For computational reasons I've been studying the Schur complement of the matrix involved in LSS fitting of temperature models. That sounds hopelessly geeky, but it's just something which is basically a measure of the number of missing values in the GHCN station data. If I can predict them, I can get a much faster algorithm. So I plotted the diagonal, and ...
A pretty dranatic marker of the effect of world wars on Met observations.
No comments:
Post a Comment