← Back to Sessions

Analysis of Shakemap Residuals for Spatially Variable Site Terms

Description: 

The U.S. Geological Survey (USGS) has produced and archived ShakeMaps for many decades as part of its earthquake monitoring efforts. The resulting ShakeMap Atlas contains thousands of recent and historical maps whose primary purpose is for calibrating the Prompt Assessment of Global Earthquakes for Response (PAGER) loss models. Here we develop spatially variable site terms (SVST), which could be considered a “partially non-ergodic” ground motion model (GMM) adjustment. Currently, ShakeMap assumes a constant spatial correlation structure across all events and regions. Given the global extent and data quantity available from ShakeMap, we investigate this assumption by looking for trends in the spatial correlation of model parameters with variables such as magnitude, tectonic environment, number of available records for a given station, and geographic region. Adding the SVST adjustment will incorporate repeatable variability into scenarios to capture physical phenomena traditionally unaccounted for by GMMs that underpin the ShakeMap interpolation method. Because real-time ShakeMaps are conditioned on recorded data for a given event, the impact of SVSTs will be less substantial, especially in areas with dense network coverage. However, these empirical factors will improve extrapolation between stations due to larger correlation lengths of the repeatable site effects. The “Did You Feel It?” (DYFI) system responses often have much greater spatial coverage than instrumental data for a large percentage of earthquakes yet take time to become available. In contrast, the SVSTs that we derive here will be available for the first version of ShakeMap. Further, DYFI responses can exhibit areas of missing data in which the most highly-damaged area is not covered because residents in these regions may be more concerned with safety than providing DYFI responses, and/or internet connectivity could be lost. These innovative SVST factors would be able to help fill in such regions.

Session: Regional-Scale Hazard, Risk and Loss Assessments - I

Type: Oral

Date: 5/2/2024

Presentation Time: 04:45 PM (local time)

Presenting Author: Annabelle

Student Presenter: Yes

Invited Presentation: 

Authors