Monitoring and Classification of Plant Species in Some Wadies of Sinai Peninsula Using Remote Sensing Applications

Abstract

This work is considered the first step of using remote sensing techniques in monitoring, assessment and classification of natural vegetation in Egypt which was not possible before using hyper spectral remote sensing technology. The unifying hypothesis of this dissertation is that vegetation associations can be differentiated using their hyper spectral reflectance in the visible to short wave infrared spectral range. The study aims to investigate and quantify spectral reflectance variation in a number of natural vegetation species; to determine whether meaningful information existed within the reflectance data that could be used for species-level identification of segregation of the natural vegetation canopy.

The present study comprised 20 species belonging to the following genera Achillea (one species), Aerva (one species), Alkanna (one species), Asclepias (one species), Astrraglus (one species), Ballota (one species), Echinops (one species), Fagonia (one species), Hyoscyamus (one species), Matthiola (two species), Origanum (one species), Peganum (one species), Phlomis (one species), Pyrethrum (one species), Stachys (one species), Teucrium (one species), Verbascum (one species), Zilla (one species), Zygophllum (one species) grown in South Sinai, Sinai Peninsula, Egypt. Macromorphological features of the studied taxa were either recorded from living specimens i.e. vegetative shoot parts, inflorescence, flowers, fruits and seeds; or compiled from literature. Micromorphological characters were recorded from each of the stem and leaf sections; the striped off lamina epidermis (ab- and adaxial) preparations. Lamina epidermis sculpture as revealed by the scanning electron microscope (SEM). ASD field spectroradiometer was used to measure spectral reflectance in the wavelength ranged from 350 to 2500 nm for 20 species. Chlorophyll content was estimated using SPAD 505 chlorophyll meter for all species. Statistical analysis was carried out to observe the correlation between chlorophyll content and remotely sensed data represented by vegetation indices VIS (NDVI, SR, MCARI, chlorophyll content and chlorophyll concentration a & b). Phenetic relationships were established by using the program NTSYS-pc 2.2 based on UPGMA clustering method for the produced data sets that were grouped into three main analyses. The first based on 317 macro / micro- morphological characters. The second analysis based on 31hyperspectral wave length characters. The third analysis based on the combination of all the used characters (348). The produced data sets were compared by measuring each of the phenograms distortion; from their relevant similarity matrices; and association to each other based on their cophenetic (ultrametric) correlation coefficients (r).