Premium
The elucidation of soil pattern in the Wyre Forest of the West Midlands, England. I. Multivariate distribution
Author(s) -
OLIVER M. A.,
WEBSTER R.
Publication year - 1987
Publication title -
journal of soil science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.244
H-Index - 111
eISSN - 1365-2389
pISSN - 0022-4588
DOI - 10.1111/j.1365-2389.1987.tb02145.x
Subject(s) - ordination , spatial distribution , principal component analysis , geography , homogeneous , multivariate statistics , stratified sampling , spatial variability , physical geography , sampling (signal processing) , soil horizon , distribution (mathematics) , forestry , soil map , spatial ecology , soil science , cartography , hydrology (agriculture) , environmental science , mathematics , statistics , geology , soil water , remote sensing , ecology , computer science , combinatorics , mathematical analysis , filter (signal processing) , biology , computer vision , geotechnical engineering
SUMMARY The soil of the eastern 6 km 2 of the Wyre Forest in the West Midlands of England was surveyed by recording the soil profile at 36 points per km 2 chosen by stratified random sampling. Ordination by principal component analysis revealed an unclustered multi‐variate distribution, which nevertheless was classified profitably into six fairly stable groups. The groups lacked spatial coherence when mapped, and a novel method of spatially weighted classification was devised to increase the spatial coherence. To be effective the spatial weight had to be very large and produced very heterogeneous groups of soil profile. Sampling was too sparse to delineate homogeneous parcels of soil, and the spatial scale of variation was not identified.