A modelling spectrum for carbon cycle studies

The concept of a spectrum of mathematical modelling covering varying degrees of inductive versus deductive modelling is used to characterise a range of studies of the global carbon cycle. Low resolution carbon cycle models are considered as examples of models from the middle of the spectrum. It is suggested that such models should be calibrated using techniques similar to ridge regression. By varying the ridge parameter it is possible to explore the consequences of decisions concerning the relative importance of deductive versus inductive aspects of the model calibration.