Estimating Fault Surface Creep Using Lidar Time Series Analysis
Abstract
While lidar has been extensively used for estimating surface displacement due to earthquakes, there is a paucity of change detection strategies that fully leverage the redundant and complex format of point clouds, especially in cases where multi-temporal lidar observations are available. Studies of fault mechanics, especially fault slip reduction and shallow slip deficit, benefit from dense near-field observations of ground displacement. Therefore, we have developed a change detection framework that directly estimates surface displacement using a time series of near-field lidar point clouds. The framework resolves a time series of ground displacements with high spatial resolution by tracking augmented planar primitives that persist over repeated lidar surveys. Compared to bi-temporal change detection, the methodology enables greater temporal and spatial consistency within a series of change detection results. We first demonstrate with a synthetic test that the proposed framework provides a higher order of detection precision and consistency compared with common methods in the literature such as iterative closest point (ICP). Then, we examine an actual fault creep time-series for a 2 km segment of the Hayward fault; validation with collocated alinement array measurements suggests sub-centimeter level agreement for fault creep, and the detection reveals 15 ± 5.2 mm, 7.7 ± 8.7 mm, and 21.1 ± 9.7 mm fault parallel displacements at 44 m from the fault trace for periods from 2015-2017, 2017-2018 and 2015-2018 respectively.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2022
- Bibcode:
- 2022AGUFM.T12E0129Z