680 Lab 4 - co-kriging

Objective: The lab's main objective is try co-kriging (in ArcGIS Geostatistical Analyst, but you can try R or Surfer instead if you like) to see whether it improves estimation or estimation variance.

Data: Just one data set this time:


  1. Estimate the (log) cadmium (prediction) surface and the prediction standard error surface using just the cadmium data.
  2. Then get both the prediction and standard error surfaces for the (log) cadmium using the (log) zinc data and co-kriging.
  3. Compare the the estimation surfaces and the estimation variance surfaces.
  4. Did co-kriging improve either surface? Was it worth it?

Deliverable: A brief (maximum 5 pages) report addressing your assessment of the comparison.

Due: in two weeks.