Earlier today I blogged on an article at Daily Caller about Dr. William Happer’s resignation from the National Security Council. I pointed out that climate models simulate much more warming than actually observed over the relevant period.

Someone responded, “Complete and utter bull^*@%. The modeling is accurate.”

Ignore the vulgarity’s telltale sign that this wasn’t by someone with a particularly scientific mindset.

Is he correct in saying, “The modeling is accurate”?

One needn’t speculate to answer that question. One need only compare the models’ output with real-world observations. The numbers are there for anyone to check. Here’s how they look when graphed:

The bold black perpendicular line in the middle shows the average simulation of the 102 models for how much warming should have occurred from 1979–2018 according to the models. All the non-bold black lines show the actual simulations of each of the 102 models. Only 6 models had simulations that were at or below the observations according to Remote Sensing Systems’ (RSS) satellite data. Only 2 models simulated less warming than shown in the reanalyses, which many consider the most reliable. Only 1 model simulated less warming than shown in the University of Alabama Huntsville’s (UAH) satellite data.

The model average, at about 0.27C, is about 35% higher than RSS, about 69% above the reanalyses, and about 108% (more than double) above UAH. 46% of the models are above the model average, and 94% are above the best of the observations.

Here’s another way of showing the comparison between the models and the observations:

Again it’s clear that the models simulate much more warming than we actually observe in the real world.

Consequently, they provide no rational basis for predictions of future temperature. Certainly they can’t justify fears of dangerous to catastrophic manmade (or other) warming. Consequently, they provide no rational basis for any policy meant to respond, whether by mitigation or by adaptation, to future temperatures.