基于遥感的光合有效辐射吸收比率(FPAR) 估算方法综述

光合有效辐射吸收比率(FPAR)是反映植被生长过程的重要生理参数,是陆地生态系统模型的关键参数,是反映全球气候变化的重要因子。基于遥感的FPAR估算方法是获取区域乃至全球尺度FPAR的有效方法。目前,主要形成了植被指数法和机理法两类方法,植被指数法是建立FPAR与植被指数的经验统计模型,简单、计算效率高;机理法则从物理模型上进行FPAR的求解与反演,机理明晰、可行性强。然而,由于FPAR本身的复杂性以及环境因素、遥感数据质量的影响,导致了估算方法面临诸多不确定性问题。为了解决这些不确定性问题以及满足生态过程深入研究的需求,将进一步注重FPAR的机理研究、先验知识的获取与积累,构建长时间序列FPAR以及高时空的FPAR算法研究。

overview on methods of deriving fraction of absorbed photosynthetically active radiation (FPAR) using remote sensing

Fraction of Absorbed Photosynthetically Active Radiation (FPAR) is an important biophysical factor for monitoring vegetation growth, as well as a critical parameter in the terrestrial ecosystem modeling and a key indicator for studying global climate change. Remote sensing technology has been proved to be an effective tool in estimating FPAR at regional and global scales, because satellite data can provide a spatially and periodic, comprehensive view of vegetation growing status. Many methods have been developed in estimating FPAR with remote sensing, which can be generally grouped into two categories. The first category of approaches are the empirical statistics models based on the relationships between vegetation indices, derived from reflectance at canopy level, and FPAR. These models are easy to use with high efficiency and much more suitable for detecting within-field spatial variability, yet they may lead to inaccurate results when applied over another place or broad scale with different land cover types. Another category of approaches for FPAR retrieval are to invert canopy reflectance models based on the BRDF (Bi-Directional Reflectance Distribution Functions) models such as the radiative transfer model and geometrical optics model, which describe the transfer and interaction of radiation inside the canopy based on physical mechanism between FPAR and vegetation canopy reflectance. These models have strong applicability and are taken as the algorithm bases among most widely used FPAR products. However, the inversion process is ill-posed due to the complexity of these physical models; the parameters and prior knowledge required by these models are hard to acquire over large areas. At the same time, other methods such as the method based on the concept of effective FPAR, which is FPAR absorbed by chlorophyll, and the method based on the airbome lidar data which is useful to characterize spatial variability of canopy structure, bring significant improvement to the two categories of methods. Due to the complexity of FPAR itself and its influencing factors, as well as the quality of remote sensing data, plenty of uncertainties existed in satellite based FPAR estimation. For statistical model, most vegetation indices are easily affected by soil background, saturation problem, atmospheric condition, and so on. These factors bring much uncertainty in the relationship between FPAR and vegetation indices. For physical models, problems including top-of-atmosphere radiance uncertainties and errors in land cover mapping are hard or even impossible to avoid. In order to deal with these uncertainties and meet the requirements of further research for terrestrial ecological process, future research focuses on FRAR retrieval based on satellite will be: further research on theoretical mechanism of FPAR estimation, seeking to minimize noise effects on vegetation indices for more accurate estimation of FPAR, improvement of the inversion methods for physically-bases models, acquisition and accumulation of prior knowledge in FPAR estimation based on systematic observation network, construction of long-term FPAR dataset based on multi-source remote sensing data, and algorithm for deriving FPAR with both high spatial and high temporal resolutions.