Year-to-year blooming phenology observation using time-lapse digital camera images

Long-term continuous phenological observation of cherry tree blooming is an important and challenging task in the evaluation of year-to-year weather and climate changes in spring in Japan. Here, (1) we performed daily field observations with a time-lapse digital camera in a deciduous broad-leaved forest in Japan from January 2004 to December 2013; and (2) we detected year-to-year variations in the blooming phenology of Prunus sargentii by visual inspection of the images and by image analysis. We found that (1) the red digital numbers (i.e., the digital intensity values of the red pixels; DN_R) extracted from the time-lapse digital camera images tended to peak in full bloom; and (2) the green excess index (GEI) (based on the red, green, and blue digital numbers) tended to be lowest in full bloom. These results indicate that DNR and GEI are useful for detecting the timing of full bloom in P. sargentii.