Climate Change: Responding to ‘No One Knows If It Is Us’
I was listening to a conversation where climate change came up, and one of the participants opined that while warming may be happening (it is tough to deny that now), we don’t know for sure that it’s the CO2 humans are pumping into the atmosphere that’s doing it. And at that point, I put in my two cents.
Exxon knew. They had an internal report with a climate model back in 1982 that made accurate predictions of CO2 concentration and temperature change for 2019, a 37-year horizon, based on their understanding of the effects of greenhouse gases.”
That pretty much shut down the conversation on that topic.
But I think that it is important to pass on this bit of knowledge in more detail, so I will do that here.
Back in 2019, I ran across a graphic from the 1982 Exxon internal memo. It showed time on the X-axis, CO2 concentration on one Y-axis, and temperature change in C on the other Y-axis. The CO2 concentration curve sweeps majestically up going from left to right, covering the late 20th century and going into the 21st, with plotting to about 2080. The temperature change curves sweeps up as well, displaced a bit. Whoever did the graph set the axes just so. Have a look…
When I saw that, something I thought would be useful would be to see just how Exxon’s engineers had done with their forecast. So I looked up the 2019 figure for CO2 concentration (IIRC, from the Mauna Loa Obsevervatory), and the 2019 temperature change figure. I straightened up the graph in CorelDraw and added lines for finer gradation to place the actuals. Plus, I got the 1982 CO2 concentration figure and added it. My marked-up version of Exxon’s graph looks like this:
It is a pretty stunning testament to the efficacy of numerical climate models. The absolute error on the CO2 concentration at the 37-year horizon is under 11 PPM. That’s under 3% error at thirty-seven years out. The temperature change prediction is similarly close; I’m estimating that at just over 3% error. (APE = abs(forecast – actual)/actual * 100)
What I knew about this from 2019 was that Exxon had produced a model internally in 1982 that made accurate long-range predictions on climate, and while they used it for their operations and planning, Exxon had since engaged in massive public relations campaigns to denigrate efforts to do climate modelling, sowing doubt about such models wherever they were offered.
A couple of months ago, I ran across a link to the whole 1982 Exxon internal memo. Reading through this is like reading a prototype IPCC report, but through something as distorting as funhouse mirrors. The people writing it up obviously had a wealth of knowledge of climate change factors, but… all of it is cast in terms of how long Exxon’s operations might go on unimpeded by pesky regulations and other responses by governmental authorities. There is even nuanced discussion of the timing of actions, as the engineers note that first the signals of climate change have to be detected from data, and then there would be a delay between detection and inevitable regulation telling them they had to stop. Yes, Exxon expected that at some point in the future, they would be required to reduce carbon emissions. They are, I think, still expecting that to happen, though nothing much has happened yet on that score.
And that brings me to Figure 9 from the 1982 Exxon internal report. The graph shows a horizontal band that represents normal climate temperature variability over time. It also shows the likely progression of temperature change given the modelled contributions of human-caused increases in atmospheric CO2 concentration, an upward departure that we all know now as the ‘hockey stick’. And the discussion in the memo is all about how detecting climate change from empirical data is difficult, so they expected this might occur around 2000, or another 18 years in the future from where they were writing this. They discussed how if detection happened later, then various of the bad consequences of global warming they knew would follow might be further off or somewhat reduced from their estimates, but if detection occurred before 2000, the obverse would hold, that things would change more quickly and be worse than in their estimates.
Michael Mann published on his detection of the ‘hockey stick’ in 1998.
Here is the 1982 Exxon internal memo Figure 9.
The Figure 9 ‘hockey stick’ isn’t exactly the same curve as the one in Figure 3. It may include the likely onset and effectiveness of regulations. If so, Exxon’s Figure 9 paints a pretty bleak picture of the future, as we will be continuing to increase CO2 concentration and associated temperature change, just not at quite the rate we were before.
To be sure, we are still barreling into the future on the twin rails laid down in the 1982 Exxon internal report, Figure 3. Maybe responses so far have contributed that 3% difference from the forecast values. But if we want to skip the worst outcomes of climate change, we’ve got to give Exxon and others what they’ve been expecting — and planning for — since the early 1980s: effective regulations that tell them to cease and desist, and to do that *now*.
For some time, I’ve been thinking that getting values off the Exxon 1982 Figure 3 graph is the first step toward a more thorough analysis of their work. I have a first pass at that, getting estimates of values from the graph for all four models (CO2 PPM solid and dashed, temperature change solid and dashed). I’ve also pulled in actual CO2 concentration values from GML (the Mauna Loa lab) via NOAA, and temperature change from NOAA, so I have the basis for at least generating some new graphs without doing all the fidgety overlay stuff on a copy of Figure 3.
It’s Exxon’s unlucky day as the media notices the blockbuster paper published in ‘Science’ on the uncanny accuracy of Exxon’s 1982 climate model.
Scientific American has a good report on this.
On the one hand, I was scooped. On the other, it’s a great paper, this is getting the attention it needs, and it is one project I can drop now.