Preliminary evaluation of two maize (Zea mays L.) growth-simulation models

Abstract

Two maize growth-simulation models developed in the U.S.A., CERES-Maize andCQRNF, were evaluated with South African data using graphical and numerical techniques. The results indicate that both models agree well with observed values of total above·ground plant dry mass, leaf area index, grain yield and soil water content. CORNF tended towards a systematic underprediction of leaf area index. CERES·Maize provided more realistic estimates of soil water content than CORNF. The future use of these models in the evaluation of cropping strategies is discussed.Twee simulasiemodelle wat in die V.S.A. ontwikkel is vir mielies, naamlik CERES·Maize en CORNF, is met Suid· Afrikaanse data getoets d.m.v. grafiese en numeriese tegnieke. Die resultate dui aan dat voorspellings van albei modelle goed met waargenome waardes van totale bogrondse droemassa, blaaroppervlakindeks, saadopbrengs en grondwaterinhoud ooreengestem het. CORNF het :n sistematiese neiging getoon om blaaroppervlakindeks te onderskat. CERES·Maize het meer realistiese beramings van grondwaterinhoud gelewer as CORNF. Die toekomstige gebruik van hierdie modelle vir gewasproduksiestrategiee word bespreek.