GEOSTATS: thanks for help on kriging problem
- Dear List Members:
Thanks to all for the very rapid response for aid in finding a simple
kriging program to verify a VB5 gridding and variance analysis code I've
been working on.
Several people recommended GSLIB. I downloaded, compiled, and ran Ktb3d
with no problems. GSLIB gave the same results as mine for the example
problem in Isaaks and Srivastava ehich used an exponential variogram
with a 50% sill. The results did differ from the values posted on
Figure 12.10 in I&S. Other programs recommended included GS+
(www.gammadesign.com) and gstat (www.geog.uu.nl/gstat/). I haven't had
a chance to try them yet.
Many thanks also to Donald Meyers from the U. of Arizona for his timely
answer to problems with the Gaussian variogram. As a final check on my
program, I was testing the topographic data set in Davis (1973). The
problem worked fine for all variograms except the gaussian which gave
complete garbage. Rather than spending hours trying to find a
non-existant bug, adding a small nugget cured the problem. Further
testing showed that a nugget as small as 6.5 (0.1%) worked with a sill
of 6500. The gaussian variogram also gave reasonable results without a
nugget for universal kriging with a linear or quadratic drift but then
the sill values were much lower (850 for the quadratic). I'll have to
look further into this ill-conditioned matrix problem.
*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!