针对棉花生长模拟模型COTTON2K在区域应用上的局限问题，探讨作物模型区域应用尺度扩展的方法，建立基于GIS的作物模型和遥感数据的同化系统。同化系统的基本原理为：以辐射传输模型SAIL-PROSPECT为纽带，通过与作物模型COTTON2K结合模拟棉花冠层的反射率，并以此作为模拟子；利用校正过的modis数据获得的棉花冠层反射率为观测子；通过寻优程序，调整作物模型的品种参数和初始状态变量，使观测值和模拟值无限接近，从而达到区域上模型参数的优化。利用优化的同化系统研究华北平原近10年棉花产量的时空变化规律并分析其驱动因素。主要研究内容包括：（1）基于田间试验和SAIL-PROSPECT模型的棉花冠层光谱反射规律研究；（2）遥感数据与COTTON2K模型的同化系统的建立；（3）华北平原棉花生产的时空分布规律。

An assimilation system for cotton growth model COTTON2K and remote sensing data based on GIS is to be set up to solve the problem of the limitation of the regional application of the crop growth model. The procedure of the assimilation is: (1) through combining the radiative transfer model SAIL-PROSPECT and cotton growth model COTTON2K, reflectance of cotton canopy is simulated. (2) the observed value of reflectance is attained by the calibrated data from Modis. (3) The variety parameters and initial state variables of COTTON2K are to be optimized through the process of searching the minimum value between simulated and observed reflectance. Thus the the parameters required by COTTON2K in regional scale are attained. Then the cotton yield in North China Plain will be simulated by the assimialtion system and the driving force of its spatial and temporal distribution will be anaylysed. The main contents of this project include (1)the research on reflectance spectrum of cotton canopy based on the fied experiments; (2) setting up the assimilation system of the remote sensing data and COTTON2K; (3) simulation and valuation of the cotton growth in North China Plain.