One must be a little careful, kriging is not a unique

interpolation method hence to say that one is going to

compare "kriging" with "IWD" does not make sense

First of all, is the mean treated as constant or a polynomial

(if the latter what terms are included, i.e., universal

kriging). Even if the mean is considered constant is it

considered known or unknown (simple vs ordinary kriging).

What variogram/covariance model(s) are used, what parameters

(including possible anisotropies). Is a unique neighborhood

used or a moving neighborhood, if the latter what conditions

are placed on it (radii, minimum/maximum number of points,

sectors, # of empty sectors?).

Note that the thin plate spline is a special case, i.e.,

a particular covariance.

Even IWD can vary, see Kane et al (Computers and Geosciences

1982) on an empirical study of the effect of the exponent.

To construct the contour map one must first interpolate onto

a regular grid, the choice of the grid mesh for interpolation

will have some effect.

Are point values interpolated or block values (one might

use either for contouring)?

What is the ultimate use of the contouring? That will have

some bearing on how to judge the comparison.

Finally the comparison will be valid only for the specific

data set and does in general not imply anything about

the comparison for a different data set.

Wanting to compare different possible results is an

interesting/useful question but it will not have a

definitive answer.

Donald E. Meyrs

Donald E. Myers

Department of Mathematics

University of Arizona

Tucson, AZ 85721

http://www.u.arizona.edu/~donaldm
--

*To post a message to the list, send it to

ai-geostats@....

*As a general service to list users, please remember to post a summary

of any useful responses to your questions.

*To unsubscribe, send email to

majordomo@... with no subject and

"unsubscribe ai-geostats" in the message body.

DO NOT SEND Subscribe/Unsubscribe requests to the list!