Abstract
Terrestrial vegetation is a crucial component of Earth's biosphere, regulating global carbon and water cycles and contributing to human welfare. Despite an overall greening trend, terrestrial vegetation exhibits a significant inter-annual variability. The mechanisms driving this variability, particularly those related to climatic and anthropogenic factors, remain poorly understood, which hampers our ability to project the long-term sustainability of ecosystem services. Here, by leveraging diverse remote sensing measurements, we pinpointed 2020 as a historic landmark, registering as the greenest year in modern satellite records from 2001 to 2020. Using ensemble machine learning and Earth system models, we found this exceptional greening primarily stemmed from consistent growth in boreal and temperate vegetation, attributed to rising CO2 levels, climate warming, and reforestation efforts, alongside a transient tropical green-up linked to the enhanced rainfall. Contrary to expectations, the COVID-19 pandemic lockdowns had a limited impact on this global greening anomaly. Our findings highlight the resilience and dynamic nature of global vegetation in response to diverse climatic and anthropogenic influences, offering valuable insights for optimizing ecosystem management and informing climate mitigation strategies.
| Original language | English |
|---|---|
| Article number | 114494 |
| Journal | Remote Sensing of Environment |
| Volume | 316 |
| DOIs | |
| State | Published - Jan 1 2025 |
Funding
This research was supported by the Oak Ridge National Laboratory (ORNL) Reducing Uncertainties in Biogeochemical Interactions through Synthesis and Computing Scientific Focus Area project and the Terrestrial Ecosystem Science Scientific Focus Area project funded through the Earth and Environmental Systems Sciences Division of the Biological and Environmental Research Office in the Department of Energy Office of Science. ORNL is supported by the DOE Office of Science under Contract No. DE-AC05-00OR22725. This work was also supported by the Bipartisan Infrastructure Law (BIL) Project Plan from USDA Forest Service (23-JV-11330180-119), Nicholas School of the Environment from Duke University, the Institute for a Secure & Sustainable Environment from University of Tennessee at Knoxville.We gratefully acknowledge the support from the Duke Computing Cluster, the ISAAC High Performance Scientific Computing Platform at UTK, and the development team of the open-access automated machine learning Python package (mljar). We also extend our thanks to NASA, ESA, Peng Cheng Laboratory, and other institutions and authors for their publicly available data, as detailed in Table S1. This research was supported by the Oak Ridge National Laboratory (ORNL) Reducing Uncertainties in Biogeochemical Interactions through Synthesis and Computing Scientific Focus Area project and the Terrestrial Ecosystem Science Scientific Focus Area project funded through the Earth and Environmental Systems Sciences Division of the Biological and Environmental Research Office in the Department of Energy Office of Science . ORNL is supported by the DOE Office of Science under Contract No. DE-AC05-00OR22725 . This work was also supported by the Bipartisan Infrastructure Law (BIL) Project Plan from USDA Forest Service , Nicholas School of the Environment from Duke University , the Institute for a Secure & Sustainable Environment from University of Tennessee at Knoxville.
Keywords
- COVID-MIP
- Ensemble
- Environmental changes
- Long-term and short-term
- Machine learning
- Multi-source remote sensing
- Pandemic
- Record-high greening
- Terrestrial vegetation
- TRENDY
