Aoxue XU

My undergraduate training was in physical geography, where I used spectroscopic methods to study carbon cycling and emerging pollutants at the molecular scale. My work centered on DOM fluorescence via EEM-PARAFAC, microplastic photochemistry, and pollutant binding behavior. It gave me a strong grounding in analytical chemistry and biogeochemistry.

The questions I cared about most sat at a different scale. I wanted to zoom out from molecular characterization and ask how human activity reshapes land use and carbon emissions. The questions became: why does land use shift the way it does across municipalities, why do carbon emissions move with particular kinds of economic activity, and whether policy can redirect any of this.

My master's thesis is supervised by Prof. Binbin Li and developed with research partners at IPAM and ICV in Brazil. It applies GIS, remote sensing, and causal inference to study how the 2018 US-China soybean trade war reshaped illegal deforestation across Brazilian municipalities.

Outside of research, I explored AI for Science. The workbench grew out of my own meta-analysis work. When I did the literature review by hand, the search, screening, and synthesis steps stood out as the bottleneck. The result is a full-stack research workbench at workbench.hazalxu.click. It integrates multi-source search, screening, retrieval, and synthesis support for end-to-end literature work.

"How do economic shocks, environmental regulation, and spatial measurement jointly shape land-use and climate outcomes?"