Rate My Professor Matthias Katzfuss

Matthias Katzfuss is a Professor in the Department of Statistics at the University of Wisconsin–Madison, a position he assumed in 2023 following a decade on the faculty at Texas A&M University. He earned his PhD in Statistics from The Ohio State University in 2011 after completing undergraduate studies in Statistics at the University of Munich, Germany. Katzfuss's research specializes in computational spatial and spatio-temporal statistics, Gaussian processes, uncertainty quantification, probabilistic machine learning, and data assimilation. His methodologies tackle massive datasets in environmental science, satellite remote sensing, climate modeling, numerical weather prediction, tornado forecasting, agriculture, and pollutant monitoring. Key innovations include Vecchia approximations for Gaussian processes, sparse inverse Cholesky factorizations, Bayesian transport maps for non-Gaussian fields, and scalable generative modeling techniques. These advancements support interdisciplinary applications, notably a 15-year NASA collaboration developing statistical software for analyzing massive remote-sensing data, improving atmospheric and surface retrievals from spectra and fusing varying-resolution measurements. His work is funded by NSF, NASA, NOAA, USDA, Sandia National Laboratory, and Jet Propulsion Laboratory.

Katzfuss was elected a Fellow of the American Statistical Association in 2024 for outstanding contributions to computational spatial statistics, environmental applications, NASA partnerships, mentoring, and professional service; he joined the ASA Board of Directors in July 2024. Additional honors include the NSF CAREER Award, Fulbright Scholarship, and Early Investigator Award from the ASA Section on Statistics and the Environment. Representative publications are 'Data-efficient generative modeling of non-Gaussian global climate fields via scalable composite transformations' (2026), 'Fast Gaussian process approximations for autocorrelated data' (2025), 'Learning non-Gaussian spatial distributions via Bayesian transport maps with parametric shrinkage' (2025), 'A General Framework for Vecchia Approximations of Gaussian Processes' (Statistical Science, 2021), 'Correlation-based sparse inverse Cholesky factorization for fast Gaussian-process inference' (2023), and 'Scalable Bayesian optimization using Vecchia approximations of Gaussian processes' (2023). With over 4,985 citations on Google Scholar, Katzfuss's scholarship has substantially shaped geospatial analysis and uncertainty quantification in statistics.