真实和有效叶面积指数及聚集指数的尺度效应

尺度效应是地球科学和定量遥感中的重要研究课题,目前的许多研究大多集中在估算尺度效应带来的误差,而对一些关键的植被结构参数是否存在尺度效应及其尺度转换方法尚存在诸多不同见解。本文针对真实和有效叶面积指数（Leaf Area Index, LAI和Effective LAI, LAI_e）以及聚集指数（Clumping Index, CI）3个植被关键结构参数,从基本概念和获取方法上分析参数的尺度效应及其尺度转换方法。从定义上看,LAI并不存在尺度效应,而LAI_e和CI则存在尺度效应,其中CI的尺度效应由LAIe引入（CI=LAI_e/LAI）。在野外实测中,LAI破坏测量法没有尺度效应,但由孔隙率模型获取3个参数的方法均具有尺度效应。异速生长方程和遥感反演方法的尺度效应取决于方法本身的线性或非线性特征。目前全球主要的LAI、LAI_e和CI遥感产品都基于非线性模型获取,其反演过程具有尺度效应。像元尺度的LAI本身并不具有尺度效应,而像元尺度的LAI_e和CI虽然具有尺度效应,但在实践中常常被忽略。因此,实际工作中应注意区分参数概念本身、野外测量、遥感反演方法以及遥感产品等所展示的不同尺度效应。

Scaling Effects of the True and Effective Leaf Area Index (LAI and LAIe) and Clumping Index (CI)

Scaling effects describe the observational differences caused by different observation standards. It is an important research topic in Earth science and quantitative remote sensing. Current studies mainly focus on estimating the errors caused by the scaling effects, but different opinions still exist about the scaling effects in some critical vegetation structural parameters and their scale transformation methods. This paper analyzes the scaling effects and the scale transformation methods of three key vegetation structural parameters, namely Leaf Area Index (LAI), Effective LAI (LAI_e), and Clumping Index (CI), based on their definitions and acquisition methods. By definition, LAI is free from the scaling effects, whereas LAI_e and CI have scaling effects. The scaling effects of CI is introduced by LAI_e (CI=LAI_e/LAI). LAI, LAI_e, and CI can be obtained through field measurement and remote sensing inversion methods. In field measurements, LAI is obtained through the direct destructive method or the indirect optical method. LAI obtained through the destructive method has no scaling effects. The indirect optical method estimates the three parameters based on the Beer-Lambert equation with the canopy gap fraction. LAI-2200, digital hemispherical photography, and photosynthetic active radiation sensors are commonly used instruments. The non-liner gap fraction model has scaling effects in deriving these parameters. Remote sensing technology uses passive optical methods, Light Detection and Ranging (LiDAR) technology, and Synthetic Aperture Radar (SAR) methods to estimate LAI, LAIe, and CI. The classic passive optical methods can be divided into the empirical vegetation index estimation methods and the physical model inversion methods. The vegetation index methods establish an empirical relationship between vegetation structural parameters and vegetation indices. The physical model inversion methods are based on physical radiative transfer models. The scaling effects of the remote sensing methods depend on the linear or nonlinear characteristics of these methods. Currently, the major global LAI, LAI_e, and CI remote sensing products are acquired from nonlinear inversion models, thus these inversion methods are subject to scaling effects. Nevertheless, the nonlinearity in the inversion process does not necessarily mean that the LAI products have scaling effects. The scaling effects of the LAI products still follow the basic LAI definition. Therefore, current remote sensing LAI products do not subject to the scaling effects at pixel level. On the other hand, both LAI_e and CI products do have scaling effects, but the scaling effects are often ignored in practice. In the validation of the remote sensing products, the scaling effects need to be considered while homogeneous areas are preferred in the validation studies. In conclusion, attention should be paid to distinguishing the different scaling effects displayed by the parameters in their definitions, the field measurement and remote sensing inversion methods, and remote sensing products. For the scale transformation, it is more important to investigate the most suitable and efficient method rather than a universal method.