Scholarlink

<p>We developed an inversion program to apply the finite volume method (FVM) to consider the topography. In the 3D magnetotelluric (MT) forward modeling, we must consider the effect of the topography because it appears as a distortion of the data. FVM is a numerical analysis method that can consider topography. We adopted FVM and the H-forming equation as a governing equation for efficient and mathematically plain implementation. We also adopted the adjoint state method (ASM) to update model parameters in the inversion. ASM can easily and efficiently calculate the gradient vector of the objective function for the inversion without calculating the Jacobi matrix, which needs large memory and computation time. We verified the program using test models. In the forward modeling, it was confirmed that the program could perform the analysis efficiently and adequately considering topography and subsurface resistivity heterogeneity. Furthermore, we conducted the inversion using a synthetic open dataset and could reconstruct the true model of the dataset. The comparison with the previous work revealed that the computation time was comparable even though we calculated it with a personal computer.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Development of a three-dimensional magnetotelluric inversion program considering topography with cell-centered finite volume method

  • 期刊论文

  • 发表日期：

    2025年1月15日

查看全文
  • Atsushi Suzuki

摘要

We developed an inversion program to apply the finite volume method (FVM) to consider the topography. In the 3D magnetotelluric (MT) forward modeling, we must consider the effect of the topography because it appears as a distortion of the data. FVM is a numerical analysis method that can consider topography. We adopted FVM and the H-forming equation as a governing equation for efficient and mathematically plain implementation. We also adopted the adjoint state method (ASM) to update model parameters in the inversion. ASM can easily and efficiently calculate the gradient vector of the objective function for the inversion without calculating the Jacobi matrix, which needs large memory and computation time. We verified the program using test models. In the forward modeling, it was confirmed that the program could perform the analysis efficiently and adequately considering topography and subsurface resistivity heterogeneity. Furthermore, we conducted the inversion using a synthetic open dataset and could reconstruct the true model of the dataset. The comparison with the previous work revealed that the computation time was comparable even though we calculated it with a personal computer.