Towards a Fluctuation Theorem in an Atmospheric Circulation Model

TL;DR

This paper investigates the distribution of trajectory divergence in an atmospheric circulation model, revealing that the probability of negative local Lyapunov exponents aligns with Fluctuation Theorem predictions, indicating potential reversibility in atmospheric dynamics over certain timescales.

Contribution

It demonstrates the applicability of the Fluctuation Theorem to an atmospheric circulation model by analyzing local Lyapunov exponents and their probability distributions.

Findings

Negative local Lyapunov exponents occur with significant probability over 10 days.

The probability of negative exponents decreases over time, consistent with Fluctuation Theorem predictions.

The effect persists across different model resolutions and vertical levels.

Abstract

An investigation of the distribution of finite time trajectory divergence is performed on an Atmospheric Global Circulation Model. The distribution of the largest local Lyapunov exponent shows a significant probability for negative values over time spans up to 10 days. This effect is present for resolutions up to wave numbers l=42 (~ 250km). The probability for a negative local largest Lyapunov exponent decreases over time, similarly to the predictions of the Fluctuation Theorem for entropy production. The model used is hydrostatic with variable numbers of vertical levels and different horizontal resolutions.