← Back to Sessions

Estimation of Ground Motion Through Deconvolution-linear Scaling-convolution Method

Description: 

Ground motions can be estimated at sites using nearby reference station records. However, the estimated records have uncertainties due to differences of site and path conditions between source and target sites. To resolve this problem, this study evaluated the ground motions estimated using the deconvolution-linear scaling-convolution method (DLSC). The DLSC method consists of three steps: (1) determine the bedrock motion from the surface motion by performing deconvolution which removes the influence of local site effect; (2) adjust the difference of the amplitude of bedrock motion between source and target sites based on the rate using a ground motion model (GMM); (3) convolve for the surface motion. To validate this approach, we collected records from 14 local events with ML ≥ 3.5 and site information at stations. Then, we evaluated the root mean square error (RMSE) between the recorded and estimated PGA values, and compared the DLSC method result with those of two other methods (referred to as Model A and B). Model A calculates PGA using the region-specific ground motion model (GMM), whereas Model B uses the GMM and interpolation of measured ground motions to calculate PGA at a site. We found that the DLSC method is effective (less RMSE than Model A and B) for the distance range less than 78 km if source station is surface-type and 84 km if source station is borehole-type. While the effective distance would change region by region, we expect that the suggested DLSC method is effective to estimate a ground motion time history at a target site nearby the seismic station.

Session: Numerical Modeling in Seismology: Theory, Algorithms and Applications [Poster]

Type: Poster

Date: 4/17/2025

Presentation Time: 08:00 AM (local time)

Presenting Author: Minyoung

Student Presenter: Yes

Invited Presentation: 

Poster Number: 122

Authors