A remotely sensed observation of an evergreen forest phenology is more difficult than that of a deciduous forest, as seasonal changes on leaves are relatively small. In this research, we estimated spring phenological transition dates of evergreen coniferous forest using GRVI (Green red ratio vegetation index), which can detect changes in leaf color. Using plantations of Japanese cedar, hinoki cypress and sawara cypress located in the Ashio mountainous region in Tochigi, Japan, as test sites, we derived GRVI time-series data from the MODIS 500 m reflectance data (MCD43A4) during a 12-year period between 2001−2012. By fitting a logistic function to the GRVI time-series data from winter to summer for each pixel, we estimated two phenological transition dates, the onset of greenup and the onset of maturity. The GRVI time-series data showed a bell-shaped seasonal pattern for each year. The averages of estimated dates of greenup and maturity onset were DOY (Day-of-year) 92 and DOY 142, respectively, and there was a 50-day difference between the two phenological dates. There were significantly negative relationships between the average of the two estimated phenological transition dates and the spring air temperature (the onset of greenup,
r=−0.73; the onset of maturity,
r=−0.74) for the 12 years. In comparison with results based on EVI (Enhanced vegetation index) from MODIS-MCD12Q2 data, our estimations showed a low degree of spatial phenological variability. Based on the statistical relationships with spring temperature, the satellite monitored GRVI could estimate the interannual of variations in onset of greenup and maturity for evergreen coniferous forest.
抄録全体を表示