Virtual Gravel Counting for Full River Surveys

Authors: Mark Fonstad*, University of Oregon, Aaron Zettler-Mann, University of Oregon
Topics: Geomorphology, Remote Sensing, Water Resources and Hydrology
Keywords: fluvial, rivers, sediment, big data, UAS, geomorphology
Session Type: Guided Poster
Day: 4/5/2019
Start / End Time: 5:00 PM / 6:40 PM
Room: Roosevelt 3.5, Marriott, Exhibition Level
Presentation File: No File Uploaded


A major practice in river studies is the survey of particle sizes and linking these to causal factors, as well as linking the size measurements to river competence and habitats. Pragmatic difficulties in measuring particle size distributions spatially has led to sparse sampling (though of high precision) within rivers, leading to fluvial theory with mediocre understanding of the spatial variations in sediment sizes other than the common downstream-fining patterns exhibited over long distances. We offer a new mapping approach for gravel-and cobble-bedded rivers that maintains much of the sample precision but embraces spatial complexity. While floating a river, we fly a UAS over the exposed bars, collecting high resolution imagery. These images are then processed into orthomosaics. The results then run through BASEGRAIN, an object detection software tool for sediment analysis to extract information on size, shape, location, and orientation for each rock. We test this approach in the gravel-dominated Sandy River of Oregon. While not yet fully automated, the approach already provides an increase of data by several orders of magnitude over traditional ground measurements. Our test of the approach displays the typical downstream fining pattern common in many rivers, but also reveals a much more complex set of smaller area particle zones. Flight altitude choice is of major importance in producing good results. Even in its infancy, this geomorphic mapping approach appears to be quite successful and transferrable to many river environments.

Abstract Information

This abstract is already part of a session. View the session here.

To access contact information login