by Ross McKitrick
Two new peer-reviewed papers from independent groups confirm that climate models overstate atmospheric warming and that the problem has gotten worse over time, not better. The papers are Mitchell et al. (2020) “Longitudinal plots of recent tropical temperature trends: Continuity model deviations in the context of internal variability” Environmental Research Letterand McKitrick and Christy (2020) “Pervasive warming bias in CMIP6 convection layers” Earth and Space Science. John and I didn’t know about the Mitchell team’s work until their paper was published, and they didn’t know about our work either.
Mitchell et al. surface view, troposphere and stratosphere over the tropics (20N to 20S). John and I look at the tropical and global lower and middle atmospheres. Both papers examined large samples of the latest generation of climate models (“Concatenated Model Comparison Project version 6” or CMIP6), i.e. the models being used for IPCC reporting. follow up and compare the model outputs with observations after 1979. John and I were able to test 38 models while Mitchell et al. reviewed 48 models. The sheer number makes one wonder why so much is needed, if science can handle it. Both papers consider “hindcasts”, which are reconstructions of recent historical temperatures in response to observed greenhouse gas emissions and other changes (e.g., aerosols and solar forced God). Across both papers, it was found that the models overcame historical warming from the near surface to the upper troposphere, in the tropics and globally.
Mitchell et al. Year 2020
Mitchell et al. , in an earlier study, examined whether the problem was with models that amplify surface warming too much as you increase altitude, or whether they are amplified vertically but start with surface warming is too much or not. The short answer is both.
In this Figure, the box/whiskers are the modeled predicted warming trends in the tropics (20S to 20N) (horizontal axis) relative to altitude (vertical axis). The place where trend strength crosses the zero line is where the stratosphere begins. Red = interior model that simulates both ocean and atmosphere. In blue: models take observed sea surface warming as given and simulate air temperature trends only. Black line: observed trend. The blue plots are still high compared to the observations, especially at 100-200hPa (the upper middle troposphere).
Overall, their findings were:
- “We found significant warming biases in the CMIP6 modeled trends, and we show that these biases are related to deviations in surface temperature (these models). this model simulates massive global warming which is not realistic).”
- “We note here in the record that from 1998 to 2014, the CMIP5 models warmed, on average 4 to 5 times faster than the observations, and in one model the warming was greater than 10 times the observations.”
- “Throughout the depths of the troposphere, there is no single realistic model that overlaps with all observed estimates. However, there is some overlap between the RICH observations and the lowest modeled trend, corresponding to the NorCPM1 model. ”
- “Focusing on CMIP6 models, we confirmed the initial findings of Mitchell et al. (2013): first, troposphere trends are modeled towards warming throughout the troposphere (and especially in the upper troposphere, around 200 hPa) and second, trends This may be associated with deviations in surface warming. Therefore, we did not see any improvement between the CMIP5 and CMIP6 samples”. (Mitchell et al. 2020)
A special prize will be given to Canadian model! “We drew attention to the CanESM5 model: it simulates the greatest warming in the troposphere, about seven times larger than the observed trends.” Government of Canada rely on The CanESM model “provides quantitative science-based information to inform climate change adaptation and mitigation in Canada and internationally.” I would be surprised if the modelers at UVic ever put warning labels on briefings for policymakers. The sticker must say: “WARNING! This model predicts atmospheric warming about 7 times greater than observed trends. The use of this model for any purpose other than entertainment is not recommended. ”
Although the above diagram looks encouraging in the stratosphere, Mitchell et al. see the model is also wrong. They predicted too little cooling before 1998 and too much after, and the cancellation effects tended to be linear. Longitudinal “fingerprints” of greenhouse gases in tropospheric warming and stratospheric cooling models. Models predict steady cooling in the stratosphere that should have continued beyond the late 1990s, but observations show no such cooling this century. The authors suggest that the problem is that the models do not correctly handle ozone-depleting effects.
The chart above focuses on the period 1998-2014. Compare the red box/beard with the black lines. The red lines are the outputs of the climate model after providing the observed greenhouse gases and other works for this time period. The predicted trends do not match the observed trend profile (black line) – there is essentially no overlap at all. They warm up too much in the troposphere and cool too much in the stratosphere. Forcing models to use prescribed sea surface temperatures (in blue), so there is a “correct” answer to the model for most of the surface area, minimizing the troposphere problem but not in the stratosphere.
McKitrick and Christy 2020
John Christy and I was earlier comparing the models with observations in the tropical troposphere, finding evidence of a warming trend in all models. This is one of the some papers I did about the tropical convection tropical bias. The IPCC cites my (and others’) work and accepts the findings. Our new paper shows that, instead of the problem fading in the latest models, it is getting worse. The aberration is observable in the lower and middle troposphere in the tropics but can also occur globally.
We tested the first 38 models in the CMIP6 group. Like Mitchell et al. we used the first stored run from each model. Below are the 1979-2014 warming trend coefficients (vertical axis, degrees per decade) and 95% error bars comparing models (red) with observations (blue). LT = lower troposphere, MT = middle troposphere. Every pattern exceeded the observed trend (blue dash line) in every pattern.
Most of the differences are significant at <5%, and the model mean (dark red) compared to the observed mean difference is very significant, meaning it is not just noise or random. Models as a group warm up too much throughout the global atmosphere, even over a period where both the column and the temperature can be observed.
We used 1979-2014 (Mitchell et al.) because that is the maximum period that all models are run with historically observed fortresses and all observation systems have available. Our results would be the same if we used the years 1979-2018, including script cuts for the last years. (Mitchell et al. report the same thing.)
John and I found that models with higher Equilibrium Climate Sensitivity (>3.4K) warmed faster (unsurprisingly), but even the low ECS group (<3.4K) showed show a warming bias. In the low group, the mean ECS was 2.7K, the average of the combined LT/MT model warming trend was 0.21K/decade and the observed counterpart was 0.15K/decade. This figure (green circle added; see below) shows a more detailed comparison.
The horizontal axis shows the warming trend of the model and the vertical axis shows the corresponding ECS model. The red squares belong to the high ECS group and the blue circles belong to the low ECS group. The filled shapes are from the LT layer and the open shapes are from the MT layer. The crosses represent the means of the four groups and the lines connecting the LT (solid) and MT (dash) layers. The arrows point to the observed mean MT (open arrow, 0.09C/decade) and LT (closed arrow, 0.15 C/decade) trends.
While the models in the blue cluster (low ECS) perform better, they still have a warming rate that exceeds the observations. If we envision a third cluster of models with a global average troposphere warming rate that overlaps with the observations, it would have to be roughly located in the region that I have outlined in color. green. The associated ECS will be between 1.0 and 2.0K.
Conclude
I understand that climate modeling is extremely difficult, and no one can fault the scientific community for finding a difficult problem to solve. But we are all living with the consequences of climate modelers stubbornly using generation after generation models with too much surface and troposphere warming, in addition to running excessive exaggeration required scenarios (e.g. RCP8.5). Back to 2005 in first report of the new US Climate Change Science Program, Karl et al. points to excessive warming in the tropical troposphere as a “potentially serious contradiction.” But instead of fixing it since then, modelers have made it worse. Mitchell et al. note that in addition to false warming trends, the deviations have broader implications because “atmospheric circulation trends depend on the temperature gradient with respect to latitude”. In other words, when the models misrepresent the tropical troposphere, it leads to potential errors in many other features of the model atmosphere. Even if the problem was initially limited to excessive warming in the tropical troposphere, it has now expanded into a more pervasive warming trend throughout the global troposphere.
If the difference in the troposphere is evenly split across the models between heating and cooling, we can attribute it to noise and uncertainty. But that’s not the case: it’s all about extreme warming. CMIP5 models heat too much at sea and too much in the tropical troposphere. Now, CMIP6 models overheat throughout the lower and mid-global troposphere. That’s bias, not uncertainty, and until the modeling community figures out how to fix it, the economic and policy-making communities are right to assume that projections of warming future is exaggerated, potentially very model dependent.
Presenter:
Karl, TR, SJ Hassol, CD Miller and WL Murray (2006). Lower-atmospheric temperature trends: Steps to understanding and correcting for differences. Summary and Product Reviews. Climate Change Science Program and Global Change Research Subcommittee
McKitrick and Christy (2020) “Pervasive warming bias in CMIP6 convection layers” Earth and Space Science.
Mitchell et al. (2020) “Longitudinal plots of recent tropical temperature trends: Continuity model deviations in the context of internal variability” Environmental Research Letter.
