作物冠层FPAR遥感反演与时间扩展方法研究

FPAR, Fraction of absorbed Photosynthetically Active Radiation, is known as an important indicator for the health condition of the plant. The remote sensing provides a promising way to obtain this parameter at large spatial scale. However, the current FPAR retrieval method from remote sensing data mainly concerned on uniform vegetated canopy, and were lack of investigation on the method to convert the FPAR value estimated from instantaneous observation to the value at a specific moment or daily average. Besides, the accuracy of FPAR was also reduced by the error caused by atmospheric correction on remote sensing data. In order to solve this problem, this study aim at developing new models and methods to improve the retrieval accuracy of the FPAR value for crop canopy..Firstly, this study will develop a new radiative transfer model for row-structured and uniform crop canopy. On basis of this new model, some new algorithms will be conducted to retrieve FPAR from canopy reflectance and apparent reflectance at TOA observed by several Chinese Gaofen (high-resolution) satellites. Furthermore, this study will investigate the way of temporal expansion for converting instantaneous FPAR to that of a specific moment or daily average. Finally, the new methods will be applied to retrieve FPAR from satellite data and validated using ground-measured data. Accordingly, this study will probably establish a new technique of retrieving FPAR from multiple high-resolution remote sensing data, and achieve the FPAR retrieval from remote sensing data without any atmospheric correction and therefore enhance the operability of FPAR retrieval from different remote sensing data.

光合有效辐射吸收比例（FPAR）是表征植被生长状况的重要生态物理参数。遥感是目前获取空间大范围尺度FPAR的关键手段。然而，现有反演算法多针对均匀冠层而发展，且缺少从单一瞬时观测数据向日平均FPAR的扩展方法，大气纠正过程的误差也降低反演精度。为了改善这一困境，本项目将以典型作物冠层为研究对象，致力于发展新模型和新方法以提高作物冠层FPAR反演精度。. 本项目将构建适合于作物行播结构与均匀结构的冠层辐射传输新模型，发展从国产高分多源卫星观测的冠层反射率与表观反射率反演作物冠层FPAR的新算法，并探索从瞬时FPAR到特定时刻FPAR、日平均FPAR的时间扩展方法；反演典型区域FPAR产品，并利用地面测量数据进行有效验证。本研究拟建立一套适合于多源高分数据的FPAR反演技术，实现无需大气纠正过程的FPAR遥感反演，增强反演过程的可行性。

吸收光合有效辐射比例因子（FPAR，Fraction of absorbed Photosynthetically Active Radiation）是指植被吸收的光合有效辐射占总入射的光合有效辐射的比例，是植被的基本生态物理参数。准确、定量地获取地表 FPAR 对农业生产、生态系统、碳循环以及区域与全球范围内的气候变化研究具有重要意义。针对现有 FPAR 反演算法对传感器类型和大气纠正过程的强烈依赖性等关键问题，本研究在辐射传输建模、FPAR 反演算法构建与验证等方面开展研究，发展了一套适合于多源遥感数据的 FPAR 反演新算法， 以提高 FPAR 反演过程的可行性和可信度，并增强不同 FPAR 产品的可比性。项目完成的主要研究工作与取得的主要成果包括：（1）在原有农田辐射传输模型SHAW的基础上，通过耦合辐射传输模型和几何光学模型，建立了能够行播冠层和均匀冠层的方向反射率与热点效应的MSHAW-BRDF模型，实现了SHAW模型从原有的单一辐射传输模型到辐射传输-几何光学混合模型的扩展。（2）在发展的 MSHAW-BRDF 模型的基础上，结合模拟叶片反射率与透过率的PROSPECT 模型与土壤样本反射率，根据多源传感器的波段特征，构建了分别从像元地表反射率（BOA-model）和表观反射率（TOA-model）反演地方时 10:30 的直射光和散射光 FPAR 的新模型。结果表明，BOA-model 和TOA-model 的算法误差分别优于 0.03 和 0.06。（3）探索了从瞬时 FPAR 到日平均 FPAR 的扩展方法。通过以上研究，形成了一个能够模拟冠层辐射与二向性反射率的新模型，建立了一套能够从多源传感器地表反射率与表观反射率反演像元 FPAR 的新算法。本研究涉及的辐射建模与反演算法等可以扩展到其他遥感数据，以期促进卫星数据的应用水平。