Thursday, September 8, 2011

Solar energy and the knowledge problem

Is it a key problem of the modern age that nobody can even have a reliable surface knowledge of more than a couple important issues, yet must have opinions on hundreds? This thought occurred to me as I attended a solar energy congress in Germany last week, then read a few columns from mainstream media pundits giving the usual talking points stating that solar is a hilarious boondoggle and always will be. People pick up a few points like these which serve to discredit an entire field of human endeavor. In the case of solar power, the main brickbat thrown is the true and obvious fact that the sun does not shine all the time, has variable input when it does shine, and is occasionally obscured by clouds. This fact is extrapolated to the conclusion that solar is useless. Similarly the cost argument is given--solar at this point is not as cheap as coal or natural gas in terms of providing a kilowatt-hour of electricity. Solar still relies on subsidies and feed-in tariffs.

But the op-ed column level of knowledge and dismissal is extremely superficial. You go to one of these big meetings devoted to a single field like solar energy and you'll see 1000 very bright people, who have devoted their working lives to this subject, discussing all areas of the field. The discussions will be at a level of detail you can't possibly evaluate at more than a surface level unless you've got a lengthy background in that area too. A person with any level of humility might take that opportunity to assess the reliability of his assumptions about a lot of things, and the confidence with which those assumptions should be held.

I've often had that thought about climate science too of course. Would an average person, placed in the audience for a few days of technical talks at a scientific conference, realize that he probably shouldn't actually dismiss an entire field of study with a few choice quotes from a Mark Steyn column? That these many, many very smart people, doing the work in which they've invested thousands of hours of work, might collectively know something? I don't know, maybe not. Political tribalism is so engrained that it encompasses many things that are far afield of political ideology per se. It's very hard to see past those cultural markers of one's tribe.

Many people at the solar conference showed a Moore's Law-like reduction in the cost of solar photovoltaic panels over the years, showing an approach to cost parity with coal/gas-fired electricity. I don't know if or when that's actually coming, and I don't understand the ins and outs of the electrical grid and how it can accept intermittent power inputs. There were also some interesting reports on Concentrated Solar Power plants, a completely different process than photovoltaics. These plants focus large amounts of direct sunlight onto a small area, to boil water and drive a conventional steam turbine. This also presents the opportunity to smooth out the intermittancy of the solar input, by storing the energy in a high heat capacity/low volatility working fluid, such as molten salt. This scheme works best in dry tropical areas, since it requires direct (unshaded, unscattered) sunlight.

Anyway, I am very far from having the level of knowledge where I'd feel comfortable evaluating the place of solar in the world's future growing energy needs. Germany subsidizes it and is as a result by far the world leader in installed capacity. You travel around Germany and you see wind and solar installations all about. My hunch is that this will all pay off in the next fifty years. No magic bullet, but renewables can probably play a growing part in the future. The problem is that I must know more than 90% of the general population on this subject, I don't know enough to intelligently support or oppose a given policy option, and yet options will be chosen. If it's impossible for the general public to be well enough informed on this and many other issues, then the basis on which things will be decided will not be accurate knowledge. It will be something else.