Visualization Tools to Communicate Riverine Erosion Hazards and Improve Flood Resiliency in Headwater Communities of the Lake Champlain Basin


February 1, 2019 to August 31, 2022


Investigators will conduct research and community outreach that will engage rural, upland communities in river corridor management practices that provide local benefits of reduced fluvial erosion, improved habitat, and reduced nutrient loading to Lake Champlain. Researchers will use the SOM (Self-Organizing Map), a machine learning algorithm that has proven helpful in making sense of sediment transport data in southern and central Vermont. Data can be more easily stored, laid out, and computed in such a way to help understand the current state of many Lake Champlain basin waterways. Research will focus on the Missisquoi River basin and the Winooski River basin, tributaries of the Lake Champlain basin.

Proposed work will automate the classification of river reaches in terms of their sensitivity to adjustment, rely on existing stream geomorphic assessment data, and integrate with an existing Geographic Information System (GIS) database. Scientists will evaluate channel-based stream power over a range of peak discharges, develop a characteristic stream power signature for reaches of various floodplain connectivity conditions, and examine data for relationships between stream power signatures and sediment regime classes. They will transfer data collected to more widely accessible state databases and technology to Vermont Agency of Natural Resources staff. Resulting visualizations can be used during outreach to communities and stakeholders to engage citizens, influence land use and river corridor planning, and improve flood resiliency.


Kristen Underwood
University of Vermont
Kristen.Underwood [at]

Mike Kline
University of Vermont
mike.kline [at]

Beverley Wemple
University of Vermont
Beverley.Wemple [at]

Donna Rizzo
University of Vermont
drizzo [at]

Resulting Publications

Analysis of Reach-scale Sediment Process Domains in Glacially-Conditioned Catchments Using Self-Organizing Maps

Published 2021
This Lake Champlain Sea Grant-funded research classified reach-scale sediment process domains using a Self-Organizing Map (SOM), identified 15 variables that explained 7 classes for 193 reaches in the Northeast US, and improved previous classifications by including degree of floodplain disconnection. SOM visualization tools provided insight into channel evolution processes and sediment regime classification framework supports adaptive river management.