Last week, the House Science Committee hosted a hearing to examine the scientific, economic, and policy issues with the recent U.S. contributions to the climate deal in Paris, Paris climate promise: a bad deal for America.

The chair of the House Science Committee, climate skeptic and conspiracy theorist Lamar Smith (R-TX), made sure that the science of climate change was on trial as well by bringing in John Christy, of the University of Alabama Huntsville and the state climatologist of Alabama, to give his views on climate science and data.

Christy testified that satellite measurements of temperature in the bulk atmosphere (up to 50,000 feet) show much less warming than the climate model simulations run for last IPCC report. The testimony was not that different from the one he gave in December in the Senate. His central argument comes from this oft-shown graph.

Screen Shot 2016-02-08 at 9.24.38 AM

Figure 1: Temperature in the bulk atmosphere (temperature mid-troposphere) reported as an average of climate model simulations (red), weather balloon data (blue), and satellite data (green). Source: House testimony of Dr. John Christy, February 2, 2016

By highlighting the divergence between the observations and the climate models, Christy argues that “the models do not accurately represent at least some of the important processes that impact the climate because they were unable to `predict` what has already occurred… as such, they would be of highly questionable value in determining policy that should depend on a very confident understanding of how the climate system works.”

Christy suggests that climate models warm up too much in response to increases in atmospheric CO2, or that the models are too sensitive to CO2. However, the distinction between models being too sensitive, unreliable for predictions, and completely wrong is important. Several members of the committee seemed to leave the hearing with the notion that satellite data disprove the role of CO2 in driving climate change, when in fact, the opposite is true.

But there is a credible case that climate model simulations are warming faster than satellite data show. The last time that Christy brought this analysis to the Hill, however, two other leading climate scientists with expertise in climate modeling (Ben Santer) and satellite data (Carl Mears) published a letter in response to the general argument that satellite data is a damning test for climate models, or disproves our understanding of CO2 and warming. The letter outlines three other explanations for the divergence seen in the satellite data. These are explanations that should be taken into account before assuming fundamental physical errors on the part of the climate models:

  • Errors in the satellite observations—Making a climate record (one that is accurate, precise, and consistent over decades) from satellite data is no small feat, and is still a work in progress. There is substantial uncertainty in long term trends estimated by satellite. I doubt that this is a strong explanation for the divergence Christy invokes, but it is important to keep that in mind.
  • Simulation input errors—The IPCC simulations that Christy used contained flawed and incomplete data on how much volcanic forcing should be applied to the modeling simulations. This error mistakenly warms the climate models, but doesn’t invalidate the CO2 hypothesis. If anything, it says that we should more carefully observe and study the agents which force the climate, and try to understand how they may impact future warming.
  • Bad luck—It could be that the variability in the real world is inconsistent with the climate models, and that difference is reflected in the satellite data. We’ve already seen that robust ocean heat uptake was most likely behind the hiatus in warming we saw in the past decades. Climate models are not designed to capture such variability, but that variability does not diminish future predictions of warming.

To what extent can each of these explain the differences between the satellite observations and climate model simulations? That is a question for valid scientific inquiry. Santer and Mears argue that some combination of these—along with errors in climate model physics—are likely behind the difference in figure 1. Alas, these possibilities were not addressed in Christy’s testimony in December or last week. And that left legislators with an incomplete picture.

There is a fair case that the Paris deal will result in little climate benefit. It has plenty of shortcomings. But the display in the House Science Committee last week did not make a convincing case that bad science is one of them.