Automatic mapping afforestation, cropland reclamation and variations in cropping intensity in central east China during 2001–2016

Accurate and automatic monitoring and assessment of vegetation changes are important to support food security, ecosystem balance and global climate regulation. Compared with urbanization and deforestation, vegetation changes such as afforestation, cropland reclamation and variations in cropping intensity were understudied. This study aimed to propose an Automatic Method for detecting Multiple vegetation Changes (AMMC) through a knowledge-based strategy. Five temporal indices were proposed in order to fully characterize different vegetation types from four aspects: vegetation abundance, temporal dispersion, primary/minor temporal continuity, and growing season length. The AMMC takes advantage of the knowledge on the expected temporal trajectories of vegetation dynamics, which could be tracked based on their corresponding trends in these temporal indices. The efficiency of the proposed AMMC method was verified with its applications in central east China using 500 m 8 day composite MODIS datasets from 2001 to 2016. An overall accuracy of 94.75% was achieved when evaluated with 3,011 reference sites. Results revealed that there were totally 7,180 km², 3,610 km² and 3,280.5 km² areas of afforestation, cropland reclamation and variations in cropping intensity in central east China, respectively. This study verified that afforestation efforts were succeeded, but “Grain for Green” project was not as expected since more cropland was reclamation implemented at less favorable biophysical conditions than cropland retirement.