Chunsheng Wang
Chunsheng Wang is a Ph.D. candidate at the International Institute for Earth System Science, Nanjing University, China. His research primarily focuses on the intersection of Earth Observation, GeoAI, and global biogeochemical cycles.
Specifically, he leverages multi-source remote sensing data and machine learning algorithms to map large-scale forest ecosystem functioning, monitor litterfall dynamics, and model soil respiration. His goal is to reduce uncertainties in traditional meteorological proxies and provide scalable data infrastructure for next-generation Earth System Models. His most recent breakthrough in evaluating global forest litterfall dynamics and its biogeochemical coupling has been published in the prestigious journal Remote Sensing of Environment.
Sessions
The organic carbon flux entering the pedosphere through forest litterfall is a critical indicator of forest ecosystem functioning and a primary driver of soil respiration (RS). However, accurately quantifying litterfall spatiotemporal dynamics at the global scale remains a major challenge due to the scarcity of high-resolution Earth Observation (EO) frameworks coupled with extensive ground observations.
Here, we present a novel GeoAI-driven approach that synthesizes 14,912 in-situ observations across 843 sites globally with multi-source remote sensing data. By leveraging machine learning algorithms, we decoupled complex biogeochemical mechanisms and generated a 500-m spatial resolution global forest litterfall product. Furthermore, we integrated these high-resolution EO derivatives into an Olson legacy model to quantify the impact of litterfall on RS across different forest biomes.
Our results reveal significant spatial heterogeneity in biogeochemical coupling, highlighting asymmetric microbial responses between tropical forests (characterized by high turnover rates) and temperate/boreal forests (exhibiting biogeochemical inertia). This study demonstrates the profound potential of integrating open Earth Observation data and machine learning to monitor global forest dynamics. Our 500-m global product provides a vital, scalable data infrastructure for next-generation Earth system models, biodiversity conservation, and forest carbon management.
(Note: This research has been recently published in Remote Sensing of Environment, 2026, https://doi.org/10.1016/j.rse.2026.115373)