| Grassland is the largest land cover type in the world which accounted for 52.17% of the total land area of the earth. Grassland vegetation is the foundation of animal husbandry development, which has the ecological functions of windbreak and sand fixation, soil and water conservation, water conservation, maintaining biodiversity and tourism ornamental etc. Grassland vegetation coverage is one of the most important parameters to describe grassland ecosystem health and ecological carrying capacity. The vegetation coverage degree can show the quality of the regional ecological environment, and indicate the regional internal ecosystem function and the ecosystem service ability. There is unique alpine meadow steppe landscape in Xinjiang Bayanbulak Grassland, which is the second largest grassland in China and an important ecological barrier of Xinjiang Tianshan Mountain region. In recent years, grassland degradation has become a serious problem because of climate changing and overgrazing, therefore, it is significant to research spatial and temporal variation characteristics of vegetation coverage and its influencing factors in Bayanbulak grassland.In this paper, the MODIS spectral estimation model of Bayanbulak Grassland vegetation coverage is established in the foundation of MODIS remote sensing data and field measured data and with the methods of regression analysis, trend analysis, buffer analysis. According to the model, the vegetation coverage in the study area is classified, the sample line and buffer zone for typical ground objects are analyzed and the temporal and spatial variation characteristics of vegetation coverage and its main influencing factors of study area from year 2001 to 2015 are studied. The main conclusions are as follows:(1) Constructing spectral estimation model of vegetation coverage. The estimation of vegetation coverage VC’s ground spectral model is a simple linear function measured by SOC NDVI, R2 is 0.865; and there is a linear relationship between MODISNDVI and measured SOCNDVI, R2 is 0.731; MODIS spectral estimation model of estimation vegetation coverage is inverted. The test shows that MODIS spectral estimation model correlation coefficient R is 0.882 with Significant correlation, the standard error SE is 11.55%, and the overall prediction accuracy is 88.92%. So it can be concluded that the established MODIS spectral estimation model based on ground spectral model has high prediction accuracy, and it can estimate vegetation coverage of Bayanbulak Grassland in growth season, which is scientific and practical relevantly.(2) Classification of grassland grades in the study area based on the built model. The remote sensing image NDVI gray scale of study area is transformed into single band vegetation coverage map based on the MODIS spectral estimation model. The experimental grassland is divided into five levels:18.87%25.61%,31.28%,13.86% and 10.38%, according to the vegetation coverage. With the foundation above the vegetation coverage analysis presented the internal differences of vegetation coverage in the study area clearly from the three aspects of grade division, surface features and zoning.(3) Sample line analysis and buffer analysis of typical features in the study area. Based on the vegetation coverage map of the study area, there are An Bn C-. D,4 lines and 4 buffers, which are corresponding with mines, roads, tourist attractions, and community. Through the analysis and conclusion, the typical features of objects influent the grassland vegetation coverage with different degrees, and the closer distance, the lower degree of vegetation coverage. The grassland influence degree in order of large to small is followed by:Mine> residential> Road> tourism attractions.(4) The analysis of spatial and temporal patterns of grassland vegetation coverage in the study area. The vegetation coverage map of year 2001-2015 is calculated by MODIS spectral estimation model. As a whole, it shows a distribution pattern of low East and West, high North and South, a weak growth trend with the time series. On the whole region scale, comparison between 2006-2010 and 2001-2005, the area of low vegetation coverage increased, the area of extremely high vegetation cover decreased, and the overall vegetation coverage decreased; Comparison between 2011-2015 and the first two periods, the low vegetation coverage area decreased while the high and extremely high vegetation coverage area increased, and the vegetation begin to recover. On the pixel scale, the 76.38% regional vegetation coverage increased and 23.62% regional vegetation coverage decreased in the study area, which the maximum value is 1.88%, and the maximum value is 1.59%.(5) The analysis of main factors affecting the temporal and spatial patterns change. The vegetation coverage changing of Bayanbulak is the interaction result of natural factors and human activities. By analyzing the correlation between the precipitation, temperature, altitude, the amount of animal husbandry and the vegetation coverage in the study area, it is concluded that the correlation R of absolute value from large to small is as follows:precipitation (0.916)> temperature (0.571)> animal husbandry (0.50)> altitude (0.115). The vegetation coverage is negatively correlated with the amount of livestock. The particular spatial distribution and variation characteristic of vegetation coverage in Bayanbulak Grassland is decided by its unique water, climate and soil conditions. Rainfall plays an important role in the growth of vegetation. The increase of vegetation coverage mainly depends on the increase of vegetation coverage, while extended vegetation growth period will increase vegetation coverage, which depends on the increase of temperature. The effect of terrain factors on vegetation cover still need further verification, and the human activities of grassland grazing and tourism behavior are main influencing factors. |
PDF Full Text Request