Streamlined subglacial bedform sensitivity to bed characteristics across the deglaciated Northern Hemisphere

co-authored by L. Miller, S.M. Principato, and S. Munevar Garcia

In this work, the development of a semi-automated slope identification tool was utilized to identify 11,628 sedimentary and bedrock streamlined subglacial bedforms across 9 sites in the deglaciated Northern Hemisphere. Morphometrics from the streamlined bedforms were correlated to underlying site topography and lithology to identify controls of bed characteristics on bedform development, extrapolated to signify different conditions of warm-based ice flow. Published by Earth Surface Processes and Landforms in 2022. Publication can be found at: All data and semi-automated identification tool is publicly available at:

Differential impact of isolated topographic bumps on glacial ice flow and subglacial processes

co-authored by L. Miller, J. S. Slawson, E.J. Mackie, and S. wang

Topographic highs (“bumps”) across glaciated landscapes have the potential to temporarily slow glacial ice flow or, conversely, increase ice flow through strain heating and subglacial meltwater production. Isolated bumps of variable size across the deglaciated landscape of the Cordilleran Ice Sheet (CIS) of Washington state present an opportunity to assess the influence of topographic highs on ice-bed interactions and ice flow organization. This work utilizes semi-automatic mapping techniques of subglacial bedforms to characterize the morphology of streamlined subglacial bedforms including elongation, surface relief, and orientation – all of which provide insight into subglacial processes during post-Last Glacial Maximum deglaciation of the landscape. Published by The Cryosphere in 2023. Publication can be found at: All data and semi-automated identification tool is publicly available at:

Spatial variability of marine-terminating ice sheet retreat in the Puget Lowland

co-authored by L.M. Miller, A.P. lepp, and r. dewitt

Preserved stratigraphy along the coastal bluffs in the Puget Lowland, Washington state provides the rare opportunity to assess temporal and spatial variability in marine ice-sheet behavior and coupled marine to subaerial landscape evolution. The outcrops record glacial and interglacial sediments that mark ice advance and retreat across the formerly marine landscape, as well as processes and conditions that were in place at the time of deglaciation; therefore, detailed assessments of the stratigraphy and sedimentological analyses elucidate ice-bed interactions at local outcrop scales across the region. Glacial isostatic adjustment (GIA) has played a major role in landscape evolution, the preservation of these outcrops, and potentially in periodically stabilizing the Puget Lobe during its final retreat from south to north into British Columbia as effective water depth was reduced due to rapid solid Earth rebound. Published by Climate of the Past in 2024. Publication may be found at: All data may be found at:

Sedimentary records of blue carbon and environmental change in coastal Virginia: an assessment of seagrass meadow sediment deposition and carbon cycling variations

co-authored by L.M. Miller, P. BERG, R. HEBERT, Z. GUO, T. WIGGINS, S. KUZMINSKI, C. WIMAN, S. MUNOZ

Seagrass meadows are capable of storing large amounts of carbon in their root systems and sediments for several decades. The potential of seagrass meadows to contribute to global carbon offsets makes understanding sediment flux, carbon storage, and disturbance event effects on these coastal ecosystems highly important across broad spatial and temporal scales. However, century-scale resilience of carbon stored by seagrass meadows and shallow sediment in which seagrass grows is not well understood. Following the die-off from disease and a hurricane event in the early 1930s, the seagrass meadows at the Virginia long-term ecological research (LTER) site in coastal Virginia have been the world’s largest seagrass restoration project beginning in the early 2000s and hosts the world’s first verified seagrass blue carbon offset project. The sedimentary records in the seagrass meadows at the Virginia LTER site can elucidate the relationships between physical disturbances across the bays, environmental changes, and carbon storage prior to the early 1930s and following restoration efforts. This work is current in prep for submission to Sedimentology.

Pairing on-and-offshore observations of paleo-ice streams to constrain and elucidate dynamics of the Cordilleran Ice Sheet (PISCES)


The deglacial history of Cordilleran Ice Sheet (CIS) marine margins over the last 50,000 years is poorly constrained despite the potential insights this record may contribute to identifying controls on ice stream stability, elucidating millennial and global-scale climate oscillations, and characterizing paleo-current dynamics across the Pacific ocean. This work proposes to use multi-proxy records to address knowledge gaps of CIS ice streams. LiDAR and bathymetric elevation data will be used to constrain geomorphic expressions of ice streams, while timing and products of ice retreat will be identified through stratigraphic and sedimentological analyses of previously collected cores from Deep Sea Drilling Project (DSDP) cruise 18, site 177. Sediment analysis, absolute radiocarbon dating, and Ramped PyrOx analysis conducted on DSDP 18-177 cores will provide detailed information about timing and transitions in the offshore depositional environment during periods of ice loss. This successful NSF EAR-PF application and accompanying documents are available upon request.

Automatic identification of subglacial streamlined bedforms using Random Forest Modeling and Topographic Position Index programming

Co-authored with e. abrahams

Utilizing data collected from McKenzie et al., 2022, this work aims to fully automate the process of identifying subglacial streamlined bedforms across formerly glaciated landscapes. Nearly 600,000 data points were used to train a Random Forest Model to characterize identified polygons as streamlined subglacial bedforms. This data is trained on features including bed topography, lithology, shape area, and shape elongation. The resulting model is over 97% accurate and the feature with the greatest influence on accuracy is bed lithology. This work is currently being cross-validated and will be tested on till-based streamlined subglacial bedforms in Wisconsin and on crystalline-bedrock forms in British Columbia before preparation for submission to Earth Surface Processes and Landforms for publication.

The Puget Lowland work is conducted on land historically cultivated and inhabited by the Skokomish, Suquamish, Squaxin, Stl’pulmsh, Steilacoom, Puyallup, Muckleshoot, and Duwamish peoples. The majority of the writing and data analysis for these works was conducted on land cultivated and inhabited by the Monacan Nation. The peoples of these Nations were custodians of the land for time immemorial before forced removal and genocide during colonization. I acknowledge their ongoing stewardship of the lands.

My other ongoing projects and interests cover topics in subglacial flux modeling, sedimentology, and other bedform mapping. If any of these mehtods or tools would be of use in your work, I’m always happy to chat and/or collaborate — do not hesitate to reach out to me via email ( if our interests align!