Estimation of Autumn, Winter, and Growing-Season Phenology of Evergreen Coniferous Forest from MODIS GRVI Time-series Data

Abstract

The objective of this study was to estimate the autumn and winter phenology and the length of the growing season of evergreen coniferous forests from MODIS GRVI (green-red ratio vegetation index) time-series data. We derived GRVI time-series data from MODIS 500 m reflectance data (MCD43 A4) for the period 2001–2012 from plantations of Japanese cedar (Cryptomeria japonica) and hinoki cypress (Chamaecyparis obtusa) in the Ashio mountainous region, Tochigi, Japan. By fitting logistic functions to the time-series data for each pixel, we detected four phenological transition dates: the onset of GRVI increase (OGI), the onset of GRVI maximum (OGMax), the onset of GRVI decrease (OGD), and the onset of GRVI minimum (OGMin). The length of the growing season for each pixel was calculated as the difference between the date of OGI and the date of OGMin. The averages of estimated dates of OGD and OGMin were DOY (day-of-year) 301 and DOY 378, respectively. These two dates were positively related to air temperature in autumn (OGD, r = 0.71) and air temperature in winter (OGMin, r = 0.86) for the 12 years. Using the statistical relationships with air temperature, we could estimate the interannual variations in the phenological transition dates in autumn and winter for evergreen coniferous forests from the satellite-monitored GRVI. The average growing season length was 287 days, or about 80% of a year. The results show that the changes in the growing season length were caused mainly by changes in the date of OGMin in winter.