Using High-Resolution Satellite Data and Phenocams to Track Species and Site-Dependent Variability in Urban Vegetation Phenology

Abstract:

Vegetation phenological response can aid in understanding the extent of anthropogenic effects on climate caused by urbanization and land-use. Past urban phenology research has been limited to only characterizing vegetation at coarse spatial scales and capturing interannual variability. Individual tree species phenology may vary in their response to temperature and water availability; however, this has not been well characterized. Site characteristics may also influence intra-annual vegetation phenology due to specific features, like pervious and impervious surfaces, that can influence temperature at the microsite-level.

This research uses high-resolution remote sensing techniques and phenological webcams (phenocams) to analyze phenological variability of urban vegetation at the scale of the individual tree species in Washington, D.C. Data collection through Planet satellite imagery and phenocams is being used to track variability in start of season (SOS) and end of season (EOS) to highlight the spatial and species differences. Phenocams have been placed throughout the study site to serve as validation data for the Planet imagery. By December 2020, there will be over 20 phenocams placed at citizen scientist houses covering a range of species and site conditions. Phenocam data have been converted to a time-series of Green Chromatic Coordinate (GCC) and Red Chromatic Coordinate (RCC) to track green-up and senescence at a daily timestep. We employ a linear mixed modeling approach to attribute the variation in phenological metrics to species and planting site conditions.

Preliminary data from satellite imagery has highlighted our ability to capture phenological time series for a large study area at the scale of the individual tree. Phenocam data, in some cases, reveals different GCC and RCC time series for trees of the same species. Due to warmer spring temperatures, there has been an earlier SOS compared to 2019 but species-specific responses are mixed. Given the availability of high-resolution datasets, the methods developed in this study can be applied to other cities globally, in temperate zones.