Topic:

Presenter: Dr. Andrew Long

In this module we investigate geostatistical methods. In particular, we are interested in the production of maps from data. The general procedure is as follows:

• examine the data, to determine what kind of interpolation (mapping) is appropriate;
• model the spatial autocorrelation of the data;
• use the model to "krig" the data (generate the map); and
• check the results.

OBJECTIVES: Those who successfully complete the module should

• understand spatial autocorrelation as a function of space, with important characteristics such as the distance over which correlation exists, the relative size of the correlation, etc.;
• understand the variogram and its relationship to kriging;
• know how to model the variogram, and plausible models;
• understand kriging interpolation and its relationship to the variogram;
• be able to use geostatistical software to
  • create and model variograms, and know why it's done (decomposing sample variance to explore spatial autocorrelation),
  • krig, and know why it's done (make better assumptions about spatial autocorrelation to get better maps than arbitrary, ad hoc inverse-distance weighting and other methods provide).

• SCENARIO 1
• SCENARIO 2