Monitoring autumn leaf phenology at the individual tree scale using PlanetScope satellite imagery and ground-based observations

Individual tree scale autumn leaf phenology marks the end of the growing season and influences ecological functions such as carbon storage of forest ecosystems. Traditional ground-based observations are time- and labor-intensive, while the coarse resolution of commonly used satellites limits their effectiveness. PlanetScope imagery, with 3 m resolution and near-daily revisit, offers new potential for monitoring phenology at the individual tree scale. However, its effectiveness remains underexplored. This study aims to assess the potential of PlanetScope imagery for monitoring autumn leaf phenology at the individual tree scale in the University of Tokyo Hokkaido Forest (UTHF), Japan. We investigated its capability to track autumn leaf phenology and evaluated the agreement of phenological metrics, the timing of start (SOA), middle (MOA), and end (EOA) of autumn phenology, using threshold-based and 3rd-order derivative methods. Our results demonstrate that PlanetScope surface reflectance (SR) and vegetation indices (VIs) effectively tracked the progression of leaf coloration, with VIs showing stronger correlations than SRs. Among SR bands, the NIR performed best (r = − 0.767). Metrics derived from the threshold-based method generally showed higher agreement. The near-infrared reflectance of vegetation (NIRv) and two-band enhanced vegetation index (EVI2) showed high agreement for MOA, where NIRv achieved higher precision (r = 0.774–0.780, RMSE = 9.2–10.8 days) but slightly lower accuracy than EVI2 (r = 0.763–0.767, RMSE = 4.8 days). Overall, these findings highlight the potential of PlanetScope imagery for individual tree monitoring of autumn phenology, while species-specific differences should be considered to further improve agreement.