Post-processing kriging with external drift estimates

Authors

  • Jorge Kazuo Yamamoto Universidade de São Paulo; Instituto de Geociências; Departamento de Geologia Sedimentar e Ambiental

DOI:

https://doi.org/10.5327/Z1519-874x2009000300006

Keywords:

Smoothing effect, Uncertainty, Interpolation variance, Pattern reproduction, Kriging with an external drift

Abstract

Kriging with an external drift has been used for estimating a poorly sampled variable based on an exhaustively sampled secondary variable. Even considering the information coming from secondary data, the estimates present smaller variance than the sample variance due to the smoothing effect. A post-processing algorithm for correcting the smoothing effect of ordinary kriging estimates is adapted for kriging with an external drift. According to results shown in this paper, the proposed algorithm not only reproduces the sample histogram but also keeps the linear relationship observed between primary and secondary variables. For very small samples, in terms of number of data points, the post-processing did not preserve the local precision as smoothed kriging does, because the correction is based on errors derived from the cross-validation. However, when available sample comprises 3% of the exhaustive data set, the post-processing algorithm gives better results than smoothed kriging.

Downloads

Download data is not yet available.

Published

2009-10-01

Issue

Section

Articles

How to Cite

Yamamoto, J. K. (2009). Post-processing kriging with external drift estimates . Geologia USP. Série Científica, 9(3), 89-101. https://doi.org/10.5327/Z1519-874x2009000300006