R help - May 2007 - a question about spatial autocorrelation in R

Dear all,

I am currently facing a problem related to the spatial autocorrelation of 
a sample of stations; these stations supply weekly data for a fixed 
time-window during the year (namely, 4-6 months per year).

For this reason I'm trying to use the R package 'spdep'
(specifically
Moran's I) in order to get rid of it.

Does anyone know how is it possible (if it is...) to estimate the spatial 
autocorrelation of this kind of data ? In other words, in the examples I 
could only see applications of Moran's I related to a
"time-independent"
distribution of data from any sample of stations.

Another question's coming up right now: if I change the sample of stations 
for a future development, is there the chance to get an estimation of the 
degree of spatial autocorrelation, given only the geographical location of 
the sampling points ?

Thanks a lot for the attention !!!

Stef
------------------------------------------------------------------------
Stefano Marchesi
Environmental Epidemiologic Unit
Environmental Protection Agency - ARPA Emilia Romagna
Via Fontanelli 23 - 41100 Modena (Italy)
tel: +39 059 433635 (int. 5035)
fax: +39 059 241312
sito web: www.arpa.emr.it/epam

On Mon, 14 May 2007 smarchesi at arpa.emr.it wrote:
> Dear all,
> 
> I am currently facing a problem related to the spatial autocorrelation of 
> a sample of stations; these stations supply weekly data for a fixed 
> time-window during the year (namely, 4-6 months per year).
Could I suggest that you read the relevant parts of Collecting Spatial 
Data by Werner G. M?ller, 3rd edition 2007? Given your second question, 
the book's treatment of experimental design is probably relevant. 

Once you've found a suitable approach, perhaps R-sig-geo would be a more 
focussed list to go forward to see if suitable functions are available?

Roger
> 
> For this reason I'm trying to use the R package 'spdep'
(specifically
> Moran's I) in order to get rid of it.
> 
> Does anyone know how is it possible (if it is...) to estimate the spatial 
> autocorrelation of this kind of data ? In other words, in the examples I 
> could only see applications of Moran's I related to a
"time-independent"
> distribution of data from any sample of stations.
> 
> Another question's coming up right now: if I change the sample of
stations
> for a future development, is there the chance to get an estimation of the 
> degree of spatial autocorrelation, given only the geographical location of 
> the sampling points ?
> 
> Thanks a lot for the attention !!!
> 
> Stef
> ------------------------------------------------------------------------
> Stefano Marchesi
> Environmental Epidemiologic Unit
> Environmental Protection Agency - ARPA Emilia Romagna
> Via Fontanelli 23 - 41100 Modena (Italy)
> tel: +39 059 433635 (int. 5035)
> fax: +39 059 241312
> sito web: www.arpa.emr.it/epam
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
> 
-- 
Roger Bivand
Economic Geography Section, Department of Economics, Norwegian School of
Economics and Business Administration, Helleveien 30, N-5045 Bergen,
Norway. voice: +47 55 95 93 55; fax +47 55 95 95 43
e-mail: Roger.Bivand at nhh.no