Accurate detection of spatio-temporal variability of plant phenology by using satellite-observed daily green-red vegetation index (GRVI) in Japan

Abstract:

To evaluate the spatio-temporal variability of biodiversity and ecosystem functioning and service in deciduous forests, accurate detection of the timing of plant phenology such as leaf-flushing, -coloring, and -falling is important from plot to continental scales. Here, (1) we detected the spatio-temporal variability in the timing of start (SGS) and end of growing season (EGS) in Japan from 2001 to 2014 by analyzing Terra and Aqua/MODIS satellite-observed daily green-red vegetation index (GRVI) with a 500-m spatial resolution. (2) We examined the characteristics of timing of SGS and EGS in deciduous forests along vertical (altitude) and horizontal (latitude) gradients and their sensitivity to air temperature. (3) We evaluated the relationship between the spatial distribution of leaf-coloring phenology derived from Landsat-8/OLI satellite-observed GRVI with a 30-m spatial resolution on 23 November 2014 and leaf-coloring information published on web sites in Kanagawa Prefecture, Japan. We found that (1) changes along the vertical and horizontal gradients in the timing of SGS tended to be larger than those of EGS; (2) the sensitivity of the timing of SGS to air temperature was much more than that of EGS; and (3) leaf-coloring information published on web sites covering multiple points was useful for verification of leaf-coloring phenology derived from satellite-observed GRVI in relation to the altitude gradient in mountainous regions.