1. Global average temperature has risen and fallen cyclically, driven by cycles in solar energy output, solar magnetic wind output, galactic cosmic ray influx (moderated by solar magnetic wind), ocean cycles, cloudiness (influenced by all of the previous cycles), volcanic cycles, and other natural causes, with both the rate and the magnitude of temperature increase and decrease considerably exceeding the allegedly anthropogenic warming of the late twentieth and early twenty-first centuries. It is impossible to rule out these natural causes as contributors to recent warming.

2. Global average temperature (GAT) has risen about 1–1.2 degrees C since about 1850, about half of that rise occurring before 1940, before anthropogenic additions to atmospheric CO2 concentration could have contributed significantly. This, too, implies that it’s impossible to rule out natural causes from the warming of the late 1970s through late 1990s (since when GAT has been relatively stable) that is the primary basis for claims of rapid anthropogenic global warming.

3. The range from nighttime low to daytime high and wintertime low to summertime high temperatures at most locations around the world is 10, 20, 30, 40, or more times the entire increase in GAT since 1850. In light of that, Dr. Richard Lindzen of MIT, one of the world’s most prominent climate scientists, says the proper response to the post-1850 warming is “So what?”

4. According to the theory that provides the basis for the computer climate models, greenhouse gas-driven warming should occur primarily toward the poles, not the equator; primarily in winter, not summer; and primarily at night, not in day. I.e., it should raise low temperatures more than high temperatures, but, because toward the poles, should not result in significant acceleration in the melt of land-based ice in the Arctic and Antarctic regions that could lead to accelerated sea-level rise. That expectation is consistent with observations. This means longer growing seasons and the expansion of cultivable land into higher latitudes, meaning more food.

5. That increasing atmospheric CO2 concentration should make GAT higher than it otherwise would be is basic physics, and critics of belief in dangerous manmade warming don’t challenge it. But it’s also basic physics that if you drop a rock and a feather at the same moment from the same height, they’ll land at the same moment—unless they’re in air, in which case the rock plummets while the feather wafts slowly downward, and if it’s windy the feather might blow up into a tree and never come down. This illustrates the fact that basic physics tells us very little about the real world, because the real world is a very complex place, and the climate system is probably the most complex thing we’ve ever studied other than the human brain and DNA. The interesting question, therefore, is not whether our CO2 (and other greenhouse gas) emissions make GAT higher, but how much higher?

6. To answer that question, first we do the basic physics (the Stefan-Boltzmann equation) about black-body radiation, from which we infer that every doubling of atmospheric CO2 (or CO2-equivalent) concentration should cause about 1 to 1.2 degrees C of increase in GAT without any feedback mechanisms. (Some scientists, due to some thermodynamic and mathematic issues, think this initial, pre-feedback warming is actually much smaller, but this is adequate to work with for this discussion. If they’re right, all that follows about climate sensitivity—how much warming comes from adding CO2 and other greenhouse gases to the atmosphere—would be strengthened. Warming to come would be less, hence all risks associated with it would be less, and cooling due to reduced emissions would be less, so the cost per unit of cooling would be higher.)

7. The computer climate models on which alarmists like those associated with the United Nations Intergovernmental Panel on Climate Change (IPCC) and various government agencies that depend on it predict post-feedback warming from doubled atmospheric CO2 (or CO2-equivalent) concentration of 1.5 to 4.5 degrees C, with 3.0 degrees remaining, in most discussions, their “best estimate.” 3.0 degrees would require that net feedbacks multiply pre-feedback warming by 250 to 300 percent.

8. According to the models, GAT should have been rising by about 0.216 degree C per decade over the relevant period. Actual observed change in GAT has been only about 0.117 degree C per decade. I.e., the models call for nearly twice the warming actually observed. (And this rests on data through 2016, which, like 1998, was abnormally warm because of an abnormally strong El Nino. If the data since the post-El Nino cooling were added, the difference would be greater.)

9. Additionally, the models say all that warming should have come only from increased CO2. If any of it came from natural causes, then their error is even greater.

10. But recent research finds that once you control for solar, volcanic, and ocean current (especially El Nino/Southern Oscillation) variability, all of the temperature variability throughout the relevant period has been explained, leaving none to be blamed on increased CO2.

11. This doesn’t mean CO2 doesn’t make GAT higher than it otherwise would be—as basic physics predicts. Instead, it helps us answer the key question, “How much?” And the answer appears to be: so little we can’t detect it.

12. It follows that if CO2 isn’t the main driver of global warming, reducing CO2 emissions won’t have much effect on future GAT—indeed, if CO2’s warming effect is so little we can’t detect it, then the effect of reducing it will also be so little we can’t detect it.

13. But even assuming the modelers’ estimates of CO2’s warming effect, full compliance with the Paris accord would reduce GAT in the year 2100 by no more than 0.17 degree C, which is far too little to have any significant effect on any ecosystem or on human wellbeing. It follows that there is no temperature benefit to be anticipated from the Paris accord.

14. Meanwhile, the cost of implementing the Paris accord would run in the range of $1 to $2 trillion per year from 2030 through the end of the century, for a total of $70 to $140 trillion dollars, making the cost per tenth of a degree C of cooling roughly $41.2 to $82.4 trillion.

15. Those trillions of dollars would, in other words, achieve no significant difference in GAT. But if instead they were spent to provide pure drinking water, electricity, sewage sanitation, nutrition supplements, infectious disease control, industrialization, better housing, etc., to the roughly 2 billion people in the world who lack those things, the improvement in human health and wellbeing would be immense.

And that—even before we get to all the arguments about the effect on the U.S. economy—is why it made perfect sense to pull out of the Paris accord. The rest of the world should follow.