Companion code to Inferring species interactions using Granger causality and convergent cross mapping by F. Barraquand, C. Picoche, M. Detto and F. Hartig. http://arxiv.org/abs/1909.00731
Linear Granger causality and convergent cross-mapping are implemented using R
. Here, we stick to time-domain approaches from packages vars
and lmtest
for Granger causality, as well as SIMoNe for regularized models, and call rEDM for convergent cross-mapping (see Hao Ye et al.'s tutorial for more information).
The analyses are organised in folders corresponding to our case studies, with 2 species interacting systems, 2 species and some added abiotic forcing, 10 and finally 20 species networks; see the Methods of the arXiv preprint for a description.
Lyapunov
computes the Lyapunov exponents for all case studies. The READMEs within each folder add more information about their structure and content.