Things we don't know - about climate

(Modifed from a draft published on

I drafted this article on November 19th, 2010. At about ten that morning the weather channel, which gets its data for Lethbridge, Alberta from environment Canada and thus ultimately from sensors less than ten kilometers from my house, said the temperature was -17C. At that same time, however, the sensors about four feet above my roof reported a temperature of -19.2C.

By coincidence, and again according to the weather channel, the all time record low for November here, -35.6C, was set on that same day in 1921.

The source number for that claim, presumably 32.1F, is actually an interpolation from various agricultural research and military facilities across southern Alberta, because the airport weather station has been moved a few times and many of the source records lost - but it should be obvious in any case that neither the thermometers in use at airports in 1921 nor the processes in place to record temperature supported anything like that level of precision.

So how cold was it here before I left that morning? there's really no way to know - and how did that compare to 1921? I don't know that either.

What I do know is that the values shown were averages taken over time; that neither instrument is predictably accurate to even one decimal place; and that the air between the two is of variable depth, variable humidity, in constant motion, and had markedly less than one chance in twenty-two of being at a real average temperature of -18.1C that day.

So how does this extrapolate to sticking a thermometer into the troposphere to estimate our planet's near ground air temperature? Well, in total the world has less than one sensor for every sixty thousand square kilometers; about three quarters of them are closely grouped in the United States, western Europe, and the militarily significant part of southeastern Russia; almost none have trustworthy time-of-readings records for more than a few years; most of the records are both short and discontinuous; most of the readings are accurate only within loose bounds; and an unknown proportion of the time series supposedly formed from instrument readings contain unknown interpolations.

There are other sources of information. For example, weather satellites have produced records for perhaps half the earth's surface since about the mid seventies - but those records too have unknown source errors; may now contain accumulated and largely undocumented differences from the source data; show significant coverage bias favoring areas important to civil aeronautics; and are generally accessible only in the form of time series whose values are derived from real measurements pertaining to the upper troposphere through calculations calibrated against the same ground sensor readings they're used to extend and correct.

In contrast many of the proxy records are both long and internally consistent - but they don't help because these are very coarse grained: whether they're based on isotope decay or tree rings, the best "rulers" these produce are location specific and marked in decadal or century intervals, not globally applicable and marked in seasons or years.

The bottom line on this is simple: I can't pretend to know the temperature within a few kilometers of my house right now to within a couple of degrees C without making basic scientific errors in everything from measurement and imagined precision to application - and when people like Jones and Hansen announce in all apparent seriousness that the entire earth is now 0.5C degrees warmer than it was during the period from 1961 to 1990 they're asking us to accept a very precise number on the basis of data that's much worse than mine and in the face of applicability, measurement, and computational ambiguities that are orders of magnitude greater.

There seem to be two arguments for not dismissing their claims as nonsense. First, that we don't need to know the atmosphere's temperature now because climate science is about change and X + 0.7 degrees will have visible effects regardless of the value of X. The Polar bear, for example, will go extinct and Manhattan will flood - except that we're pretty sure the medieval warming period was just one of many such in history and not only did the polar bear make it through those embarrassingly undead, but what's known of civilizational history in estuaries and around tidal basins from the Thames to the Yellow does not suggest the existence of longer term human noticeable flooding during any of those extended warm periods.

Second there's the Foundation myth: the belief that it's possible to predict the direction and extent of motion of something like a collection's center of mass (or the chartrist's Dow Jones average) without knowing anything about the motion of the individual units involved - or, in other words, that we can predict where a herd of cattle will go when stampeded without needing to know where they started, how many there were, what frightened them, much about the land they're on, the direction each animal starts in, or even whether they're actually cattle.

The Frank Slide took place on April 29th of 1903, about an hour's drive from here when an estimated 90 million tons of limestone tipped off Turtle Mountain to bury the people, their town, and the railway beneath an estimated two kilometer rubble run-out. This slide hasn't moved much since, has been extensively studied, is comprised of materials for which the basic physics of motion and energy transfer are well understood - and yet the best we can do in terms of placing its center of mass in 3D is plus or minus about twenty meters - roughly on the same order of accuracy as predicting yesterday's temperature in Lethbridge to within one degree.

Basically the Foundation idea is intuitively obvious and makes for great science fiction, but the reality of any analysis aimed at actually making it work is that you need a secure grip on starting conditions, an understanding of the physics of change, strong boundaries on the range of change, and a small enough data set to make the simulation computationally feasible - so if you've ever wondered why the best known climate models come down to thirty or forty years of encrusted tinkering you now know: these models are continually adjusted to predict their own inputs, but cannot reliably predict excessions because the underlying climate science does not meet any of the conditions required for this kind of modeling to work.

So what do we know? We know that many of the people warning us of the horrible consequences of human caused global warming haven't been the disinterested scientists they've pretended to be - basically from Hansen and Jones to Gore and Waxman most of the more deeply committed have shown themselves deeply corrupted. That's sad, but even sadder is the hidden reality: that knowing Mann and Bradley made up the hockey stick to defend a lie doesn't tell us anything about global climate change - it just tells us things we didn't want to know about them.

Most people, of course, know the numbers don't work but rationalize accepting alarmist conclusions anyway because they think that "greenhouse science" - the belief that increasing greenhouse gases in the atmosphere will cause traumatic global warming - is settled; and so see the lack of response to increasing atmospheric CO2 in weather data as a reflection on the quality of the data, not the theory.

Basically these people assume the wolf to justify the alarm: picturing Gore et al as yelling "Wolf!" because "greenhouse science" proves the wolf - and then excusing the business of rather obviously drawing improbable conclusions from inadequate data as laudable and necessary moral sacrifice by experts committed to rousing the rest of us to action.

Unfortunately the science on greenhouse gas effects is not only not settled, the claims made for it seem rather more likely to be wrong than right. Specifically, the supporting proposition for the usual assertion (that human action is causing negative climate change) is that the planetary atmosphere will trap more solar energy, thus causing atmospheric heating, when it contains relatively more greenhouse gases, particularly CO2, then when it contains relatively less is asserted by many but completely unsupported by data or experiment.

The classic pop sci demonstration involves heating two jars: one containing normal air, the other CO2 enriched air, and discovering that the latter holds more heat longer than the former - but this is an artifact of density, not climate magic: if you use an inert gas like argon instead of CO2 to raise the amount of matter in the target jar, you get exactly the same effect.

Charlatanry aside, the main issue is that air in a jar does not behave like air in an atmosphere and thus that what happens in the jar is not a guide to what happens in the atmosphere. The key difference is that the material in the CO2 enriched jar is of a fixed mass, in a fixed state, and there is no expectation that its energy absorption and retention rates will change during the experiment. Imagine glimpsing the earth from some significant distance and it can look just like that: a gravitational container filled with air and a bit of heavier stuff in the center. But up close, time passes and things happen: water and greenhouse gases move into and out of the atmosphere, mixing occurs at different rates both vertically and horizontally, some surfaces are net radiators, others net absorbers - overall the longer term energy balance seems to work, but many of the specifics and nearly everything about the rates of change involved, are neither understood in the science nor modeled in the jar experiment.

Or, to put it another way, the biggest difference between the jar experiment and reality is that in the real world there's only one jar: i.e. the CO2 introduced into the planetary test jar comes from the planetary test jar. Thus it's true that the materials in the planetary jar change state over time - trees grow, coal burns - but because the total mass in the atmosphere is very nearly constant, the assumption that the input energy is roughly constant means that the total amount of heat energy the entire system can hold in long term equilibrium against the space around it has to be close to a constant too - and thus that short term (decadal) fluctuations can occur, but that a glaciated world cannot become tropical, or a tropical one glaciated, without significant changes in energy input.

So if we can't believe in the data, the people, or the "settled science", what can we believe? Perhaps that a hypothetical Canadian Canute party offering a credible commitment to end winter would win in a landslide? Or, more seriously, that all the fuss about whether or not humans are influencing global climate change has allowed the alarmist lobby to insert an obvious falsehood into the public consciousness on this issue: the belief that even minor global warming will produce terrible harm when what we know of both history and biology says the contrary is far more likely to be true.

By 10 PM on the evening of November 19th, for example, it was about -27 here with the wind chill dropping that down to an effective -40 something: an environment just as much the opposite of the green and fecund jungle most of the earth's life has evolved in as the driest deserts in north Africa, central Asia, and Australia.

Come spring the area around here will go green with rain and erupt with life: people in our parks, ducks on our lakes, fawns in our coulees - and the water cycle effects that might well go with even a few degrees increase in "average" atmospheric temperaure worldwide might do the same for the roughly one third of the earth's potentially arable land that's now too dry or too cold for agriculture.

So there's something else we don't know: why do "greens", people who profess to favor life and bio diversity in all its forms, so strongly oppose change most likely to strongly favor life and bio-diversity?

The obvious answer, that many of the leaders involved are merely using environmentalism as a handy bludgeon for the achievement of unrelated political or monetary goals, may well be correct, but is merely an ad hominem argument allowing us to dismiss them while telling us nothing about either the desirability or reality of anthropogenic global warming.

So when you get down to it, what we know about global warming is pretty much nothing: we've no baseline, so don't know if it's happening; we've no cost/benefit evaluation so don't know whether it would be net positive or net negative; if it is happening we don't understand its causation and if it isn't we don't understand why not; so really the only thing we're pretty sure of is that the people jumping up and down screaming that they have the answers are either deluded or charlatans.

Questions? Comments? Suggested fixes? murph at winface com, please

Paul Murphy wrote and published The Unix Guide to Defenestration. Murphy is a 25-year veteran of the I.T. consulting industry, specializing in Unix and Unix-related management issues.