Limits and Potentialities of Studying Dryland Vegetation Using the Optical Remote Sensing

Abstract

In optical remote sensing studies, the reflectance of the vegetation canopy in arid and semiarid areas is affected by the optical properties of the vegetation elements, their arrangement in the vegetation canopy and the optical properties of the surrounding environment. The study of vegetation and surrounding environment parameters presents significant peculiarities in arid areas. Low vegetation cover leads to a small contribution of vegetation reflectance in the total pixel reflectance relative to the other materials. Most types of dry ecosystem shrubs do not differ enough from one another to allow discernment of vegetation type. Vegetation in arid and semiarid areas adapts its structure and phenology to the harsh environment, which affects the overall brightness and temporal and spatial interspecies spectral variability. Moreover, the surrounding environment in dry ecosystem influences the reflectance of the vegetation by multiple scattering and nonlinear mixing and variable spectral composition of soil surface. Many remote sensing techniques are insensitive to nonphotosynthetic vegetation, which can be a major component of total cover in dry ecosystem areas. Spectral mixture analysis (SMA) appears to be the most promising technique to obtain information on vegetation cover, soil surface type and vegetation canopy characteristics. The empirical signature libraries of the world’s dominant vegetation types could be upgraded for use with SMA

Mapping Cropland in Smallholder-Dominated Savannas: Integrating Remote Sensing Techniques and Probabilistic Modeling

“…Results from other remote sensing studies in semi-arid environments suggest that fraction cover estimates made using unmixing methods are less affected by soil background than NDVI [32][33][34][35][36][37].There are several methods for estimating sub-pixel fractions of land cover, including multi-resolution approaches and spectral mixture analysis (SMA). Gessner et al [38] introduced a method of scaling up from an initial classification of high-resolution data to medium and coarse resolution imagery through a process of sampling and the use of random forest regression trees.…”

“…Spectral unmixing has been suggested to be among the most promising techniques for obtaining data on surface cover in savanna environments where there is a large presence of non-photosynthetic vegetation (NPV), and reflectance is affected by scattering, mixing, and variability in soil composition [34]. The technique has been recommended for semi-arid study sites where there is a high frequency of mixed pixels and reflectance is dominated by the soil background [35].…”

Monitoring desertification in a Savannah region in Sudan using Landsat images and spectral mixture analysis

“…The spectral response in remote sensing from open canopies is a function of the number and type of reflecting components, their optical properties and relative proportions (Adams et al, 1995). SMA appears to be the most efficient technique to obtain information on vegetation cover, soil surface type and vegetation canopy characteristics in semiarid areas because the scale of variability of the principle landscape elements in semiarid areas is larger than the pixel size in most of the remote sensing satellite imageries (Adams et al, 1995;Dawelbait and Morari, 2008;Okin and Robert, 2004).…”

“…Two main approaches are usually followed (Dawelbait and Morari, 2008;Smith et al, 1990): a) calculation of vegetation indices; b) image classification.…”

Land Degradation Monitoring in the Ordos Plateau of China Using an Expert Knowledge and BP-ANN-Based Approach

“…NDVI and albedo were selected to develop the land degradation detection model. The NDVI is defined as [9,11]:…”