Energetics, skeletal dynamics and long-term predictions in Kolmogorov-Lorenz systems

TL;DR

This paper analyzes the energetics and skeletal dynamics of Kolmogorov-Lorenz systems, introducing a new long-term energy-based prediction method and providing insights into regime transitions in chaotic models.

Contribution

It presents a novel energetic approach to understanding regime transitions and introduces a new long-term prediction method based on maximum energy shells in Kolmogorov-Lorenz systems.

Findings

Constructed a skeletal dynamics model reproducing Lorenz map behavior.

Identified the role of energy and Casimir maxima in classifying orbital motion.

Proposed a new energy-based long-term prediction method.

Abstract

We study a particular return map for a class of low dimensional chaotic models called Kolmogorov Lorenz systems, which received an elegant general Hamiltonian description and includes also the famous Lorenz63 case, from the viewpoint of energy and Casimir balance. In particular it is considered in detail a subclass of these models, precisely those obtained from the Lorenz63 by a small perturbation on the standard parameters, which includes for example the forced Lorenz case in Ref.[6]. The paper is divided into two parts. In the first part the extremes of the mentioned state functions are considered, which define an invariant manifold, used to construct an appropriate Poincare surface for our return map. From the experimental observation of the simple orbital motion around the two unstable fixed points, together with the circumstance that these orbits are classified by their energy or Casimir maximum, we construct a conceptually simple skeletal dynamics valid within our sub class, reproducing quite well the Lorenz map for Casimir. This energetic approach sheds some light on the physical mechanism underlying regime transitions. The second part of the paper is devoted to the investigation of a new type of maximum energy based long term predictions, by which the knowledge of a particular maximum energy shell amounts to the knowledge of the future (qualitative) behaviour of the system. It is shown that, in this respect, a local analysis of predictability is not appropriate for a complete characterization of this behaviour. A perspective on the possible extensions of this type of predictability analysis to more realistic cases in (geo)fluid dynamics is discussed at the end of the paper.