This is a condensed training course that reviews use of geostatistics for soil mapping. The focus is on using different versions of regression-kriging (variable covariates) to map continuous soil properties and factor-type variables (soil classes). Wrapper functions built from gstat and geoR will be customized for the purpose of soil mapping and then used to automate production of soil maps.The objective of this course is to review the state-of-the-art geostatistical tools and promote use of analytical techniques for automated mapping of environmental variables.
T. (Tom) Hengl, G.B.M. (Gerard) Heuvelink
MSc and PhD level students working primarily with soil data. GIS teams working with various environmental data.
Soil mapping (advanced)
R programming (advanced)
GIS modelling (basic)
Spatial prediction techniques a review (1.5 hrs)
Regression-kriging (1.5 hrs)
Fitting regression models using spatial data (1.5 hrs)
Fitting and interpreting variograms (1.5 hrs)
Introduction to gstat / geoR (1.5 hrs)
Exercise: spatial predictions and spatial simulations (1.5 hrs)
Exercise: spatial predictions and uncertainty assessment using participants’ data (1.5 hrs)
Summary discussion (1.5 hrs)
Software / materials:
Participants typically use their own laptopSoftware: