Spectral analysis and the astronomical theory of climatic change

Abstract

THE astronomical theory¹ attributes long term changes in climate to changes in the Earth's orbital geometry. This theory has become increasingly accepted in recent years. To this acceptance, workers associated with the CLIMAP project have added exceptionally strong support² using spectral analysis of deep-sea core data. They have linked periodicities in the Earth's orbital parameters with periodicities in deep-sea core parameters related to sea-surface temperature, ice volume and oceanic salinity. Their argument rests on the assumption of a linear response between input (radiation changes caused by changes in orbital geometry) and output (the climate-dependent parameters mentioned above). At the same time they have attributed the anomalously high variance associated with the longest periodicity (∼ 100,000 yr) to a nonlinear response. There is no contradiction here because the response function, the climate ‘black box’, most probably has both linear and nonlinear components. I examine here the question of nonlinear response in more detail.