Weather modelling using a multivariate latent Gaussian model

Preprint
Publisher Website
Google Scholar
  • 1 January 2001
    • preprint
      • Published in RePEc
Abstract
We propose a vector autoregressive moving average process as a model for daily weather data. For the rainfall variable a monotonic transformation is applied to achieve marginal normality, thus defining a latent variable, with zero rainfall data corresponding to censored values below a threshold. Methodology is presented for model identification, estimation and validation, illustrated using data from Mynefield, Scotland. The new model, a VARMA(2,1) process, fits the data and produces more realistic simulated series than existing methods dur to Richardson (1981) and Peiris and McNicol (1996).
All Related Versions

This publication has 0 references indexed in Scilit: