Key points are not available for this paper at this time.
Abstract Crop models habitually run ahead of the measurements needed for rigorous calibration and validation. The treatment of key physiological processes is often very uneven, and the availability of microcomputers encourages modelers to develop complex schemes providing output that is unqualified by errors or uncertainties. Crop models are potentially powerful tools, but they should be used in a more disciplined way.
J. L. Monteith (1996) studied this question.