From Academic Research to Applied Satellite Data
From forest ecology research to hands-on data work at a climate company,
I bring both perspectives to every forest and environmental data challenge.
Why Forests and Satellite Data
Forests are too vast to survey on the ground alone, and too complex to read from space without ground truth. My work for the past decade has been building the bridge between the two: training satellite imagery on field data, then refining it with machine learning to capture forest and land use change at a scale and accuracy neither can reach alone. Off the clock, I’m still chasing that same appreciation for trees. Hiking and rock climbing through the Appalachian forests is how I spend most weekends.
Experience
Forest Carbon GIS Analysis, Anew Climate
- Deliver the spatial analysis for IFM (Improved Forest Management) carbon projects, using remote sensing, machine learning, statistics, and programming to generate high-quality carbon credits
- Help develop ACR dynamic baselines that improve the accuracy and reliability of IFM projects
Postdoctoral Researcher, Purdue University
- Automated tree species distribution modeling across North America using machine learning and forest inventory data
- Collected and processed inventory data from 60+ countries to build an AI-based forest growth model for estimating carbon sink capacity
Ph.D., Forestry and Natural Resources, Purdue University
- Developed and applied data-driven methods for classifying and mapping forest types across North America
- Used machine learning classification to map planted and natural forest distribution in East Asia
- Demonstrated that forest types shift their ranges differently than their constituent tree species do
M.S., Forestry and Natural Resources, Purdue University
- Designed and led an empirical study on soil greenhouse gas flux in agricultural land
- Led a meta-analysis on soil responses to manipulated precipitation changes
B.S., Department of Bioenvironmental Sciences, Nagoya University
- Studied the fundamentals of forest ecosystems and environmental data
- Completed an exchange program at Southern Illinois University Carbondale
Publications
Mycorrhizal symbioses and tree diversity in global forest communities
This paper used global forest inventory data to show that ectomycorrhizal tree dominance is linked to lower local tree diversity in warm, moist regions, but not at high latitudes or in arid climates.
Forest types outpaced tree species in centroid-based range shifts under global change
This study found that forest types are shifting their geographic ranges roughly three times faster than their individual tree species (86.5 vs. 28.8 km per decade), driven by changing species composition within each forest type.
Mapping planted forests in the Korean Peninsula using artificial intelligence
The study compared deep learning (UNet) and machine learning (Random Forest) models for mapping planted forests across North and South Korea from Sentinel-2 imagery, each trading off differently between recall and precision.
Spatial database of planted forests in East Asia
This paper built the first spatial database of planted forests across East Asia — about 949,000 km², 87% of it in China — with 95% accuracy, to help quantify their role in climate change mitigation.
Dynamics of forest ecosystems under global change: Applications of artificial intelligence in mapping, classification, and projection
This dissertation applied AI to classify, map, and project forest types across North America and East Asia — the work behind several of the publications on this page.
Co-limitation towards lower latitudes shapes global forest diversity gradients
The study mapped global tree species richness at high resolution from ~1.3 million forest plots, showing that temperature drives the overall diversity gradient, while soil, topography, and land use explain why the tropics are especially rich.
Spatiotemporal trends of black walnut forest stocking under climate change
This research trained machine learning models on 1.4 million+ tree records to project how climate change will reshape black walnut stocking by 2080 — gains in the northern part of its range, declines in the south.
Reviews and syntheses: Soil responses to manipulated precipitation changes – an assessment of meta-analyses
This review synthesized 16 meta-analyses covering 42 variables, finding that belowground carbon and nitrogen cycling speed up under wetter conditions and slow under drier ones, while soil microbial communities stay relatively resilient.
Potential of biochar to mitigate the effects of increased precipitation variability on soil greenhouse gas emissions and soybean growth
This thesis found that biochar reduced CO2 and N2O emissions and boosted soybean yield under normal rainfall, but increased CO2 emissions once precipitation became more erratic.