Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Master → Chine → 2018 → A Study of Vegetation Patterns and Numerical Simulation in the Plain Desert Temperate Zone

Inner Mongolia Agricultural University (2018)

A Study of Vegetation Patterns and Numerical Simulation in the Plain Desert Temperate Zone


Titre : A Study of Vegetation Patterns and Numerical Simulation in the Plain Desert Temperate Zone

Auteur : 陈程;

Grade : Master’s Theses 2018

Université : Inner Mongolia Agricultural University

This paper takes the typical vegetation of desert grassland of Siziwangqi Banner as the research object.In the research area of desert steppe of Siziwangqi banner,the hyperspetral data of establishment species Stipa breviflora,dominant species Artemisia frigida and Cleistogenes anatipestulata and degraded indicator species Salsola were collected by the GaiaSky-mini airborne and terrestrial hyperspectral instrument.Then these hyperspectral curves were denoised,linearly differentiated,envelope removed,in order to extract feature bands and parameterize feature information.Otherwise the characteristics and general patterns of desertification grassland vegetation were obtained by comparative analysis.For hyperspectral data,select the best band combination for vegetation index calculation,combine the pixel dichotomy model,estimate vegetation coverage,and perform linear regression analysis with measured coverage.Invert the model and replace the measured values with grassland degradation.The result will offer possibilities for remote sensing monitoring.The main findings are as follows :(1)The original spectral reflectance of typical vegetation in the desert grassland of Siziwangqi Banner is low in the visible light range from 400 to 550 nm,showing a growing trend,and does not exceed 0.2.The reflectance at 550 nm to 680 nm gradually decreases,and the reflectance at near-infrared wavelengths ranges from 680 to 1000 nm.The rate is higher,not higher than 0.55.(2)The red edge effect of Artemisia frigida was the most obvious,the absorption peak area of Cleistogenes anatipestulata was the largest,and the estimation accuracy of the degree of coverage of Salsola was the highest,four implants can be identified by spectra.(3)Estimation coverage of vegetation index,SAVI and NDVI performed best,and the results were closest to the measured values.This research is the first use of hyperspectral remote sensing technology to monitor the desert grassland of Siziwangqi Banner,and the results are of great significance to the identification and degradation of typical vegetation in Siziwangqi Banner

Mots clés : Desertification grassland; Spectral analysis; Linear differential; Envelope removal; Vegetation index; Coverage;

Présentation (CNKI)

Page publiée le 17 mai 2019