Multivariate geostatistics for soil classification
| Autor: | Roberto Bruno, Paola Dosi, Giuseppe Patrizi, Giuseppe Raspa, Nicola Filippi |
|---|---|
| Předmět: | |
| Zdroj: | Scopus-Elsevier Quantitative Geology and Geostatistics ISBN: 9780792321576 |
| Popis: | This paper deals with soil classification for semi-detailed mapping (scale 1:25000) in an area situated in the Emilia Romagna Region (Northern Italy). The available data consist of field and laboratory measurements of more than 750 soil samples spaced at various distances. For this study six variables were considered, namely three grain sizes, pH and 2 calcium contents. Principal Component Analysis was carried out in order to further reduce the number of variables. The three first factors extracted more than 80% of the global variance and seemed to synthesize adequately the soil characteristics and to be relevant to soil genesis. Geostatistical structural analysis of the factors showed nested structures at different scale. Thus it was possible to estimate their spatial components at the sample locations through a simplified version of factorial kriging. Then, a cluster analysis was applied sequentially over the groups of samples consecutively identified on the basis of specific combinations of the components of the factors. The procedure seems to help the paedologist to identify criteria for classifying and mapping soils at detailed scale. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|