The orientations of planar rock layers are critical parameters for understanding the response of rocks to deformation and tectonic stress. These data are typically measured in-situ using a structural compass or digital clinometer. However, recent advances in remote-sensing approaches such as satellite imagery, LiDAR scanners, and uncrewed aerial vehicles allow the generation of 3D digital surface models from which orientations can be extracted. Additionally, new edge-detection capabilities potentially support the automatic extraction of candidate layers from imagery, suggesting that automated synthesis of bedding orientations may be close at hand.

Extracting bedding orientations from 3D input data is straightforward but their uncertainty is subject to errors in input data, viewing geometry, and the plane-fitting process that can be hard to model jointly. In Quinn and Ehlmann (Earth and Space Science, 2019) we presented a PCA-based plane fitting method that is tuned to be resilient to errors in all dimensions and produces a closed-form error distribution that can be parameterized in spherical coordinates and plotted on a stereonet. This approach is scale-independent and has been used to establish bedding orientation uncertainty in orbital and rover-based research on Mars, but its usage for structural modeling on Earth is underexplored.

Here, we present a scalable workflow for collecting orientation data for comparison with ground-based measurements. The Mapboard GIS iPad app is used to trace candidate beds atop Mapbox commercial satellite imagery and DEMs. 3D bedding traces are extracted for orientation modeling and stored in a PostgreSQL database. Bedding orientations from the Macrostrat, Rockd, and StraboSpot platforms (which derive from published geologic maps and user-contributed observations) are targeted for comparison. The results show good correspondence between in-situ and remotely measured orientations, even using commercial DEMs with > 1 m/pixel resolution.

This approach exhibits the growing alignment between remote sensing capabilities, analytical tools, and geological data synthesis platforms. Implementation at regional scale allows validation against standardized imagery, elevation, and structural orientation datasets. Moving forward, similar extraction approaches could be applied to more localized datasets (e.g., UAV-based photogrammetry or LiDAR scans) or paired with layer detection algorithms. Large, easy-to-extract orientation datasets would simplify structural modeling and interpretation of remotely-sensed geological models from outcrop to regional scale.