I have mapped the x,y coordinate and stem location of each tree on a 50m x
50m plot, and measured its diameter. There are a total of ~1000 stems of
all different sizes and 7 different species. Different species occur in
clumps, while others are randomly distributed across the plot. One species
in particular makes up ~80% of the density and cross sectional stem area.
I would like to use variogram analysis to describe the spatial patterns of
stem size and stem occurrence, i.e. what is the range of autocorrelation of
size and stem presence on this plot.
Approaches which occur to me:
1. to address the occurence issue, assign each stem a "1", and randomly
choose an equal number of non stem points and assign them "0", and do an
indicator variogram-type analysis. Isn't this measuring the spatial
dependence of "stem presence"? A problem I forsee is that the nugget gets
artificially lowered if the stems are not evenly distributed throughtout
the plot--many (0-0)^2 's. Would this, however, alter the range of
autocorrelation, my goal?
2. to address the size class issue: just use the x,y coordinates of each
tree, and assign stem area to the z value. I suppose I am really answering
the question "are trees that are growing close together more similar to one
another than one would expect by chance *where trees are growing*(not "in
general on the plot")?"
3. divide the plot up into grid cells, or randomly overlay subplots onto my
big plot, and calculate stem density and total stem xc area within each
plot, giving the x,y coordinates the center of each subplot
This seems problematic too, although I read a paper where they did
something like this.
*I actually did numbers 1 and 2 and obtained variograms with beautiful
forms, but feel a tad uncomfortable with what I've done (nagging guilt
about being a simpleton).
Does anybody have any opinions/other suggestions on how to attack this
P.S. I have all of my data entered into ArcView and Arc Info...
*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!