Saturday, October 29, 2011

How junk science is done: Start with a bogus model and fiddle with the parameters until you get the desired answer

Global warming: Middle East's vital wet winters are disappearing - CSMonitor.com

The team estimates that average precipitation from November through April in the region between 1971 and 2010 fell 6.8 percent below the average from 1902 to 1970.

The team began the hunt for causes.
...The team began using computer models to assess the effect of this ocean warming on the Mediterranean's winter precipitation.

The team found that warming all the oceans by a uniform 0.5 degrees Celsius (about 1 degree Fahrenheit) would dry out the eastern Mediterranean.

But the oceans haven't warmed uniformly. The greatest warming has come to tropical oceans. So the team focused next on warming the tropical oceans uniformly in their virtual world. The team got a Mediterranean-wide drying and a wetter northern Europe.

Still, neither of these experiments produced the strong positive North Atlantic Oscillation-like signal with enhanced drying to the south and heavier precipitation to the north.

But by adding another 0.5 degrees C to the Indian Ocean alone, model produced the strong north-south differences in precipitation associated with a positive phase of the North Atlantic Oscillation.

Overfitting

With four parameters I can fit an elephant and with five I can make him wiggle his trunk.

John von Neumann

No comments: