Identification of Regional Variability Component by Geostatistical Analysis of Stream Sediments

Abstract

This work proposes a particular application of kriging analysis to stream sediments, seeking to separate a bidimensional spatial component of variability from a mono-dimensional one. The base model considers the stream sediment values as the sum of two components: a regional one, defined on the entire x-y space; the other one, referred to the mono-dimensional space of the meandric co-ordinate, representative of the specific stream flow variability. The regional component of sample values, called local contribution, can be obtained given the model adopted, namely the “regularising” property of flow component. Then, a geostatistical structural analysis of local contribution can be made and, finally, a continuous mapping of the regional component can be obtained by kriging. This approach has been tested on the case study of a stream sediment sampling, executed in the Reno river basin (Northern Italy) in order to obtain heavy metals distribution.