The Department of Statistics at Texas A&M University hosted the 2025 EnviBayes Workshop on Complex Environmental Data in October, bringing to campus the research and ideas of a distinguished group of scholars who study environmental Bayesian statistics. Fifty-four participants from major U.S. national laboratories and leading research institutions attended the workshop.
With three plenary talks, 21 invited presentations and contributions from the 2023 and 2024 EnviBayes Student Paper Award winners, a platform was created for these scholars to discuss emerging challenges in the field. “Sessions tackled cutting-edge and next-generation challenges of environmental Bayesian statistics, from scalable Bayesian methods and interpretable deep learning for massive remote sensing data to advanced modeling for air pollution and complex environmental simulations,” said Dr. Rajarshi Guhaniyogi, a Texas A&M statistics professor.
The talks highlighted the rapid evolution of environmental statistics and its increasing importance in areas such as climate research, pollution monitoring and ecological forecasting.
A workshop highlight was the poster session, a showcase of work from researchers representing institutions from across the country, such as the University of Texas at Austin, Colorado State University, Johns Hopkins University, Penn State University, the University of Connecticut and Texas A&M. As a result, the workshop fostered a sense of community and mentorship, prompting active discussions among established faculty, postdoctoral scholars and graduate students.
Texas A&M graduate student Jacob Andros also participated in the workshop, contributing as both a presenter and a student researcher. Andros’ research focuses on developing efficient and scalable statistical models for large, complex spatial datasets, particularly those that do not meet the assumptions of traditional regression methods. While he did not present a poster, Andros delivered a conference talk on distributed computing for Gaussian process regression and its application to modeling storm surge data.
Andros said he attended the workshop because much of his research intersects with environmental data and large-scale spatial modeling. “It allowed me to see more examples of complex environmental datasets that others are dealing with in their research,” he said. He added that the workshop helped clarify priorities in the field and helped him with ideas for extending his own work. Although he spent much of the conference assisting with logistics, Andros said the poster session stood out. “It primarily showcased research from other students,” he said, noting that student presentations were often easier to follow and encouraged more direct, meaningful conversations.
In keeping its commitment to supporting the next generation of statisticians, the workshop also provided partial financial support to all registered graduate students to ensure access for emerging scholars and strengthening opportunities for professional development.
The Organizer: Dr. Rajarshi Guhaniyogi
Guhaniyogi’s research spans Bayesian parametric and nonparametric methodology, high-dimensional environmental data modeling, deep learning, tensor regression, spatial and spatio-temporal analysis, and distributed Bayesian inference. He has active projects in environmental science, neuroscience, and large-scale machine-learning applications.
Reflecting on the workshop, Guhaniyogi said the event achieved its goal of bringing together researchers working with both environmental challenges and Bayesian statistics. “This year’s meeting successfully attracted many international leaders in the field,” he said, with participants from major national laboratories such as Los Alamos, Sandia and Lawrence Berkeley, as well as academic institutions including the University of California Irvine, Colorado State University, Johns Hopkins University, Penn State University, and the University of Missouri.
He added that the workshop’s structure successfully fostered meaningful engagement. “To ensure meaningful interaction, I maintained an optimal conference size and avoided parallel sessions,” Guhaniyogi said. “I believe these deliberate efforts allowed us to fully accomplish the major aims of the conference.”
Guhaniyogi was proud of the informal atmosphere they created with the poster sessions as a standout part of the event. “This kind of engaged dialogue is precisely what workshops like this should aim to provide,” he said, believing that the format encouraged collaboration and exchanging ideas across career stages.
His research has been supported by the National Institutes of Health, the National Science Foundation, the Department of Energy and the Office of Naval Research. His commitment to organizing EnviBayes 2025 reflects his ongoing mission to advance innovative Bayesian methodology and foster community among environmental statisticians at all career stages.
EnviBayes continues as the flagship conference of the Environmental Bayesian Society, under the International Society for Bayesian Analysis, one of the world’s largest statistical associations. The conference is held every two years, with host institutions selected through a competitive proposal process. Colorado State University hosted EnviBayes in 2023, but the location for the 2027 conference has not yet been announced. Students interested in environmental Bayesian statistics, conference participation or future hosting opportunities should monitor updates through the International Society for Bayesian Analysis and the American Statistical Association, where official announcements and opportunities will be posted.