摘要
该文对作物生长模型模拟的数据和观测数据之间的同化进行了研究。采用农业技术转移决策支持系统(DSSAT)中的小麦生长模型对叶面积指数(LAI)进行了模拟,应用插值方法解决模型模拟的连续值和地面观测的离散值的时间尺度的匹配问题,结果表明,与最邻近点插值法、线性插值法、三次样条插值法和立方插值法的插值结果相比,表明立方插值法对实测的LAI的插值效果较好。其次是应用卡尔曼滤波法对内插后的LAI进行了同化,试验结果表明,内插方法与卡尔曼滤波法相结合是一种具有稳定性和可用性的方法,经过同化后的LAI数据较接近真实情况,其效果好于DSSAT模型模拟的LAI。
An approach of assimilating ground measured leaf area index (LAI) into the CERES-Wheat model was developed in this paper. The CERES-Wheat model was used to simulate LAI under the Decision Support System for Agro-Technology Transfer (DSSAT) shell. Interpolation methods were used to solve the matching problem of the time scale between dynamic LAI simulated by the CERES-Wheat model and discrete LAI observed on ground. By comparisons, the interpolated LAIs by using cubic interpolation method based on measured LAI data were better than those of using the nearest neighbor interpolation, linear interpolation and cubic spline interpolation methods, respectively, which were assimilated by using Kalman Filter later. The experimental results showed that combination of the interpolation method and the Kalman Filter algorithm had stable and usable results. Assimilation results of LAI are close to their simulated and measured ones, which are better than the LAI data simulated by the model alone.
出处
《农业工程学报》
EI
CAS
CSCD
北大核心
2010年第S1期176-181,共6页
Transactions of the Chinese Society of Agricultural Engineering
基金
国家高技术研究发展计划课题(2007AA12Z139)
国家科技支撑计划课题(2006BAD10A01)
北京市自然科学基金项目(6062019)资助
关键词
模型
遥感
监测
同化算法
卡尔曼滤波
models, remote sensing, monitoring, assimilation algorithm, Kalman Filter
