Automating measurement of extensional fault/fold scarp geometry and resurfacing in Reykjanes, Iceland: Implications for fault scaling and growth

Madeline F. Schwarz, Karl J. Mueller, & Stephen D. Sheehan

Published September 11, 2022, SCEC Contribution #12457, 2022 SCEC Annual Meeting Poster #104

Evaluating the slip distribution along individual faults informs fault displacement hazard analyses by constraining the propagation history of active extensional fault systems. The abundance of high-resolution topographic data enables widespread mapping and measuring of fault extents and vertical separations. We evaluated different signal processing, remote sensing, and machine learning techniques to develop a standardized, versatile workflow for automating measurement and characterization of active fault and fold scarps from high-resolution topographic data. Our method utilizes weighted-overlay preprocessing schemes coupled with unsupervised image segmentation and classification algorithms to detect and extract fault scarps from 2-meter digital elevation data into vectorized fault databases. Slip profiles were generated from algorithmically-defined scarp extents. We applied the approach to mapped fault scarp segments within the Vogar fissure swarm in southwest Iceland. The algorithm correctly identified 48 fault scarp segments out of 58 extracted features. We compared the automated results to 55 hand-drawn scarps using the scarp-extent surface area. Results demonstrate an algorithm accuracy of ~76%, a misclassification rate of ~42%, and a percent error of 29%. We conducted slip-distribution analyses on 32 algorithmically-defined scarps. Slip profiles show a predominance of plateau-shaped distributions as well as evidence of fault linkage. Weighted-overlay analysis can effectively extract geological structures from topographic data and unsupervised algorithms are capable of distinguishing continuous, obvious fault scarps from other landforms. Algorithm-defined scarp widths and scaling profile shapes indicate normal fault growth is influenced by trishear and cover thicknesses from resurfacing events. Pairing these surface measurements with subsurface models furthers our understanding of fault and fold growth in three dimensions.

Key Words
Automated fault mapping, Slip distribution

Citation
Schwarz, M. F., Mueller, K. J., & Sheehan, S. D. (2022, 09). Automating measurement of extensional fault/fold scarp geometry and resurfacing in Reykjanes, Iceland: Implications for fault scaling and growth. Poster Presentation at 2022 SCEC Annual Meeting.

Related Projects & Working Groups
Earthquake Geology