Crop modelling: current status and opportunities to advance

Abstract

Systems approaches are an appropriate means to meet the complex challenges facing managers of agricultural systems. Crop modelling can play a significant part in systems approaches by providing a powerful capability for scenario analyses. Crop modelling has developed extensively over the past 30 years and a diverse range of crop models are now available. It is argued, however, that the tendency to distinguish between, and separate, the so-called “scientific” and “engineering” challenges and approaches in crop modelling has constrained the maturation of modelling. It is considered that effective crop modelling must combine a scientific approach to enhance understanding with an applications orientation to retain a focus on prediction and problem-solving. The major issue in effective crop modelling is seen as how to achieve the appropriate balance between simplicity and complexity in combining the biology, physics, and prediction requirements for each specific task. Avoidance of unnecessary complexity and maintenance of transparency are considered as guiding principles.
A modular and generic crop template is presented as the appropriate vehicle to advance the science of crop modelling by facilitating the substantive dialectic needed to determine how to best model crops at the functional component or sub-module level. Such a template provides a workable means to compare modelling approaches at component level. This component comparison cannot be achieved through whole model inter-comparison studies. Within existing crop models, algorithms for sub-modules range from purely descriptive to derived from sound understanding. It is argued that seeking approaches to derive crop responses, rather than just describe them, is the means to advance crop modelling. This may involve quantifying the operative functional controls that determine crop responses. This gives rise to derived relationships or emergent properties and forges the connection between modelling and research. This approach is discussed in relation to modelling biomass allocation.

Given a comprehensive crop model with robust predictive capability, there are many opportunities for applications ranging from research to crop improvement. In the controlled environments of protected cultivation, seeking optimal combinations of environmental control and crop management strategies to maximise profitability is feasible using optimisation algorithms in conjunction with greenhouse climate control and crop models. It is argued that a participatory approach that includes managers as partners in this process is required to effect change.