Modelo probabilístico de pronostico por punto.

Abstract

We try a rnodel in adaptive sequential strategy, with Bayesian statistics and mínimum square cost function. We use isohipsas inforrnation of at rnost ten days in sequence and forecast the isohipsa value just above the station reporting the data, we get a good approxirnation in forecast at all atrnospheric levels with a rninirnal cornputer effort.