Geospatial Analytics Ph.D. students will share current research in a series of short talks.

Hosted by: Rachel Kasten, CGA’s Student Services Specialist

Student presenters:

> Eli Horner — Quantifying and Partitioning Uncertainty through Space and Time in Ecological Forecasts using the Sobol Method; co-advised by Dr. Chris Jones and Dr. Ross Meentemeyer

Summary: Whenever we make predictions about the future, we need to consider the uncertainty in those predictions. The more we know about where predictive uncertainty is high and about what uncertainty sources contribute most to that uncertainty, the better we can prioritize model development and data collection to make our models better. My work and presentation focuses on quantifying the uncertainty through both space and time for predictions of the spread of sudden oak death in Oregon and then partitioning that uncertainty, that is, breaking it down into its parts to tell us how much of our overall uncertainty comes from each uncertainty source, along with how these patterns change over both space and time.

> Gwen Kirschke — Running a Field Campaign: Challenges and Rewards; advised by Dr. Elsa Youngsteadt

Summary: In this presentation, I will describe a field campaign I ran at the Rocky Mountain Biological Laboratory in summer 2025. During this campaign, I lead a 6-person field team that collected ground cover, floral abundance, and floral resource information in 480 quadrats across 8 sites, concurrent with flights by the NEON Airborne Observation Platform, which collected high-resolution hyperspectral imagery. I am using these field data to train models predicting floral resource abundances throughout the subalpine valley where we worked. In order to test whether resulting floral resource products improve our ability to predict the ecologically relevant measure of bee productivity, an undergraduate researcher and I measured solitary bee nesting rates at 6 sites in the same domain, and collected traditional, transect-based measures of floral resource availability to use as a comparison. I will also speak generally about the process of planning, funding, and conducting field work, and why I have found it to be a valuable part of my research program, despite the challenges.

> Titilayo Tajudeen — Assessing Coastal Forest Retreat under Current and Future Climate and Sea Level Rise Projections; advised by Dr. Katie Martin

Summary: Coastal forests are increasingly threatened by saturated soils and rising salinity due to sea-level rise, saltwater intrusion, and storm surges. In response to increasing salinization and flooding, healthy coastal forests that rely on freshwater are transforming into landscapes dominated by dead or dying trees, bordered by salt-tolerant shrubs and grasses, eventually becoming marshes or open water. Quantifying the rate and pathways of these landscape changes is essential for understanding how coastal ecosystems deteriorate and for predicting future transformations. The speed and direction of wetland shifts will depend on the rate of saltwater intrusion, sea level rise, and factors such as soil type, landforms, and hydrologic connectivity. Although some studies have documented the conversion of these areas, limited research has examined the lateral and vertical retreat rates of forests, and our understanding of the future progression of forest-to-marsh loss is still limited. To improve our understanding of these impacts, I analyzed long-term trends in forest retreat from remote sensing data to determine how quickly the ecosystem is shifting both inland and vertically and the factors contributing to these losses.