Could you please tell me how to perform Constrained Analysis of Principal Coordinates (CAP, Anderson and Willis, 2003) , which only based on environmental data, to test whether land use type (represents different land use intensities) influenced the environmental characteristics of the streams, using R software?
The details are as follows:
###########
data(varespec)#species data data(varechem)#environmental data ## Basic Analysis vare.cap
If I understand correctly, the model you wish is:
env_vars~land_use_type
In this case, Bray Curtis dissimilarity makes little sense, it assumes all variables are on the same scale (as in community ecology). Redundancy analysis (RDA, i.e. capscale with method="euclidean" and predictors present, otherwise it's just PCA) is probably more what you want. Since your variables are probably not on the same scale, you will most likely have to standardize them:
env_vars
Thank you very much Mr.Adeux. I'm so happy. You give me hope.
I think what you said is very reasonable, but I have a question: in the model of “mod
In the vegan example you provide, community data (varspec, your environmental variables) are present in another object than predictors (varchem, land use intensity for you).
Lets imagine all your data is in one file, which you imported in R and named proj, you have land use intensity factor in your first column (named land_use_type), and 5 environmental variables (hence 6 columns total), then you could run:
proj[,2:6]
I accept your advice and sincerely appreciate your help.@Guillaume Adeux
Environmental variable analysis is just my first step.The second step of fish community data analysis is also the core.
Haha, it must have been a tough road for me.I keep trying. Thank you.@Guillaume Adeux
I used Research Gate for the first time to ask questions and added the article information in the attachment.
If you are interested, you can also pay attention to it.
I hope it will help you too.
Their Fig 1 is essentially what you are doing with your env. variables whereas in Figure 2 response matrix is community data (absence/presence with Sorensen dissimilarity and relative abundance with Bray curtis). Here data is not standardized as we did with env. variables. The analysis corresponding to Fig1a would be:
capscale(vegdist(your_community_data,method="bray",binary=T)~land_use_intensity,dist="euclidean",your_data)
In Fig 2b community data is hellinger transformed and then a bray curtis dissimilarity is computed:
capscale(decostand(your_community_data,method="hellinger")~land_use_intensity,dist="bray",your_data)
You are not only handsome, but also excellent!
Thank you very much!@Guillaume Adeux
I'll try to run my data.
The problem has not been solved, and it is not as simple as it seems.
Feel free to share your thoughts.
- Badges
- Science topic