Simulation
of uptake of water and N requires demand and supply functions for both. Simulating crop response requires assessment
of the stress associated with any shortage.
Demand for water is calculated from very well tested evapotranspiration
formulae based on the physics of the process.
In day timestep models, stress is assessed
using the ratio of the daily demand to daily supply, the latter based on consideration
of the physics of water transport in soil and plant. Stress occurs when the demand rate exceeds supply
rate. The use of a stress index that
affects plant processes has been very successful in simulating the effects of
water stress without considering the details involved. Treatment of N in models has been very
similar to their treatment of water, despite most water departing from the
plant the same day it is taken up, while most N is retained. Demand is set to meet minimum and maximum
concentrations of N that change with ontogeny or biomass, with the optimum
concentration being about midway between these.
N supply is calculated daily from soil processes. N-stress occurs and affects other processes
when the crop is at lower than optimum N concentration. Calibration of such a model is data intensive
and empirical. We propose an alternative
approach that is both more mechanistic and simpler to implement. Plant N is assigned into three pools by
priority. First priority is to
structure. N in this pool is not
translocatble elsewhere. Second priority
is to green tissue, and is assigned per unit green area. Third priority is labile storage. By assuming that specific leaf N
concentration is constant, N effects are expressed through their effects on
light interception with no effect on light use efficiency. We describe implementation in models of
wheat, potatoes and maize and compare simulations with independent data.