Search

The geophysical research community has developed a relatively large amount of numerical codes and scientific methodologies which are able to numerically simulate through physics the extreme behavior of the Earth systems (for example: volcanoes, tsunamis earthquakes, etc). Furthermore, nowadays, large volumes of data have been acquired and, even near real-time data streams are accessible. Therefore, Earth scientist currently have on their hands the possibility of monitoring these events through sophisticated approaches using the current leading edge computational capabilities provided by pre-exascale computing infrastructures. The implementation and deployments of 12 Digital Twin Components (DTCs), addressing different aspects of geophysical extreme events is being carried out by DT-GEO, a project funded under the Horizon Europe programme (2022-2025). Each DTC is intended as self-contained entity embedding flagship simulation codes, Artificial Intelligence layers, large volumes of (real-time) data streams from and into data-lakes, data assimilation methodologies, and overarching workflows which will are executed independently or coupled DTCs in a centralized HPC and/or virtual cloud computing research infrastructure.

The EU Center of Excellence for Exascale in Solid Earth (ChEESE) develops exascale transition capabilities in the domain of Solid Earth, an area of geophysics rich in computational challenges embracing different approaches to exascale (capability, capacity, and urgent computing). The first implementation phase of the project (ChEESE-1P; 2018¿2022) addressed scientific and technical computational challenges in seismology, tsunami science, volcanology, and magnetohydrodynamics, in order to understand the phenomena, anticipate the impact of natural disasters, and contribute to risk management. The project initiated the optimisation of 10 community flagship codes for the upcoming exascale systems and implemented 12 Pilot Demonstrators that combine the flagship codes with dedicated workflows in order to address the underlying capability and capacity computational challenges. Pilot Demonstrators reaching more mature Technology Readiness Levels (TRLs) were further enabled in operational service environments on critical aspects of geohazards such as long-term and short-term probabilistic hazard assessment, urgent computing, and early warning and probabilistic forecasting. Partnership and service co-design with members of the project Industry and User Board (IUB) leveraged the uptake of results across multiple research institutions, academia, industry, and public governance bodies (e.g. civil protection agencies). This article summarises the implementation strategy and the results from ChEESE-1P, outlining also the underpinning concepts and the roadmap for the on-going second project implementation phase (ChEESE-2P; 2023¿2026).

Destination Earth initiative pursues the implementation of a digital model of the Earth. With the aim to help understand and simulate the evolution and behavior of the Earth system components, to aid in better forecasting the impacts on human system processes, ecosystem processes and their interaction. The current state of the art technologies in numerical computations (HPC), data infrastructures (involving data storage, data access, data analysis), enable the possibility of developing numerical clones mimicking Earth¿s geophysical extreme phenomena.A Digital Twin for GEOphysical extremes (DT-GEO),is a new EU project funded under the Horizon Europe programme (2022-2025), with the objective of developing a prototype for a digital twin on geophysical extremes including earthquakes, volcanoes, tsunamis, and anthropogenic-induced extreme events. It will enable analyses, forecasts, and responses to ¿what if¿ scenarios for natural hazards from their genesis phases and across their temporal and spatial scales. The project consortium brings together world-class computational and data Research Infrastructures (RIs), operational monitoring networks, and leading-edge research and academia partnerships in various fields of geophysics. It mergesthe latest outcomes from other European projects and, Centers of Excellence. DT-GEO will deploy and test 12 Digital Twin Components (DTCs). These will be self-contained entities embedding flagship simulation codes, Artificial Intelligence layers, large volumes of (real-time) data streams from and into data-lakes, data assimilation methodologies, and overarching workflows for deployment and execution of single or coupled DTCs in centralized HPC and virtual cloud computing Ris. (DT-GEO: A Digital Twin for GEOphysical extremes, project ID 101058129) How to cite: Carbonell, R., Folch, A., Costa, A., Orlecka-Sikora, B., Lanucara, P., Løvholt, F., Macias, J., Brune, S., Gabriel, A.-A., Barsotti, S., Behrens, J., Gomes, J., Schmittbuhl, J., Fred

We present an accelerated full-waveform inversion based on dynamic mini-batch optimization, which naturally exploits redundancies in observed data from different sources. The method rests on the selection of quasi-random subsets (mini-batches) of sources, used to approximate the misfit and the gradient of the complete data set. The size of the mini-batch is dynamically controlled by the desired quality of the gradient approximation. Within each mini-batch, redundancy is minimized by selecting sources with the largest angular differences between their respective gradients, and spatial coverage is maximized by selecting candidate events with Mitchell’s best-candidate algorithm. Information from sources not included in a specific mini-batch is incorporated into each gradient calculation through a quasi-Newton approximation of the Hessian, and a consistent misfit measure is achieved through the inclusion of a control group of sources. By design, the dynamic mini-batch approach has several main advantages: (1) The use of mini-batches with adaptive size ensures that an optimally small number of sources is used in each iteration, thus potentially leading to significant computational savings; (2) curvature information is accumulated and exploited during the inversion, using a randomized quasi-Newton method; (3) new data can be incorporated without the need to re-invert the complete data set, thereby enabling an evolutionary mode of full-waveform inversion. We illustrate our method using synthetic and real-data inversions for upper-mantle structure beneath the African Plate. In these specific examples, the dynamic mini-batch approach requires around 20 per cent of the computational resources in order to achieve data and model misfits that are comparable to those achieved by a standard full-waveform inversion where all sources are used in each iteration.

We present a general concept for evolutionary, collaborative, multiscale inversion of geophysical data, specifically applied to the construction of a first-generation Collaborative Seismic Earth Model. This is intended to address the limited resources of individual researchers and the often limited use of previously accumulated knowledge. Model evolution rests on a Bayesian updating scheme, simplified into a deterministic method that honors today's computational restrictions. The scheme is able to harness distributed human and computing power. It furthermore handles conflicting updates, as well as variable parameterizations of different model refinements or different inversion techniques. The first-generation Collaborative Seismic Earth Model comprises 12 refinements from full seismic waveform inversion, ranging from regional crustal- to continental-scale models. A global full-waveform inversion ensures that regional refinements translate into whole-Earth structure.

We successfully apply the semi-global inverse method of simulated annealing to determine the best-fitting 1-D anisotropy model for use in acoustic frequency domain waveform tomography. Our forward problem is based on a numerical solution of the frequency domain acoustic wave equation, and we minimize wavefield phase residuals through random perturbations to a 1-D vertically varying anisotropy profile. Both real and synthetic examples are presented in order to demonstrate and validate the approach. For the real data example, we processed and inverted a cross-borehole data set acquired by Vale Technology Development (Canada) Ltd. in the Eastern Deeps deposit, located in Voisey's Bay, Labrador, Canada. The inversion workflow comprises the full suite of acquisition, data processing, starting model building through traveltime tomography, simulated annealing and finally waveform tomography. Waveform tomography is a high resolution method that requires an accurate starting model. A cycle-skipping issue observed in our initial starting model was hypothesized to be due to an erroneous anisotropy model from traveltime tomography. This motivated the use of simulated annealing as a semi-global method for anisotropy estimation. We initially tested the simulated annealing approach on a synthetic data set based on the Voisey's Bay environment; these tests were successful and led to the application of the simulated annealing approach to the real data set. Similar behaviour was observed in the anisotropy models obtained through traveltime tomography in both the real and synthetic data sets, where simulated annealing produced an anisotropy model which solved the cycle-skipping issue. In the real data example, simulated annealing led to a final model that compares well with the velocities independently estimated from borehole logs. By comparing the calculated ray paths and wave paths, we attributed the failure of anisotropic traveltime tomography to the breakdown of the ray-theoretica