Earth system dynamics
Nonlinear dynamics, extreme events, tipping behavior, and land-atmosphere feedbacks, with an emphasis on mechanisms that can be recovered from incomplete observations.
Research
I study a changing Earth system by combining climate science, geospatial observations, physical understanding, and machine learning.
Research directions
Climate impacts
Vulnerability and resilience of ecosystems and urban infrastructure under climate change.
Spatiotemporal intelligence
Physics-aware and data-driven modeling, uncertainty quantification, and interpretable AI.
Featured papers
Climate systems
Generative deep learning improves reconstruction of global historical climate records
Recovering long-term global climate records to better understand historical warming, hydroclimate variability, and climate extremes.
arXiv:2602.16515, 2026
Forest ecosystems
Decadal sink-source shifts of forest aboveground carbon since 1988
Tracing multi-decadal forest carbon dynamics to reveal emerging sink-source shifts and the changing role of tropical and boreal forests in the carbon cycle.
Nature Communications, 2026
Urban systems
Worldwide rooftop photovoltaic electricity generation may mitigate global warming
Assessing the global climate benefits of rooftop photovoltaics and their role in urban energy transitions and warming mitigation.
Nature Climate Change, 2025
Earth system modeling
Iterative integration of deep learning in hybrid Earth surface system modelling
Framing how deep learning and Earth system knowledge can be iteratively integrated to improve environmental modeling and scientific understanding.
Nature Reviews Earth & Environment, 2023