The von S affair

RC had a recent post about some errors in von S’s Science ’04 article. Now Salon.com has an article on it too, being nice to us. Jolly good.

The main point made is a statistical one – read RC for that. A second one is the error in the initialisation procedure for von S’s model run (not sufficiently spun up back to pre-ind). This causes a drift (down) during the sim, which over-estimates the variance. But von S has also said According to our computer models temperatures fluctuated more strongly and faster. 900 years ago they were almost as warm as today. But now we know why his model says that…