Climate and thermodynamic systems of maximum dissipation

Abstract

THE Earth–atmosphere is a classic example of a closed, dissipative and nonlinear thermodynamic system which is subject to both regular and irregular impulses causing significant departure from steady state. It is closed because it exchanges energy (solar and thermal radiant energy) but not mass with its environment. It is dissipative because the net input of radiant energy occurs mainly in regions of high temperature towards the Equator and the net output occurs mainly in regions of low temperature towards the poles. It is nonlinear basically because of the multiplicity of internal feedbacks and because of the importance of advective processes. It has steady-state character in the sense that the annual mean radiant energy input is very close to the annual mean output, and parameters such as the annual mean temperature do not vary significantly from one period to another. The regular seasonal variation in solar position ensures significant departure from the steady state so defined, and there are also significant irregular departures arising (for instance) from variations of solar input and IR output caused by variations in the amount and distribution of cloud. Recently I have shown1 that the overall Earth–atmosphere climate system seems to have adopted a format whereby the total thermodynamic dissipation associated with the horizontal energy flows in the atmosphere and ocean is a maximum. ‘Format’ in this context refers to the annual average geographic distribution of cloud, surface temperature and the horizontal energy flows. The practical significance of this is that, if one could accept it as a general principle governing climate behaviour, one could use it directly as a means of a priori prediction of climate and climate change without needing detailed analysis of the internal workings of the system. I could not explain previously why the Earth–atmosphere system should be so constrained. This note points out that the Earth–atmosphere has characteristics such that it might be expected to obey such a constraint. Furthermore, these characteristics are sufficiently general that the same principle of selection of steady-state mode of maximum dissipation may apply to a broad class of non-linear systems.