Geoscience Reference
In-Depth Information
Fig. 7.21
Variogram estimator (blue circles), population variance (solid line), and spherical,
exponential, and linear models (dotted and dashed lines).
variograms
ignoring the direction of the separation vector
h
. In a more
thorough analysis, the variogram should be discretized not only in
distance but also in direction (directional bins). Plotting
directional
variograms
, usually in four directions, we are sometimes able to observe
dif erent ranges (
geometric anisotropy
), dif erent scales (
zonal anisotropy
),
and dif erent shapes (indicating a trend). h e treatment of anisotropy
requires a highly interactive graphical user interface, which is beyond the
scope of this topic (see the VarioWin sot ware by Panatier 1996).
•
Number of pairs and the lag interval
- When calculating the classical
variogram estimator it is recommended that more than 30 to 50 pairs of
points be used per lag interval (Webster and Oliver 2001). h is is due to
the sensitivity to outliers. If there are fewer pairs, the lag interval should
be increased. h e lag spacing does not necessarily need to be uniform but
can be chosen individually for each distance class. It is also possible to
work with overlapping classes, in which case the
lag width
(
lag tolerance
)
must be dei ned. However, increasing the lag width can cause unnecessary
