Abstract: [Objective] The purpose of this study is to analyze the relationship between measured spectrum and photosynthetic light utilization efficiency (LUE) of the sunlit leaves of the main tree species in the Maoershan Forest Farm. The definition of photochemical reflectance index (PRI) is modified to obtain a stable relationship between the PRI and LUE of the sunlit leaves, which provides an important basis for accurately estimating gross primary productivity (GPP) and net primary productivity (NPP) by the carbon cycling model. [Method] The data were collected from 10: 00-15: 00 AM in August 2016. LUE was measured by Li-Cor6400 photosynthetic apparatus, and leaf reflectance was measured in laboratory by the SVC HR-1024i. The correlation and regression analysis of PRI and LUE were carried out by SAS software. The relations between multi-band reflectivity and LUE were analyzed by stepwise regression analysis. The higher correlation bands with LUE were selected to modify PRI, and the relationship between modified PRI and LUE was established in coniferous and broadleaved species. [Result] 1) The correlation coefficient between traditional PRI and LUE was low. The Pearson correlation coefficient was 0.234 (P=0.079) for all sunlit leaves of coniferous and broadleaved species, among which 0.121 (P=0.389) for broadleaved species and 0.467 (P=0.174) for coniferous species. Further regression fitting for conifers showed a lower correlation (R²=0.218 2; P=0.174), and poor applicability of traditional PRI for this study. 2) By stepwise regression analysis of multiple-band reflectivity, the multiple regression models of the reflectivity of five bands were obtained after screening, and the correlation was improved (R²=0.402 2, P=0.04). The contribution of reflectance at 518 nm to the LUE was the largest with R²=0.133 4. Based on the definition of traditional PRI, the PRI index was modified by replacing the reflectivity at 531 nm with the reflectivity at 518 nm, and a new regression model was established between modified PRI and LUE. 3) For all sunlit leaves of coniferous and broadleaved species, the modified PRI was well related to LUE (R²=0.496 9, P<0.001; RMSE=0.024 7); for sunlit leaves of broadleaved species, R² and RMSE were 0.4045(P<0.001) and 0.0236, respectively; for sunlit leaves of coniferous species, R² and RMSE were 0.5385(P<0.001) and 0.018 9, respectively. The results derived from the new model were significantly improved compared with the results of using the multiple regression model constructed by the traditional PRI and LUE. [Conclusion] Compared to the traditional PRI, the modified PRI index can improve the relationship between the PRI and LUE of all sunlit leaves of coniferous and broadleaved species in Maoershan region.

Key words: sunlit leaves, modified photochemical reflectance index, photosynthetic light use efficiency, stepwise regression