US4 | Deep-time Digital Earth
Deep-time Digital Earth
Co-sponsored by ILP
Convener: Chengshan Wang | Co-conveners: Jennifer McKinley, Hans Thybo, Patricio Guillermo Villafañe, Monica Munassa Ribeiro Petreque Chamussa Juvane
| Thu, 18 Apr, 14:00–18:00 (CEST)
Room E1
Thu, 14:00
Incomplete evidence for the long-term evolution of Earth and difficulties of integrating existing data, in particular in digital form, limit the geoscientific understanding of Earth’s past and future. A wealth of data exists in archives and publications, which is not readily accessible to scientists. Compilation of such data in digital databases with existing digital data can mark a qualitative step forward to geosciences, in particular by combining it with efficient new data structures, data extraction methods and processing software.

This challenge is being explored by the new Deep-time Digital Earth (DDE) programme, launched by IUGS and around 30 other geoscience organisations, with the vision to foster a deep-time data driven research paradigm. DDE will harmonize deep-time earth data, share global geoscience knowledge, and advance geoscience understanding and research through development of an open-access on-line digital infrastructure.

This interdisciplinary Union-wide session includes contributions from various geoscience disciplines, particularly those where research progress is dependent on the availability and access to complex, global-scale datasets and models in deep time. The session presents contributions from application of Big Data analyses through Cloud Computing, and studies based on building, extending and using online platforms, together with discussion of secure and ethical data sharing in a digital infrastructure aiming at solving Deep-time issues.

Session assets

Orals: Thu, 18 Apr | Room E1

Chairpersons: Jennifer McKinley, Patricio Guillermo Villafañe
On-site presentation
A.M. Celâl Şengör

Deep time was discovered by James Hutton on a scientific basis in 1788. However, deep time became deeper since then we know now the earth is older than 4.5 Ga. Since that time, our earth has been cooling considerably. In the Archean the heat flow out of our earth was 6 times more than today. That means strict uniformitarianism cannot be applied for the entire earth history. The question becomes whether the fundamental processes on earth have changed in any fundamental way. Things we do not have actualist examples of the Archean environments today, these forces are to resort to the first principles of physics and chemistry. We have to ask ourselves what mechanism may have dominated heat loss of the planet in the Archaean and the early Proterozoic that governed the tectonic regime of the earth. What mechanism has governed the tectonic regime of the earth in the Archean and the early Proterozoic? The most efficient way is to lose heat with convection. Therefore, we can assume that the convection has always dominated terrestrial tectonics. However, there is a problem; in the Archean the mafic crust was thicker, but the lithosphere was thinner. This arises a doubt whether the subduction was possible under such circumstances. Geochemical work showed that wet melting was going on the earliest Archean. This indicates some sort of subduction must have been going on. On Venus, shortening amounts of hundred kilometers or more in the margins of the Tesserae seems to be accomplished by the décollement folding and thrusting. Then the question is what happened to the underthrust crust? Similar process on early earth may have produced wet melting by transporting ocean water in to the earth’s mantle. This tells us that plate tectonics must have been operative since the earliest time on earth. However, at the time spreading centres and subduction zones must have been much longer than they are today at consequently plates were slower. Continent grow by smashing island arcs against one another including their large subduction accretionary complexes. The principal task in deep earth research today is to be able to track the transition from the tectonics of the very hot earth to the present one. 

How to cite: Şengör, A. M. C.: Deep Time and Uniformitarianism, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-20408,, 2024.

Virtual presentation
Manuel Pubellier, Benjamin Sautter, Yang Song, and Chengshan Wang

Geological information is crucial for the civil society via public authorities, and is produced by scientists from academia and national geological surveys. This information is useful for better use of resources (water, energy, and mineral), adaptation to the climate change, and of course natural hazards. In the meantime, the Earth has become smaller, information technology has improved considerably, and allows to cross-analyze data. If communication is made easier for research and dissemination of the geological knowledge, formats have become the key for scientific exchanges.

The Commission for the Geological Map of the World (CGMW) was created for this exact purpose in the early 20th century in an attempt to bring scientists together, but it took a century to acknowledge that we are more than ever accountable for the preservation of the Earth, and forced to understand the importance of the linkage between geology and global change. As early as 1964, the UNESCO bi-annual program encouraged specialists to develop and use a uniform terminology and classification for the different Earth sciences. CGMW is a rank A UNESCO organization and is affiliated to IUGS and IUGG. The spearheading product of CGMW had been the Geological Map of the World at 1:35M scale for decades, and the opportunity of elaborating a large Geological Map of the World at scale 1:5M – (World5M), was proposed in 2018, as an IUGS Big Science Program DDE (Deep time Digital Earth).

This program, co-funded by the Chinese Academy of Geological Sciences (CAGS), aimed at harmonizing global digital Earth data and securing compatible and interoperable databases, and was rapidly considered as essential for international maps, which can be accessible by a data platform aligned with the vision and mission of the IUGS Big Science Program. This project, (1) integrates at a scale 1:5M the geological maps of continents and oceans which have been produced under supervision of the CGMW, (2) established the legend system and adequate architecture for the map database, and (3) constructed a new seamless and digital geological map of the world at the scale 1:5M. In reality, the geology presented in each small scale continental geological map, differs drastically in terms of stratigraphic cuts and trans-boundary connections of structural continuities. In addition, the databases were constructed originally in different ways with contrasting semantics and data standards, thus requiring a robust collaborative work in Geology and Geomatics. Among the outcomes of the project, is the possibility to integrate selected geological features over a large coverage with a similar resolution. This has been already used in new mapping syntheses, making them more informative.

How to cite: Pubellier, M., Sautter, B., Song, Y., and Wang, C.: Changes in the world of geological mapping, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5788,, 2024.

On-site presentation
Francois Robida, Zhang Minghua, Harvey Thorleifson, and Mark Rattenbury

DDE's ambition is to make better use of data from the earth sciences to produce new knowledge using new digital tools such as AI. The earth sciences have always relied on observations to produce knowledge. Scientists have always had a duty to ensure the quality of their data and to enable it to be reused by other researchers, by describing it as precisely as possible.

In the DDE approach, which makes use of vast quantities of digital data from a variety of sources, it is particularly important to be able to ensure the quality and relevance of the data used, in order to have confidence in the results produced by this approach. These data may have been produced by the scientific community on different continents, at different times, often for specific purposes.

When data is made available to researchers on platforms, it must comply with the FAIR (Findable - Accessible - Reusable - Interoperable) principles. For these principles to be implemented, the international community needs to agree on common terminologies and ontologies produced by the scientists themselves. This approach will provide machines with the reference systems needed for the reasoning used in AI.

As part of DDE, a Task Group on Standards (DDE/STG) has been set up to adopt common standards and a common approach to producing and updating these standards. This task group draws on the expertise of the Commission for the Management and Application of Geoscience Information (IUGS/CGI) and the Committee on Data of the International Science Council (CODATA). This presentation will provide an update on this work.

How to cite: Robida, F., Minghua, Z., Thorleifson, H., and Rattenbury, M.: Standards: a key tool for DDE, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4558,, 2024.

On-site presentation
Simon Hodson

This presentation, as part of a DDE-convened Union Symposium at EGU, will discuss avenues to pursue to enable greater Interoperability and Reusability of Deeptime Digital Earth Data, particularly in cross-domain research scenarios.

CODATA is the Committee on Data of the International Science Council (ISC).  Consequently, an important part of its mission is to engage with International Scientific Unions and related initiatives, on data issues.  The ISC has entrusted CODATA to develop a programme of activity: ‘Making Data Work for Cross-Domain Grand Challenges’.  After some exploratory work, the flagship activity is the WorldFAIR project which focuses on the implementation of the FAIR principles both within and across 11 different domain and cross-domain case studies.  Other related work includes the recommendations on FAIR vocabularies, with the International Union for the Scientific Study of Populations (report, and in relation to the ISC-UNDRR Hazard Implementation Profiles.  Similarly, CODATA is working with a number of International Scientific Unions, notably IUPAC, around the Task Group on Digital Representation of Units of Measure.

The common threads of this work are both to encourage the adoption and implementation of the FAIR principles, and to explore the requirements for better enabling cross-domain research.  Such work is of paramount importance: the major global scientific and human challenges of the 21st century (including climate mitigation and adaptation, disaster risk reduction, the interplay of society, the economy and energy policy) can only be addressed through cross-domain research that seeks to understand complex systems through machine-assisted analysis at scale.  Our capacity for such analysis is currently constrained by the limitations in our ability to access and combine heterogenous data within and across domains.

CODATA has recently concluded a Memorandum of Understanding with the Deeptime Digital Earth initiative.  This agreement indicates a number of shared interests.  Particularly important is collaboration among DDE, IUGS CGI (Commission for Geosciences Information), and the wider CODATA and FAIR communities on the further development and representation of key terminologies.  Additionally, through a case study approach, DDE, IUGS CGI, CODATA and other partners plan to explore the applicability of the WorldFAIR methodology and the use of FAIR Implementation Profiles to understand FAIR requirements, progress and alignment.  Finally, the applicability of the emergent Cross-Domain Interoperability Framework (CDIF) will be explored, and further refinements and recommendations made. 

This presentation will describe the context for this collaboration and outline the specific activities.  It will be an important opportunity to socialise the community to this initiative, to get feedback and advice on the approach and to invite collaboration and expert input from the wider EGU community.

How to cite: Hodson, S.: Cross-Domain Interoperability and Deeptime Digital Earth Data, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-21592,, 2024.

On-site presentation
Zhenhong Du
Big Data and AI have transformed Earth sciences, enabling data-driven discovery. While digital infrastructures have simplified access for most, deep-time geoscience faces challenges with scattered heterogeneous data and traditional theoretical methods. To overcome this, we propose the Deep-time Platform—a one-stop online research platform for geoscientists. Utilizing cloud computing and advanced tech, it offers open access to deep-time geoscientific data, knowledge, models, and computing power. The Deep-time Engine ensures seamless coordination. The Platform is aimed at enabling and empowering global geoscientists’ collaborative innovation and discoveries. The Deep-time Platform is a significant advancement in geoscientific exploration, fostering global collaboration and promoting a data-driven research paradigm within the framework of open science.

How to cite: Du, Z.: One-Stop Online Research Platform for Geoscientists, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5184,, 2024.

On-site presentation
Dietmar Müller

Plate tectonic reconstructions have progressed substantially over the last decade, incorporating evolving plate boundaries as well as plate deformation. These technological innovations have spurred the construction of a variety of new regional and global plate models. They are acting as a catalyst for the emergence of a new generation of geodynamic models as well as new approaches for studying the flux of material, including volatiles, from the surface to the deep Earth and vice-versa. Plate models with dynamic plate boundaries have evolved to reach further back in geological time, extending into the Proterozoic. Uncertainties in Proterozoic reconstructions are difficult to quantify, but the availability of the GPlates software accompanied by a variety of open-access data sets has enabled the community to develop alternative models, exploring a range of possible interpretations of available geodata to reconstruct plate motions and plate boundary evolution. The emergence of deep-time plate models has opened numerous opportunities for Earth system analysis, including an improved understanding of the evolution of Earth's mantle structure through time, quantifying solid Earth carbon degassing, and linking biodiversity evolution to plate tectonic and surface processes. Three additional developments are significant in the context of emerging spatio-temporal deep-time data analysis: (1) the availability of large open-access geological, geochemical and geochronological databases; (2) the spread of shared open-source software and workflows aiding data analysis; and (3) the rapid recent rise of open AI tools to extract new knowledge from a complex, hyperdimensional data volume through space and time. The pyGPlates and GPlately python libraries have particularly played an enabling role for allowing the analysis of plate models as well as geodata attached to tectonic plates. Together, these developments are catalysing the emergence of entirely new approaches to study deep time Earth system evolution. The applied drivers of deep-time geodata science are to a large extent tied to rapid climate change, the need to better understand potential future trajectories of Earth's surface environments and to enable a transition to renewable energy generation and an electrified transport sector. This transition demands a significant increase in exploration for and discovery of critical minerals below the well-explored surface. It is estimated that at least 384 new mines for graphite, lithium, nickel and cobalt alone are required to meet demand for battery energy storage by 2035. Deep-time Earth models allow the connection of traditional mineral exploration data to evolving tectonic and surface environments, providing the opportunity to build new approaches for better understanding the emplacement and preservation of mineral deposits. The societal and economic need to rapidly reduce our dependence on fossil fuel and to better understand the complex feedbacks between deep Earth, the hydrosphere, atmosphere and biosphere is invigorating the entire field of geology. I will briefly outline some emerging applications of deep-time Earth system analysis and provide an outlook for the future.

How to cite: Müller, D.: Deep time plate configurations and their emerging applications, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4849,, 2024.

Coffee break
Chairpersons: Hans Thybo, Monica Munassa Ribeiro Petreque Chamussa Juvane
On-site presentation
GeoGPT, the large earth science language model system
Jian Wang
On-site presentation
Robert M. Hazen, Shaunna M. Morrison, and Anirudh Prabhu

Minerals and the rocks that hold them are the oldest objects that we can hold in our hands. Each specimen is an information-rich time capsule, waiting to be opened. Every sample contains scores of attributes, including trace and minor elements, stable and radiogenic isotopes, solid and fluid inclusions, optical and magnetic properties, structural defects, twinning, exsolution, zoning, and more. These diagnostic attributes are enhanced by information on the ages, associations, and geological contexts of specimens. Collectively, these characteristics of minerals tell stories of the origins and evolution of planets and moons.

Mineral informatics[1], a field championed by the Deep-time Digital Earth Program, advances understanding of planetary evolution by pursuing three complementary objectives. First is to develop comprehensive open-access data resources for minerals and rocks. Several vital platforms, including,, and, provide large and growing resources for researchers. Ongoing work will build these essential research tools, while advancing a culture of FAIR data.

A second objective of our work is to establish a new mineral classification system based on mineral informatics that highlights formation processes and evolutionary stages of minerals [2-6]. Traditional approaches to classifying minerals ignore this history. The International Mineralogical Association has catalogued >6000 mineral species, each with a unique combination of idealized chemical composition and crystal structure. This essential scheme facilitates identification of different condensed crystalline building blocks of planets and moons. However, the IMA system lacks contexts of time and process. We have introduced, and are now completing, a new and complementary approach to mineral classification called the “evolutionary system of mineralogy.” Our system differs from that of the IMA in three important ways. (1) We split many IMA species based on their varied modes of formation and age of earliest occurrence. For example, diamond (carbon in the diamond crystal structure in the IMA system) occurs in at least 5 distinct mineral “kinds” in the evolutionary system, including stellar diamond and impact diamond. (2) We lump varied IMA species that form continuous solid solutions through the same process, for example, combining different species of the tourmaline group into a single kind. (3) We include  amorphous  solids, such as obsidian and limonite, which are important in crustal processes. 

A third goal of the mineral informatics program is to develop and apply advanced methods of data analysis and visualization to better characterize evolving mineral systems through more than 4.5 billion years of planetary history. To this end, we have incorporated network analysis, cluster and analysis and community detection, association analysis, and other methods to quantify the changing diversity and distribution of minerals through deep time, while estimating total mineral diversity and predicting new localities of critical mineral resources.

References: 1. Prabhu et al. (2023) Am.Min., 108, 1242-1257; 2. Hazen R.M. et al. (2008) Am.Min., 93, 1693-1720; 3. Hazen R.M. & Morrison S.M. (2022) Am.Min., 107, 1262-1287; 4. Hazen, R.M. et al. (2023) In: Bindi and Cruciani [Eds.], Celebrating the International Year of Mineralogy. NY: Springer, pp.15-37; 5. Hazen R.M. et al. (2022) Am.Min., 107, 1288-1301; 6. Hazen R.M. (2019) Am.Min., 104, 468-470.

How to cite: Hazen, R. M., Morrison, S. M., and Prabhu, A.: Mineral Informatics: A Key to Deep-Time Data-Driven Discovery in Earth and Planetary Sciences, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4641,, 2024.

On-site presentation
Shuzhong Shen, Yukun Shi, Yingying Zhao, Shuhan Zhang, and Huaichun Wu

There has been long-standing debate about what environmental factors are the main drivers of biodiversity changes. During the last hundreds of years, there are convincing signs that rising and falling CO2 and temperatures are affecting both marine and terrestrial biodiversities. However, this has been rarely tested with high-resolution biodiversity changes in deep time. It is also difficult to judge whether the current diversity loss caused by global changes is a long-term tendency or a catastrophic event because the observatory records are too short to predict the future. Fusuline foraminifers were a major group of Carboniferous through Permian marine microorganisms for ~91.8 million years (Myr) during which they experienced changes from icehouse to greenhouse climates. Here we use a high-resolution analysis of fusuline diversity with an average resolution 40 thousand years (kyr) to analyze their speciation and extinction dynamics at multiple temporal levels during this interval of major climatic shifts. This new database encompasses 1391 species from 293 published stratigraphic sections worldwide using constrained optimization method (CONOP). Our results show a symmetric diversity pattern with a peak between 295.24 Ma and 293.57 Ma in the middle of the lifespan of Fusulinida and temporally coincident with the apex of the Late Paleozoic ice age (LPIA). The shift from icehouse to greenhouse climates led to the decline of fusuline diversity. Major disruptions in fusuline diversity are found during the late Moscovian-Kasimovian interglacial event, the post-LPIA long-term warming and the middle-late Guadalupian extinction before the clade was eliminated by the end-Permian mass extinction (EPME). Each of the events of large diversity loss are linked to global warming, probably induced by massive release of greenhouse gases from intensive volcanism. The high temporal resolution also allows us to interrogate the finer-scale patterns revealing that species richness, origination, and extinction rates were paced by long-term astronomical forcing, including ~1.0 Myr obliquity and ~2.1 Myr eccentricity cycles. This highlights the substantial role of astronomically forced climate variability on the rhythms of biological evolution. Our study suggests climatic forcing of long-term changes and catastrophic events in fusuline diversity, with global cooling fueling foraminifera diversifications. This pattern is consistent with the late Cenozoic diversifications of recent foraminifera before the mid-Pliocene warming period.

How to cite: Shen, S., Shi, Y., Zhao, Y., Zhang, S., and Wu, H.: Climate changes driving fusuline macroevolution, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6960,, 2024.

On-site presentation
Bilal Haq

Earth’s sedimentary record holds physical and chemical clues that allow us to decipher sea level variations in the deep past in fair amount of detail and time resolution, even though many issues remain that stifle precision. On the longer time scales (multiple million years) sea level trajectories can also be reconstructed from paleogeographic data (marine floodings of continents) and geophysical modeling (sea floor geodynamics and basin volume changes). The last few decades have seen significant advances in our understanding of the behavior of sea level at both regional (eurybatic) and global (eustatic) spatial scales and updated sea-level curves have been published for the entire Phanerozoic, albeit with different degree of accuracy and resolution that generally decrease further back in deep time due to increasing uncertainties about timing and amplitudes of eustatic variations. One major recent discovery made through geophysical modeling has been the long-wavelength and relatively slow (multiple million years) warping of continental margins due to mantle driven dynamic topographic changes that significantly affect amplitudes estimates and often go undetected. On the positive side, for some parts of the Mesozoic and for all of the Cenozoic oxygen-isotopic data (δ18O) of marine benthic foraminifera have proven useful in constraining both the timings and (to a lesser extent) the amplitudes of sea-level rises and falls. Digitizing sea-level variations in deep time poses a challenge, as multiple streams of parallel data have to be correlated with meaningful precision, each dataset having its own inherent degree(s) of uncertainties. Nonetheless, For most portions of the Phanerozoic sea-level variations data, especially the timing of sea-level withdrawals,  has been arrived at through calibration to several different scales (i.e., magnetic reversal, biochronologic, isotopic), as well as absolute time. The Cenozoic (the last 66 Myr.) has also been fine tuned through astronomical (orbital) cyclostratigraphy. Nevertheless deep time eustatic history should be regarded as work in advanced progress and, with periodic revisions (as new bio-chronological data and new technologies become available), it can be kept current for enhanced utility. In this presentation, the advances made in recent years and the state of the art of our current understanding of sea level in deep time will follow the discussion of the technological limitations.




How to cite: Haq, B.: Decrypting Sea Level Variations in Deep Time, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6059,, 2024.

On-site presentation
Isabel Montañez, Gabriel Bowen, Daniel Breecker, Bärbel Hönisch, Kate Huntington, and Dana Royer

Paleo-CO2 reconstructions are integral to understanding the evolution of Earth system processes and their interactions given that atmospheric CO2 concentrations are intrinsically linked to planetary function. Furthermore, past periods of major climate change provide unique insights into the response of land-atmosphere-ocean interactions to warming-induced climate change, particularly for times of pCO2 comparable to those projected for our future. How well the past can inform the future, however, depends on how well paleo-CO2 estimates areconstrained. CO2 estimates exist for much of the past half-billion years (the Phanerozoic), but proxies differ in their assumptions and degree of understanding, and there is substantial uncertainty and inconsistency in existing paleo-CO2 estimates. Here, we introduce a community initiative, CO2PIP, focused on advancing the science of paleo-CO2 reconstruction through critically evaluating and modernizing existing records and building a statistically robust multi-proxy atmospheric CO2 record for the Phanerozoic. CO2PIP builds on the previous work of the Cenozoic CO2 Proxy Integration Project (CenCO2PIP) Consortium (Hönisch et al., 2023) and takes a multi-step approach to building the next generation Phanerozoic CO2 record. We are building a standardized paleo-CO2 proxy data repository that includes all metadata and updated chronology and meets FAIR (findable, accessible, interoperable, reusable) data standards. Existing terrestrial-based CO2 estimates are being modernized through additional analyses and measurements, and a set of forward proxy system models are being developed to provide a quantified representation of proxy sensitivities to environmental and ecophysiological conditions and processes that govern the CO2 signals. Ultimately, statistical inversion analysis of the simulated and modernized proxy datasets will be used to produce quantitative, data-driven CO2 reconstructions for individual records and to generate a robust, quantitative reconstruction of atmospheric CO2 concentrations through the Phanerozoic. Digital infrastructure for presenting and archiving the CO2 compilation and project outputs ( ensures full accessibility to the scientific community and the public.

Hönisch, B. Royer, D., Breecker, D. O., et al., 2023, Towards a Cenozoic history of atmospheric CO2. Science, v. 382 (6675), DOI: 10.1126/science.adi5177).

How to cite: Montañez, I., Bowen, G., Breecker, D., Hönisch, B., Huntington, K., and Royer, D.: CO2PIP Consortium for Advancing paleo-CO2 reconstruction and Building the Next-Generation Phanerozoic CO2 Record, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13387,, 2024.



  • A.M. Celâl Şengör, ITU Eurasia Institute of Earth Sciences, Türkiye
  • Manuel Pubellier, Ecole Normale Superieure, France
  • Francois Robida, France
  • Simon Hodson, CODATA, the Committee on Data of the International Science Council, France
  • Zhen Hong Du, Zhejiang Unversity, China
  • Dietmar Müller, Univ. of Sydney, Australia
  • Robert Hazen, Carnegie Institution for Science, United States of America
  • Shuzhong Shen, Nanjing University, China
  • Bilal Haq, Sorbonne University, France
  • Isabel Montañez, University of California, Davis, United States of America