[Optimal model of photosynthesis-light response curve in canopy of planted Larix olgensis tree.]

Rectangle hyperbola model (RH), nonrectangle hyperbola model (NRH), exponential model (EM), modified rectangle hyperbola model (MRH), and modified exponential model (MEM) were applied respectively for modeling the photosynthesis-light response curves (PLC) based on four types of curves (photosynthesis-light response curve of inhibition, PLC_i, photosynthesis-light response curve of saturation, PLC_s, photosynthesis-light response curve of unsaturation, PLC_u andphotosynthesis-light response curve in weak light, PLC_w) in canopy of 15 years old planted Larix olgensis tree in the Maoershan Forest Farm, Heilongjiang Province. The major photosynthetic physiological indexes including maximum net photosynthetic rate (P_{n max}), dark respiration rate (R_d), light compensation point (LCP), light saturation point at saturated light intensity (LSP) and apparent quantum yield (AQY) were calculated. All the five candidate models were comprehensively compared by the model goodness-of-fit and the precision estimations of photosynthetic physiological indexes in the four types of curves. The results showed that MEM model was only suitable for fitting the PLC_i. MRH model showed the best goodness-of-fit for PLC_i and PLC_s(R_a²was 0.9986 and 0.9978, respectively). Meanwhile, NRH model expressed the best fitting result in PLC_u and PLC_w (R_a² was 0.9996 and 0.9963, respectively). MRH had the lowest mean absolute value of relative error (MAPE) when evaluating P_{n max} in different types of curves (0.1%s and PLC_u, as well as R_d of PLC_u(MAPE was 1.8%, 0.1%, and 3.9%, respectively). RH model represented better prediction effect for LCP of PLC_i and R_d of PLC_s(MAPE was 1.0% and 2.7%, respectively). EM was more sui-table for estimating LCP of PLC_w(MAPE was 0.2%). Comprehensive analysis showed that MRH model not only proved well model goodness-of-fit and accurate estimation of photosynthetic physiological indexes, but also showed the highest stability when fitting different types of curves. As a result, MRH model was finally selected as the optimal canopy photosynthesis-light response model of planted L. olgensis tree.