烟草光合-蒸腾速率日变化估算模型研究

摘要

In order to rapidly and accurately estimate the growth status of tobacco plants under natural conditions,and to im-prove the precision management level of tobacco production, the author set up standard models to simulate the daily variation in photosynthesis-transpiration rate of tobacco at vigorous growth stage by using principal component analysis and neural network a-nalysis, and verified the universality of these models. The results showed that the established multiple regression models about the factors(photosynthetic active radiation,leaf surface temperature,stomatal conductance,and intercellular carbon dioxide concen-tration)regulating the leaf photosynthesis-transpiration rate of tobacco variety Yuyan No. 10 had high fitting precisions,and their correlation coefficients were more than 0.87??. These models were used to simulate the photosynthesis-transpiration parameters of Qinyan 96,and the correlation coefficient between actually-measured values and fitted values was above 0.91??. The constructed simulation model by using neural network analysis had a higher fitting precision than that by using principal component analysis.%为了快速、准确估测自然条件下烟草生长状况,提升作物精细管理水平,建立了烟草光合蒸腾速率日变化标准模型;利用主成分分析与神经网络分析方法拟合旺长期烟草的光合蒸腾速率日变化的过程,并对该模型的通用性进行了验证.研究表明,对豫烟10号光合蒸腾速率的调控因子叶片截获的光合有效辐射(PAR)、叶片表面温度(T)、气孔导度(gs)和胞间CO2浓度(Ci)建立的多元回归模型的相关系数均在0.87??以上,模型的拟合精度较高.利用该模型拟合秦烟96的光合蒸腾生理参数,实测值与预测值的相关程度也达到0.91??以上.与主成分分析建立的拟合模型相比,用神经网络进行演算的拟合模型精确度更高.