Modeling crop nitrogen requirements: a critical analysis

Four approaches to simulate N requirements of wheat were tested using data collected at the Auzeville experiment station of TNRA, near Toulouse, France, from an experiment providing a wide range of soil N available for crop uptake. In these approaches, crop N requirements are expressed in terms of characteristic plant N concentration curves (maximum, critical, and minimum), representing expected concentration for a given crop N status throughout the growing season. Modeling approaches were evaluated for their ability to discriminate between N-limited and non-limited wheat plots as well as to properly represent the upper and lower limits of observed plant N concentrations. Best results were obtained using the growth dilution concept to represent the characteristic curves, while others based on temperature sums, growth stages, or fraction of the growth cycle were less satisfactory. Simulation of crop growth and N uptake based on N requirements estimated using the growth dilution concept resulted in a relationship between biomass at harvest and N uptake that correctly described an upper boundary for all observed data points. However, simulated and observed crop N uptake on a plot by plot basis resulted in low agreement. This was attributed to uncertainty in the measurement of initial soil N and crop N uptake, and the effect of other growth reducing factors (e.g. diseases) and possibly physical and/or chemical restrictions to field N uptake normally not accounted for by crop growth models. (C) 1997 Elsevier Science B.V.