B4 2 After, 3D Deformation Field From Matching Pre- To Post-Event Aerial LiDAR Point Clouds, The 2010 El Mayor-Cucapah M7.2 Earthquake Case
Abstract
Aerial LiDAR surveys reconstruct with amazing fidelity the sinuosity of terrain relief. In this research we explore the 3D deformation field at the surface after a big earthquake (M7.2) by comparing pre- to post-event aerial LiDAR point clouds. The April 4 2010 earthquake produced a NW-SE surface rupture ~110km long with right-lateral normal slip up to 3m in magnitude over a very favorable target: scarcely vegetated and unaltered desert mountain range, sierras El Mayor and Cucapah, in northern Baja California, close to the US-México border. It is a plate boundary region between the Pacific and North American plates. The pre-event LiDAR with lower point density (0.013-0.033 pts m-2) required filtering and post-processing before comparing with the denser (9-18 pts m-2) more accurate post event dataset. The 3D surface displacement field was determined using an adaptation of the Iterative Closest Point (ICP) algorithm, implemented in the open source Point Cloud Library (PCL). The LiDAR datasets are first split into a grid of windows, and for each one, ICP iteratively converges on the rigid body transformation (comprising translations and rotations) that best aligns the pre- to post-event points. Perturbing the pre- and post-event point clouds independently with a synthetic right lateral inverse displacements of known magnitude along a proposed fault, ICP recovered the synthetically introduced translations. Windows with dimensions of 100-200m gave the best results for datasets with these densities. The simplified surface rupture photo interpreted and mapped in the field, delineates very well the vertical displacements patterns unveiled by ICP. The method revealed block rotations, some with clockwise and others counter clockwise direction along the simplified surface rupture. As ground truth, displacements from ICP have similar values as those measured in the field along the main rupture by Fletcher and collaborators. The vertical component was better estimated than the horizontal having the latter problems in flat areas as expected. Hybrid approaches, as simple differencing, could be taken in these areas. Outliers were removed from results. ICP detected extraction from quarries developed between the two dates of LiDAR collection and expressed as a negative vertical displacement close to the sites. To improve the accuracy of the 3D displacement field, we intend to reprocess the pre-event source survey data to reduce the systematic error introduced by the sensor. Multidisciplinary approach will be needed to make tectonic inferences from the 3D displacement field revealed by ICP, about the processes at depth expressed at surface.
- Publication:
-
AGU Spring Meeting Abstracts
- Pub Date:
- May 2013
- Bibcode:
- 2013AGUSM.G31A..03H
- Keywords:
-
- 1209 GEODESY AND GRAVITY / Tectonic deformation;
- 1988 INFORMATICS / Temporal analysis and representation;
- 1994 INFORMATICS / Visualization and portrayal