Carbon Cycle and Computer Models
So many processes have to be considered in the carbon cycle that it is extremely difficult to keep them in mind, and impossible to calculate without building a computer model to simulate them. Scientists interested in the carbon cycle have built a number of such models over the years. Such models can have between 50 and 100 interacting equations describing all the different processes of the carbon cycle that are relevant to the problem of how carbon dioxide changes through geologic time.

To what extent should the answers generated from such models be trusted? All one can say is this: Models are the best we can do, everything else is ballpark back-of the envelope stuff. This means we should use models to educate ourselves about possibilities, realizing that their output produces probabilities not measurements.