This example illustrates the Variogram Cloud Arcview-XGobi link on data that are precipitation values from sampling sites in the northeastern United States for the third and fourth quarters of 1990. It is used in conjunction with the forest health monitoring data introduced in the SCDF link example to assess stress on the forests. In addition to precipitation, several other data sets were used including satellite images of the region. For more details on this see Majure and others (1996B).
In this example, a variogram cloud has been plotted for precipitation values for the third quarter of 1990. This plot shows that a group of points seems to lie above the main group of points.
When several of these "outlying" points are brushed in red, red lines are used to connect the pairs of sampling locations that correspond to each point brushed in the variogram cloud. This indicates that all of the brushed points have a single sampling location in common.
By zooming in on the site in question (below), and displaying elevation contours, it becomes obvious that the site in question sits on top of a mountain, which causes it to have higher precipitation values than other sites in the area.
With this link it is also possible to examine if spatial dependence changes depending on the angle between sampling locations. This is important because spatial dependence can be directional, that is, according to angular distance rather than (or as well as) Euclidean distance between points. Assessing directional dependence can be done with linked brushing after cloning a new XGobi window and brushing in a plot of the cosine versus the sine of the angle between two points. Since angle is a modular measurement (0=360) we use the plot of cosine versus sine to maintain the modularity by representing the values as a full circle for brushing. An example of this technique can be seen in Symanzik, Majure, and Cook (1996C), in the Multivariate Variogram Cloud link, and in the video segment on the Variogram Cloud link.
Last Revision: Fri Dec 20 11:47:49 CST 1996