Defense: "Magali: An open tool for the magnetic microscopy community"

Data

Horário de início

14:00

Local

Sala de aula P205 - Bloco Principal - IAG/USP

Defense

Student: Yago Moreira Castro
Program: Geophysics
Title: Magali: An open tool for the magnetic microscopy community

Advisor: Prof. Dr. Leonardo Uieda - IAG/USP

Judging Comitee:

  1. Prof. Dr. Leonardo Uieda – Presidente e Orientador - IAG/USP
  2. Prof. Dr. André Luis Albuquerque dos Reis – ON/MCTI
  3. Prof. Dr. Gelvam André Hartmann - Unicamp

Other Members:

  1. Prof. Dr. Filipe Altoé Temporim - UFOP
  2. Profa. Dra. Susanne Taina Ramalho Maciel - UnB
  3. Prof. Dr. Carlos Alberto Mendonça - IAG/USP

Abstract: 

Paleomagnetic studies rely on the magnetization preserved in ferromagnetic minerals to reconstruct the Earth’s ancient magnetic field. While bulk rock measurements have long been the standard, advances in magnetic microscopy now enable the investigation of magnetization at the scale of individual grains. These methods can provide more accurate paleomagnetic information by resolving the contributions of stable single-domain and pseudosingle-domain particles, but their adoption has been hindered by the limited availability of accessible open-source software and standardized data formats. We present Magali, a Python library designed to provide robust, open, and reproducible tools for magnetic microscopy inversion. The package builds upon a novel workflow for source detection, position estimation, and dipole moment inversion, enhancing the methodology with a Levenberg-Marquardt nonlinear inversion scheme. This scheme leverages gradient information to improve convergence and reliability compared to earlier derivative-free inversion methods, refining both source positions and dipole moments. By integrating established image-processing techniques with efficient linear and nonlinear inversion routines, Magali enables the semi-automatic estimation of dipole source parameters from Quantum Diamond Microscope (QDM) data and other magnetic microscopy datasets. Its fully opensource implementation, adherence to FAIR data principles, and emphasis on reproducibility will provide the paleomagnetic community with a flexible and accessible platform to advance grain-scale magnetic studies.

 

Keywords: magnetic microscopy, paleomagnetism, open-source, Python