Abstract: A field experiment for the AD268 with large leaf and dwarf structure and the XY335 with small leaf and high structure at three planting densities of low (3 plants·m⁻²), medium (6 plants·m⁻²) and high (9 plants·m⁻²) was conducted. This study integrated the vertical distribution of canopy photosynthetically active radiation (PAR), a 3D canopy model, and a canopy photosynthesis model to analyze responses of the extinction coefficients estimated by the logarithmic approach (k_log) and the regression approach (k_nls) to cultivar and density and their differences, and to evaluate the effects on simulating canopy light utilization. The results showed that there was a difference between k_log and k_nls, and this difference was more obvious for AD268 and aggravated with increasing density. The fitness based on k_nls for canopy light distribution (R²=0.86) was better than k_log (R²=0.77). The daily CO₂ assimilation, above-ground dry matter accumulation and radiation use efficiency of AD268 simulated by k_log and k_nls did not differ significantly at low densities, but the difference became larger with increasing density. The critical densities at which the difference become larger were 0.7-2.5plants·m⁻².

Key words: Crop model, Parameter estimation, Radiation use efficiency, High photosynthetic efficiency, Ideotype