Geostatistical characterization of the soil of Aguascalientes, México, by using spatial estimation techniques

Abstract

Four spatial estimation techniques available in commercial computational packages are evaluated and compared, namely: regularized splines interpolation, tension splines interpolation, inverse distance weighted interpolation, and ordinary Kriging estimation, in order to establish the best representation for the shallow stratigraphic configuration in the city of Aguascalientes, in Central Mexico. Data from 478 sample points along with the software ArcGIS (Environmental Systems Research Institute, Inc. (ESRI), ArcGIS, ver. 9.3, Redlands, California 2008) to calculate the spatial estimates. Each technique was evaluated based on the root mean square error, calculated from a validation between the generated estimates and measured data from 64 sample points which were not used in the spatial estimation process. The present study shows that, for the estimation of the hard-soil layer, ordinary Kriging offered the best performance among the evaluated techniques.