Vegetation Phenology Extraction and Analysis in Urban Area using Landsat Vegetation Fraction Time Series

Considering diverse vegetation types and complex land cover changes in urban and surrounding areas, a new method was developed for phenology extraction in urban and surrounding areas using Landsat time series and global generic endmembers from standardized spectral mixing analysis. NDVI and EVI time series from Landsat images were also generated for comparison. Landsat time series images over Beijing area of three different years, i.e., 1990, 2000 and 2014, and total 83 scenes were used in this study. The results of phenology extraction from vegetation fraction showed slightly better than those from EVI, and much better than those from NDVI. It also showed that the characteristics of vegetation phenology in urban and surrounding areas have obvious differences in spatial distribution and among different vegetation types. The method provides an effective tool for phenology extraction and analysis in urban areas.