Premium
Use of Geostatistics in Designing Sampling Strategies for Soil Survey
Author(s) -
Di H. J.,
Kemp R. A.,
Trangmar B. B.
Publication year - 1989
Publication title -
soil science society of america journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.836
H-Index - 168
eISSN - 1435-0661
pISSN - 0361-5995
DOI - 10.2136/sssaj1989.03615995005300040028x
Subject(s) - geostatistics , kriging , loam , sampling (signal processing) , spatial variability , soil science , anisotropy , geology , soil water , soil survey , variogram , mathematics , statistics , physics , filter (signal processing) , quantum mechanics , computer science , computer vision
Semivariograms are used to quantitatively assess spatial variability of depth to mottles, depth to gravels, and thickness of loamy sand and/or coarser‐textured layers, which are definitive criteria for classification of soils derived from alluvium on the Canterbury Plains near Lincoln College, New Zealand. The three properties vary anisotropically with the anisotropy ratio being highest for depth to mottles ( k = 5.84), lowest for thickness of loamy sand and/or coarser‐textured layers ( k = 1.58), and intermediate for depth to gravels ( k = 2.43). Directions of maximum variation for depth to mottles and depth to gravels are NE‐SW across an abandoned channel hollow. This pattern is reflected in the soil map of the study area and in the smaller‐scale soil map of the adjacent region. Such variation reflects the past regional drainage patterns of channels flowing predominantly in a NW‐SE direction. The appropriate field configuration of a sampling scheme for future survey of similar adjacent soils would be rectangular with a sample spacing in the direction of least variation k (anisotropy ratio) times that in the direction of maximum variation. Suitable sample numbers and sampling intervals to achieve desired levels of precision in the direction of maximum variation are determined. These are obtained from graphs showing relationships between kriging standard error, sample spacing and sample number. This geostatistical approach is more efficient than conventional statistical methods in designing sampling strategies: less samples are needed for kriging than for the conventional method to achieve the same level of precision.