An Unconventional Approach to Geostatistical Estimation

Abstract

After a cursory discussion on sample requirements for estimation purposes for uncorrelated data and spatially correlated data for which an experimental semivariogram can be obtained at a reasonable cost and for which no doubts arise about the semivariogram model, I will dwell at a greater length on an alternative “ad hoc” procedure for the latter case. Full information of the semivariogram is not required with this procedure, which would provide good approximation of the estimation variance more directly and at a relatively low sampling and computing cost, and without the uncertainties of the choice of the semivariogram model.