Exciting news! We're transitioning to the Statewide California Earthquake Center. Our new website is under construction, but we'll continue using this website for SCEC business in the meantime. We're also archiving the Southern Center site to preserve its rich history. A new and improved platform is coming soon!

Monitoring Fault Creep on the Hayward Fault using Structure from Motion

Jerlyn L. Swiatlowski, & Gareth J. Funning

Published August 14, 2018, SCEC Contribution #8509, 2018 SCEC Annual Meeting Poster #117

Fault creep is an ongoing process that can be a problem for urban areas that lie along the fault trace. A prime example is Fremont, CA, on the creeping Hayward Fault, where the mean creep rate is ~6 mm/yr. Expressions of fault creep can be seen through offsets in man-made features such as fences, sidewalks, and buildings. Of these features, sidewalk offsets become more visible over the years due to fault movement. Since most sidewalk offsets are along public streets, they can be recorded photographically for years. Monitoring sidewalk offsets allows us a to measure the local creep rate of the Hayward fault and monitor how it is affecting the surrounding urban areas. Here we present creep rate results using photogrammetry to observe and measure creep from multiple sidewalks along the Hayward fault in Fremont.

Since 2009, we have been measuring the creep rate in Fremont, using digital cameras. Originally, we took photos with the camera set up horizontally on a tripod at a fixed height above, and centered over, the offset feature. In 2015, we moved to using Structure from Motion (SfM), a photogrammetric technique that uses multiple 2D images to create 3D point clouds. The main requirements for this technique is to have a fixed focal length camera, take an even distribution of photos around the sidewalk, and have a known reference scale within each photo. Using the Agisoft PhotoScan software, we produce 3D point clouds of each offset sidewalk, using a reference object photographed along with the sidewalk for scale. This newer approach eliminates the need to align and level the camera, and reduces the total time to take photos at each site.

We have 3 years of photographs, and point clouds derived from them, for 36 sidewalk sites in Fremont. We use the Iterative Closest Point (ICP) method to compare point clouds from the same sidewalk from different years. First, we roughly align the sidewalks and crop out the same are. Then we align the point clouds using only points from one side of the offset, by identifying as many features as possible that are the same between each set of photographs and marking them as anchors to fine tune the alignment. The final step is then to estimate the displacement between the two point clouds on the side that was not aligned – the result of creep during the time interval. Preliminary results show that measuring displacement using SfM and ICP is possible, and we can resolve creep at rates as low as 3 mm/yr using the method.

Key Words
Fault creep, Photogrammetry, Structure from Motion, Hayward Fault

Citation
Swiatlowski, J. L., & Funning, G. J. (2018, 08). Monitoring Fault Creep on the Hayward Fault using Structure from Motion. Poster Presentation at 2018 SCEC Annual Meeting.


Related Projects & Working Groups
Tectonic Geodesy