Experience has shown that the climate stubbornly refuses to conform to what has been predicted by computer climate models. I thought it was because the climate change alarmists were fudging the data, so its explained by the usual garbage in-garbage out formula. It turns out because computers have their idiosyncrasies and because there are several complex mathematical and statistical problems that remain unsolved, garbage out is not only possible but often certain even if nobody put garbage in.
One fun example Dr. Christopher Essex explains in the video below is called Machine Epsilon. We all know that you add any number to 1, no matter how small the number you add, the result will be a number greater than one. No so with computers. In a computer program a very small number added to one renders just 1, not greater than 1. If this is a done a few hundred thousand times the error is magnified. He explains how this applies and what occurs. Fascinating stuff.
He then goes on to much more difficult unsolved equations and problems that are little more difficult to follow in every detail but still quite understandable by anyone on a more general level. It all adds up to why computer models of future climate are wrong even if the climate alarmists were not faking the data. I still think they are playing games with their data and assumptions because, as Dr. Essex says, many of them don’t know that computer modeling cannot accurately predict future climate, at least not with what is currently known.
The first 20 minutes is an introduction in which Dr. Essex presents some humorous conclusions that the alarmists have relied on, and then gets into the real science of the problems with computer modeling of climate.