Abstract: Using remote sensing ecological index (RSEI) to evaluate regional ecological changes can obtain the ecological environment changes quickly, efficiently, and objectively. In this study, RSEI of four ecological factors, greenness, humidity, heat, and dryness were calculated using the Landsat data and the spatial and temporal patterns of ecological environment in the Gurbantunggut Desert from 2006 to 2017 were qualitatively and quantitatively evaluated, based on the weights determined by principal component analysis. The results showed that humidity and greenness played a positive role in the eco-environment quality of the Gurbantunggut Desert, whereas heat and dryness had negative effects. Normalized vegetation index (NDVI) representing the greenness was the top contributor among those four factors. The RSEI of the Gurbantunggut Desert declined from 2006 to 2017, with a 20.1% decrease of the mean value from 0.294 in 2006 to 0.243 in 2017. The ecological environment of the study area showed a deterioration. The eco-environmental quality of the central part of the Gurbantunggut Desert was relatively stable. The eco-environment quality of the dense vegetation areas in the northeast and the irrigated areas in the south was getting better, but that of southern and northwestern parts of the desert was getting worse.

Key words: ecological environment, principal component analysis, remote sensing ecological index (RSEI), Gurbantunggut Desert