Detection of the different characteristics of year-to-year variation in foliage phenology among deciduous broad-leaved tree species by using daily continuous canopy surface images

Highlights

• We took downward images of the canopy surface in a deciduous forest for 10 years.

• We examined year-to-year variation in foliage phenology for each tree species.

• We examined threshold values of a RGB-index (GEI) for detecting phenological change.

• Species-specific year-to-year variations in the timing of leaf-fall were observed.

• Threshold values of GEI were affected by the greenness of understories.

Abstract

Clarification of species-specific year-to-year variations of the timings of the start of leaf-expansion (S_LE) and the end of leaf-fall (E_LF) is an important and challenging task because these timings may alter spatial and temporal variations in ecosystem services such as carbon stock and climate control. Although many previous studies have applied automatically captured digital camera images to observe the timings of S_LE and E_LF, the evaluation of the long-term variation in both timings of each tree species based on image analysis has not yet been sufficiently investigated. In this study, we investigated the year-to-year variation in the timings of S_LE and E_LF for multiple deciduous broad-leaved tree species in a cool-temperate deciduous broad-leaved forest in Japan by using long-term and daily hemispherical (“fish-eye”) canopy surface images from 2004 to 2013. We found that (1) differences in the characteristics of year-to-year variations in the timing of E_LF among the tree species were more apparent than those of the timing of S_LE among the tree species, (2) the threshold value of the camera-based index (green excess index) for detecting the timing of E_LF varied depending on the spatial and temporal distribution of understories and the visual distortion of the fish-eye images, and (3) the phenological sensitivity of the timing of E_LF to air temperature was lower than that of the timing of S_LE. Our results indicate that it might be helpful for ecologists to use daily continuous canopy surface images for monitoring of species-specific characteristics of spatial and temporal changes in foliage phenology in mixed-species deciduous broad-leaved forests.

Introduction

Changes in plant phenology, such as advanced or delayed timings of the start of leaf-expansion (S_LE) and the end of leaf-fall (E_LF) (e.g., Cong et al., 2013, Jeong et al., 2011, Zeng et al., 2011), may alter ecosystem functions such as water, energy, and carbon exchanges between the land surface and the atmosphere in terrestrial ecosystems by changing physical properties (e.g., solar albedo and evapotranspiration) and biological processes (e.g., CO₂ uptake by photosynthesis) (Barford et al., 2001, Churkina et al., 2005, Lawrence and Slingo, 2004a, Lawrence and Slingo, 2004b, Myneni et al., 1997). Long-term trends of advanced S_LE and delayed E_LF showed different characteristics among each region at a continental scale. For example, the long-term trend of delayed E_LF (e.g., 2.2 days per decade from 1986 to 2005 in China [Chen and Xu, 2012], and 1.6 days per decade from 1959 to 1993 in Europe [Menzel and Fabian, 1999]) was slower than the long-term trend of advanced S_LE (e.g., 4 days per decade from 1986 to 2005 in China [Chen and Xu, 2012], and 2.0 days per decade from 1959 to 1993 in Europe [Menzel and Fabian, 1999]) in temperate climate regions across Europe and China, whereas the long-term trend of delayed E_LF (1.6 days per decade from 1953 to 2000 [Matsumoto et al., 2003] and 2.7 days per decade from 1961 to 2011 [Nagai et al., 2013a]) was faster than the long-term trend of advanced S_LE (0.9 days per decade from 1953 to 2000 [Matsumoto et al., 2003]) in Japan. These local phenological changes may alter ecosystem services, such as carbon stock and climate control, on a regional to a global scale (Peñuelas et al., 2009).

Sensitivity of foliage phenology under environmental changes shows different characteristics among temperate deciduous tree species (Fu et al., 2013, Morin et al., 2010, Vitasse et al., 2009a, Vitasse et al., 2009b). For instance, Vitasse et al. (2009a) reported that the phenological sensitivity of leaf unfolding to temperature for oak was stronger than that for beech in southern France. These differences in species-specific phenological response to environmental changes may affect the competition for light among the tree species, and it may also therefore affect the distribution of tree species in temperate mixed-species deciduous forests (Kramer et al., 2000). Urbanski et al. (2007) suggested that the change of spatial distribution of tree species might alter the net ecosystem production in a mixed deciduous forest. These reports suggest that consideration of the differences in the species-specific phenological response to environmental change in each region would provide us with useful information to predict the response of ecosystem services to rapid environmental change.

Recently, automatically captured digital camera images have been applied to continuous phenological observations because such observations make it possible to collect automated data at high temporal resolution in multiple ecosystems with low operating costs (e.g., Alberton et al., 2014, Crimmins and Crimmins, 2008, Migliavacca et al., 2011, Richardson et al., 2007, Sonnentag et al., 2012). Temporal changes in RGB color information (the digital intensity value of red, green, and blue), which is extracted from the images, allow researchers to detect foliage phenological changes such as leaf-expansion and leaf senescence (e.g., Henneken et al., 2013, Ide and Oguma, 2010, Nagai et al., 2013b, Zhao et al., 2012). Some studies have used the greenness index for estimation of the green-up date (Ahrends et al., 2008, Ahrends et al., 2009, Ide and Oguma, 2010, Richardson et al., 2007). Although the image-analysis methods for the detection of the timing of phenological changes, such as leaf-expansion, leaf-coloring, and leaf-fall, have been developed by such previous studies (e.g., Nagai et al., 2011), the evaluation of the long-term variation in the timings of these phenological events in each deciduous tree species based on this type of image analysis has not yet been sufficiently investigated.

In this study, (1) we investigated the year-to-year variation in the timings of S_LE and E_LF for four tree species (redvein maple: Acer rufinerve, Erman's birch: Betula ermanii, Miyama cherry: Prunus maximowiczii, and white oak: Quercus crispula) in a cool-temperate deciduous broad-leaved forest in Japan by using daily continuous canopy surface images from 2004 to 2013, (2) we also investigated the differences of the phenological sensitivity in the timings of S_LE and E_LF to air temperature among the tree species, and (3) we discussed the usefulness and problems of using canopy surface images for phenological observations in a forest ecosystem. Our aim was to provide robust evidence of the usefulness of daily continuous canopy surface images for continuous observation of the interannual variation in foliage phenology among various tree species within a deciduous broad-leaved forest.