opetchey/multifarious_response_diversity

Measuring the response diversity of ecological communities experiencing multifarious environmental change

This repository contains data, analysis, and supplementary materials for the empirical example described in the manuscript titled "Measuring the response diversity of ecological communities experiencing multifarious environmental change ".

Contents

Data: Contains the datasets used in the empirical example shown in the Appendix.

Experiment: Includes the RMarkdown file (Creating the data for Response diversity in the context of multifarious environmental change.Rmd) used for the analysis and generating results mentioned in the main text.

Management: Information regarding project management and author contributions.

Second version: Houses the R code to reproduce figures as well as the Supplementary material accompanying the main text.

old: Contains deprecated or older code used in exploratory analysis.

r: Stores custom R functions used in the analysis.

Usage Data: The raw datasets are located in the Data folder.

Experiment: Run the Creating the data for Response diversity in the context of multifarious environmental change.Rmd file in the Experiment folder for reproducing the analysis.

Second version: Refer to the Appendix for supplementary details and the "make_figure_v2" R file for reproducing figures.

r: Contains custom functions used in the analysis.

multifarious_response_diversity's Issues

Using other function for simulating performance data?

If we used other functions, as well as the current one, and find that we can still measure response diversity, we gain some confidence that the measurement method is robust. Need to decide if we do this.

Some other functions are mentioned in Frank's comment in issue #1.

Mathematical quality

I started some work on simulating the dependence of a rate on two environmental variables, when they act additively and non-additively. I've started by using the Eppley function (see the report for details). The additive version is fine. But while I have added an interaction, the maths is rather low quality. E.g. changing the interaction term changes multiple things, and not just one (e.g. the optimum). I might be able to make something nicer, but probably someone with more mathematical skills could do so more efficiently and probably produce something better. I thought to first ask Shyamolina if she could help. And after that a professor here, Reinhard Furrer. I've worked with him before and would be happy to again. If either of them help with this, then I think we must add them to the project team (meaning among other things they will be co-author).

Interested in your thoughts. (I think that all current team members should approve addition of any new team members.)

Gaussian GAMs only?

Getting the partial derivatives is more straightforward for Gaussian GAMs due to the identity link function. Get’s harder with non-Gaussian.

Owen proposes that we constrain this project to Gaussian ones. But would also mean we need real data that are compatible.

Real data?

Is it possible to include an example using real data?

Calculating partial derivatives

Seems that gratia::derivatives may not, currently, be able to do what we need.
gavinsimpson/gratia#101
https://stats.stackexchange.com/questions/495775/first-derivative-of-fitted-gam-changes-according-to-specified-model-distribution

Indeed, gratia::derivatives does not work when it is given a gam that includes a term that is something like an interaction term, e.g. gam(y ~ ti(x1) + ti(x2) + ti(x1, x2)
(https://www.rdocumentation.org/packages/mgcv/versions/1.8-40/topics/te states that ti is good when setting up a gam with main effects and interaction structure.

I know next to nothing about the inner workings of gams, so would either have to put in a lot of time. Perhaps one of you knows enough / wants to learn enough. Or we get someone to help.

Please give permission to make repo public

Dear @gavinsimpson
Dear @SamRPJRoss
Dear @FrancescoPola

Please would you give me permission to make this repo public? Doing so will make everything public, including conversations in the issues. I reviewed the conversations and see no problem making them public, but I feel it would be good for all of us to give permission to make them and everything else public.

Best wishes
Owen