My Second Reply to Senator Hatch

Below is my reply to this letter from Senator Orrin Hatch (R-Utah).  I’ve collected all my posts about Orrin Hatch here.

Dear Senator Hatch,

Thanks for your March 30 reply to my criticisms about your stance on climate change.  If you don’t mind, I’d like to keep trying to influence your views about this important issue.  As I told you before, I am a Republican who shares your aversion to increased government control, and a few years ago I might have been categorized as a “climate change skeptic” myself.  I wasn’t an activist, by any means, but I bought into some of the same objections about uncertainty in models, and so on, that you bring up.  When I really started looking into the issue, however, I had enough related expertise to realize that the skeptics I had been listening to were conveniently leaving out important information, throwing out red herrings, and focusing on grey areas having to do with details that wouldn’t really change the overall picture.  I’m hoping I can help you see what I did.

Many of the objections you bring up are telltale signs that whomever you’ve been listening to has been distracting you with arguments that, at best, are rabbit trails that lead away from the main issues, and at worst, are flatly untrue.  (I don’t mean this as a criticism of you, since I’ve been distracted by some of the same arguments in the past.)

The most important scientific issue is overall climate sensitivity–i.e., if the rate of energy input or output from the Earth is changed (by altering solar input, greenhouse gas concentrations, and so on,) how much will the average temperature change?  One can apply basic physics to calculate that if you double the CO2 concentration in the atmosphere, the temperature will rise by about 1.2 °C, and nobody (besides a few non-scientist crackpots) disputes this figure.  When the feedbacks in the system kick in, however, they can cause the temperature change to vary from this initial figure.  I gather from your website that you already understand this point, but here’s where you go wrong.  What really matters is not whether the models have all the individual feedbacks (like water vapor and clouds) exactly right, but whether the magnitude of the feedbacks as a whole is about right.  And this overall climate sensitivity is something that has been empirically estimated in a number of ways.  So in fact, it is “backed up with something more than hypothetical computer modeling,” as you put it.  Climate sensitivity has been estimated for the 20th century, for the last millennium, for the last several hundred thousand years, and even for the last several hundred million years, using different methods of estimating greenhouse gas concentrations and solar input, among other things.  Obviously, all these different methods are going to have different degrees of uncertainty, but the fact is that they all come up with central estimates that are pretty close to that of the IPCC–about 3.0 °C for a doubling of CO2.  They consistently indicate that the system is dominated by positive feedbacks.  If you would like to have a look at a good summary of this evidence, with references to the primary literature, check out the following URL.

http://www.skepticalscience.com/climate-sensitivity-advanced.htm

It’s true that there’s considerable uncertainty in the magnitude of the cloud feedback, for instance, but improving the representation of individual feedbacks in models is only likely to significantly affect the timing of temperature change, not the overall climate sensitivity, which has been empirically estimated.  Since you brought up the fact that a few scientists think the Sun has a much bigger role in climate change than the IPCC models indicate, you will no doubt be interested to read in the article I linked that Tung and Camp (2007) estimated climate sensitivity by looking at the response of the climate to the 11-year solar cycle.  And guess what?  The sensitivity they estimated was the equivalent of about 2.3-4.1 °C for a doubling of CO2, which is quite similar to the IPCC estimate of 2.0-4.5 °C.

A second tactic often used by climate contrarians is to throw out statistics that seem to contradict model projections, when in reality they either don’t really contradict the projections, or aren’t really the kind of thing the models are designed to project with any confidence.  Your response seems to indicate that you’ve been listening to these kinds of arguments, as well.  You argue, for example, “The IPCC models did not predict that NOAA’s satellite data would show Utah’s annual average temperature declining in the last 15 years, with a steeper decline in the last 10 years.  If human CO2 were the major climate driver, this would not be occurring.  At least, this is what many climatologists (including IPCC lead authors) have told me.”

To understand how your “IPCC lead authors” have misled you, you need to understand something about modeling complex systems.  That is, the models will ALWAYS be oversimplified in some respects, but that doesn’t mean they can’t be pretty good at some things.  Think of a bathtub, for instance.  It’s pretty easy to measure the average rate water is coming in from the spout, and it’s pretty easy to show that the rate water exits the drain is closely approximated by a simple function of the size of the drain hole and the height of the water above it.  Therefore, I can make a simple model of my bathtub that will be able to predict the average water level with striking precision if I know the average flow rate from the spout.  But if you want me to make a model to predict how my rubber duck will meander around the tub, then that’s a different matter!  For that, I’d have to take into account the exact shape of the tub, the positioning of the spout and drain, the hydrodynamic properties of my rubber duck, the exact starting position of the duck, and so on.  And even then, I probably couldn’t do a very good job, because the water isn’t delivered to the tub in a perfectly uniform stream–it varies chaotically.  So my model would have to be vastly more complicated, and the best I could probably end up doing is to give a not-very-precise range of possible outcomes for my duck.

It’s the same way with climate models.  If you want a model that can pretty accurately project the global average temperature evolution, given changes in solar input (analogous to water coming into the tub from the spout) and greenhouse gas concentrations (analogous to the size of the drain hole), then that’s pretty do-able.  But if you want to know exactly how ocean and atmospheric circulation patterns will change to affect specific localities like Utah, that’s WAY more complicated.  For one thing, the resolution of the best IPCC models is poor enough that a place the size of Utah would be represented by very few pixels.  At this course resolution, geographic features like mountain ranges, which have big effects on local climate, can’t be represented very well.  Consequently, nobody has ever claimed that GCMs should be that good at projecting climate change in a small, mountainous place like Utah.  You can read about some of these model limitations in the IPCC report.

http://www.ipcc.ch/publications_and_data/ar4/wg1/en/ch11s11-es.html

http://www.ipcc.ch/publications_and_data/ar4/wg1/en/ch1s1-5-1.html

Your choice of time periods for the trends you cited was similarly suspect.  Weather is chaotic at short time scales, so it isn’t very predictable.  But the chaotic fluctuations get averaged out over time, so that the average temperature in a certain locality over a year is much more predictable than the temperature on a certain day.  Likewise, the average temperature over the entire globe is much more predictable than that of a particular locality.  So how long does it take before all those chaotic fluctuations average out, and we’re talking about changes in “climate,” rather than changes in “weather”?  Climatologists generally look at time periods of at least 20-30 years, so 10-15 year time periods are still in the range that can vary a bit unpredictably, especially if you are talking about a small locality like Utah!

Let’s pretend your climatologist informants hadn’t made the mistake of using Utah (!!!) temperature trends to make their point.  What about the global temperature trend?  Well, the global temperature trend has been positive for the last 10-15 years, but it’s true that it hasn’t been as strongly positive as it was before that.  And you’re right–this is probably due in part to decreased solar activity.  If we look at shorter time periods (say 7 years), there have been some in the last while that had negative temperature trends.  Does this show that the models are wrong about what CO2 does in the atmosphere?  It turns out that the models aren’t good at predicting exactly when such short-term downturns will come, but they absolutely do predict that such downturns will come.  Take a look at Fig. 10.5 in the latest IPCC report, for instance.  (See the URL below.)  The individual lines in this figure show projections of different models (subjected to a 3-yr running average).  Note that the lines go up for a few years, and then go down for a few years at a time.  They mostly go up, but the point is that downturns lasting several years are completely consistent with the models.

http://www.ipcc.ch/publications_and_data/ar4/wg1/en/figure-10-5.html

Whoever was telling you that a short-term downturn in average temperature in a locality like Utah, of all places, is good test of whether “human CO2 [is] the major climate driver,” wasn’t being entirely forthright.  The fact is that something like that could be consistent with CO2 being a major climate driver, or with the opposite scenario, so it doesn’t really constitute a test of any real hypothesis.

The tendency to focus on questions that don’t get at the central issues manifests itself again when you say you’re holding out for “a conclusive explanation for the observed causal relationship problem of the Vostok ice cores”.  Why?  We know that if the temperature starts increasing for some reason (changes in solar input, or whatever) the ocean will expel some of its dissolved CO2, and it takes hundreds of years for the deep ocean to circulate up to the top to interact with the atmosphere.  That’s just basic geochemistry.  There are also other plausible processes that could explain why temperature would go up first, and then CO2 afterward, e.g., colder periods are drier, and so there is more dust in the air.  This dust contains iron, and can fertilize the Southern Ocean.  The algae and plankton that live there would thrive, using up more CO2, and sink to the bottom of the ocean, sequestering it.  There are still other processes that could have contributed, as well, but it’s hard to tell at this point which ones would have been the dominant causes.  But even if it’s impossible to tell what were the exact processes involved in creating the time lag between temperature and CO2 rises, this question doesn’t really address what adding or removing CO2 in the atmosphere did to enhance temperature changes during the glacial-interglacial cycles.  To address that, you have to do an energy balance calculation.  Luckily, we have a good idea how solar input was changing over this time period, we know how greenhouse gas concentrations changed, and we have a good idea what the extent of the ice sheets was (which would have affected the reflectivity of the Earth’s surface).  Yet again, if you plug all these numbers into the relevant energy balance equation, you come up with a climate sensitivity of about 3 °C for a doubling of CO2.  (See the article on climate sensitivity I linked above.)

The vocal fringe you’ve been listening to is constantly whining about their papers being rejected due to bias, but in my experience, every one of these charges I’ve checked into has turned out to be bogus.  Those rejected papers really were terrible, in other words.  For instance, Roy Spencer wrote a whole book because he couldn’t get one of his papers published, but when I took apart the model his work was based on, I found out that he was using a nonsensical statistical technique that allowed him to obtain almost any answer he wanted.  And he did.  Following are links to my analysis of Spencer’s model, as well as a page at the Skeptical Science website where charges of bias in the review process based on out-of-context quotations from the “Climategate” e-mails are examined.

https://bbickmore.wordpress.com/2011/02/25/roy-spencers-great-blunder-part-3/

http://www.skepticalscience.com/Peer-review-process.htm

I ask you again to step back and look at the trend I’ve been pointing out throughout my correspondence with you.  If a single contrarian paper is published, it’s enough for you to pronounce AGW dead, even though you don’t have the expertise to assess it.  If less than 1% of the IPCC scientists dissent about the IPCC’s conclusions, it’s enough for you to dismiss the overwhelming consensus of experts.  If any natural variations have ever occurred, then you just assume recent changes have been naturally caused, too, even if no natural cause can plausibly be identified for the present trend.  If someone proposes action that won’t immediately and completely solve the problem, you justify yourself in neglecting to take any action at all.  If you are told that there are still significant grey areas about the details of climate dynamics, you take that to mean we can’t have any real confidence in the overall results of climate models.  After all, you buy into the assertion that the climate models themselves are all we have to constrain overall climate sensitivity, which is flatly untrue.  You think local, short-term temperature trends should be an adequate test of whether GCMs are accurately projecting what they are supposed to, which is also false.  You make a big deal about the timing of temperature and CO2 changes in the glacial/interglacial cycles, even though overall climate sensitivity (the really important issue) isn’t really addressed by this.  Finally, you continue to avoid admitting a scientific consensus exists on the issue, but at the same time buy into the idea that the skeptics are being shut out of the peer-reviewed literature?  How could that happen, unless there really were a strong consensus?

When you add all this up and boil it down, it’s hard to avoid the following conclusions.  1) Even if you’ve tried to get up to speed on the science of climate change, there are still very large gaps in your understanding of the basics.  2) Some of the scientists you are getting your information from have taken advantage of these gaps to mislead you about the state of the evidence regarding the fundamental scientific questions.  3) You have been susceptible to this manipulation by this tiny minority of climate scientists because it’s more convenient, in terms of our shared political ideology, to believe there is no climate change problem to solve, since most of the solutions proposed so far have involved greater government regulation.  Or at least, you find it convenient to “keep an open mind,” which allows you to do essentially nothing while avoiding head-to-head confrontations with the majority of climate scientists.

You confirmed this last conclusion when you said, “As a policymaker, I am always wary of any effort to centralize control of human activity.  And I am genuinely skeptical that turning control of human carbon emissions over to centralized powers is useful or desirable at this time.”

My challenge to you is to be a real conservative, rather than simply a right-wing ideologue.  A real conservative confronts problems, and tries to find solutions that involve the least possible government interference.  Right-wing ideologues try to pretend there are no problems for the government to solve.  Don’t believe a tiny minority of scientists just because you are pre-disposed to want reasons to doubt the consensus.  There are always reasons for doubt, so a few grey areas in the science should not be sufficient to allow us to ignore clear evidence that we are subjecting ourselves and our children to grave risks.