Andres Schmidt

Andres Schmidt is an Assistant Professor developing machine learning approaches for environmental challenges, including wildfire risk assessment and carbon cycle modeling. His research integrates process-based models with neural networks to analyze biogeophysical processes.

Core methodologies:

• Bayesian-regularized neural networks for environmental prediction
• Geostatistical clustering for regional risk assessment
• Deep learning for post-fire recovery tracking
• Carbon flux optimization using observational networks

Recent work (2016-2025) shows strong focus on wildfire applications (7 papers) and carbon-water cycle modeling (5 papers). Technical innovations include Bayesian-regularized ensembles for water contamination prediction, modular geostatistical frameworks for risk assessment, and convolutional networks for infrastructure recovery monitoring.

Data integration: Leverages flux tower networks, aircraft campaigns, satellite imagery, and sensor arrays to train models across scales from Oregon forests to Indonesian peatlands.