Vegetation phenology is an important parameter in models of global vegetation and land surfaces, as the accuracy of these simulations depends strongly on the description of the canopy status. Temperate forests represent one of the major types of vegetation at mid-high latitudes in the Northern Hemisphere and act as a globally important carbon sink. Thus, a better understanding of the phenological variables of temperate forests will improve the accuracy of vegetation models and estimates of regional carbon fluxes. In this work, we explored the possibility of using digital camera images to monitor phenology at species and community scales of a temperate forest in northeastern China, and used the greenness index derived from these digital images to optimize phenological model parameters. The results show that at the species scale, the onset dates of phenological phases （Korean pine, Mongolian oak） derived from the images are close to those from field observations （error 〈 3d）. At the community scale the time series data accurately reflected the observed canopy status （A^2=0.9） simulated using the phenological model, especially in autumn. The phenological model was derived from simple meteorological data and environmental factors optimized using the greenness index. These simulations provide a powerful means of analyzing environmental factors that control the phenology of temperate forests. The results indicate that digital images can be used to obtain accurate phenologicai data and provide reference data to validate remote-sensing phenological data. In addition, we propose a new method to accurately track phenological phases in land-surface models and reduce uncertainty in spatial carbon flux simulations.
