Yet Another Amazing Detail About Climate Science

They’ve got incredible supercomputers to play with, but still our climate scientists build up global anomaly charts from only a handful of stations (1,000 or so if the guys at Climate Monitor are not mistaken, or one on average for every area the size of Bangladesh).

You’d think world-rescuing climate specialists would have computers and bandwidth enough to handle data from a million stations by now, or a billion even. But they say they don’t even need that.

Why not? Because they have discovered that 1,200-km grids are ok. Why are they ok? Because

temperature anomaly patterns tend to be large scale

How so? Here’s the root of it all, from 1987:

The 1200-km limit is the distance at which the average correlatiom coefficient of temperature variations falls to 0.5 at middle and high latitudes and 0.33 at low latitude

Who would believe that it all depends on a correlation coefficient of…0.5?

0 Replies to “Yet Another Amazing Detail About Climate Science”

  1. The 0.5 number is r. The coefficient of determination, r-squared, is the proportion of variability in a data set that is accounted for by the statistical model. Thus r^2 = 0.25 would be the important number here (and it is a pretty disappointing number at that).

  2. A coefficient of 0.5 lies in the “probably something there” category. Coefficients of 0.33 are a dime a dozen and can be ignored. This ain’t science; it’s Madame Gazonga’s crystal ball.

  3. Dear Omnologos. When I were a youth, I was taught to read a map by studying the ground and relating it to the map. The more ground identified, the more certain one’s position. Could it be that W-R CS’s having drawn a map, are not over-anxious to identify so much ground that the map will need to be re-drawn?

Leave a Reply - Lascia un commento

This site uses Akismet to reduce spam. Learn how your comment data is processed.