I am working with Environmental modeling specifically with the interaction between forest -soil-water and environmental parameters
I found that CCA could be a better option to use Canonical correspondence analysis in R software. But, Facing a challenge with the R- Language or R_ script to the canonical correspondence analysis Particularly “CCA triplot”.
I know, it is hard. My knowledge of R is very limited. But you might be able to find a manual online or in the program? Or email staff at different universities :)
Yes, R is not an easy trip. You have to invest some time to get familiar with the software. But you might find some help for a start by viewing the following presentation, based on CCA of Mexican plants. I hope you have at least some basic experience with R:
- http://evol.bio.lmu.de/_statgen/Multivariate/11SS/cca.pdf
- You may find the underlying data here: http://evol.bio.lmu.de/_statgen/Multivariate/13SS/MexicanPlants.txt
Use a recent version of R. Install the package vegan and load it according to the presentation text. Installation:
- https://cran.r-project.org/web/packages/vegan/index.html from CRAN or a mirror site near to you.
- https://cran.r-project.org/web/packages/vegan/vegan.pdf as short manual
The presentation will lead you through all the necessary steps including two triplots.
Good Luck!
Before talking about Canonical Correspondence Analysis (CCA), we must first talk a little about Correspondence Analysis (CA)*. What is the CA? Similarly to Principal Component Analysis (PCA)**, it is an multivariate ordination analysis which uses variables from a single data set to find new variables or axes (orthogonal components) that are most able to explain the data set. However, the PCA assumes that observations (for example, species distributions) are linearly (or monotonically) related to each other and to the gradients (for example, environmental variables). On the other hand, the CA expects observations (for example, species distributions) to have unimodal distributions along the gradients (for example, environmental variables). Notably, this is usually the case for species distributions (see the figure below). A species usually shows highest abundance not in the lowest or highest value of an environmental gradient, but usually near an intermediate level. This is why the CA (and also CCA) are so used by ecologists.
https://cran.r-project.org/web/packages/vegan/vegan.pdf as short manual
What is your objective? Is it a sort of detecting relationships? I will share you some if you could tell me why you want to use CCA.
Dear Kiflay, thank you so much for your help
i want to see the species richness vs topographical, slope, climatic factors as well as detecting the relationship and also need to do regression model with different factors.
my email is [email protected]
- Badges
- Science topic