Think they’re still using the same thermometer in exactly the same location as 1895? If you do, I’ve got a bridge to sell you.

Awakening once again to NPR, I heard crime news – the serial killer believed dead in Florida and some new development in the Ira Einhorn case, possibly an extradition fight. Science was making news this morning, too. Evolutionary biologists have come up with a neat new explanation of the Cambrian explosion, which has been a thorn in their side forever. Apparently, the whole earth flopped over on its side 500 million years ago and somehow made random genetic mutations at the cellular level happen faster(?) Funny we hadn’t heard about this before. You’d think an earth flop-over would have been discovered by the guys who know so much about tectonic plates. I also heard an extended NPR segment on the Greenhouse Effect – a.k.a. Global Warming – which has made the usual invisible transition from hypothesis to scientific fact. As a result, the president has decided we all need to worry about this. It sounded like a remedial seminar on the subject had been conducted at the White House, with the Pres taking on the role of simple-minded questioner while various scientists played the role of patronizing know-it-all. I got the impression we’re all supposed to be feeling guilty because we still get in our cars and drive to work.

There must be some evidence in support Global Warming, but the only one the mass media like to cite is far from convincing to me. This has to do with a reported rise in average temperatures of one degree (Fahrenheit, I think) during the last hundred years. No expert in climate, I’m willing to concede their argument that one degree has pretty serious implications. It’s how they get to the one degree that leaves me a little skeptical.

Let’s think about this for a minute. What is the ‘average’ temperature on earth right now? Yes, I mean at this very moment. One hundred two degrees, as the thermometers in Arizona might report? Fifty below, as the ones in Antarctica would claim? Neither, obviously.

It’s not as if there’s one definitely correct number that represents the answer to this question. The word ‘average’ always means that we’re going to perform some calculation. To begin with, the discipline of mathematics gives us at least three different definitions of what an ‘average’ is. The ‘mean’ is the arithmetic average, which we calculate by adding up all individual instances of something and then dividing that total by the number of instances. The ‘median’ is a function of counting – we take all individual instances of something, then count up from the bottom until we reach the halfway point. The ‘mode’ is the most common number found in all individual instances – we gather together all the instances of something and see which value occurs most often.
I apologize. I know this is boring, but it’s got to be important. The scientists are talking about the melting of glaciers, the flooding of thousands of miles of coastline, the forced migration of major populations, the devastation of our agricultural equilibrium, and dozens of other effects of their one degree ‘average increase.’ So there’s a quite valid reason for asking whether they’re as certain as they sound.

Back to the math. All the definitions of ‘average’ assume that that there is some finite number of instances to be used as the basis for calculation. In the case of temperature on earth, this is not strictly true. The atmosphere is made of gases, not subject to counting like dollars or stones. It must be that we can artificially create enough instances by the act of measuring to eliminate the difference between gases and stones. How do we do that? Is it sufficient to record the airport temperature of New York, Chicago, and Los Angeles, add those temperatures together and divide by three? Probably not. Maybe we need to add Paris, London, Tokyo, Moscow, Sydney, and Little America in the Antarctic. Would that do? Again, probably not. That leaves out a lot of places, and measurements in the city are tricky anyway, because artificial structures like asphalt paving have a tendency to soak up additional heat. So we’d better add in a bunch of pure countryside and farmland – put some of our thermometers in fields, forests, mountains, ocean-top oil rigs, deserts, prairies, and plateaus. Still, this doesn’t tell us much about how to weight the number of instances we measure, so that we balance arctic and Antarctic cold properly against tropical and temperate zones. And even then, we’re taking a lot for granted – having read Admiral Byrd’s Alone, I’ve learned that temperatures vary pretty considerably only a hundred or so miles apart in the Antarctic.

I suppose we’re going to have to concede that whatever number of instances we record, the ‘average’ number we arrive at is not necessarily going to be objectively ‘right.’ Because no matter how many thermometers you have out there, say one hundred thousand, you’d get more accurate data if you put another million in the spaces in between the hundred thousand, and more accurate data still if you put another hundred million in between those. It doesn’t take a weather wizard to know that the temperature can be at least a little bit different one hundred yards from where you’re standing now. Which would be the right number for the location listed under the name of your home town? Is that in the shade? In the sun? Or somewhere in between. You decide.

Considering all this, it looks as if we’re computing some theoretical average which we must assume bears some definite relationship to the objectively ‘right’ number we can’t measure. Which is another way of saying we’re sure the amount of our unmeasurable and uncorrectable error will never change. Everyone happy so far?

But the Global Warming hypothesis depends on far more than our theoretically correct though ‘not right’ average temperature on earth at this moment. The one degree change we’re looking for has occurred over one hundred years. This must mean that our theoretically correct number is actually determined by the number of instances – and the standard of measurement precision – that was already established in the year 1897.
Eighteen hundred and ninety seven. William McKinley was President of the United States. The automobile was a curiosity that frightened the horses. The continents of the world were connected by steamship travel and the telegraph. Charles Lindbergh hadn’t been born. There weren’t any airports anywhere. The North and South Poles hadn’t been discovered yet. But the worldwide temperature recording system was already in place.

This means, for example, that the New York City measurement has to be coming, year after year, not from the state-of-the-art instruments at LaGuardia, but from a thermometer that’s been religiously maintained on the lefthand tower of the Brooklyn Bridge. I hope nobody accidentally broke and replaced that thermometer at any point during the last hundred years, or moved it to the righthand tower, or forgot to record the readings while they were away on vacation for a month, or ever made up any readings because they got behind or just didn’t care enough during that ugly divorce in nineteen- ought-seven. Because the one degree change we’re after is less than two percent of the theoretical average, which is already just a bit flimsy as a computation strategy. Bad data would ruin everything. Equipment changes, human carelessness, or changes in measurement location might invalidate the numbers completely, and that would never do because we’re talking semi-apocalypse here.

You have to admire the discipline of science. To think that they were able to assemble all the thermometer readers all over the world in 1897 and train them to be unfailingly accurate and reliable is pretty impressive. To think that over the whole hundred years, no Tibetan shepherd ever said, ‘oh, about thirty-two degrees,’ when – thanks to his untreated nearsightedness – he was inclined to guesstimate a likely reading for those pesky western meteorologists. Amazing.

But the most astounding thing of all is that this degree of accuracy has been achieved in a field whose practitioners claim is not an exact science. Meteorologists who can’t tell us for sure if the tornado they’ve sighted is going to mow down my hometown or the City of South Bend, Indiana, are certain they know what the average temperature on earth will be forty years from now. This is made all the more miraculous by the statistical concept of standard deviation – meaning the amount of normal built-in variability – which is pretty high when it comes to temperature. That’s why we continue to set record highs and lows in temperature on individual days in every single year. Christmas in New York can be as warm as sixty-five degrees Fahrenheit or as cold as ten below zero. It’s this kind of variability that makes it difficult even to compare seasonal averages. Was last summer five percent cooler than this summer? In my neck of the woods we had more cool days last year but hotter hot spells. How should I compare this year to last year in terms of average. Who the hell knows for sure?

All we do know for sure is that it’s one degree hotter in summer, on average, last summer aside, than summer was, in general, a hundred years ago. Or is it winter that’s getting warmer instead? Like the one a few years back when the northeastern U.S. got raked by five ice storms of a severity not seen since they began taking weather measurements. Which reminds me – how long has that been? Of course. About a hundred years. The temperatures on earth have been pertinent to the Global Warming question since the end of the last ice age about ten thousand years ago. This means we’re depending on data from one percent of the relevant time period to calculate the standard deviation. And the standard deviation we come up with has to be so dead-reliable that it can be used to verify a less-than-two percent change in ‘average’ temperature.

Scientists like thought experiments. I have one I’d like to try on them. Ask a friend to record the mileage of all (or most) trips he takes in his automobile during the last week in December. Then calculate the percentage change in length of trip, up or down, from the beginning of the week to the end of the week, and use this number to project the average length of an automobile trip on January second. Now: would you bet your life that this prediction will be accurate within one mile? Really?

There’s always the possibility, I guess, that scientists are citing the temperature change ‘evidence’ to us because we’re too stupid to understand the real evidence. I know they’ve been busy calculating the number of tons of carbon dioxide in the air, and they’ve got their chemistry down cold – except, of course, when the number of variables gets too large. Which is the only reason their projections about how much impact atmospheric events like volcanic eruptions have on the earth get a little overstated at times. Or am I wrong about that? Was I mistaken when I heard the dire prediction that the area surrounding Mount St. Helen would be a wasteland for decades? But maybe what I’m wrong about is the extent to which the area has already recovered from the devastation of the eruption.

You see, not being a scientist, I can’t prove anything. My duty is therefore to shut up and nod vigorously when the scientists talk. And then to feel ashamed and fearful because I’m not doing anything to prevent the environmental catastrophe I’m causing by driving to work, buying a Christmas tree once a year, and exhaling carbon dioxide every day. I know I should prefer the worldwide depression that would follow the prudent shutting down of the entire fossil fuel industry and all the markets and products and jobs that flow from it. I know I should.

One of the scientists at the President’s Global Warming Nursery School said that those of us who don’t care about the Greenhouse Effect are like passengers on a bus bound for disaster: we think there’s nothing to be afraid of as long as the bus is surrounded by fog. Whose fog, buster? Ours or yours? And does the bus driver have the foggiest idea where he’s taking us? Sorry for asking.