RTS InTrain: Retrogressive thaw slump inventory and machine learning training-data development

RTS InTrain aims to bring the permafrost science community together to address the challenges and opportunities in machine learning (ML) training data development for remote sensing (RS) analyses. Specifically, we aim to identify the resources, common interests and needs required to apply RS-ML approaches to identify and characterize retrogressive thaw slumps (RTS). The action group will address multiple objectives of the IPA by developing training data creation protocols, pooling and standardize existing training datasets (incl. addressing the time-dependent nature of RTS), and expanding cyberinfrastructure to make data accessible and discoverable for the science community and the public. The long-term products from the working group will allow for pan-Arctic assessments of permafrost thaw at spatial and temporal scales relevant to society. This will allow stakeholders to identify and mitigate permafrost-related problems. We will bring together those interested in the data creation, analysis, or application of the information at regular intervals to share knowledge and exchange experiences.

Contact: Anna Liljedahl (aliljedahl@woodwellclimate.org)