当前位置: X-MOL 学术Seismol. Res. Lett. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
lsforce: A Python‐Based Single‐Force Seismic Inversion Framework for Massive Landslides
Seismological Research Letters ( IF 3.3 ) Pub Date : 2021-07-01 , DOI: 10.1785/0220210004
Liam Toney 1, 2 , Kate E. Allstadt 1
Affiliation  

We present an open‐source Python package, lsforce, for performing single‐force source inversions of long‐period (tens to hundreds of seconds) seismic signals. Although the software is designed primarily for landslides, it can be used for any single‐force seismic source. The package allows users to produce estimates of the three‐component time series of forces exerted on the Earth by a landslide with postprocessing options to estimate the trajectory of its center of mass. Green’s functions for a user‐selected 1D Earth model are obtained automatically from the Incorporated Research Institutions for Seismology Synthetics Engine webservice or can be computed for custom 1D Earth models using Computer Programs in Seismology. lsforce implements the two most commonly used source parameterizations: a fully flexible, high‐resolution approach and a more stable but lower‐resolution method of overlapping triangle sources. Regularization options include a blended zeroth‐, first‐, and second‐order semiautomated Tikhonov regularization scheme, as well as additional optional constraints on start times, end times, and on the sum of forces. Uncertainty due to data selection can be assessed using either a leave‐one‐out approach or a modified jackknife technique that randomly excludes subsets of the data for multiple re‐inversions. Numerous built‐in plotting methods allow for easy quality control and assessment of results. In this article, we briefly outline the theory and methodology, describe our implementation, and demonstrate the usage of lsforce using the well‐studied 28 June 2016 Lamplugh rock avalanche in Alaska. Despite the rapidly increasing prevalence of landslide single‐force inversions in the landslide and seismology literature over the past decade, to our knowledge this is the first open‐source code for performing such inversions.

中文翻译:

lsforce:基于 Python 的大规模滑坡单力地震反演框架

我们提出了一个开源 Python 包 lsforce,用于执行长周期(数十到数百秒)地震信号的单力源反演。尽管该软件主要是为滑坡设计的,但它也可用于任何单力地震源。该软件包允许用户通过后处理选项来估计滑坡对地球施加的力的三分量时间序列的估计值,以估计其质心的轨迹。用户选择的一维地球模型的格林函数是从 Incorporated Research Institutions for Seismology Synthetics Engine webservice 自动获得的,或者可以使用地震学中的计算机程序为自定义一维地球模型计算。lsforce 实现了两个最常用的源参数化:一个完全灵活的,高分辨率方法和重叠三角形源的更稳定但分辨率较低的方法。正则化选项包括混合的零阶、一阶和二阶半自动 Tikhonov 正则化方案,以及对开始时间、结束时间和力总和的附加可选约束。由于数据选择的不确定性可以使用留一法或改进的折刀技术进行评估,该技术随机排除数据的子集以进行多次反演。许多内置的绘图方法可以轻松控制质量和评估结果。在本文中,我们简要概述了理论和方法,描述了我们的实现,并使用经过充分研究的 2016 年 6 月 28 日阿拉斯加的 Lamplugh 岩石雪崩演示了 lsforce 的使用。
更新日期:2021-06-28
down
wechat
bug