Re: AI-GEOSTATS: spatial autocorrelation
- On Tue, 4 Sep 2001, Sue Rodriguez-Pastor wrote:
> Hello,This is not necessarily true, there has been some use of moran's I in the
> I'm helping someone with some spatial analyses for some data with plant
> species data across a set of natural pools. We are trying to determine
> whether there is spatial autocorrelation, but am not sure what test to
> At first I was advised that perhaps using the Moran's I index would work,
> but I have had some folks suggest that Moran's I works best for adjacent
> polygons or raster data. I am using an ArcView polygon data layer.
metapopulation literature. See for instance the chapter by Michael Gilpin
in the 1997 Hanski and Gilpin book on metapopulations. Also Epperson has
written quite a few papers on population genetic models using moran's I
for disjunct populations. You do need to be very careful how you define
your wieght matrix.
>As above be careful. Geostatistics assummes an underlying continous model
> Another recommendation was to use semivariograms included in
> Geostatistical Analyst with ArcView 8.1, which I will attempt in the
> coming week.
where as in this situation you have sites on the nodes of a graph. The
unsampled locations are really a problem because any measure of
connectivity in your system would need to include all the ponds.
What is your overall question you are trying to answer? Is it related to
the persistance of the metapopulation? Is it about dispersal? More
information on what you are really trying to ask biologically might help
to suggest an approach.
> I have a few questions.CH3
> My first question is: What is the appropriate test that should be employed
> to determine whether there is spatial autocorrelation given this scenario?
> There are approximately 100 natural pools that
> have varying abundances of the plant in question. However, only 32 pools
> were sampled. It was not a random sampling...out of the 100 pools, these
> were the only ones that were not damaged by construction. The other pools
> are still functional, however, and do contain the plant being studied.
> On one end of the site most of the pools were sampled, whereas on the other
> end of the site only a few pools were sampled.
> So my second question is: Can I really test for spatial autocorrelation if
> we do not have any data for the damanged pools? My concern is that we will
> have a sampled pool with high plant density next to 4 or 5 non-sampled
> pools...will we get the wrong outcome?
> I don't know if this makes sense...it's difficult to explain virtually!
> If anyone has any suggestions I'd really appreciate it!
> Thanks very much in advance,
> Sue Rodriguez-Pastor
> Dept. of EPO Biology
> University of Colorado
> email address: smrp@...
> * To post a message to the list, send it to ai-geostats@...
> * As a general service to the users, please remember to post a summary of any useful responses to your questions.
> * To unsubscribe, send an email to majordomo@... with no subject and "unsubscribe ai-geostats" followed by "end" on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list
> * Support to the list is provided at http://www.ai-geostats.org
N Nicholas Lewin-Koh
/ \ Dept of Statistics
N----C C==O Program in Ecology and Evolutionary Biology
|| || | Iowa State University
|| || | Ames, IA 50011
CH C N--CH3 http://www.public.iastate.edu/~nlewin
\ / \ / nlewin@...
| || Currently
CH3 O Graphics Lab
School of Computing
National University of Singapore
The Real Part of Coffee kohnicho@...
* To post a message to the list, send it to ai-geostats@...
* As a general service to the users, please remember to post a summary of any useful responses to your questions.
* To unsubscribe, send an email to majordomo@... with no subject and "unsubscribe ai-geostats" followed by "end" on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list
* Support to the list is provided at http://www.ai-geostats.org