Why Climate Models Run Hot

By Rud Istvan – Re-Blogged From http://www.WattsUpWithThat.com

EPA administrator Pruitt wants to “Red Team” the catastrophic anthropogenic global warming (CAGW) consensus best reflected in the IPCC assessment reports (AR). At its core, CAGW rests on just three propositions:

1. CO2 is a ‘greenhouse’ gas retarding radiative cooling. This should not be in serious dispute since Tyndall experimentally proved it in 1859.

2. The Earth is warming. Although the details are in dispute because of temperature data quality problems and ‘adjustments’, the general fact is not. The Earth has been intermittently warming since the Little Ice Age (LIA) ended. For example, the last Thames Ice Fair was in 1814.

3. CO2 and its knock-on effects caused the recent warming, and climate models (such as the CMIP5 archive for IPCC AR5) predict this will continue to catastrophic levels. This is an extremely dubious proposition.

This guest post addresses proposition 3. It does so in a short sound bite ‘abstract’ useful for debating warmunists, and then in a typical WUWT full climate science guest post. It is a modest Red Team contribution.

Sound bite ‘abstract’

Climate models have run hot since 2000. Except for the 2015-16 now fully cooled El Nino blip, there has been no warming this century except by Karlization or (newly) Mearsation. Yet this century comprises about 35% of the total increase in atmospheric CO2 since 1958 (Keeling Curve). The climate models went wrong on attribution. The warming ~1920-1945 is essentially indistinguishable from that of ~1975-2000. AR4 figure SPM.4 said the earlier period was mostly natural (because not enough change in CO2). The CMIP5 archive assumes the latter period is mostly CO2 (and other GHG). That assumption is fatally flawed; natural variation did not magically cease in 1975.

Fully documented post

CMIP5 climate models have run hot since before 2000, and the divergence of CMIP5 from observations is highly statistically significant. Details are in Dr. Christy’s 29 March 2017 Congressional written testimony (available on line), from which Figure 2 provides sufficient up-to-date evidence.


This divergence is rooted in the attribution problem between natural and anthropogenic warming. It is unavoidably inherent in CMIP5 models for a very basic reason.

To properly model essential climate features like convection cells (thunderstorms), a grid cell needs to be less than 4km on a side. The finest resolution in CMIP5 is 110km at the equator; the typical resolution is 280km. This is because halving grid size requires an order of magnitude more computation. So adequately simulating such atmospheric processes from first principles is computationally intractable. Details are in my 8/9/2015 WUWT guest post “The Trouble with Climate Models”.

The solution is to parameterize such processes (for example, put a number on the probability of how many thunderstorms per grid cell per time step –a conceptual rather than actual example as parameters are a bit more complicated). Parameters are obviously just guesses. So they are tuned to best hindcast compared to observations; for CMIP5 the ‘experimental design’ was from yearend 2005 back three decades to 1975.[1]

Parameter tuning implicitly drags the attribution problem into CMIP5.

Dr. Richard Lindzen, MIT professor emeritus, first made the observation that the period of warming from ~1920-1945 is essentially indistinguishable to that from ~1975-2000. This is readily apparent visually, and is also true statistically.


IPCC AR4 WG1 figure SPM.4 and the accompanying text make clear that the earlier period (circled in blue) was mostly natural; there simply was insufficient change in atmospheric CO2 to explain the rise in temperature without natural variability. A portion of figure SPM.4 (readily available via the IPCC) is reproduced below as sufficient evidence.


The IPCC intent of AR4 WG1 figure SPM.4 was to convince policy makers that the second warming (circled in red) had to be AGW. But that IPCC logic is fatally flawed. The SPM did not tell policy makers about model parameter tuning, which clearly drags natural variation into the model parameter tuning period assumed by IPCC to be AGW. So the warming is falsely attributed only to CO2. Note also the subtle “cheat” in Fig SPM.4 of models using only natural forcings. The issue is not guessed natural forcings. We do not know why natural variation occurs, only that it does (no model of ENSO periodicy, for example). Natural forcings are not the issue; only the resulting natural warming variation is. Natural temperature variation, not ‘forcings’, is the proper statement of the attribution problem. The AR5 WG1 SPM makes IPCC’s erroneous and unscientific belief explicit:

§D.3 This evidence for human influence has grown since AR4. It is extremely likely that human influence has been the dominant cause of the observed warming since the mid-20th century. [Bold mine]

Natural variation did NOT stop in the mid-20th century. And that is why CMIP5 models now run hot.


Leave a Reply

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

WordPress.com Logo

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

Google photo

You are commenting using your Google 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