Experimental Constraints on Grain-Size Dependent Seismic Attenuation and Dispersion in Synthetic Dunite

Hitank Kasaundhana, Jian Yangb, Tongzhang Qua,c, Kathryn Haywarda, Hayden Millera, Ulrich Fauld, and Ian Jacksona

a: Research School of Earth Sciences, Australian National University, Canberra, Australian Capital Territory, Australia

b: China University of Petroleum, East China

c:  University College of London

d: EAPS, Massachusetts Institute of Technology, Cambridge, MA, USA

We investigated the grain-size sensitivity of viscoelastic relaxation in pure, dry, melt-free olivine through forced-oscillation experiments on seven synthetic dunite specimens with a mean grain size from 4 to 150 µm. All specimens were confirmed to be dry and melt-free within tested temperature range. Experiments were conducted in the ANU attenuation apparatus at 0.2 GPa, 400–1300 °C, and oscillation periods from 1–1000 s. Within the anelastic absorption band of a generalized Burgers model, the background dissipation obeys Q-1d-αm where α~ ¼ relates to the period dependence of dissipation and m ≈ 2.8 to the grain-size sensitivity of the Maxwell relaxation time. Extrapolation of this model to mantle conditions predicts attenuation far lower than seismological observations, implying that sub-solidus relaxation is enhanced by grain-boundary impurities, or by pre-melting, or dislocations, or minor melt. Preliminary results from San Carlos olivine suggest that impurities can significantly increase dissipation relative to pure olivine.

Exporting Earth’s Ears: Core Secrets of Other Worlds

Yun-Ze Cheng1, and Hrvoje Tkalčić1

1 Research School of Earth Sciences, The Australian National University, 142 Mills Road, 2601 ACT, Canberra, Australia.

The deep interior of terrestrial bodies, particularly the core, is fundamental to understanding their evolution. For the Moon, the core's state influences its early magma ocean crystallization, the longevity of its dynamo, and potential mantle overturn. To provide new constraints on the lunar core, we hypothesize that the far-side deep moonquakes lack identifiable S-arrivals as indicators of an S-wave shadow zone imposed by the lunar core, rather than a partial-melt layer atop the core. Our simulations of synthetic core-diffracted phases show that a larger lunar core is required to explain the observation, leading to the Large Core Model (LCM) for the Moon. Furthermore, we propose using the coda-correlation technique on future mission data to identify weak core-reflected phases. We also apply radiative transfer theory to model scattering structures, which may help constrain the core in other planetary bodies like Mars.

Linking Ocean Island Basalt Geochemistry with Mantle Plume Dynamics

Shihao Jiang, Thomas Duvernay, Mark J. Hoggard, Rhys Hawkins, Ian H. Campbell, D. Rhodri Davies

The hypothesis that the lithosphere mechanically limits mantle plume upwelling and associated decompression melting - known as the lid effect - is supported by trends in ocean island basalt geochemistry. However, the lid-effect concept remains too general to account for specific geochemical variations, such as REE concentrations, observed across a range of OIB suites. We use 3D coupled geochemical-geodynamical simulations of plume decompression melting to explore how melt-region geometry and melt composition vary with plume excess temperature and lithospheric thickness, and how plume-lithosphere interactions influence these processes. We find that, for a single plume, melt extraction across the domain of a single-lithology plume can reproduce much of the geochemical diversity observed within individual islands. We also identify a melt-flux filter that only plumes generating high melt fluxes produce sufficient melt volumes capable of percolating through thick lithosphere.

Modelling GIA with G-ADOPT

William Scott, Mark Hoggard, Thomas Duvernay, Sia Ghelichkhan, Angus Gibson, Dale Roberts, Stephan C. Kramer, and D. Rhodri Davies

Robust models of viscoelastic Earth deformation under evolving surface loads underscore many problems in geodynamics and are particularly critical for paleoclimate and sea-level studies through their role in Glacial Isostatic Adjustment (GIA). A long-standing challenge in GIA research is to perform computationally efficient inversions for ice-loading histories and mantle structure using a physically realistic Earth model that incorporates three-dimensional viscosity variations and/or complex rheologies. For example, recent geodetic observations from melting ice sheets appear inconsistent with long-term sea-level records and have been used to argue for transient rheologies, generating debate in the literature and leaving large uncertainties in projections of future sea-level change. Here, we extend the applicability of G-ADOPT (a Firedrake-based finite element framework for geoscientific adjoint optimisation) to these problems. Our implementation solves the equations governing viscoelastic surface loading while naturally accommodating elastic compressibility, lateral viscosity variations, and non-Maxwell rheologies (including transience). Crucially, G-ADOPT enables automatic derivation of adjoint sensitivity kernels, allowing gradient-based optimisation strategies that are essential for high-dimensional inverse problems. Using synthetic Earth-like experiments, we illustrate its capability to reconstruct ice histories and recover mantle viscosity variations, providing a roadmap towards data assimilation and uncertainty quantification in GIA modelling and sea-level projections.

Sensitivity Kernels for Features in the Coda-correlation Wavefield

Zhi Wei, Sheng Wang, Thanh-Son Pham, and Hrvoje Tkalčić

Research School of Earth Sciences, Australian National University, Canberra, Australian 

We calculate finite-frequency sensitivity kernels for different prominent seismic features clearly appeared in the coda-correlation wavefield. These kernels directly illustrate the sensitivity of measurements, such as traveltime, to variations in model parameters, such as P- and S-wave velocities. Under the theory of how the seismic features on the correlogram construct, we perform the forward simulation of the coda-correlation wavefield bypassing the traditional cross-correlation operation between station pairs. We assume uniformly distributed sources for the coda-correlation wavefield in our simulation, since the real coda-correlation wavefield is constructed by stacking correlations of very late coda waves from thousands of seismic events. We compute the adjoint coda-correlation wavefields by using the features in particular waveform windows. With the time-reversed coda-correlation wavefield and adjoint coda-correlation wavefield, we obtain the sensitivity kernels for different features. We analyse the distribution of the sensitivity kernels of different features in this research. These kernels directly show the fundamental mechanic how these features construct in the coda-correlation wavefield. This research paves the way to use the coda correlation wavefield for tomography of different parts of the deep Earth based on full-waveform inversion.

Trans-Conceptual inference of competing theories

(using the data to decide the physics)

Malcolm Sambridge

In all inverse problems we make conceptual assumptions regarding the problem definition. These can be about the way the model is parameterised, the number of unknowns, the statistical for the observational noise and the numerical approximations involved in solving the forward problem which makes predictions from the model. In the latter case one might be faced with a choice between an approximate, computationally cheap, forward solver that does not fully account for the physics and a more accurate, but numerically costly one.  When is it safe to use the simpler (approximate) theory?

Recently it has been shown that the data themselves can be used to decide these questions, a problem we describe as Trans-conceptual inference (Sambridge et al, 2025). An example in that paper demonstrated the application of trans-conceptual Bayesian sampling to seismic tomography, where travel times were used to decide between competing model parameterization classes and physical forward theories simultaneously. They showed that in their experiments the Bayesian evidence based on 100 seismic rays strongly favoured the (correct) ray theory over a straight ray approximation.

Here we perform a combined analytic and numerical study to examine the question of how Bayesian evidence between competing ray theories depends on measurement noise, number of ray paths and knowledge of seismic structure. We conclude that in many realistic situations travel time data contain sufficient information to easily identify ray theory as the correct forward model. As a consequence the straight ray approximation is insufficient to represent the physics and introduces significant error into any analysis where it is used.

Adaptive Ambient Noise Tomography for Natural Hydrogen Exploration: A Case Study from the Gawler Craton, South Australia

Shixian Dong, Chengxin Jiang, Caroline M. Eakin, Louis Moresi, Meghan S. Miller, Graham Heinson, H2EX Ltd 

Natural hydrogen is emerging as a low-cost, clean, and renewable energy source of increasing interest. However, standardized exploration approaches are still lacking, particularly in cratonic regions without sedimentary cover where conventional basin-focused techniques are not directly applicable. This study investigates a ~2,500 km² area in eastern Eyre Peninsula, South Australia, within exploration permit PEL 691 held by H2EX Ltd. The region overlies the Gawler Craton and is transected by several major fault systems, which may provide migration pathways and storage spaces for natural hydrogen. 

In 2024, two one-month passive seismic surveys were conducted using dense nodal arrays. The first regional deployment consisted of 150 stations with an average spacing of 3–4 km. A subsequent survey targeted key fault zones with 100 additional stations, reducing the spacing to 1–1.5 km in critical segments. To integrate datasets across these contrasting scales, we developed a Poisson–Voronoi adaptive tomographic inversion method. This approach preserves imaging coverage in sparsely sampled areas while achieving high resolution in densely instrumented zones, as validated through extensive resolution tests. 

The resulting shear-wave velocity model resolves crustal structures from the surface to ~15 km depth, delineating major fault zones, crystalline basement features, and localized sedimentary layers. These results provide new insights into the structural controls on natural hydrogen generation, migration, and accumulation in crystalline terranes. More broadly, the study demonstrates an effective seismic imaging strategy for resource exploration in geologically complex settings, offering methodological and theoretical support for the future development of natural hydrogen as a strategic energy resource.