Limitations and Failures of the Milankovitch Theory

Although variations of the Milankovitch theory can account for some aspects of climate change, there are serious failures that require attention. In particular, we will discuss the status of the "causality problem", the apparent fact that major shifts in climate occur prior to the Milankovitch driving force, and the spectral peak problem, in which the spectral shapes predicted by the Milankovitch theory do not match those in the spectrum and the bispectrum of the data. The standard Milankovitch theory ascribes all climate change to the same mechanism: summer insolation in the northern hemisphere. Variations in cloud cover are ignored, even though the net forcing of clouds is (at the present time) approximately 30 Watts per square meter -- substantially greater than the rms variations in insolation (18 Watts per square meter). Thus changes in cloud cover could be more important than changes in the standard Milankovitch parameters. We will discuss mechanisms that link variations in the orbital inclination of the Earth to changes in cloud cover, and how these address the causality problem, the spectral problems, and several other failures of the Milakovitch theory. How can we reconcile the failures of the Milankovitch approach with its obvious successes (e.g. in accounting for the 23 kyr cycle in sapropels, and the atmospheric oxygen variations in the Vostok data)? The answer is that climate is multi-dimensional. Insolation certainly affects climate. But we should not make the logical mistake of therefore assuming it accounts for the 100 kyr cycle of glaciation that has dominated for the past 800 kyr.