Coherence resonance in paleoclimatic modeling
Aurelien Bosio, Pietro Salizzoni, Carlo Camporeale
Abstract
Abstract Through a unified mathematical framework, the stochastic behavior of three celebrated low-order lumped models, previously proposed for paleoclimate simulations, is considered. Due to the coherence resonance mechanism, the feedbacks between noise and the dynamical system reproduce the hallmark of the Pleistocene climate, i.e. the 100 ky pulsation, in a range of the model parameters that is unexpectedly wide and far from the original modeling setting. In this way, the issue of arbitrary coefficient tuning of lumped approaches in paleoclimatology can be partially bypassed. A stability analysis of the considered dynamical systems allowed the parameter space to be exploited, in order to separate the deterministic-dominated region from the stochastic-dominated region. Noise intensity is varied and the closeness in the parameter space to Hopf bifurcations and/or bistable conditions is investigated in order to understand what conditions make the models prone to coherence resonance with a 100-ky pulsation, with or without the forcing induced by varying astronomical parameters.