← Back to Gallery

Automated Framework for Remote Sensing-based Rupture Slip Profile Processing: An Application to the M7.7 Mandalay, Myanmar, Earthquake Surface Rupture

We present an automated processing framework for extracting key parameters from surface rupture slip profiles, including the 3-component displacement vector, rupture width, and number of individual ruptures. This approach significantly reduces processing time compared to existing methods. Current models often cannot fully characterize fault ruptures—some do not estimate fault zone width, while others fail to accurately represent displacement. Improving the efficiency and completeness of this extraction is vital for both Probabilistic Fault Displacement Hazard Analysis and fault characterization. Moreover, existing methods typically rely on manual processing of each fault profile, which limits scalability and often requires dataset downsampling.

Our automated process addresses these limitations using a flexible functional form that models both the rupture offset and the linear segments. This form allows for fault-normal, fault-parallel, and vertical displacements while honoring constraints such as common rupture location and width across components. We fit this idealized profile shape to observed data using a stochastic gradient descent algorithm that optimizes the functional parameters to minimize misfit. To improve model convergence, we iteratively freeze and unfreeze parameter subsets and explore many initial values sets. Additionally, the method detects and accommodates multiple rupture offsets within a single profile by summing the functional form, yielding a more accurate characterization. The primary outputs of the model include fault displacement, rupture width, rupture center location, and slope characteristics.

We applied this approach to the M7.7 Mandalay, Myanmar, earthquake, using pixel correlation of imagery from Sentinel-2, Landsat-8, and Landsat-9 satellites. Preliminary results show that the automated method produces displacement estimates consistent with those from manual analysis, while significantly reducing processing time. Ongoing efforts focus on fully automating the characterization of displacement uncertainty and generating complete uncertainty distributions for displacement parameters.

Session: Late-breaking on Recent and Future Large Earthquakes [Poster]

Type: Poster

Room: Exhibit Hall

Date: 4/15/2025

Presentation Time: 08:00 AM (local time)

Presenting Author: Sravani Boggaram

Student Presenter: No

Invited Presentation: 

Poster Number: 154

Additional Authors