Detecting Causality in Complex Ecosystems

Summary

Writing in Science, George Sugihara, Robert May and five colleagues introduce a method “that can distinguish causality from correlation” in complex, nonlinear systems, including tangled cases “not covered by the current Granger causality paradigm”, the usual statistical test. They try it on simple models and on real ecosystems, among them “the controversial sardine-anchovy-temperature problem”.

Rationale

We called tracking cause and effect “maybe impossible”. Sugihara’s team shows that in some tangled systems it can be done, given the right data – data cultural change rarely leaves.