GEOSTATS: Multivariate spatial autocorrelation and Mantel's tests
- Hello all:
Of all the listservs I'm subscribed to, this one seemed the most appropriate to
I'm looking for a way to calculate spatial autocorrelation for a multivariate
datatset. What I have is a table containing X, Y, Z1, Z2, Z3...Zn information
that was derived from a point-in-polygon
analysis of species distribution points over categorical GIS data layers (i.e.
Soil type, Forest type, Topography, etc.) and I'm looking for the lag distance
that best minimizes the effects of spatial autocorrelation.
I have some apps that calculate either global stats (Moran's) or variograms, but
they are univariate! What is the best approach? Should I go through the long
process of calculating the best lag distance for
each individual layer and use the average best lag? Alternatively, does anyone
have a pre-compiled app for Windows or DOS that will perform the appropriate
calcs for this type of multivariate data?
My readings have also pointed me towards "Partial Mantel's test on multiple
Of course I have several questions concerning this:
1: How do I derive a distance matrix for the categorical GIS layers (predictor
2: How do I derive a distance matrix for geographic location (predictor
3: How do I derive a distance matrix for the species distribution data in
question (response variable)?
4: Where can I find an app that performs these calcs?
I have access to most GIS apps (except ArcINFO) and several stats apps (mainly
Thank you for your time and help with these (confusing newbie) questions.
*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!