"Integrating machine learning into large-scale aerial and satellite remote sensing to observe and understand changing coasts"
USGS Pacific Coastal and Marine Science Center
Dan has a PhD in nearshore oceanography and 15 years’ experience of coastal and river research. Since 2020, Dan has been contracting for the USGS Pacific Coastal and Marine Science Center in Santa Cruz, California. His work principally involves making novel measurements from imagery, for monitoring coastal change. For the last 10 years, he has utilized machine learning more and more in this effort, and now leads various software development projects centered around data labeling, model training and the integration of trained models into geospatial and mapping workflows. Prior to rejoining the USGS in 2020, he lived and worked in Arizona for 8 years, including stints as a Research Professor at Northern Arizona University, and a Research Geologist at the USGS Grand Canyon Monitoring and Research Center. During that time, his focus was on making novel measurements of aquatic environments using sonar for applications in ecology and sediment monitoring, and he still keeps a toe in that underwater world.
The world’s coastlines are spatially highly variable, coupled-human-natural systems that comprise a nested hierarchy of component landforms, ecosystems, and human interventions, each interacting over a range of space and time scales. Understanding and predicting coastline dynamics necessitates frequent observation from imaging sensors on remote sensing platforms. Machine Learning models that carry out supervised (i.e., human-guided) pixel-based classification, or image segmentation, have transformative applications in spatio-temporal mapping of dynamic environments, including transient coastal landforms, sediments, habitats, waterbodies, and water flows. This talk will summarize recent and ongoing efforts by the U.S. Geological Survey “Remote Sensing Coastal Change” project to effectively combine aerial and satellite remote sensing and Machine Learning-based image segmentation to observe and understand changing coasts, using imagery collected at a variety of scales. This work includes the development of the “Doodleverse”, a growing set of software tools, models, and data sets, that can be repeatedly repurposed for different measurement targets in the coastal zone, and well beyond.