I have measured phylogenetic diversity of 3 plots in my preliminary data using R platform. For Faith's PD, I compare the values against a null model and interpret the plot as having high PD if it shows values greater than null. Should the extent to which this value is greater than null be statistically significant (p-values), or should I interpret them concentrating on the standardized effect sizes (z-values)?
Also, the null value of mean pairwise distance (MPD) for all plots is the same. Why is this so? I understand that MPD is an averaging index, but so is MNTD (mean nearest taxon distance) and that gives me different null values for each plot. How should I interpret these values and what is their ecological meaning?
First of all, the observed PD values have to be significantly higher or lower than the null distribution (P-value) before you can say that the observed values differ (significantly) from the null distribution. However, when the observed PD is higher than the null PD, you can't just say that you have a 'high PD'. You're observed PD is just higher than the null PD. This means that you found more phylogenetic diversity in your plot than would randomly expected. In the literature, you can find different explanations of the deviations from the null distribution (e.g. lower than null may indicate phylogenetic conservatism). In my paper "Impact of land-use intensity on the conservation of functional and phylogenetic diversity in temperate semi-natural plant communities" in Biodiversity & Conservation, I did more or less the same analysis. You can find more information and the interpretation of the results there.
What do you mean with: "the null value of mean pairwise distance (MPD) for all plots is the same". Which null model did you use? And you calculated the null values 999 times (or more)? MPD is the average distance between two random individuals in a sample, while MNTD is the average distance to the closest heterospecific individual for all individuals in the sample (if abundance weighted).
Koenradd, Thanks a bushel-full for that. I will read your paper.
As for the null model MPD values, the null model was "richness" meaning it would maintain the same richness of a given plot while picking species randomly from the species pool. All three plots had same null, while for MNTD they differed. I think it has something to do with the way MPD is calculated. I used the picante package for R for calculations. All models were run 10000 times.
Yes, Alex, I have. The nulls were run to correct for that correlation. I was a little confused if I should use nulls for MPD, because MPD already corrects for it by being an averaging index. Any thoughts on that?
- Badges
- Science topic
- Similar topics
- Geoscience
- Asia
- India