Analysis Of Spectral Features Of Main Desert Plants In The Lower Reaches Of Tarim River

Tarim River Basin has natural resources and fragile’s ecological.Domestic and overseas scholars studythe vegetation recovery of bank forests from the physiological and ecological points.But the reports ofdesert plants from the point of hyperspectral properties were till rare. The study used HR-768spectroradiometers to measure the hyperspectral data of the main desert plants in the lower reaches ofTarim River,,the relationship between the hyperspectral and the change of moisture gradient.and desertplant identification by hyperspectral. Providing the basis to built desert plants spectral library and monitor.The main conclusion are as followings:The difference bands of desert plants (Populus euphratica,Tamarix hispida,Lycium ruthenicum,Glycyrrhiza inflate, Karelinia caspia, Alhagi sparsifolia, Halimodendron halodendron) spectral were750nm~1070nm,1180nm~1270nm,1650nm~1790nm. With the decline of groundwater, The leaf moisturewas decreased with groundwaterv decreasing, spectral reflectance was increased in Moisture absorptionband（980nm,1180nm,1455nm）.In the vicinity of band1455nm was the most hypersensitive band because of the correlation betweendesert plants and water content was the best.The near-infrared wavelengths could identify the different tree species.The identification accuracieswere90.4%,91.7%,92.9%.The second derivative reflectance analysis identification model has the bestperform in accuracy.The relationship between the water content of the samples and the reflectance spectrum have analyzedby using correlation coefficient method. The result showed that the band1466~1646nm was the mosthypersensitive band because of the correlation between Populus euphratica and water content was the best.According to the water absorption deepth in the bands of1466nm～1646nm to establish the linearmodel which was used to forecast the water content of Populus euphratica: Y=7.309X1474-5.77X1482-0.629（R~2=0.954）.The model precision is98.41%.Using the methods of first derivative reflectance, logarithm first derivative reflectance, continuumremoval to process the Tamarix spp.original spectrums. Using the method of correlation coefficient todetermined that the band1300nm～1620nm was the most hypersensitive band.Using three methods toestablish the linear model which was used to forecast the water content of Tamarix spp. Y=0.5X1459+0.451（R~2=0.840）；Y=58.108X1428+0.463（R~2=0.928）；Y=-1.031X1428+0.818（R~2=0.942）.The continuumremoval spectrum linear model used to forecast the Tamarix spp.water content has the best perform inaccuracy. The model precision is95.94%.The red edge parameter and NDVI of Populus euphratica and Tamarix spp. In different seasons. Theresults show that the reflectivity of blue bands and red bands were decreasing from spring to summer.Green peak moved to blue bands.The reflectivity of near infrared band increases. The red edge parameterand NDVI increased greatly. The reflectivity of blue bands and red bands were rising from summer toautumn. Green peak moved to red bands. The reflectivity of near infrared band was decreasing. The rededge parameter and NDVI decreased greatly.