.

It’s not like climate scientists consult the oracle of the basic model every day, or even once a year… they don’t need to. They were taught it their climate larval stage, often long before they’d written one paper. The basic model shows that the warming of the 1980s and 90s was “caused” by CO2. Then the GCMs are tuned to that trend and that assumption.

The silent basic model is thus the whole nub of the problem. People have been arguing about the parameters when they should also have been talking about the equations that connect those parameters  — the architecture.  They added the separate warming forces before they did the feedbacks, but they should have done the feedbacks first and then added them all together after.

My favorite line below: “Once two numbers are added, there is no way of telling what the original numbers were just from their sum.” The models can’t figure out what was man-made and what was natural…

In the basic model the atmosphere responds the same way to all forms of warming, always the water vapor amplification and never simple rerouting.

Good news here, we’re in a low-acronym-zone today. This post is ASR and OLR-free, low on WVELS. No equations!

The solar model starts again soon…



19. Comments on Conventional versus Alternative

Dr David Evans, 11 November 2015, David Evans’ Basic Climate Models Home, Intro, Previous, Next.

This post completes the first two parts of this series — problems with the conventional basic climate model, and fixing them with the alternative basic climate model. Here are just some general comments, tying together some of the main ideas. There are hardly any acronyms, and no equations.

After this the series will embark on its third and final part, an hypothesis about the main cause of global warming. Kindly note that whether the third part of the series eventually proves to be right or wrong has no bearing on the correctness of these first two parts about climate model architecture.

This post is only about basic models, not GCMs, except where it specifically states otherwise.

Two Major Errors in the Conventional Architecture

The conventional basic climate model is just based on a radiation balance. It adds the radiation imbalances due to the various influences on climate, which for basic models are principally the changes in absorbed sunlight and the changes in CO2. Then a response, consisting of the Planck sensitivity and feedbacks, is applied to the sum of the radiation imbalances, in order to calculate the surface warming — see Fig. 2 of post 3. This response calculates the surface warming required to bring the net radiation back into balance. Simple — but simplistic, and the problems of climate science stem from it.

Once two numbers are added, there is no way of telling what the original numbers were just from their sum. Hence the response in the conventional model has no way of distinguishing the source of the radiation imbalance — a W/m2 of increased absorbed sunlight or a W/m2 of decreased outgoing heat appear identical to the response. Perhaps owing to the state of  climate knowledge in 1896 when this model architecture arose with Arrhenius, the response is the “solar response” — the response of the Earth to extra absorbed sunlight. It can be largely estimated from the Stefan-Boltzmann equation, and the main feedbacks to surface warming involving moist air. So, the simple radiation balance architecture induces the conventional model to apply the solar response to all climate influences; one size fits all. This is the first major error.

The feedbacks are part of the sole conventional response, so the same feedbacks are also applied to every radiation imbalance, regardless of source. It needn’t be this way (see Fig. 1 of post 5), but in the conventional basic climate model all feedback are in response to surface warming. But this is absurd: solar warming mainly heats the surface, which is quite unlike blocking some heat from escaping to space from the upper troposphere by increasing CO2. Why should they have the same feedbacks? Couldn’t there be feedbacks to increased CO2 that are not responses to increased surface warming? This is the second major architectural error.

The feedbacks to surface warming are well-known and presumably thoroughly researched by now. However the rerouting feedback proposed in post 7 is apparently new, which would appear puzzling at first because it is fairly obvious. The rerouting feedback is just that the extra heat in the upper troposphere due to increased CO2 causes more emission of heat to space from water vapor — so some of the heat blocked from escaping via emissions by CO2 molecules just reroutes out to space via water vapor molecules instead. It is a response to increased CO2 concentration, or more precisely, a response to the warming in the upper troposphere caused by increased CO2. It is not a response to surface warming, or it would be triggered by increased absorbed sunlight that warmed the surface. It presumably went unnoticed because it is in the blindspot of the conventional architecture — in which only feedbacks to surface warming exist. How many other feedbacks are there in that blindspot?

The Dam Analogy

To explain how the conventional climate architecture arose, and how to fix it, consider the dam analogy.

The amount of heat on Earth is like the amount of water in a dam. There is one inflow to the dam—a river of absorbed sunlight from the Sun (sunlight reflected by clouds and ice does not heat the Earth). Water flows out of the dam through four pipes, one for each of the main sources of emission of heat to space. The pipes are only partly full; they could carry more if the water level in the dam rose. When the dam is in its normal “steady state”, neither filling nor emptying, the inflow from the Sun is equal to the total flow through the outlet pipes.

Dam Analogy for the Climate System

– Response to More Sunlight

What happens if the absorbed sunlight steps up to a new level? More water would flow into the dam, so the water level would rise. When steady state was resumed, the total outflow would match the new inflow and there would be more water in the dam — more heat on Earth.

Hotter objects emit more heat, and it’s the same at the outlet pipes. More water through the surface pipe implies more emissions from the surface, which means that the surface must be warmer, which means a higher “global temperature” (the average temperature of the air at the surface, where we live).

– Response to More Carbon Dioxide

What happens if the concentration of atmospheric CO2 increases? This is like impeding the flow of heat to space through the CO2 pipe with a partial blockage. (Btw, the flow in the CO2 pipe has declined about 4% since 1750. That is, from about 21% of total emissions to about 20%.)

The input to the dam is unchanged, so the total outflow remains the same when steady state resumes. So the effect of increasing CO2 is to redistribute the heat radiating to space—less from CO2, more from the other pipes.

The conventional basic climate model dates back to 1896, when climate data was sparse. People could estimate the solar response almost entirely from lab-based data. But those lab-based principles could not be used to directly estimate what would happen in the whole atmosphere (with clouds, rain and convection) if the radiation to space was merely redistributed.

So a fateful piece of reasoning was applied (in what has become a ten trillion dollar mistake — climate spending worldwide is about US$1.5 trillion annually): blocking an outflow from the dam was assumed equivalent to increasing the inflow by the same amount. The amount of water in the dam would be the same in either case. Switching from blocking outflow to extra inflow does not alter the net inflow, so radiation is still balanced. So it appears logical, doesn’t it?

So the conventional basic climate model calculates the surface warming due to increased CO2 as equal to the surface warming due to increased absorbed sunlight, where the increase in absorbed sunlight is the same as the reduction in emissions of heat to space by CO2. This is where the conventional architecture comes from.

It’s effectively the same in the large computerized climate models—the GCMs. While the GCMs treat an increase in absorbed sunlight differently to an increase in CO2 by taking many more factors into account, the end results are similar. The GCMs apply mainly the same feedbacks to extra CO2 as to extra absorbed sunlight, and calculate a similar surface warming for a given radiation imbalance (compare Fig.s 2 and 3 of post 17).

GCMs are a bottom-up model, trying to take everything into account and consequently drowning in uncertainty. But they are tuned to reproduce the warming of the 1970s to 1990s — which is assumed to be entirely due to increasing CO2, because the rate of observed warming is roughly the same as that calculated by the basic climate model. So, ultimately, the GCMs are tweaked to match the basic model.

But hang on! How can redistributing the outflow between the pipes be equivalent to adding more water into the dam? The amount of outflow is different! More sunlight mainly heats the surface, while extra carbon dioxide blocks some heat from being radiated to space from the upper atmosphere. They seem pretty different.

Generations of climate scientists have convinced themselves this logic is correct. What if they got it wrong? What if there is more to it than merely balancing the radiation?

The dam analogy instead suggests that if the CO2 pipe is blocked a little then the water would just back up a fraction then flow out the other pipes. The response of the heat is to reroute through all the other pipes. In particular, the rerouting feedback suggests it mainly reroutes through the water vapor pipe. (This is the opposite to what happens in the conventional models, both basic and GCMs, where water vapor amplification decreases the energy flow through the water vapor pipe as CO2 increases.)

The Alternative Model

An alternative basic model was developed in posts 11 to 18 that fixes these two major architectural errors in the conventional basic model. It allows for rerouting, and instead of applying the solar response (the climate response to increased absorbed sunlight) to the influence of CO2, it applies a response specifically for CO2. It also applies a radiation balance, so the radiation still balances.

(Imagine a plumber asking: you sure you want me to connect the CO2 influence to the solar response, or do I connect it to the CO2 response? “What CO2 response?” replies the conventional climate scientist, “Just apply the solar response to everything”.)

Confining our attention to just climate influences that change absorbed sunlight or CO2, the two models are exactly the same — except that the conventional model applies the solar response to the CO2 influence, whereas the alternative model applies a specific CO2 response to the CO2 influence. They differ by just one connection.

Basic climate models -- the crucial difference is where the CO2 connection goes

There is far more climate data available now than in 1896. When the alternative model is fitted with the data, it finds a much lower sensitivity to CO2—the UN’s IPCC overestimated future warming by a factor of five to ten. Less than 20% of the global warming of the last few decades was due to CO2.

Looking more closely at the data and the differences between the conventional and alternative models, there is no getting away from the centrality of the hotspot to the climate debate (though the supporters of the conventional model do not like to talk about it, and the public conversation oh-so-mysteriously omits it). When the Earth’s climate responds to increased solar radiation (“the solar response”) it causes the water vapor emission layer (WVEL) to ascend, thereby causing the hotspot. In the last few decades we’ve had strongly rising CO2; the conventional models apply the solar response to the influence of CO2 and consequently show a huge, prominent hotspot (like Fig. 2 of post 17). That hotspot, or more precisely the ascent of the WVEL, is the water vapor amplification that causes more than half of the warming in the conventional models, by constricting the flow of heat to space through the water vapor pipe as the surface warms. So it is extremely germane to the alarm over CO2.

But when the correct instruments for the job measure what is happening (30 million radiosondes, from the 1950s), they find no hotspot at all. Even when the satellite channels are optimized to find the hotspot, they cannot find it. The data seems reasonably clear: the WVEL did not ascend (see post 17 for more details). This is a crushing blow for the conventional models. Whilst the 30 million radiosondes might be in a conspiracy to fool the poor climate scientists, a simpler explanation is that the WVEL actually fell slightly in reality, and the solar response should not be applied to the influence of CO2 — the CO2 response causes the WVEL to fall, outweighing the surface warming and solar response which causes the WVEL to rise.

The alternative model finds the CO2 sensitivity is likely less than 0.15 °C per W/m2, compared to the solar response of 0.54 °C per W/m2 (respectively λC and SB of Fig. 1 of post 13). The solar response is much stronger than the CO2 response — that is, the surface warming caused by a given radiation imbalance is much greater if the solar response is applied than if the CO2 response is applied. So the one size fits all approach of the conventional architecture, applying the solar response to the CO2 influence, greatly exaggerates the warming effect of increasing CO2. The CO2 response is of course weaker because it includes its own response-specific feedbacks, such as the rerouting feedback, which are systematically precluded from the conventional model by its architecture.

The conventional model seemed to work for temperature (though not the hotspot) when the world was warming in the 1980s and 1990s, but has failed since then. A model that has the wrong architecture would act like this — correct sometimes by accident (especially if tuned to fit the data), but failing a lot of the time too, and no amount of hammering on the model can make it work all the time. It’s just wrong.

Parameters versus Architecture

Skeptics are usually skeptical because of empirical evidence that disagrees with the climate models. ‘Nuf said.

But why do the models get it wrong? Previously skeptics have questioned the three parameter values in the basic model — the total feedbacks, the reduction in radiation to space from CO2 when the CO2 concentration doubles, and the Planck constant (in order of decreasing doubts). But until now skeptics appear to have accepted the architecture of the conventional basic climate model — how those parameters are arithmetically combined to estimate the equilibrium climate sensitivity to CO2, the ECS.

Joanne and I were in that camp too until recently, thinking the problems probably lay mainly with the water vapor feedbacks as evidenced by the missing hotspot.

But it turns out it was the architecture that was badly flawed. Merely fixing the architecture, as in this series, brings the ECS into line with the empirical evidence and resolves the hotspot data (yes the hotspot is missing, because the CO2 response pushed the water vapor emissions layer down).

Establishment climate scientists are mainly believers because they believe in “the basic physics”. But they have assumed that the basic physics were applied correctly. Yes, the basic physics may well be about right, but it was applied incorrectly — the model architecture was wrong. It makes all the difference in the world.

Most everyone involved with the CO2 theory has been barking up the wrong tree  for years — or at least failing to bark up a very fruitful tree. Until now most everyone focused on the parameter values instead of the architecture, which could explain why this series is apparently novel in a mature field that has been fought over so intensely.

Of course the really productive action in climate research is in non-CO2 causes, mainly the province of the skeptical blogs. And that’s where we’ll be turning in the next blog post.


It appears the world has been led astray by the ideas of a primitive model, a first attempt at modeling the climate that doesn’t really work. This applies to the big computerized models (the GCMs) too, because they basically have the same two major architectural flaws as the basic climate model, and they are indirectly tuned to give roughly the same results.

Fixing the model architecture has the potential to bring most establishment climate scientists and skeptics onto the same page. Of course this is just a futuristic pipe dream and won’t happen any time soon, but the potential is there. And let’s not even consider the politics of it all; I’m just talking about the science.

It’s taken a few decades to be sure the conventional model doesn’t work, because climate changes only slowly. Good news: simply fixing the model implies the CO2 will never be much of a problem.

A religion based on a modeling mistake — let’s hear it for the modern world!

9.3 out of 10 based on 87 ratings