I've written before about my interest in A.E. Douglass, a turn-of-the-last-century astronomer and all-around scientist (the disciplines were so less clearly defined a hundred years ago) who almost accidentally developed dendrochronology.

In the process, Douglass revolutionized two sciences - archaeology and paleoclimate studies - but he was after something entirely different. He was studying the sun, and wanted to know whether the 11-year sunspot cycle was a permanent feature. He figured that trees, by putting on fat rings in wet years and thin ones in dry years, might show signs of a periodicity linked to the solar cycle. The archaeology was nowhere near his mind when he got started, and the climate stuff was only a means to his astronomical end.

It didn't take him long to grasp the significance of what he was up to, though, once he realized that he could statistically identify synchronicity in tree rings in sites some distance from one another. (His first sites were in Flagstaff, Ariz., and then a hundred or more miles away in Prescott, Ariz.) He eventually became convinced that, if he could just nail down the cycles, he'd have a tool that could be used to make climatological forecasts.

Douglass believed by the end of his career that he had found the solar cycle. I'm not clear enough on the statistical tools he used, the way he did the spectral analysis to pick the cycle out of the quite noisy data, to understand whether he was right. It's not one of those things that's immediately obvious, like "every 11th year is wet" or something, relying instead on more sophisticated statistical analysis to pick signal out of a lot of noise.

But in the years since, other people with more modern tools have continued to poke at the question, and there are now clear signs that there is a climate signal caused by solar variability. It's not strong enough to be useful in telling us whether this winter will be wet or dry, but it's there.

I've been reading some of Douglass's old papers from the early 20th century, but until this weekend I hadn't spent much time in the contemporary literature. Following a troll down a rabbit hole for the last couple of days over at Quark Soup, however, has yielded some really interesting stuff I didn't know about the question. I know, it's better to let trolls be trolls, but this one was raising some interesting questions about the role of solar irradiance in global climate change.

Turns out the troll was full of crap, misquoting the literature to try to argue that changes in solar output were sufficient to explain 20th century warming. The literature he was citing didn't support the assertion, arguing instead that there was some solar warming over the 20th century, but that it was insufficient to explain the warming we've actually seen. Even this point clearly remains controversial, but whatever. The literature was nevertheless pretty interesting.

There's a great summary of the state of the science done earlier this year by John Eddy of the National Solar Observatory in Tucson. Eddy chaired a NASA-sponsored panel looking at what we know and don't know about this stuff.

My favorite bit from Eddy's paper is how Sir William Herschel thought he'd found a correlation between the sunspot cycle and the price of wheat. Herschel argued that fewer sunspots was linked to climatological changes that included poor growing conditions. (The link to the markets is a dicey one: "Demographic pressures, civil unrest and other social changes all played a significant part in cereal prices at the time," notes William James Burroughs in his book Weather Cycles: Real Or Imaginary. Burroughs is talking about a different attempt to make the connection, but the point's the same.)

It's really only been since 1978 that, using newly launched satellites, we've been able to measure the solar irradiance - how much energy the sun is actually dumping into our atmsophere. There's still some debate, apparently, over whether the data collected over that relatively brief interval (a little more than two solar cycles so far, not much to go on) shows a long-term warming trend.

If it does, it doesn't appear sufficient to explain the warming measured over the 20th century (this was at the heart of the argument with the troll). But there are some clearly measurable connections between the solar cycle and temperatures here on Earth. According to Eddy, it's found in measurements of both ocean and air temperature, though the details of the mechanisms involved are still a bit murky. The timing of the ocean numbers, for example, is a bit of a puzzle because of the damping and feedbacks involved. But whatever those details, the 11-year cycle can definitely be seen.

Which means Douglass, and all the other early pioners looking for sun-climate connections, were right.

(Hat tip to Henri Grissino-Mayer, whose Ultimate Tree Ring site is a web treasure, the work of a smart and dedicated and very generous guy.)