Potential of C-band Synthetic Aperture Radar Sentinel-1 time-series for the monitoring of phenological cycles in a deciduous forest

Annual time-series of the two satellites C-band SAR (Synthetic Aperture Radar) Sentinel-1A and 1B data over five years were used to characterize phenological cycle a temperate deciduous forest. Six metrics start (SOS), middle (MOS) end (EOS) budburst leaf expansion stage in spring, (SOF), (MOF) (EOF) senescence autumn extracted using an asymmetric double sigmoid function (ADS) fitted ratio (VV/VH) backscattering at co-polarization VV (vertical–vertical) cross polarization VH (vertical-horizontal). Phenological also derived from other four vegetation proxies (Normalized Difference Vegetation Index NDVI Sentinel-2A 2B images, situ measurements measurements, Leaf Area LAI litterfall temporal dynamics). These estimated compared observations obtained by visual ground, achieved binoculars three inter-calibrated observers, on bi-weekly basis during weekly senescence. We observe decrease coefficient (?0) development phase spring increase phase, contrary what is usually observed various types crops. In vertical polarization, ?0VV shows very little variation throughout year. S-1 VV/VH provide good description seasonal allowing estimation metrics. Estimates provided dates MOS criterion differ approximately 8 days average (mean deviation) observations. During estimates MOF are later deviate about 20 while differences order 2 4 between based time-series, respectively. A deviation 7 days, comparable that budburst, those determined monitoring litterfall. While emergence described or explain (or ?0VH), senescence, decorrelated better explained dynamics. This behavior resulted hysteresis phenomenon relationships LAI. For same NDVI, response different depending considered. study high potential offered Sentinel-1 for detection forest phenology, thus overcoming limitations caused cloud cover optical remote sensing phenology.