荒漠河岸林典型植被胡杨叶片气孔导度模拟研究  

Simulation of Leaf Stomatal Conductance of Populus euphratica Forest in Ejina Desert

作　　者：李端 李继彦 王佩将 原黎明 LI Duan;LI Jiyan;WANG Peijiang;YUAN Liming(School of Geograph Science,Taiyuang Normal University,Jinzhong Shanxi 030619)

机构地区：[1]太原师范学院地理科学学院,山西晋中030619

出　　处：《安徽农学通报》2023年第7期82-89,共8页Anhui Agricultural Science Bulletin

基　　金：山西省高等学校科技创新项目(2022L414,2022L411);山西省高等学校人文社会科学重点研究基地项目(20210141)。

摘　　要：胡杨是西北内陆干旱区生态系统的重要组成部分,气孔在维持胡杨正常的生理活动和其生态系统功能的发挥有着非常重要的作用。本文以2019年7—8月份额济纳荒漠河岸胡杨林叶片气孔导度为研究对象,结合胡杨生理指标以及环境因子实测数据,基于Jarvis模型、BWB模型和Bp神经网络模型3种模型,分别对胡杨叶片的气孔导度值模拟,并将模拟结果与实测结果进行比较,确定胡杨气孔导度的最适模型。结果表明,使用BWB模型模拟的气孔导度,其RMSE、RE和MAE值分别为0.071、0.453和0.058,这些相关数值均小于其他2种模型,且其决定系数R^(2)值为0.636,在3种模型中属于最大值,可见使用BWB模型的模拟效果最好;使用Bp神经网络模型模拟的效果次之,其RMSE、RE、MAE和R^(2)值依次为0.111、0.888、0.090和0.318,决定系数值处于居中位置;相比而言,使用Jarvis模型模拟的效果相对较差,其RMSE、RE和MAE值分别为0.838、4.958和0.788,这些数值均大于其他2个模型,且其决定系数R^(2)值为0.261,在3个模型模拟中最小。因此,对荒漠河岸林典型植被胡杨叶片气孔导度的模拟研究表明,BWB模型的模拟精度较高,由其模拟的叶片气孔导度值更接近基于Li-COR 6400便携式光合作用系统测定的真实值。Populus euphratica is an important part of the ecosystem in the inland arid area of northwest China.Stomata plays a very important role in maintaining the normal physiological activities of P.euphratica and the functioning of its ecosystem.In this paper,the stomatal conductance of leaves of P.euphratica forest in the Ejina desert riverbank in July and August 2019 was taken as the research object.Combined with the measured data of physiological indexes and environmental factors of P.euphratica,the stomatal conductance of leaves of P.euphratica was simulated based on three models,namely the Jarvis model,BWB model,and Bp neural network model,and the simulated results were compared with the measured results.Determine the optimum model of stomatal conductance of P.euphratica.The results showed that the RMSE,RE,and MAE values of stomatal conductance simulated by the BWB model are 0.071,0.453 and 0.058,respectively.These related values are all smaller than the results simulated by the Jarvis model and BWB model,and the R^(2) value is 0.636,which is the maximum value among the three models.It can be seen that the BWB model has the best simulation effect;Bp neural network model was used to simulate the second effect,the RMSE,RE,MAE and R^(2) values were 0.111,0.888,0.090 and 0.318,and the determination coefficient was in the middle position.In comparison,the simulation effect of the Jarvis model is relatively poor.Its RMSE,RE and MAE values are 0.838,4.958 and 0.788,respectively.These values are all larger than the other two models,and its coefficient of determination R^(2) is 0.261,which is the smallest among the three models.Therefore,for the simulation of leaf stomatal conductivity of P.euphratica forest in the Ejina desert riverbank,the simulation accuracy of the BWB model is higher,and the simulated leaf stomatal conductivity value is closer to the real value determined based on Li-COR 6400 portable photosynthesis system.

关 键 词：气孔导度 Jarvis模型 BWB模型 BP神经网络模型 胡杨 

分 类 号：S792.11[农业科学—林木遗传育种]