
New paper led by Shaoyu Wang hashtag #RSE 📢 We developed a deep-learning based GPP model, #GPPnet, which only requires multi-channel #Sentinel2 reflectance and incoming light. No land cover, no VPD, no temperature. This simple model predicted canopy photosynthesis for both C3 and C4 (without separate models), captured drought/heatwave effects, light use efficiency and interannual variability. Why and how multi-channel reflectance can explain variations in canopy photosynthesis remains an open question—and an exciting avenue that deserves future study. Because all input data are available in real time on #GEE, we envision monitoring canopy photosynthesis any place, any time. Ultimately, integrating data-driven (GPP-net), semi-empirical (NIRvP), and process-based models (BESS) will enable a wide range of applications, from near-real-time crop monitoring to global estimates of photosynthesis. @DechantBenjamin @HelinZhang11301 @FengHuaize @IjeonghoN @ChanghyunChoi13 @snucals1 @SeoulNatlUni GPP-net: a robust high-resolution GPP estimation network for Sentinel-... sciencedirect.com/science/ar…
























