Scott Adams (of Dilbert comic fame) recently did a video about climate persuasion titled “Scott Adams solves the climate debate and saves the world (really)”, available here on periscope or here on twitter that has sparked folks to distill their arguments for and against climate change. After hearing about this and faced with the constant barrage of media stories about the latest inevitable climate doom story I thought it would be appropriate for me summarize five reasons why I think that the political agenda to transform the energy system of the world is not supported well enough by sufficient scientific evidence to proceed.
The majority of the claims that anthropogenic carbon dioxide emissions are the cause of the observed warming and that we only have a short time to do something or else are based on projections from global climate models (GCM). As noted below I do have relevant experience, education and background to inform my opinion. However, I believe that anyone who does research based on their personal experience, background and education can reach their own informed opinion of these claims if they actively try to get both sides of the story.
First a bit of background of these models. If you want details, Dr. Judith Curry did a detailed overview that includes a summary description. For my purposes all you need to know is that these models are a variation upon the meteorological models that provide predictions that everyone uses when you make decisions based on weather forecasts. There are differences but they use the same physical relationships such as momentum, conservation of heat and conservation of mass.
I have M. S. and a B. S. degrees in meteorology and in my fourth semester of Weather Analysis & Forecasting the laboratory assignment was to break off into teams and write a simple weather forecast model. I have been inside this kind of model but most readers also have some mathematical background that is relevant. In particular you may remember in algebra that if you have three equations you can solve for three unknowns but if you only have two equations you cannot solve them. The problem in meteorological models is that you have more unknowns than you have equations so model developers have to improvise. In particular, instead of using direct relationships for every single factor that affects weather or climate forecasts, meteorologists use parameters to simulate the effects of some atmospheric processes.
The first reason that I am skeptical of any GCM results is the use of parameters which can be thought of as “fudge factors’. Model developers necessarily have to account for some things that cannot be modeled directly. John von Neumann allegedly summed up the problem stating that “With four parameters I can fit an elephant, and with five I can make him wiggle his trunk”[1]. In other words, he could develop a mathematical model that described an elephant simply by fudging the parameters. Everyone who makes a decision based on a weather forecast has learned that you can trust a forecast for tomorrow better than one several days away. In the 43 years since I graduated the forecasts have become more reliable for dates further in the future because weather forecasters have constant feedback and have been able to adjust the parameters in the meteorological models to improve forecasts based on observations. There is only one global climate system and forecasts made today for 100 years away cannot be checked until 100 years have passed. One insurmountable problem is that the parameters and their use in GCMs cannot be verified as correct in our lifetimes.
Another issue with the parameters is that the focus is on just one of these parameters. Richard Lindzen commented on this:
“Here is the currently popular narrative concerning this system. The climate, a complex multifactor system, can be summarized in just one variable, the globally averaged temperature change, and is primarily controlled by the 1-2% perturbation in the energy budget due to a single variable – carbon dioxide – among many variables of comparable importance. This is an extraordinary pair of claims based on reasoning that borders on magical thinking.”
My final difficulty with parameters in GCMs is that they are used to model clouds. Dr. Curry explains that in order to solve the physical equations in a global climate the world has to be divided up into a three-dimensional grid. The equations are calculated for each grid cell and repeated to generate a forecast. My particular problem is that the grid cell size needed in order to do these calculations are on the order of 100 km horizontally, the vertical height is often 1 km and they do the calculations every 30 minutes or so. As a result, the models cannot simulate clouds. That single parametrization is a big enough driver of climate that this model component alone could dominate the GCM projections. This uncertainty is well understood in climate science by those who have worked with these models. However, the problems with parameterization is not well understood and its ramifications on the policy decisions is poorly understood by most of those who advocate eliminating fossil fuel use.
My second reason for not trusting these models is related to my experience in air pollution meteorology and work I did as a consultant to EPA evaluating the performance of air quality dispersion models. Complexity-wise those models are orders of magnitude simpler than climate models and they are simple enough to be directly verified. All air quality dispersion models are based on results from field studies that injected tracers into stack effluents, measuring the downwind concentrations in a test array under a wide array of meteorological conditions and then developing coefficients for pollution dispersion. I worked on a project where we compared the known emissions and observed ambient concentrations to model projections at power plants. The results showed that the models in use at the time adequately predicted maximum concentrations so EPA was comfortable that they were working correctly. The frightening result in my mind is that it was not uncommon for the model to predict a maximum concentration close to the maximum observed value but the meteorological conditions for the predicted maximum would be different than the meteorological conditions for the observed maximum.
Consider the differences between the GCMs and air quality models. Air quality models use parameters that are based directly on observations, incorporate known emissions, and have been extensively verified in field studies but those models could get the right answers for the wrong reasons. GCMs use parameters that are based on model developer opinions, have to estimate future emissions, and cannot be tested in the climate system. We are supposed to believe the models are getting the right answers for right reasons. I don’t think that is a reasonable assumption for modeling that is the basis for converting the entire economy away from fossil fuels.
My third reason for not accepting the commonly held belief that catastrophe is inevitable is that the projections that make that claim are only one answer in a wide range of potential outcomes. Consider for example that the Intergovernmental Panel on Climate Change (IPCC) does not give a single value for the sensitivity of atmospheric temperature to carbon dioxide. Instead they give a range of potential warming for a doubling of atmospheric concentrations of carbon dioxide of between 1.5 and 4.5 degrees C. Nic Lewis does a nice job discussion climate sensitivity here. The GCM projections cover a wide range of potential outcomes from benign to catastrophic. Without going too deep I want to point out that the damage claims for increased carbon dioxide depend on the shape of the distribution of this sensitivity. Damage affects society when costs are greater and those estimates are strongly affected by the probability of extreme damages. The first problematic aspect of this issue is that the projections for high impacts rely on a relatively high probability of extreme impacts. Although recent research has shown that the likelihood of extreme outcomes is lower than previously thought, those results have not been incorporated into the damage estimates. If they were considered, then the costs currently claimed would be reduced.
More problematic to me than the range of possible model outcomes is the use of the worst-case representative concentration pathway as the business as usual scenario. The IPCC developed a set of four pathways to concentrations representing the range of radiative forcing (i.e. greenhouse effect) in the literature for 2100. The range of possible future atmospheric forcing levels runs from relatively low levels of carbon dioxide to the highest forcing they thought possible. The problem is that these concentrations had to be related back to emission scenarios and the worst case representative concentration pathway with a forcing of 8.5 watts per meter squared is so high that the emission scenario necessary to get that level is not credible. Many of the really scary projections that get the headlines and dominate the narrative why we need to reduce carbon dioxide emissions use RCP 8.5 as business as usual. We already know that the likelihood of that future emission scenario is extremely unlikely. When coupled with models that give a range of outcomes from benign to problematic, I can only conclude that while it is not impossible that there could a catastrophic impact the probability is so low that it should not drive policy decisions.
My fourth reason for not trusting the models that claim carbon dioxide is the primary driver of the recently observed warming is that even for the limited results we have from the models that can be compared to climatic system they don’t do well. Dr. Curry has explained that inconsistency well:
Between 1910 and 1940, the planet warmed during a climatic episode that resembles our own, down to the degree. The warming can’t be blamed on industry, she argues, because back then, most of the carbon-dioxide emissions from burning fossil fuels were small. In fact, Curry says, “almost half of the warming observed in the twentieth century came about in the first half of the century, before carbon-dioxide emissions became large.” Natural factors thus had to be the cause. None of the climate models used by scientists now working for the United Nations can explain this older trend. Nor can these models explain why the climate suddenly cooled between 1950 and 1970, giving rise to widespread warnings about the onset of a new ice age.
The final reason that I believe that the political agenda to transform the energy system of the world is not supported by sufficient evidence to be credible is the suggestion that climate change is easy to solve because renewables are a viable solution. As convinced as I am that the climate science does not support this agenda I believe the suggestion that wind and solar can solve our energy problems is even more of an exaggeration. In fact that issue is the primary driver why I blog and have written so much about New York’s climate change plans. However, don’t listen to me, listen to Bill Gates who states “The idea that we have the current tools and it’s just because these utility people are evil people and if we could just beat on them and put (solar panels) on our rooftop—that is more of a block than climate denial,” Gates said. “The ‘climate is easy to solve’ group is our biggest problem.” Another problem is that the renewable “solution” very likely has very significant environmental impacts that are generally ignored. Michael Shellenberger had a Ted talk “Why renewables can’t save the planet” that addresses this issue.
To sum up. The rationale used to justify the need to convert the energy system of the world is that carbon dioxide will cause inevitable catastrophe and we can be saved if only we implement renewable wind and solar which will be easy to do. I don’t believe the science supports inevitable catastrophe because those projections are based on global climate models. Those models use too many fudge factors that can give too many results that can never be tested, much simpler models that are based entirely on observations can give the right answer for the wrong reason and the model results to date do not adequately predict the one climate experiment we can test. Most of the catastrophic outcomes that dominate the political and media narrative depend on an emissions scenario that is not credible. I do not believe that diffuse and intermittent solar and wind can be used to replace reliable and affordable electric power much less generate enough energy to convert transportation, heating, and industrial use of fossil fuel to electricity.
[1] Attributed to von Neumann by Enrico Fermi, as quoted by Freeman Dyson in “A meeting with Enrico Fermi” in Nature 427 (22 January 2004) p. 297
Pragmatic Environmentalist of New York 