Model Selection for Fitting Photosynthetic Light Response Curves of Cassava

Purpose In order to provide references for cassava’s cultivation techniques and physiological ecology research, the best photosynthetic response curve model for cassava was studied.

Method The photosynthesis light response curve was measured in three growth stages using the Li-6400 portable photosynthesis system. Four light response curve models, including exponential function model, rectangular hyperbolic model, non-rectangular hyperbolic model and modified rectangular hyperbolic model, were applied to fitting its light response curve.

Results All of the determinant coefficients of four models in three growth stages were larger than 0.990, which meant that above four models could fit well the photosynthesis light response curves. The determinant coefficients of modified rectangular hyperbolic model was maximum, MSE, MAE and AIC were minimum, the estimated values such as the maximum of net photosynthetic rate (P_nmax), light saturation point (LSP) and light compensation point (LCP)and dark respiration rate (R_d) by the modified rectangular hyperbolic model were most close to the measured values, so the modified rectangular hyperbolic model was the most proper photosynthetic response curve model for cassava in three growth stages. The fitted initial quantum yield, P_nmax, LSP and LCP were 0.048−0.090, 21.75−22.37, 2 140.25−2 866.15 and 61.48−95.23 μmol/(m²·s), respectively in three growth stages of cassava by modified rectangular hyperbola model.

Conclusion Modified rectangular hyperbola model should be used to estimate photosynthetic light-response curve parameters for cassava.