非均匀植被BRDF模型中的分布格局尺度问题

Climate models require a realistic description of canopy radiation with reasonable computational efficiency, and the accurate retrieval of land surface biophysical parameters from multiangle satellite observations also need a fundamental understanding of the radiative processes inside the vegetation canopy. Many vegetation bidirectional reflectance distribution function (BRDF) models have been developed for these two purposes, and the recent researches are focus on their applications to complex heterogeneous canopy. In some models (e.g. the Monte Carlo ray tracing models), the complex canopy structure is explicitly described at the stand and leaf level with a three dimensional canopy scenes, while in the others (e.g. the geometric optical models), relative simple statistical descriptions are used. The models use explicit canopy structure description show good performance in simulating the reflectance over heterogeneous canopy, but the lack of priori information of the detailed canopy structure hinder their use in both land surface process studies and land parameters remote sensing retrieval applications. The models use statistical canopy structure descriptions have the advantage of computer efficient and flexible for inversion, however, still need to improve the simulation accuracy. It is a challenging task to develop a canopy BRDF model which is compute efficient but also able to provide accurate simulations. We analysis the statistical description used in present geometric optical models, and found the use of first-order statistics is one of the main reason impact the model's simulation accuracy. The first-order statistics only provide the averaged distribution information of the whole region, and fall to provide the distribution pattern in variant scale, and the latter is important to estimate the canopy structural impact on canopy radiation. The use of second-order statistics will overcome this short coming in statistical description, but how to incorporate the second-order statistics into the canopy radiative transfer models, no research has been done in this direction. Therefore, in this project, we will explore a new method to estimate the impact of canopy structure on canopy radiative transfer process with second-order statistics, and develop a canopy BRDF model which can provide reliable BRDF simulations, and also computer efficient enough for land surface process studies and flexible for inversion.

植被冠层的辐射传输过程的研究，既是陆面过程研究的需要，又是定量遥感反演植被参数的需要。而关于复杂的非均匀植被的辐射传输问题,是研究的难点所在。目前的植被BRDF模型采用不同的方式来描述冠层复杂的植被结构，而各有其优缺点。一些模型采用非常精确方式来描述植被结构，有利于保证模拟精度，但缺点是过于复杂，计算量大、而且模型所需要的精确结构信息在实际应用中很难获得，因此既不适于陆面过程参数化研究的需要、也不适合遥感反演应用的需要。而目前常用的统计描述方法又偏于简单，难以保证模型的模拟精度，尤其是对复杂结构的非均匀植被的模拟。针对这一问题，本项目拟通过建立一个基于二阶统计描述方法的、结构明晰、计算迅速、能适用于复杂非均匀植被冠层模拟的几何光学模型，来解决目前对于复杂非均匀植被的辐射传输过程研究中，缺少合理有效的统计描述方法，导致模型的应用范围受到局限的问题。