摘要
以油料向日葵为研究对象,采用田间试验对北方盐渍化地区土壤-植物-大气连续体中不同界面的能态进行了研究。揭示了盐渍化地区SPAC系统内的水分迁移机理。通过相关性分析得出叶水势的影响因素有:土壤水分、盐分,净辐射,大气湿度,大气温度。在分析影响因素的基础上,建立了基于BP神经网络的盐渍化地区油料向日葵叶水势动态模型,通过网络训练,确定网络的拓扑结构为7∶24∶1,模型的平均相对误差为13.8%,符合田间试验要求。
Taking oil sunflower, which is local main economic crop, as research object, energy state of different interfaces of Soil- Plant-Atmosphere Continuum in saline soil of north china was researched through field experiment. The result revealed the water translocation mechanics of SPAC system. Influencing factors of leaf water potential includes soil water content, soil salt, net radia tion, atmosphere humidity and atmosphere temperature. On the basis of analysis on the influencing factors, dynamic model based on BP neural network for leaf water potential of oil sunflower in saline soil was established. Through network training, the topology structure was chosen as 7 ： 24 ： 1, and the average absolute relative error was 13.8%.
出处
《节水灌溉》
北大核心
2006年第5期22-25,28,共5页
Water Saving Irrigation
