Curriculum Vitæ

About

The Software Underground is a non-profit professional society for people at the intersection of geoscience and computing. I have been active in the Software Underground community since its inception in 2014 and was very happy to join the Board in 2022.

About

pyOpenSci is a non-profit looking to connect, curate, and improve open-source scientific software in the Python language. I joined their Advisory Council in 2022 to help develop community norms and guidance around tooling and techniques in software engineering.

About

JOSS is a developer friendly, diamond open-access journal for research software packages. I joined the editorial team in 2019 as Topic Editor for Geoscience. I reluctantly left the editorial team in 2022 because of the increased workload brought on by my new admin roles at the University of Liverpool.

About

EarthArXiv is a non-profit preprint server for the Earth Sciences. I joined their Advisory Council in 2019 and participated in discussions about how to setup and improve the platform, as well as in moderating submissions. I left the Council in 2022 when my 3 year term was up.

Thesis: Forward modeling and inversion of gravitational fields in spherical coordinates

Advisor: Valéria C. F. Barbosa

About

After my Master's degree, I stayed at the Observatório Nacional for my PhD, also with Valéria C. F. Barbosa. In 2016, I defended my thesis, which was submitted for publication in 3 parts:

• Modeling the Earth with Fatiando a Terra (2013)
• Tesseroids: forward modeling gravitational fields in spherical coordinates (2016)
• Fast non-linear gravity inversion in spherical coordinates with application to the South American Moho (2017)

During my PhD, I presented the following yearly seminars:

• leouieda/seminario-on-2012
• leouieda/qualify (qualification exam)
• leouieda/seminario-on-2014
• leouieda/seminario-on-2015

Abstract

We present methodological improvements to forward modeling and regional inversion of satellite gravity data. For this purpose, we developed two open-source software projects. The first is a C language suite of command-line programs called Tesseroids. The programs calculate the gravitational potential, acceleration, and gradient tensor of a spherical prism, or tesseroid. Tesseroids implements and extends an adaptive discretization algorithm to automatically ensure the accuracy of the computations. Our numerical experiments show that, to achieve the same level of accuracy, the gravitational acceleration components require finner discretization than the potential. In turn, the gradient tensor requires finner discretization still than the acceleration. The second open-source project is Fatiando a Terra, a Python language library for inversion, forward modeling, data processing, and visualization. The library allows the user to combine the forward modeling and inversion tools to implement new inversion methods. The gravity forward modeling tools include an implementation of the algorithm used in the Tesseroids software. We combined the inversion and tesseroid forward modeling utilities of Fatiando a Terra to develop a new method for fast non-linear gravity inversion. The method estimates the depth of the crust-mantle interface (the Moho) based on observed gravity data using a spherical Earth approximation. We extended the computationally efficient Bott's method to include smoothness regularization and use tesseroids instead right rectangular prisms. The inversion is controlled by three hyper-parameters: the regularization parameter, the density-contrast between the real Earth and the reference model (the Normal Earth), and the depth of the Moho of the Normal Earth. We employ two cross-validation procedures to automatically estimate these parameters. Tests on synthetic data confirm the capability of the proposed method to estimate smoothly varying Moho depths and the three hyper-parameters. Finally, we applied the inversion method developed to produce a Moho depth model for South America. The estimated Moho depth model fits the gravity data and seismological Moho depth estimates in the oceanic areas and the central and eastern portions of the continent. We observe large misfits in the Andes region, where Moho depth is largest. In Amazon, Solimões, and Paraná Basins, the model fits the observed gravity but disagrees with seismological estimates. These discrepancies suggest the existence of density-anomalies in the crust or upper mantle, as has been suggested in the literature.

Thesis: Robust 3D gravity gradient inversion by planting anomalous densities

Advisor: Valéria C. F. Barbosa

About

I did my Master's degree in Geophysics at the Observatório Nacional in Rio de Janeiro, Brazil, under the supervision of Valéria C. F. Barbosa. I started in March 2010 and defended my dissertation in October 2011. The method that we developed is implemented in the software Fatiando a Terra. The dissertation was later published as the paper:

• Robust 3D gravity gradient inversion by planting anomalous densities (2012)

During my MSc, I presented the following yearly seminars:

• leouieda/seminario-on-2010
• leouieda/seminario-on-2011

Abstract

We have developed a new gravity gradient inversion method for estimating a 3D density-contrast distribution defined on a grid of rectangular prisms. Our method consists of an iterative algorithm that does not require the solution of an equation system. Instead, the solution grows systematically around user-specified prismatic elements, called "seeds", with given density contrasts. Each seed can be assigned a different density-contrast value, allowing the interpretation of multiple sources with different density contrasts and that produce interfering signals. In real world scenarios, some sources might not be targeted for the interpretation. Thus, we developed a robust procedure that neither requires the isolation of the signal of the targeted sources prior to the inversion nor requires substantial prior information about the nontargeted sources. In our iterative algorithm, the estimated sources grow by the accretion of prisms in the periphery of the current estimate. In addition, only the columns of the sensitivity matrix corresponding to the prisms in the periphery of the current estimate are needed for the computations. Therefore, the individual columns of the sensitivity matrix can be calculated on demand and deleted after an accretion takes place, greatly reducing the demand for computer memory and processing time. Tests on synthetic data show the ability of our method to correctly recover the geometry of the targeted sources, even when interfering signals produced by nontargeted sources are present. Inverting the data from an airborne gravity gradiometry survey flown over the iron ore province of Quadrilátero Ferrífero, southeastern Brazil, we estimated a compact iron ore body that is in agreement with geologic information and previous interpretations.

About

In the fourth year of my BSc degree, I went on an international exchange program to York University to study in their Geomatics Engineering degree. I spent the year learning about geodesy, gravimetry, positioning, and least-squares adjustment, all of which I still use to this day. I also had a great time in Toronto and got to make a bunch of international friends.

Thesis: Cálculo do tensor gradiente gravimétrico utilizando tesseroides

Advisor: Naomi Ussami

About

My Bachelor's degree in Geophysics is from the Universidade de São Paulo, Brazil, where I studied from 2004 until 2009. I did an undergraduate research project and eventually my thesis under the supervision of Naomi Ussami. This was when I started development of the software Tesseroids and the research that lead to the paper which is the first part of my PhD thesis:

• Tesseroids: forward modeling gravitational fields in spherical coordinates (2016)

Abstract

The GOCE satellite mission has the objective of measuring the Earths gravitational field with an unprecedented accuracy through the measurement of the gravity gradient tensor (GGT). The data provided by this mission could be used to study large areas, where the flat Earth approximation can have its limitations. In these cases the modeling could be done with tesseroids, also called spherical prisms, in order to take the Earths curvature into account. The GGT caused by a tesseroid can be calculated with numerical integration methods, such as the Gauss-Legendre Quadrature (GLQ). In the current project, a computer program was developed for the direct calculation of the GGT using the GLQ. The accuracy of this implementation was evaluated by comparing its results with the result of analytical formulas for the special case of a spherical cap. Next, the developed program was used to calculate the differences in the GGT caused by the flat Earth approximation. These differences reach are up to 30% in the Tzz component for a 50 deg x 50 deg x 10 km model. Finally, the computer program was used to calculate the effect caused by the topographic masses on the GGT at 250 km altitude for the Paraná basin region. In regions of large topographical variations, the components of the GGT due to the topographic masses have amplitudes of the same order of magnitude as the GGT components due to density anomalies in the interior of the crust and mantle.

Award: IES\R3\213141

Amount: GBP 10,500

Abstract

The magnetization that is locked in certain minerals at the time of their formation is one of the few gateways we have to the Earth's distant past. By measuring the magnetization of certain rocks we are able to determine properties about the Earth's magnetic field in the past, which provides crucial information about our planet's climate history and the movement of the plates that make up the outermost layer of the Earth. For decades, researchers have only been able to make measurements of the average magnetic field of each rock sample, which can lead to large uncertainties in our estimates or even having to discard entire samples. Recent advances in technology are allowing us to make measurements in such detail that we may soon be able to distinguish the magnetic fields of the individual minerals that make up the rock sample. This new technology opens the door to using methods normally applied to national or continental scale geophysical surveys to micrometer scale data. There is still much to be explored and refined before this can be achieved. This collaboration will bring together experts from both large scale geophysics and micrometer scale paleomagnetism to explore the possibilities and help define future directions of research.

Award: 2020 Fellow

Amount: GBP 3,000

Abstract

The SSI has a yearly fellowship program to fund the organization of communities around scientific software (creating of local user groups, workshops, hackathons, etc). My original plan for the Fellowship was to run some Software Carpentry workshops in Liverpool and to develop lesson material for transitioning from coder to open-source software maintainer (based on our AGU workshops). Due to the COVID19 pandemic, I wasn't able to execute these plans. Instead, in 2022 I organized the Geo+Code event as part of my Fellowship.

Award: 1948602

Amount: USD 757,597

Abstract

The Generic Mapping Tools (GMT) is open source software infrastructure used in the Earth, ocean, and planetary geosciences. GMT supports other software platforms and delivers data processing and visualizations (graphs, charts, maps) that promote new discoveries and their dissemination to society. The primary goal of this project is to transition GMT to a governance structure that includes a broader and more diverse community of developers. This project will (1) recruit and train new and diverse developers, (2) build a broad and sustainable developer community, and (3) modernize, simplify, and strengthen the GMT software. The first two tasks are critical social activities while the last is technical, involving code hardening, interoperability improvements, interface modernization, documentation completion, and data upgrades. This project will provide training in cutting-edge computational software development and data analysis, along with engaging undergraduate students in scientifically challenging tasks related to the GMT project. The project will also facilitate ongoing user training and developer workshops. This project is based on a vision for the future of GMT that incorporates how governance, communications, developer recruitment and training will evolve in the next decade. The project will design and implement a sustainable model for GMT maintenance and curation, and execute a series of essential technical improvements. These improvements will address a) automated testing and verification of results, b) development of GMT-powered software libraries in other languages (Python, MATLAB, Julia) and c) the recruitment of new and enthusiastic developers so that GMT may continue to be maintained and evolve in a changing computational landscape. In addition, GMT products such as coastline maps will be revised with modern high-resolution data as well as technical documentation of how updates are produced. Finally, this project will advance GMT to a sustainable environment that results in lower maintenance, greater confidence in GMT products, and a more engaged community of users and developers.

Award: 1829371

Amount: USD 174,975

Abstract

Research under the EarthScope umbrella involves both data processing and display of results. The Generic Mapping Tools (GMT) is a set of scientific analysis and plotting modules that are widely used in the Earth Sciences, particularly because they allow for specialized processing and imaging suitable for the geosciences. GMT is a long-lived (~30 years) and robust NSF-funded software toolset that originated in the ocean sciences but has spread to all aspects of the geosciences and beyond. This project will add numerous enhancements to GMT, including a much simplified usage syntax called modern mode. In addition, new modules for simplifying the creation of animations, 3-D stacking and interpolation of grids, conversion of imagery for placement in Google Earth, and greatly improved interactive documentation will be built during the project. The Generic Mapping Tools (GMT) scripting has remained remarkably similar over its 30 years of existence. Because thousands of GMT classic scripts have been written, making major changes has been problematic, even if such changes would simplify its usage. The proposed modern mode will solve this dilemma as it introduces two new commands (begin and end) that starts and end a modern mode session. Hence it is not possible to accidentally enter modern mode without the use of those commands, thus allowing all existing GMT scripts to run unchanged. Once in modern mode, the required syntax will be simplified to avoid repetitive options and the user will no longer be responsible for assembling a composite PostScript file. The default output format will be PDF but more than one format may be selected, including standard raster image formats. Improved interactive documentation will be added to showcase modern mode (while maintaining the documentation for classic mode). To strengthen EarthScope science there will be new modules for 3-D grid stacking and interpolation, KML image quadtree building for Google Earth, and a new animation module that vastly simplifies the task of building scientific animations. It is anticipated that UNAVCO will assist with annual workshops aimed at EarthScope scientists using GMT in shell, MATLAB, or Python environments.

Abstract

This internal grant funded a full-time graduate technician for the Exploration Geophysics Lab (LAGEX) on a 4-year contract. The technician was responsible for maintaining the computer hardware we had, working on our open-source software Fatiando a Terra, and aiding in the teaching done in the lab.