Basu_Bundick_2017

[jpfeifer]

1. No, the approximation point for perturbation is still the deterministic steady state.
2. At third order, you get a term g_{x\sigma\sigma} (in SGU (2004) notation) that affects the eigenvalues of the solution. But under pruning, you prevent this term from affecting the eigenvalues. That can change the persistence of the system.