SPATIOTEMPORAL PRECIPITATION MODELING IN RURAL WATERSHEDS

The natural variability of precipitation in agricultural regions both in time and space is modelled using extensions of Box & Jenkins (1976) methodology based on the ARMA procedure. This broad class of aggregat e regional models belongs to the general family of Space-Time Autoregr essive Moving Average (STARMA) processes. The paper develops a three-s tage iterative procedure for building a STARMA model of multiple preci pitation series. The identified model is STMA (I-3). The emphasis is p laced on the three stages of the model building procedure, namely iden tification, parameter estimation and diagnostic checking. In the param eter estimation stage the polytope (or simplex) method and three furth er classical nonlinear optimization algorithms are used, namely two co njugate gradient methods and a quasi-Newton method. The polytope metho d has been adopted and the developed model performed well in describin g the spatio-temporal characteristics of the multiple precipitation se ries. Application has been attempted in a rural watershed in southern Canada.