Use of empirical global radiation models for maize growth simulation

• Published in: Agricultural and forest meteorology 2004-11, Vol.126 (1), p.47-58
• Main Author: Pohlert, Thorsten
• Format: Article
• Language: English
• Subjects: Agricultural and forest climatology and meteorology. Irrigation. Drainage; Agronomy. Soil science and plant productions; Biological and medical sciences; Crop growth model; Fundamental and applied biological sciences. Psychology; General agronomy. Plant production; Global radiation models; Missing value problem; Sensitivity study; WOFOST
• ISSN: 0168-1923; 1873-2240
• DOI: 10.1016/j.agrformet.2004.05.003

Global radiation is one of the essential environmental factors which controls both photosynthesis and evapotranspiration. Consequently, it is a driving input parameter for various agro-ecological models. However, it is difficult for model users to obtain high quality and long-term series of records because of the low spatial density of irradiation station networks and missing values within recorded series. The scope of this study is to analyse the use of the (1) Ångstrøm, (2) the Bristow and Campbell (3) and the Allen global radiation models, respectively, for maize growth simulation with the crop growth model WOFOST (Model for WOrld FOod STudies). Two locations in the temperate regions, Wageningen (The Netherlands) and Córdoba (Argentina), as well as one site in the tropics, Los Baños (The Philippines), were selected for this study. The simulated maize yields from the runs with measured data were compared with the runs of two scenarios: (i) daily global radiation is completely estimated for the full season; and (ii) daily global radiation is estimated to close missing values ranging from one to four weeks within each growth period. This study pointed out, that all of the models can be applied to close incomplete global radiation series for maize growth simulation with WOFOST at temperate locations. In contrast the models (2) and (3) may not be used to generate daily data for a full season at tropical locations, as the simulated yield distributions differ significantly ( *P < 0.05). Due to poor prediction skill of the models (2) and (3) at the tropical location, this approach fails for both closing incomplete series and generating data for a full season for maize growth simulation.