A TEST OF THE COMPUTER-SIMULATION MODEL ARCHWHEAT1 ON WHEAT CROPS GROWN IN NEW-ZEALAND

ARCWHEAT1 is a computer model which simulates the growth and development of wheat (Triticum aestivum L.) crops without water or nutrient limitations. Previously, it had been calibrated and tested for crops in the UK and western Europe. This paper presents the results of a more rigorous validation in which predictions from the model were compared with measurements on crops grown at three locations in New Zealand, during three seasons, under conditions of optimum water and nutrients. The conditions of the experiments were outside the model's development and calibration range. Simulation of phenological development, PAR interceptance, PAR-use efficiency and above-ground dry matter production was very good, with root means square deviations (RMSD) of simulations from observations of less than 10%. Simulation of grain growth and yield was only fair, with RMSDs between 20 and 30%. Predictions of final shoot number, grain number and mean grain biomass were poor. The model validation exercise has highlighted areas where further research is needed to provide definitive descriptions of the response of wheat crops to their environments.