Towards the quantitative control of crop production and quality. I. The role of computer models in soil and plant research

Abstract

1.	A model contains a heirarchy of sub‐models
2.	At each level the sub‐models contain: Facts Mechanistic relations Empirical relations often derived from the output of lower order sub‐models.
3.	Model inputs often contain stochastic elements whose effect on output can be assessed statistically.
4.	When testing models, empirical inputs should, if possible, be derived from independent experiments.
5.	The data from successive years of experiment may be used to update the model parameters.
6.	When the sensitivity of the output to its inputs has been determined, the model may be simplified, and models of different crops combined into a comprehensive model.
7.	Predictive models with a practical aim should be distinguished from theoretical models which must be rigorously tested at their extremes if they are to be corroborated.