Dynamic Change Of Vegetation Coverage By Monitoring Remote Sensing In Mu Us Sandland

Using the main data sources were 1990, 2000 and 2007 (during two period) Landsat TM images. The MuUsSandLand that dynamic changes about vegetation cover of qualitative and quantitative in 17 years was analysised. Used computer classification based on the NDVI and visual interpretion results of image classification combined,after extraction of vegetation data. Calculate the landscape pattern index for each year, which is analyzed. Used transfer matrix of research into landscape about different period, The period (2000-2007) as a basis transfer for the use of vegetation markov forecast model to research the future in 21 years vegetation. Finally the vegetation changes caused by natural and human factors were studied.The results show that:(1) Qualitative study result about vegetation cover that subject vegetation is low ,the area from 33176.7369km2 in 1990 reduced to 30671.6454km2 in 2007. Overall positive transfer (9.31%) is more than negative transfer area (1.84%).So vegetation is whole improve and local deterioration. (2) Quantitative study result is that vegetation cover was sharp decline in former phase(1990-2000) and in the later phase (2000-2007)was tended to be stable.(3) The dynamic changes of landscape patterns of vegetation was that Vegetation was tended to be stable after overall decline . The degree of fragmentation tendency was complete-fragmentation-whole complete and local fragmentation. The landscape type diversity equalization of development was not appeared. whole landscape was controlled by small number patch types.(4) The result of Markov forecast model was the whole tendency is not good obvious. Therefore for ecological environment must add governance will tend to be good.(5) Analysis show that climate change and human activity had close relationship with vegetation changes.