Shin Nagai

Review: Development of an in situ observation network for terrestrial ecological remote sensing: the Phenological Eyes Network (PEN)

The Phenological Eyes Network (PEN), which was established in 2003, is a network of long-term ground observation sites. The aim of the PEN is to validate terrestrial ecological remote sensing, with a particular focus on seasonal changes (phenology) in vegetation. There are three types of core sensors at PEN sites: an Automatic Digital Fish-eye Camera, a HemiSpherical SpectroRadiometer, and a Sun Photometer. As of 2014, there are approximately 30 PEN sites, among which many are also FluxNet and/or International Long Term Ecological Research sites. The PEN is now part of a biodiversity observation framework. Collaborations between remote sensing scientists and ecologists working on PEN data have produced various outcomes about remote sensing and long-term in situ monitoring of ecosystem features, such as phenology, gross primary production, and leaf area index. This article reviews the design concept and the outcomes of the PEN, and discusses its future strategy.

Assessing the use of camera-based indices for characterizing canopy phenology in relation to gross primary production in a deciduous broad-leaved and an evergreen coniferous forest in Japan

Abstract Recent studies have reported that seasonal variation in camera-based indices that are calculated from the digital numbers of the red, green, and blue bands (RGB_DN) recorded by digital cameras agrees well with the seasonal change in gross primary production (GPP) observed by tower flux measurements. These findings suggest that it may be possible to use camera-based indices to estimate the temporal and spatial distributions of photosynthetic productivity from the relationship between RGB_DN and GPP. To examine this possibility, we need to investigate the characteristics of seasonal variation in three camera-based indices (green excess index [GE], green chromatic coordinate [rG], and HUE) and the robustness of the relationship between these indices and tower flux-based GPP and how it differs among ecosystems. Here, at a daily time step over multiple years in a deciduous broad-leaved and an evergreen coniferous forest, we examined the relationships between canopy phenology assessed by using the three indices and GPP determined from tower CO2 flux observations, and we compared the camera-based indices with the corresponding spectra-based indices estimated by a spectroradiometer system. We found that (1) the three camera-based indices and GPP showed clear seasonal patterns in both forests; (2) the amplitude of the seasonal variation in the three camera-based indices was smaller in the evergreen coniferous forest than in the deciduous broad-leaved forest; (3) the seasonal variation in the three camera-based indices corresponded well to seasonal changes in potential photosynthetic activity (GPP on sunny days); (4) the relationship between the three camera-based indices and GPP appeared to have different characteristics at different phenological stages; and (5) the camera-based and spectra-based HUE indices showed a clear relationship under sunny conditions in both forests. Our results suggest that it might be feasible for ecologists to establish comprehensive networks for long-term monitoring of potential photosynthetic capacity from regional to global scales by linking satellite-based, in situ spectra-based, and in situ camera-based indices.

What makes the satellite-based EVI–GPP relationship unclear in a deciduous broad-leaved forest?

Recent studies have suggested that gross primary production (GPP) of terrestrial vegetation can be estimated directly with the satellite-based Enhanced Vegetation Index (EVI). However, the reported EVI–GPP relationships showed wide variability, with the regression functions showing widely scattered data. In the present study, we examined the possible reasons for this variability in the EVI–GPP relationship using daily EVI values from satellite and field measurements and daily flux-based GPP in a cool-temperate deciduous broad-leaved forest in Japan. The variability appears to be caused by noise due to cloud contamination in the satellite data as well as the different seasonality of EVI and GPP, especially during the leaf-expansion period. Our findings indicate that improvement of cloud screening and consideration of the leaf-expansion period are critical when applying the EVI–GPP relationship.

In situ examination of the relationship between various vegetation indices and canopy phenology in an evergreen coniferous forest, Japan

We examined the relationship between four vegetation indices and tree canopy phenology in an evergreen coniferous forest in Japan based on observations made using a spectral radiometer and a digital camera at a daily time step during a 4 year period. The colour of the canopy surface of Japanese cedar (Cryptomeria japonica) changed from yellowish-green to whitish-green from late May to July and turned reddish-green in winter. The normalized difference vegetation index (NDVI), enhanced vegetation index (EVI) and plant area index (PAI) showed no seasonality. In contrast, the green–red ratio vegetation index (GRVI) increased from March to June and then decreased gradually from July to December, resulting in a bell-shaped curve. GRVI revealed seasonal changes in the colour of the canopy surface. GRVI correlated more positively with the evaluated maximum photosynthetic rate for the whole forest canopy, A max, than did NDVI or EVI. These results suggest the possibility that GRVI is more useful than NDVI and EVI for capturing seasonal changes in photosynthetic capacity, as the green and red reflectances are strongly influenced by changes in leaf pigments in this type of forest.

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

Abstract Clarification of species-specific year-to-year variations of the timings of the start of leaf-expansion (SLE) and the end of leaf-fall (ELF) 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 SLE and ELF, 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 SLE and ELF 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 ELF among the tree species were more apparent than those of the timing of SLE among the tree species, (2) the threshold value of the camera-based index (green excess index) for detecting the timing of ELF 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 ELF to air temperature was lower than that of the timing of SLE. 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.

Examination of the extinction coefficient in the Beer–Lambert law for an accurate estimation of the forest canopy leaf area index

Leaf area index (LAI) is a crucial ecological parameter that represents canopy structure and controls many ecosystem functions and processes, but direct measurement and long-term monitoring of LAI are difficult, especially in forests. An indirect method to estimate the seasonal pattern of LAI in a given forest is to measure the attenuation of photosynthetically active radiation (PAR) by the canopy and then calculate LAI by the Beer–Lambert law. Use of this method requires an estimate of the PAR extinction coefficient (k), a parameter needed to calculate PAR attenuation. However, the determination of k itself requires direct measurement of LAI over seasons. Our goals were to determine (1) the best way to model k values that may vary seasonally in a forest, and (2) the sensitivity of estimates of canopy ecosystem functions to the errors in estimated LAI. We first analyzed the seasonal pattern of the “true” k (k p) under cloudy and sunny conditions in a Japanese deciduous broadleaved forest by using the inverted form of the Beer–Lambert law with the true LAI and PAR. We next calculated the errors of PAR-based LAIs estimated with an assumed constant k (LAIpred) and determined under what conditions we should expect k to be approximately constant during the growing period. Finally, we examined the effect of errors in LAIpred on estimates of gross primary production (GPP), net ecosystem production (NEP), and latent heat flux (LE) calculated with a land-surface model using LAIpred as an input parameter. During the growing period, cloudy k p varied from 0.47 to 1.12 and sunny k p from 0.45 to 1.59. Results suggest that the value of LAIpred was adequately estimated with the k p obtained under cloudy conditions during the fully-leaved period (0.53–0.57). However, LAIpred was overestimated by up to 0.6 m2 m–2 in May and November. The errors in LAIpred propagated to errors in modeled carbon and latent heat fluxes of –0.21 to 0.32 g C m–2 day–1 in GPP, –0.09 to 0.19 g C m–2 day–1 in NEP, and –3.2 to 3.9 W m–2 in LE, which is close to the measurement errors recognized in the tower flux measurement. LAIpred estimated with an assumed constant k can be useful for some ecosystem studies as a second-best alternative if k is equated to the value of k p measured under cloudy conditions especially during the fully-leaved period.

Relationship between spatio-temporal characteristics of leaf-fall phenology and seasonal variations in near surface-and satellite-observed vegetation indices in a cool-temperate deciduous broad-leaved forest in Japan

We examined the relationship between the spatio-temporal distribution of leaf litter for each species and the seasonal patterns of in situ and satellite-observed daily vegetation indices in a cool-temperate deciduous broad-leaved forest. The timing and distribution of leaf-fall revealed spatio-temporal relationships with species and topography. Values of the normalized difference vegetation index (NDVI), enhanced vegetation index (EVI), and green–red vegetation index (GRVI), measured both in situ and by satellite, and those of the in situ-measured leaf area index (LAI), rapidly declined at the peak of leaf-fall. At the late stage of leaf-fall, in situ-measured values of NDVI, EVI, and LAI declined but those of GRVI changed from decreasing to increasing. The peak timing of leaf-fall, when 50–73% of the leaf litter had fallen, corresponds to LAI = 1.80–0.81, NDVI = 0.61–0.54, EVI = 0.29–0.25, and GRVI = 0.01 ∼ ‒0.07. Although the distribution of leaf litter among species displayed spatial characteristics at the peak of leaf-fall, spatial heterogeneity of amount of leaf litter at the peak timing of leaf-fall was less than that at the beginning and end. These facts suggest that the criterion for determining the timing of leaf-fall from vegetation indices should be a value corresponding to the peak of leaf-fall rather than its end. In a high-biodiversity forest, such as this study forest, the effect of spatial heterogeneity on the timing and patterns of leaf-fall on vegetation indices can be reduced by observing only the seasonal variation in colour on the canopy surface by using GRVI, which consists of visible reflectance bands, rather than that of both leaf area and colour of the canopy surface by using NDVI and EVI, which consist of visible and near-infrared reflectance bands.

Spatial Scale and Landscape Heterogeneity Effects on FAPAR in an Open-Canopy Black Spruce Forest in Interior Alaska

Black spruce forests dominate the land cover in interior Alaska. In this region, satellite remote sensing of ecosystem productivity is useful for evaluating black spruce forest status and recovery processes. The fraction of absorbed photosynthetically active radiation (FAPAR) by green leaves is a particularly important input parameter for ecosystem models. FAPAR_1d is computed as the ratio of absorbed photosynthetically active radiation (APAR_3d) to the incident photosynthetically active radiation at the horizontal plane above the canopy (PAR_1d, FAPAR_1d = APAR_3d/PAR_1d). The parameter FAPAR_1d is scale dependent and can be larger than 1 as a result of laterally incident PAR. We investigated the dependence of FAPAR_1d on spatial scale in an open-canopy black spruce forest in interior Alaska. We compared FAPAR_1d with FAPAR_3d( = APAR_3d/PAR_3d), the latter of which considers incident PAR as actinic flux (spheradiance) (PAR_3d). Our results showed the following: 1) landscape scale FAPAR_3d(30×30 m²) was always larger (0.39-0.43) than FAPAR_1d (0.19-0.27) due to the landscape heterogeneity and incident PAR regime, and 2) at the individual tree scale, FAPAR_1d was highly variable, with 34% (day of year [DOY] 180) to 52% (DOY 258) of , whereas FAPAR_3d varied across a much narrower range (0.2-0.5). The spatial-scale dependence of the ratio of PAR_3d to PAR_1d converged at the pixel size larger than 5 m. Thus, a 5-m or coarser resolution was necessary to ignore the lateral PAR effect in the open-canopy black spruce forest.

Utility of information in photographs taken upwards from the floor of closed-canopy deciduous broadleaved and closed-canopy evergreen coniferous forests for continuous observation of canopy phenology

Abstract Hemispherical photographs taken on forest floors are used to monitor seasonal changes in canopy openness or leaf area index in ecological studies. Those analyses usually use black and white images converted from the original colour images. Photographs taken by downwards-facing cameras installed on towers are used to provide detailed information on leaf expansion, maturation and senescence of various tree species through the analysis of red, green and blue ‘digital numbers’ (DNRGB) extracted from those images. To examine the usefulness of colour information encoded in upwards hemispherical photographs in monitoring canopy phenological characteristics, we examined the consistency of DNRGB values between downwards and upwards images in deciduous broadleaved and evergreen coniferous forests in Japan. In the deciduous broadleaved forest, the DNRGB values in the upwards images were able to detect canopy phenology almost as well as those in the downwards images. However, we found the effects on DNRGB of (1) the spatial heterogeneity among observed points, (2) low-vegetation (before the beginning of leaf-expansion and after the end of leaf-fall period) and (3) white balance settings. In the evergreen coniferous forest, in contrast, the DNRGB values in the upwards images did not capture canopy phenology. These different results may be attributable to the light attenuation characteristics in the canopies due to the geometries of leaves and branches. Thus, the DNRGB values obtained from upwards images are almost as good as those of downwards images for monitoring detailed canopy phenology in deciduous broadleaved forests with a closed canopy.

Using ordinary digital cameras in place of near-infrared sensors to derive vegetation indices for phenology studies of high arctic vegetation

To remotely monitor vegetation at temporal and spatial resolutions unobtainable with satellite-based systems, near remote sensing systems must be employed. To this extent we used Normalized Difference Vegetation Index NDVI sensors and normal digital cameras to monitor the greenness of six different but common and widespread High Arctic plant species/groups (graminoid/Salix polaris; Cassiope tetragona; Luzula spp.; Dryas octopetala/S. polaris; C. tetragona/D. octopetala; graminoid/bryophyte) during an entire growing season in central Svalbard. Of the three greenness indices (2G_RBi, Channel G% and GRVI) derived from digital camera images, only GRVI showed significant correlations with NDVI in all vegetation types. The GRVI (Green-Red Vegetation Index) is calculated as (GDN − RDN)/(GDN + RDN) where GDN is Green digital number and RDN is Red digital number. Both NDVI and GRVI successfully recorded timings of the green-up and plant growth periods and senescence in all six plant species/groups. Some differences in phenology between plant species/groups occurred: the mid-season growing period reached a sharp peak in NDVI and GRVI values where graminoids were present, but a prolonged period of higher values occurred with the other plant species/groups. In particular, plots containing C. tetragona experienced increased NDVI and GRVI values towards the end of the season. NDVI measured with active and passive sensors were strongly correlated (r > 0.70) for the same plant species/groups. Although NDVI recorded by the active sensor was consistently lower than that of the passive sensor for the same plant species/groups, differences were small and likely due to the differing light sources used. Thus, it is evident that GRVI and NDVI measured with active and passive sensors captured similar vegetation attributes of High Arctic plants. Hence, inexpensive digital cameras can be used with passive and active NDVI devices to establish a near remote sensing network for monitoring changing vegetation dynamics in the High Arctic.