Chemometric investigations on the representativity of soil sampling |
| |
Authors: | Jürgen Einax Bernd Machelett Sabine Geiß and Klaus Danzer |
| |
Institution: | (1) Department of Chemistry, Friedrich Schiller University, O-6900 Jena, Federal Republic of Germany;(2) Institute for Plant Nutrition and Ecotoxicology, O-6900 Jena, Federal Republic of Germany |
| |
Abstract: | Summary The characterization of the heavy metal contents of a soil area is carried out by narrow screening sampling and subsequent analysis. It is possible by means of the univariate variance analysis to detect significant alterations within the tested area. The position and extent of multivariate changes of the features can be characterized more exactly with the help of multivariate statistical methods such as the combination of cluster and discriminant analytical methods or the principal component analysis. The maximum possible distance between two sampling points is determined by calculation of the autocorrelation function of the features. The number of soil samples, which is necessary for precisely determining the heavy metal contents of the investigated area, can be derived therefrom. With the help of the multivariate autocorrelation function the distance for the representative sampling of the multivariate loads by heavy metals can be described. |
| |
Keywords: | |
本文献已被 SpringerLink 等数据库收录! |
|