•
Breeding is simply running the nonlinear model a
second time, from perturbed initial conditions
