Canegrowers require cost effective nitrogen (N) management procedures
which ensure that N does not limit productivity or profitability. The N cycle
is complex, with many crop and soil processes interacting with variable weather
regimes on a diverse range of soils. The prospects for producing models to aid
nitrogen research and management is considered in this paper. Modelling concepts
are introduced by examining two basic classes of model, within which a wide
variation in scope and complexity is possible.
Static models do not explicitly contain time as a variable. Examples considered here include multiple regression equations relating cane yield to applied fertiliser, but specified at a number of empirical levels. The economic attraction for erring on the high side of N fertiliser application rates, in the face of both uncertainty and risk in the shape of the response curves, is demonstrated with these models.
Dynamic models simulate variation in crop or soil properties and processes
over time. Examples considered include CENTURY - a model of soil organic
matter cycling configured for sugarcane production systems with monthly time
steps, and APSIM - a framework linking N and water models to a sugarcane
crop module that operates on a daily time step. Priority areas for the application
of such models include research on maintenance of soil fertility levels in the
long-term and on reducing losses of N through processes such as leaching.