Interpolation – Kriging, Cokriging, Conditional Simulation, and IDW

GS+ provides four types of Interpolation – Kriging, Cokriging, Conditional Simulation, and Inverse Distance Weighting. Output is written to ASCII files that can be read for mapping by GS+, ArcView®, Surfer®, or other mapping or GIS programs.

Kriging provides optimal interpolation of points across a spatial domain for which autocorrelation has been documented and measured with variograms. GS+ provides both block and punctual kriging, and allows the user to choose the most appropriate variogram model to use for the interpolation.

Cokriging is a type of kriging that allows one to better estimate map values using a secondary variate sampled more intensely than the primary variate. If the primary variate is difficult or expensive to measure, then cokriging can greatly improve interpolation estimates without having to more intensely sample the primary variate.

Conditional Simulation provides optimal interpolation whereby measured data values are honored at their locations. Other interpolation methods will smooth out local details of spatial variation, which can be a problem when you are trying to map sharp spatial boundaries such as contamination hotspots or fault lines.

Inverse Distance Weighting (IDW) provides classical interpolation based on nearest neighbor weighting. It is a simple interpolation method used in mapping programs that do not use geostatistics, and assumes spatial dependence among points close to one another (without measuring it).

The Interpolation Grid allows the user to define the boundaries of the interpolated area and the intensity (grid spacing) at which the interpolation will proceed.

Polygon Outlines define irregular map boundaries and special areas to exclude from kriging. An unlimited number of polygons can be defined by an unlimited number of vertices (x-y boundary points).

Polygon Maps display the areas that will be included or excluded from kriging. Exclusive and inclusive polygons are colored differently, and polygons can be nested within one another.

Cross Validation Analysis allows one to test different variogram models; bootstrapping provides comparisons of the actual value of every point sampled vs. its estimated value when removed from the data set.