Periodic Reporting for period 1 - TANGO (A probabilistic framework for assessing polar wander – Constraining paleolongitude in deep time)
Période du rapport: 2021-08-01 au 2023-07-31
Many of the most outstanding Earth science questions require an understanding of paleogeography and its temporal evolution through the incessant motion of the tectonic plates. Plate motions of the last ~130 Ma are well-resolved from magnetic isochrons and hotspot tracks, but as those records are progressively destroyed by subduction, they cannot be used in deeper time. Before 130 Ma, plate motions can only be quantified through the study of paleomagnetism, but two key components have so far prevented us from using this tool to its full potential. First, the paleomagnetic record contains substantial noise that is not integrated into the analysis. Second, owing to the inherent axial symmetry of the Earth's magnetic field, paleolongitude cannot be constrained and remains the greatest uncertainty in pre-Cretaceous plate reconstructions. It is of the utmost importance to consider what shortcomings may be present in our present methods, and how these may be improved. This fellowship aimed to reshape the current state-of-the-art in paleomagnetic data analysis by assessing polar wander with rigorous uncertainty estimates to accurately determine paleolongitude, addressing a long-standing challenge in geophysics.
In summary, the project TANGO developed an innovative and cross-disciplinary approach to assess polar wander with rigorous uncertainty estimates, allowing paleolongitude to be constrained in deep time. This approach addresses one of the most long-standing problems in geophysics and opens a new frontier in paleomagnetic and paleogeographic research. The project has surpassed initial expectations by generating outcomes for every work package and producing more publications than originally envisioned. A summary of the 2-year fellowship is as follows:
- five published papers of new work (Geophysical Research Letters, Computers & Geosciences, JGR: Solid Earth, Earth-Science Reviews);
- three published software packages;
- three papers in preparation;
- one article featured as an Editor’s Highlight in JGR: Solid Earth, which has also been published on Eos.org;
- one invited talk at the EGU General Assembly 2023;
- two sessions convened at EGU 2024 and AGU 2023;
- one research visit to the University of California, Berkeley, and another to Utrecht University in the Netherlands;
- two blog posts (EGU and IAGA).
The project aimed to provide open access to research data and to make research Findable, Accessible, Interoperable, and Reusable (FAIR). TANGO's main research output was twofold: computer software and technical research outputs, including publications. The research data and the publications from the TANGO project were deposited on the Zenodo platform, an EU-supported portal for big data management with extended digital library capabilities for open access and open data. More specifically, the research data and publications are located on the Zenodo-curated OpenAIRE platform (https://www.openaire.eu/) a major EC-supported initiative for fostering open science in Europe. A Zenodo Community has been created for the project (https://zenodo.org/communities/tango/) where all source code and documents are, and will continue to be, accessible through a single collection. The community is already active and is being updated and enriched with standard Zenodo metadata, including the Grant Number (101025975) and the Project Acronym (TANGO).
- five published papers of new work (Geophysical Research Letters, Computers & Geosciences, JGR: Solid Earth, Earth-Science Reviews);
- three published software packages;
- three papers in preparation;
- one article featured as an Editor’s Highlight in JGR: Solid Earth, which has also been published on Eos.org;
- one invited talk at the EGU General Assembly 2023;
- two sessions convened at EGU 2024 and AGU 2023;
- one research visit to the University of California, Berkeley, and another to Utrecht University in the Netherlands;
- two blog posts (EGU and IAGA).
The project aimed to provide open access to research data and to make research Findable, Accessible, Interoperable, and Reusable (FAIR). TANGO's main research output was twofold: computer software and technical research outputs, including publications. The research data and the publications from the TANGO project were deposited on the Zenodo platform, an EU-supported portal for big data management with extended digital library capabilities for open access and open data. More specifically, the research data and publications are located on the Zenodo-curated OpenAIRE platform (https://www.openaire.eu/) a major EC-supported initiative for fostering open science in Europe. A Zenodo Community has been created for the project (https://zenodo.org/communities/tango/) where all source code and documents are, and will continue to be, accessible through a single collection. The community is already active and is being updated and enriched with standard Zenodo metadata, including the Grant Number (101025975) and the Project Acronym (TANGO).
Tectonic plate motions are integral in shaping many aspects of Earth, including its thermal, geochemical, tectonic, biological, climatic, and magnetic evolution. Understanding the history of plate movements is therefore directly relevant to a wide variety of research areas. Paleomagnetic data offer a key constraint for plate movements, providing the only quantitative tool for plate modeling in pre-Jurassic times. Paleomagnetic data are typically analyzed and communicated in the form of so-called apparent polar wander paths (APWP), which represent the time-dependent position of Earth's spin axis relative to a given block of lithosphere. However, the standard methods for analyzing and grouping paleomagnetic data have been limited in how they propagate and quantify uncertainties, and the consequences of these limitations were not well understood. This fellowship addressed these challenges and reshaped the current state of the art by developing a multidisciplinary approach that combines paleomagnetic data analysis, machine learning, and statistics to constrain polar wander through deep time. The research demonstrated that even with substantial noise, polar wandering can be assessed with unprecedented temporal and spatial resolution.
The ability to generate high-resolution APWPs presents exciting new opportunities. Full tectonic plate motions (i.e. including longitude) from paleomagnetic data have long been considered an alluring but ultimately intractable problem. Paleomagnetic Euler pole (PEP) analysis presented a unique means to recover such information, but the feasibility of the methodology in light of the noise accompanying paleomagnetic data cast doubts on its fidelity. By developing an unsupervised learning method for PEP analysis, the tools here presented mark an important step forward. Consequently, a reinvigorated paleomagnetic Euler pole analysis has shown that it can retrieve paleo-kinematic information from assemblies of paleomagnetic data with sufficiently high resolution. While conventionally constructed APWPs cannot meet these requirements, this new framework may provide APWPs that can.
The project aimed to provide open access to research outputs and to make research Findable, Accessible, Interoperable, and Reusable (FAIR). In accordance, the open software and the scientific outputs from the TANGO project were deposited on the Zenodo platform, curated by the OpenAIRE platform (https://zenodo.org/communities/tango/).
The expertise of the contributors and the research design have granted the researcher valuable insights and expertise in paleomagnetic data analysis, machine learning, and mantle dynamics. An outstanding aspect of this project is the extensive network of collaborators that has been fostered. The scientific collaborations established during this fellowship contribute to its enduring legacy and strength.
The ability to generate high-resolution APWPs presents exciting new opportunities. Full tectonic plate motions (i.e. including longitude) from paleomagnetic data have long been considered an alluring but ultimately intractable problem. Paleomagnetic Euler pole (PEP) analysis presented a unique means to recover such information, but the feasibility of the methodology in light of the noise accompanying paleomagnetic data cast doubts on its fidelity. By developing an unsupervised learning method for PEP analysis, the tools here presented mark an important step forward. Consequently, a reinvigorated paleomagnetic Euler pole analysis has shown that it can retrieve paleo-kinematic information from assemblies of paleomagnetic data with sufficiently high resolution. While conventionally constructed APWPs cannot meet these requirements, this new framework may provide APWPs that can.
The project aimed to provide open access to research outputs and to make research Findable, Accessible, Interoperable, and Reusable (FAIR). In accordance, the open software and the scientific outputs from the TANGO project were deposited on the Zenodo platform, curated by the OpenAIRE platform (https://zenodo.org/communities/tango/).
The expertise of the contributors and the research design have granted the researcher valuable insights and expertise in paleomagnetic data analysis, machine learning, and mantle dynamics. An outstanding aspect of this project is the extensive network of collaborators that has been fostered. The scientific collaborations established during this fellowship contribute to its enduring legacy and strength.