Suppose I have a variable (Leaf area index) that shows periodicity in its
semivariograms (at several lags). And this periodicity can be seen in the
E-W direction in a satellite image. There is also a N-S trend (values
increasing to the south). When modeling an omnidirectional variogram, a
nested model including a spherical and hole effect structures fits
adequately the sample variogram. The rose variogram indicates that the
azimuth angle = 160 degrees is the major direction of continuity; so 160 and
70 degrees are the angles selected to approach the problem with anisotropy.
The directional variograms also show periodicity but the hole model effect
cannot be applied in more than one direction (Deutsh and Journel,
Geostatistical Software Library and user's guide, page 25). So each
directional variogram is modeled with a nested model including two spherical
models; this model does not represent the periodicity of the variable. When
performing ordinary kriging or conditional simulation the use of the
omnidirectional model gives better results (visually and statistically)than
the use of the anisotropic model.
Is there a way by which I can include the periodicity in the anisotropic
model or do I just have to accept the omnidirectional approach as the best
one in this case, because it represents better the periodic behavior?
Thanks in advance,
*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!