[R] geoR, plot of variog4 lines incomplete

[Paulo Justiniano Ribeiro Jr]

Mike

The source of the problem is not obvious to me
from what you've reported
and I could ot reproduce in a few tests with other data.

Would it be possoble to send me the data-set?

Cheers
P.J.


On Fri, 2 Jun 2006, Michael Fuller wrote:

> Date: Fri, 2 Jun 2006 14:54:46 -0400
> From: Michael Fuller <mmfuller at tiem.utk.edu>
> To: r-help at stat.math.ethz.ch
> Subject: [R] geoR, plot of variog4 lines incomplete
> 
> I'm using R for Mac OSX version 1.14 (2129) and the geoR package version 
> 1.6-5 (the current version in the R repository). I'm running R in OS 10.4.6 
> on a Mac G4 iBook (933MHz, 640 MB DDR SDRAM). I searched the R archive and 
> did not find a posting on this issue.
>
> I want to use the variog and variog4 functions of geoR to characterize the 
> pattern of spatial autocorrelation of tree density on a forest plot. But I 
> get incomplete line plots when I plot the output of variog4, compared to the 
> plot for the output of variog. I'm using the same geodata for each plot.
>
> I've attached 2 graphs to illustrate. The first is a plot of a 90-degree 
> directional variogram created using variog:
>> trees90.vgm<-variog(trees.geo,max.dist=100,direction=pi/2)
>> plot(trees90.vgm)
>
> The second is a plot of a multidirectional variogram created using variog4:
>> trees.vgm4<-variog4(trees.geo,max.dist=100)
>> plot(trees.vgm4)
>
> The graphs are also available at:
> www.tiem.utk.edu/~mmfuller/R_plots.pdf
>
> You can see that the 90 degree line in the multi-line plot from variog4 
> (dotted green line) does not include all the data shown in the single 90 
> degree variogram (circles) plotted using variog. The lines for other 
> directions in the variog4 plot are also incompletely drawn.
>
> Any suggestions as to what's causing this plotting problem?
>
> The raw data is a file containing 3 columns x,y,z. The x,y values are the 
> coordinates of quadrats placed end to end in a grid on a forest plot. The z 
> values are the density of trees in each quadrat. Each quadrat is 15m square 
> and the plot dimensions are 220x780m. The data are arranged in columns 
> (coordinates 20 to 240) and rows (coordiantes 765 to 150) representing the 
> data posting of each quadrat.
>
> I create geodata objects using:
> trees.geo<-read.geodata("tree.dat",header=TRUE)
>
> Here is a data sample:
> x	y	z
> 20	765	0.1556
> 35	765	0.1556
> 50	765	0.0356
> 65	765	0.0622
> 80	765	0.0933
> 95	765	0.0400
> 110	765	0.1111
> 125	765	0.1244
> 140	765	0.0800
> 155	765	0.0667
> 170	765	0.0711
> 185	765	0.0889
> 200	765	0.0667
> 215	765	0.0844
> 20	750	0.1067
> 35	750	0.0800
> 50	750	0.0578
> 65	750	0.0622
> 80	750	0.0578
> 95	750	0.0933
> 110	750	0.0844
> 125	750	0.1200
> 140	750	0.0711
> 155	750	0.0711
> 170	750	0.0711
> 185	750	0.0578
> 200	750	0.0489
> 215	750	0.0933
> etc
>
> I have repeated the analysis using different quadrat sizes and the problem of 
> incomplete drawing of lines is unaffected.
>
> Any suggestions are appreciated!
> Mike

Paulo Justiniano Ribeiro Jr
LEG (Laboratório de Estatística e Geoinformação)
Departamento de Estatística
Universidade Federal do Paraná
Caixa Postal 19.081
CEP 81.531-990
Curitiba, PR  -  Brasil
Tel: (+55) 41 3361 3573
Fax: (+55) 41 3361 3141
e-mail: paulojus at est.ufpr.br
http://www.est.ufpr.br/~paulojus