Hyper-spectral characteristics of rolled-leaf desert vegetation in the Hexi Corridor, China

Abstract

Desert plants survive harsh environment using a variety of drought-resistant structural modifications and physio-ecological systems. Rolled-leaf plants roll up their leaves during periods of drought, making it difficult to distinguish between the external structures of various types of plants, it is therefore necessary to carry out spectral characteristics analysis for species identification of these rolled-leaf plants. Based on hyper-spectral data measured in the field, we analyzed the spectral characteristics of seven types of typical temperate zone rolled-leaf desert plants in the Hexi Corridor, China using a variety of mathematical transformation methods. The results show that: (1) during the vigorous growth period in July and August, the locations of the red valleys, green peaks, and three-edge parameters, namely, the red edge, the blue edge, and the yellow edge of well-developed rolled-leaf desert plants are essentially consistent with those of the majority of terrestrial vegetation types; (2) the absorption regions of liquid water, i.e., 1400-1500 and 1600-1700 nm, are the optimal bands for distinguishing various types of rolled-leaf desert plants; (3) in the leaf reflectance regions of 700-1250 nm, which is controlled by cellular structure, it is difficult to select the characteristic bands for differentiation rolled-leaf desert vegetation; and (4) after processing the spectral reflectance curves using a first-order differential, the envelope removal method, and the normalized differential ratio, we identify the other characteristic bands and parameters that can be used for identifying various types of temperate zone rolled-leaf desert plants, i.e., the 510-560, 650-700 and 1330-1380 nm regions, and the red edge amplitude. In general, the mathematical transformation methods in the study are effective tools to capture useful spectral information for species identification of rolled-leaf plants in the Hexi Corridor.

The Impact of NPV on the Spectral Parameters in the Yellow-Edge, Red-Edge and NIR Shoulder Wavelength Regions in Grasslands

“…Wavelength of the strongest absorption in the transition between yellow-edge and red-edge (650-690 nm [23,33]; wavelength of the lowest reflectance between 600 nm to the R segmentation in this study).…”

“…The slope of the linear regression between the reflectance and wavelength in red-edge (680-760 nm [23,29]; from R segmentation to NIR segmentation in this study).…”

“…The spectral reflectance of the vegetation in the 600-900 nm region appears as three broken-lines, representing yellow-edge, red-edge and NIR shoulder (Figure 2). Previous studies reported slightly different wavelength regions of yellow-edge and red-edge (e.g., 560-640 nm [23] or 550-580 nm [29,44] for yellow-edge, and 680-760 nm [23,29], 670-780 nm [18,34,39] or 670-750 nm [44] for red-edge). Therefore, it is important to capture the accurate turning points between yellow-edge and red-edge, and between red-edge and NIR shoulder before extracting the spectral parameters in these three wavelength regions.…”

“…For hyperspectral remote sensing, the spectral characteristics of green, red, red-edge and near-infrared (NIR) regions are sensitive to grassland vegetation growth [18,19] and are widely used in monitoring green cover [9], leaf area index (LAI) [20], biomass [21] and ecological characteristics [22]. The parameters of yellow-edge [23], red-edge [24], NIR shoulder [20] and various vegetation indices (e.g., the most popular vegetation index [25]: NDVI (Normalized difference vegetation Index)) [20] indicate the spectral characteristics of the four wavelength regions. However, the estimation accuracy of LAI or green cover is unstable in grasslands, especially the grassland with large amounts of dead materials [25].…”

“…Although, comparing to vegetation indices, red-edge parameters are less sensitive to the changes in specific biophysical factors (vegetation type, canopy structure and soil cover) and environment conditions (site condition, atmospheric effects, irradiance and solar zenith angle) [29,39]. Wei et al's research indicates that the red-edge parameters are affected by the underlying surface in arid grasslands [23]. However, it remains unclear exactly how dead materials, a large component of native grasslands, influence red-edge parameters.…”