Articles
CROP MODELLING IN THE USA
Article number
248_46
Pages
365 – 372
Language
Abstract
The whole agricultural research infrastructure in the USA is organized for dealing with results from field plot experiments and similar analytical, reductionist research.
In many ways this infrastructure is unsuitable for crop modelling.
For the first 15 years after computers became widely available, modellers experienced a lot of opposition, but this diminished as other researchers and administrators began to realize the great potential of modelling.
In the last five years, crop modelling has become an acceptable and even a fashionable research activity in the USA. Instead of defending their work from criticism and hiding their problems, modellers are now sharing their code more freely and discussing their problems more openly.
The desire and need to share code has led to attempts to standardize model structure, input data and data format.
It has also prompted efforts to start a peer-reviewed electronic journal to "publish" all forms of agricultural software.
There is a realization that mechanistic models which describe causal relationships are more versatile than empirical models which use fitted regression equations.
One crop simulator, and an expert system that drives it, is already in use on USA cotton farms.
In many ways this infrastructure is unsuitable for crop modelling.
For the first 15 years after computers became widely available, modellers experienced a lot of opposition, but this diminished as other researchers and administrators began to realize the great potential of modelling.
In the last five years, crop modelling has become an acceptable and even a fashionable research activity in the USA. Instead of defending their work from criticism and hiding their problems, modellers are now sharing their code more freely and discussing their problems more openly.
The desire and need to share code has led to attempts to standardize model structure, input data and data format.
It has also prompted efforts to start a peer-reviewed electronic journal to "publish" all forms of agricultural software.
There is a realization that mechanistic models which describe causal relationships are more versatile than empirical models which use fitted regression equations.
One crop simulator, and an expert system that drives it, is already in use on USA cotton farms.
Authors
B. Acock
Keywords
Online Articles (59)