If the air's make-up changes, does the Earth's temperature?

That time, that place

Nineteenth-century geology had turned up something strange: traces of glaciers across Europe. The Earth had once been far colder than now.

Why? A weaker sun, a changed orbit, or a changed atmosphere?

The last had a clue behind it. In 1859 Tyndall showed experimentally that carbon dioxide and water vapour let visible light straight through while catching the infrared that carries heat away.

Why this question

Arrhenius set out to put a number on it.

If atmospheric carbon dioxide halved, how much colder would the Earth be? If it doubled, how much warmer? Answering meant working latitude by latitude and season by season, allowing for cloud and the reflectivity of snow.

There were no computers. He did it by hand. It took over a year, reportedly fourteen hours a day, and he later wrote that it was the dullest stretch of his life.

The result: double the carbon dioxide and the global mean rises by five or six degrees. Modern models give something around three — remarkably close for a hand calculation from a century earlier.

What was found

And he took it as good news.

To a Swede a warmer Earth sounded welcome. He wrote of better harvests in northern countries and fewer famine years. At the rate coal was being burned, doubling would take some three thousand years — no cause for haste.

It took a little over one hundred. He had no way to put the growth rate of coal consumption into the calculation.

The field largely ignored the result. The main objection was that the oceans would simply absorb the extra carbon dioxide. When this was actually measured in the 1950s, the ocean's uptake turned out to have limits.

From 1958 Keeling began measuring atmospheric carbon dioxide daily in Hawaii. That curve took up Arrhenius's calculation as observation, sixty years on.

Terms on this card

The greenhouse effect