Crop Modeling: From Infancy to Maturity

Abstract Crop modeling, the computerized simulation of dynamic crop systems, was born about 30 years ago, when systems analysis and modern computers presented a new technique to crop scientists. Since then, crop modeling has gone through a number of developmental stages, similar to those of living organisms. From its infancy, crop modeling seemed to promise a well‐behaved, elegant surrogate for ambiguous and cumbersome field experimentation. Indeed, some of the earliest models proved to be among the most notable achievements to date. During the juvenile stage that followed, there was an impressive increase in complexity and computer sophistication, accompanied by some of the growing pains of childhood. Greater expectations led to more and more detailed descriptions of the functioning of the biotic and abiotic components of cropping systems. The results were often trivial, and the big payoff tended to recede into the future, but the need for predicting future crop performance for management and hypothesis testing, together with progress in crop science and computer technology, spurred crop modeling. The next phase, adolescence, a period marked by intense activity, confusion, and excessive confidence‐sometimes challenged by doubt‐appears to be extending into the present. Not only is the original promise turning out to be elusive, but widely accepted guidelines for scientific modeling, such as greater reductionism, universality, and validation, are being questioned. Maturity may be emerging as expectations become pragmatically adjusted to reality. Crop modeling, like advanced ecological modeling, is proving to be more a heuristic tool than a surrogate for reality. In academic, research, and applied roles, such models can be of great value when used as aids to reasoning about the functioning and response of crop systems under many relevant, nontrivial scenarios.