Geoscience Reference
In-Depth Information
On being smarter than a screwdriver
People are much smarter than they are given credit for, when given half a chance, a bit of
encouragement, and are not trumped by fashion. If everyone involved put as much effort
into learning and explaining over the last 25 years, as they put into avoiding, enabling,
and groupthink, we would not be living the life Kafka. But no matter how Kafkaesque our
approach to this problem is, an inexplicable toe hold remains when the nonsense is alleged
to be backed by computers.
Computer cachet is a pernicious part of the superstitious cloud of confusion stopping
us from thinking. Computers aren't oracles either. They are just tools, like a screwdriver.
The first rule when using a screwdriver is to be smarter than the screwdriver. You can hurt
yourself if you aren't. It's the same for computers. The question is sometimes put to me as
to whether computer climate models are ready for policymaking. Climate models are the
best we have, but they are far from good enough. Even many experts are unaware of the
extent of their limitations.
Computers can only hold a finite number of numbers. Computer scientists call that a
'finite representation'. Because of that you can get garbage out even when you don't put
garbage in. It alters arithmetic; it alters the equations themselves, and it means important
physics has to be faked because modern computers are far too slow and their
representations are far too small for climate. No climate model fully employs the known
physics. They are empirical. But climate forecasting is not an empirical problem. If
one has a computer large enough, it is easy to estimate how long a typical modern
computer would take to do one 10-year forecast without some of this fake (empirical)
physics. With a Kolmogorov microscale of about a millimetre for air, one gets numbers
like 100,000,000,000,000,000,000 years. That is longer than it took for Douglas Adams'
famous fictional computer, Deep Thought, to answer the cosmic question.
Icouldelaborate, butinatime ofneo-shamanism, moral turpitude experts, andWWM,
does it matter? The climate fervor was never about 'science'. That word is just a gimmick
and a weapon. Policymakers can get little from computer climate models if they fail
to grasp their deep, unredeemable limitations, while being distracted by para-scientific
agendas. It's no good wondering whether climate models are ready for policy, when most
policymakers are not ready for models. They are not smarter than this screwdriver.
