Geostatistics analysis is a powerful tool which has been used in studies of spatial dependence. With respect to the temporal dependence, few analysis are performed with this kind of methodology. In this study, the geostatistics technique was used to adjust a model to a series of temporal values of precipitation and its performance to predict data was also evaluated. The study was carried out at the Rio Itapemirim Basin in a series of monthly precipitation from 1940 to 2006. The adjusted model was compared to the Box and Jenkins model. The best geostatistics model adjusted was the one which used with 96 neighbors in the prediction and mean relative absolute error of 80.46. Geostatistics methodology showed better results for estimating the monthly precipitation in relation to the SARIMA model (2,1,1)(0,1,1)12.
Bacia do Rio Itapemirim; Box & Jenkins; time series
