Abstract: [Objective] Fractional vegetation cover(FVC)can reflect the status of land use/cover enougth, especially the arid areas where are sensitive to vegetation change degree significantly. Xinjiang belongs to arid areas, because of the influence of human activities and climate change vegetation cover in Xinjiang has changed in recent years.Therefore, it is necessary to monitor and investigate the change of Xinjiang vegetation cover dynamically, which would provide a reference for improving regional ecological environment. [Method] Based on the NDVI data from 1998 to 2012 of Xinjiang and the FVC data obtained by using dichotomy, this paper studied the characteristics of vegetation changing trend and its spatial clustering distribution through Sen + Mann Kendall trend analysis and spatial auto-correlation analysis. [Result] The results show that: 1) The vegetation cover declines slightly among 15 years in Xinjiang. 2) Sen + Mann Kendall analysis reflects the spatial distribution characteristics of FVC changing trend in Xinjiang. Improvement areas of vegetation cover are mainly distributed in the north slope of Tianshan Mountain, and degradation areas are mainly distributed in the ecotone of mountains and plains, degradation degree is particularly severe in Ili region. Meanwhile, the areas with very low vegetation coverage remain unchanged basically, and degradation areas are mainly distributed around the low coverage region. 3) The spatial auto-correlation analysis further verifies the results of Sen + Mann Kendall analysis. Vegetation cover exhibits apparent agglomeration phenomenon in Xinjiang.Partial correlation analysis shows that vegetation coverage is given priority to "high-high concentration" and "low-low concentration".[Conclusion] It can be seen that the relatively high vegetation coverage areas are easier to be improved, while low coverage areas or areas without coverage are more difficult to be improved, and present obvious degradation. According to the clustering characteristics of vegetation coverage, vegetation coverage presents obviously "H - H gathered" and "L-L gathered" pattern. This phenomenon has potential links with the regional climate, water resources distribution and the influence of human activities. In order to understand the driving mechanism of vegetation change in arid areas, we can analyze the affecting factors on the changes of vegetation cover in the future. Due to the influence of human activities on vegetation coverage change in a short period is more significant, the space analysis can provide corresponding guidance for improving vegetation in arid areas. [Other] This paper explains the characteristic of vegetation changes from the perspective of vegetation spatial clustering. On the one hand, it validate the results of Sen + Mann Kendall trend analysis, on the other hand, it would provide a basis for the overall analysis of characteristics of the vegetation changes.

Key words: fractional vegetation cover(FVC), Sen + Mann-Kendall trend analysis, spatial auto-correlation analysis, spatial distribution characteristics, Xinjiang