So here is a quick English translation of my mail by 9/10/2000
There were effectively substantial changes between the current version and
the version which allowed to realise RLQ graphs in the 97' topic
documentation (Thema44.pdf). Check one can take out there in the following
way.
The Tables:TabMeanVar program allows to do the equivalent of the figure 17
of the documentation (average position of species along a gradient of
samples). It is enough to use the following parameter setting (I use the
name as in the doc but the filenames will be evidently those you will have
chosen):
Input table File : AuFau (faunistic table)
X-position file : MoyCond (File that contains conditional means for each
species obtained from MatAlg: Diagonal Inner Product see p.56))
Y-position file : A.&&L1 (File that contains sample scores from RLQ
analysis)
Use ranking on X and ordination on Y.
Copy-Paste in a drawing software. A small rotation of 90 ° clockwise is
enough for obtaining the figure 17.
(One can do a " species labelling " with graph1D: labels by using the file
MoyCond and species names. Attention to do a rotation to 180 ° clockwise to
put them in correspondence with the previous graph. Use the commands
"vertical or horizontal mirror" on your graphic software).
The equivalent of the figure 18 can be obtained with the module TabCat. The
option TabCat: Values allows to position the values of the weights of
variables by using the average positions of the species. It is necessary to
use the same dialogue as on the page 57 of the documentation:
---.cat file : AuEsp.Cat (table of trait categories)
Option: Row score X : MoyCond (File that contains conditional means for each
species obtained from MatAlg: Diagonal Inner Product)
Option: Row weight file : AuEsp.cmpl (File that contains the weights of
species in the MCA of traits)
To obtain a summary of these distributions, it is necessary to use TabCat:
MeanVar with the same dialogue entry. One obtains the position of the trait
categories at the average of their distribution among the species plus the
amplitude of distribution. One verifies that these results are in accordance
with those obtained in the Figure 18 of the documentation (without
histograms).
By superimposing the results of TabCat: Values and TabCat: MeanVar one has
the completeness of the operation of averaging.
I hope this is helpful. Note that the graphical presentation of data along
one axis is well documented in Thema27.pdf . Have a look also to the
documentation of modules.
Best wishes
----------
De : "Byungok Ahn" <byungok.ahn@uni-essen.de>
À : "adelist" <adelist@biomserv.univ-lyon1.fr>
Objet : RLQ graphics
Date : Mar 21 aoû 2001 1:19
Hi ADElist
Being much indebted to Sylvain Dolédec´s advice I was able to run a RLQ
analysis.
I have plotted the first ordination axis of the RLQ analysis of the
environmental variables (Doc. Fig. 16).
But I still have problems with presentation of the RLQ results for the
species abundance (Fig. 17, especially, how can I label the species
positioned at the weighted average (circles) of sites where they occur?) and
species traits (Fig. 18). I just try to follow the explanation of Sylvain
Dolédec (ADElist 09/10/2000) but I fail in getting the graphic which is
close to Fig. 18 (I am very poor at French!). Please note that I used FCA
instead of MCA to analyze the species traits.
Can someone tell me how to plot these graphics?
All the very best,
Byungok Ahn
--------------------------------
University of Essen
Institute of Ecology
Department of Hydrobiology
D-45117 Essen
Germany
Phone: +49/2 01/1 83-43 93
Fax: +49/2 01/1 83-44 42
This archive was generated by hypermail 2b30 : Thu Jul 04 2002 - 09:39:13 MEST