| Vegetation changes,such as the loss of vegetation,may cause soil erosion and climate anomalies,which will seriously affect the human being’s living environment and even the balance of the ecosystem.Therefore,it is of great significance to monitor vegetation changes.Phonological information,interannual change information,and disturbance information,which can effectively monitoring the vegetation change process continuously,implicated in remote sensing time series data.The paper carry out a research concerning the method of vegetation change process monitoring and vegetation change type recognition using MODIS and Landsat datasets from 2001 to 2016,extracting the variation information of impervious surface percentage,and analyzing the spatiotemporal analysis of Chinese urban expansion in early 21st century.The main contents are as follows:1.A method was proposed for characterizing vegetation changes process based on temporal similarity trajectory.The temporal similarity trajectory was used to reveal the process of vegetation dynamic change.The Logistic model was used to fit the temporal similarity trajectory.The vegetation change region is determined by goodness of fit,and the change time,variation amplitude and rate of change can be obtained from the fitting parameters.The overall accuracy reached 90.90%and the kappa coefficient was 0.7905 based on the accuracy assessment from the reference sites.The area of vegetation change in China accounted for 11.7%of the area of China’s mainland during the 2001-2016,and the time of change was mainly concentrated in 2004 to 2007,accounting for nearly 70%of the change area.2.The change trend of the multi-dimensional remote sensing index was constructed to identify vegetation change types.The types of land cover can divided into four basic cover types:vegetation,water body,bare soil and impervious surface.Based on the annual trend of remote sensing index from four dimensions of Optimized Soil Adjusted Vegetation Index(OSAVI),Land Surface Water Index(LSWI),Ratio Index for Bright Soil(RIBS)and Biophysical Composition Index(BCI),a change characteristic relationship table of vegetation changing to impervious surface,bare soil and water body has been established.The overall accuracy reached 88.45%and the kappa coefficient was 0.7672 based on the accuracy assessment from the reference sites.And the Google Earth image shows that the change type and time of the monitoring have a good agreement with the actual change.3.The information extraction of percentage change of impervious surface was presented.The percent impervious surface variation information was extracted based on identification results from Landsat data.The spatial distribution of percent impervious surface change in China was obtained using the random forest algorithm,which combined with the multi-year trends of OSAVI,LSWI,RIBS and BCI.From 2001 to 2016,the area of urban expansion in China accounted for 3.68%of the area of China’s mainland.4.The spatiotemporal analysis of urban expansion in China was analyzed from 2001 to 2016 based on the two aspects of socio-economic factors and urban development trajectories.The speed of economic growth and urban expansion area of China has a strong correlation,and their correlation coefficient reaches 0.82.The speed of construction has slowed as construction time has increased in the areas where urban expansion has been taking place for more than three years.This paper analyzed the ratio of land utilization and land productivity of China’s cities combined with population flow data and economic growth data,and the result show that Beijing,Shanghai,Guangdong and Tianjin are much higher than the national average regardless of land yield or the urban land use rate.Finally,four different patterns of urban expansion were proposed according to the city’s development trajectory. |
