Spatial data analysis of finite strain data across a thrust sheet using R

Abstract

In this paper, the finite strain data (FSD) collected from a Sheeprock thrust sheet in Utah are spatially analyzed. The steps and special manner of spatial analysis of these data are examined due to the importance and influence of data characteristics on spatial prediction accuracy. To do so, the exponential and spherical parametric variogram models with and without nugget effects are fitted by maximum likelihood and restricted maximum likelihood methods on original and normally transformed data. We also fitted other models to classical and robust empirical variograms by ordinary least squares and weighted least squares methods. Next, these models are applied on data and edited data by using robust kriging. Finally, the accuracy of the obtained results are compared with Mukul's model by cross-validation in order to choose the best variogram model for FSD and also specify the importance of attention to characteristics of the data in spatial data analysis. © 2008 Elsevier Ltd. All rights reserved.