[Spatiotemporal dynamics of vegetation cover based on trajectory change detection: a case study in Dapeng Peninsula of Shenzhen].

By using the second-time developed ArcEngine component at pixel level, this paper studied the spatiotemporal dynamics of vegetation cover in the Dapeng Peninsula of Shenzhen, China in 1986-2007, and analyzed the characters and causes of the dynamics. To quantify this dynamics, the NDVI changes in 1986-2007 were extracted from 10 time-series TM/ETM+ remote sensing images, and the results showed that from 1986 to 2007, there were four trajectories of vegetation cover change in the Peninsula, including stable (a), stable-rising-stable (aba), stable-descending-stable (aca), and stable-descending-stable-rising-stable (acaba). The area with these four types occupied 71.54% of the total. Among the four types, type "a" was most common, occupying 1/3 of the study area, mainly in the mountains; and type "acaba" was the typical one, which was closely related to the deforestation and reforestation after the human disturbances of original vegetation. The areas at higher elevation or steeper slopes exhibited smaller vegetation change, mainly because of the constrained human disturbances. Timing of the vegetation cover change showed a relative stability in the mid-90s of 20th century, but a dramatic change after 2003, coinciding with the growth of Shenzhen City.