This program enables to calculate modified test of significance for analysis of mean Ellenberg indicator values (EIVs) with other variables derived from species composition. The test of significance is based on modified permutation schema (see more deatails here).
MoPeT_v1.2.zip (last update: 24.10.2013)
MoPeT is based on R program - you need to have R installed on your computer, but you don't need any knowledge of R program. All data can be loaded from spreadsheet format. Optionally, you can run the program also from JUICE software (http://www.sci.muni.cz/botany/juice/) - in this case you need to have both JUICE and R installed on your computer. JUICE is a free program for editing vegetation data, developed by Lubomír Tichý from Masaryk University Brno, Czech Republic.
R.exefile is located - typically it's in
c:\Program Files\R\R-2.X.X\binfolder1), where
R-2.XX.Xis installed version of R (e.g. R-2.14.1)
MoPeT_vX.X.batexecutable file (vX.X in the name is the current version of the script you are using) - the interface of the MoPeT application should open.
Use this function if you want to test the differences between mean EIVs for groups of samples, assembled together e.g. by cluster analysis. One-way ANOVA is calculated, returning p-values based on modified permutation test.
Use this function to calculate significance of correlation of mean EIVs with external variable, which is either derived from species composition (e.g. species richness of samples, scores of samples along ordination axes etc.) or is measured environmental variable. Modified permutation test is based on comparing real correlation coefficient with null distribution of correlation coefficients, generated by modified permutation of mean EIVs. Pearson, Kendall and Spearmann correlation coefficients are available.
Calculates regression of mean EIVs with external variable, which is either derived from species composition (e.g. species richness of samples, scores of samples along ordination axes etc.) or is measured environmental variable. Regression can be calculated either with mean EIVs as dependent or independent variable, modifying the setting of
Dependence of variables in regression (dependent ~ independent). Modified permutation test is based on comparing real F-value of the regression with null distribution of F-values generated by modified permutation of mean EIVs.
This option is suitable, if you aim to calculate relationship of mean EIVs with ordination axes from unconstrained ordination analysis. It is based on an algorithm used in
vegan package in R, written by Jari Oksanen, which calculates relationship between ordination axes and mean EIVs using multiple regression, where mean EIVs is dependent variable and sample scores on ordination axes (1, 2 or more) are independent variables. Advantage of this analysis, compared to calculating correlation of mean EIVs with each axis separately, is lower number of tests (in correlation, number of tests = number of mean EIVs × number of ordination axes, in multiple regression number of tests = number of mean EIVs). Significance is based on F-values as testing criteria, and is calculated by comparing multiple regression F-values of real mean EIVs with null distribution of F-values generated by modified permutations. Importance of individual mean EIVs can be judged by comparing their R2. If you want to use absolute values of R2 or R2adj, note that these are biased (overestimated) due to the similarity issue between mean EIVs and ordination axes (if external variables are ordination axes); not-biased values are represented by variable
R2.adj.iv (but the theory at the background is not yet elaborated, so the use of these values is not recommended).
Forest vegetation plots, located in even distances along transects following the steep valley slopes of Vltava river valley (close to Zlatá Koruna, Czech Republic) and collected during 2001-2003 (Zelený & Chytrý 2007). Each transect started at the valley bottom and end up at the upper part of the valley slope. Plots are of the size 10×15 m. In each plot, all species of tree, shrub and herb layer were recorded and their abundance was estimated using 9-degree ordinal Braun-Blanquette scale (these values were consequently transformed into percentage). At each plot, various topographical and soil factors were measured or estimated - soil pH is reported here. The dataset contains 27 transects with alltogether 97 samples.
All files are in plain text format, with cells separated by tabulators. First row usually contains variable names (or species names in case of vegetation matrix), first column contains plot numbers or species names. See preview for details.
The zip file needs to be unzipped first and than the file with *.wct extension can be opened in JUICE (use
File > Open in JUICE menu, to click on the icon of the file is not enough). Species Ellenberg indicator values are already initiated, i.e. they are assigned to species (you may check this in JUICE menu Indicator values > Initiation). The table is separated into 5 groups, resulting from cluster analysis (Ward algorithm and Euclidean distances).
please write me (firstname.lastname@example.org), I will try to fix it. I really appreciate constructive feedback!