Novos estimadores de semivariância no estudo da estrutura de dependência espacial entre amostras.

Abstract

Geostatistic is characterized by the study of the spatial dependence among the samples of regionalized variables. The spatial dependence structure among the samples of variables is verified by the semivariance function analyze, witch may be estimated by the semivariograms, given by the Classical Semivariance Estimator of Matheron, by the Cressie & Hawkins Estimator and by the Pairwise Relative Semivariance Estimator. In 1994 LI & LAKE proposed two new estimators called New1 and New2, which according to the authors are more efficient than the others. In this study were simulated data sets with spatial dependence structure kneed, by the Monte Carlo Simulation and the simulated data were applied in the five estimators of the semivariance function and compared with the simulated semivariances, which is given by the theoretical model. In this comparation, the Matheron, New1 and New2 estimators estimated more similar values to the estimated.