Kriging and variogram models

摘要: Geostatistics is a popular class of statistical methods for estimating, or predicting, the value continuous spatial process at unobserved locations given set known locations. Spatial prediction this sort typically performed using method as kriging, which provides estimates that are optimized over predictors linear combination observed values process. Beyond providing point locations, variances kriging (or standard errors) readily available measures uncertainty in predictions. An underlying assumption covariance structure known. Thus, before can be performed, spatial-dependence must determined Rather than estimating function directly, an alternative description variogram estimated from data. In article, we review both and methods. addition, illustrate these rainfall dataset provide brief introductions to some extensions geostatistical