https://arxiv.org/api/Ldv8ILqRUNrgWvQRPesm28iPXHg 2026-04-03T08:36:51Z 8544 135 15 http://arxiv.org/abs/2603.00383v1 2026-02-27T23:51:36Z

Simulating long-term, fully dynamic sequences of earthquakes and aseismic slip (SEAS) on geometrically complex fault networks remains computationally demanding due to the cost of resolving elastodynamic interactions. Although high-performance computing improves feasibility, simulations remain expensive, particularly for multicycle evolution, motivating the widespread use of quasi-dynamic approximations based on radiation damping. Here we present an efficient numerical framework for fully elastodynamic SEAS simulations on complex fault networks. The method adopts a divide-and-conquer strategy in which elastodynamic self-effects and fault-to-fault interactions are treated separately using boundary integral formulations tailored to each interaction type. Self-interactions along planar faults are computed using a non-replicating spectral boundary integral formulation that eliminates periodic-image artifacts, while interactions between arbitrarily oriented faults are evaluated through a fully dynamic space-time boundary integral representation accelerated by hierarchical matrices (H-matrices). A key advance is a selective H-matrix compression strategy based on fault-wise assembly of independent binary trees, enabling low-rank approximation of long-range interactions while preserving near-field accuracy and excluding self-effects from the hierarchical structure. Additional efficiency arises from physics-informed truncation of elastodynamic histories using mode-dependent time windows and causality-based kernel truncation. Benchmark multi-fault simulations validate accuracy against reference uncompressed solutions. The method reduces interaction complexity from O(N^3) to O(N^2 log N), yielding up to three orders of magnitude speedup and an order-of-magnitude memory reduction for typical problem sizes (~3e10 degrees of freedom), enabling fully dynamic SEAS simulations on workstation hardware.

2026-02-27T23:51:36Z Federico Ciardo Pierre Romanet http://arxiv.org/abs/2504.16009v2 2026-02-27T16:14:42Z

This work presents a macroscopic model for the flow of two immiscible and incompressible fluids within inhomogeneous porous media. At the pore scale, the flow is governed by the full Navier-Stokes equations while the phase interface evolution is described by the Cahn-Hilliard equation. Applying the volume averaging method, we rigorously derive upscaled equations that characterize the Darcy-scale behavior of the two-phase system. The derivation yields unclosed terms originating from spatial derivations, which are subsequently closed by modeling them as functions of averaged quantities and specific transport coefficients. These coefficients are evaluated by solving localized closure problems defined on representative elementary volumes (REVs). A key contribution of this study is the formal incorporation of wetting behavior into the averaged chemical potential. We further discuss the theoretical distinctions between the proposed framework and standard empirical two-phase Darcy models. Finally, numerical simulations of the upscaled equations are performed, demonstrating the model's capability to capture essential two-phase flow characteristics in porous media.

2025-04-22T16:16:19Z 26 pages,8 figures Chunhua Zhang Peiyao Liu Cheng Peng Lian-Ping Wang Zhaoli Guo http://arxiv.org/abs/2603.00214v1 2026-02-27T15:42:05Z

LLM agents are increasingly used for code generation, but physics-based simulation poses a deeper challenge: natural-language descriptions of simulation models are inherently underspecified, and different admissible resolutions of implicit choices produce physically valid but scientifically distinct configurations. Without explicit detection and resolution of these ambiguities, neither the correctness of the result nor its reproducibility from the original description can be assured. This paper investigates agentic scientific simulation, where model construction is organized as an execution-grounded interpret-act-validate loop and the simulator serves as the authoritative arbiter of physical validity rather than merely a runtime. We present JutulGPT, a reference implementation built on the fully differentiable Julia-based reservoir simulator JutulDarcy. The agent combines structured retrieval of documentation and examples with code synthesis, static analysis, execution, and systematic interpretation of solver diagnostics. Underspecified modelling choices are detected explicitly and resolved either autonomously (with logged assumptions) or through targeted user queries. The results demonstrate that agent-mediated model construction can be grounded in simulator validation, while also revealing a structural limitation: choices resolved tacitly through simulator defaults are invisible to the assumption log and to any downstream representation. A secondary experiment with autonomous reconstruction of a reference model from progressively abstract textual descriptions shows that reconstruction variability exposes latent degrees of freedom in simulation descriptions and provides a practical methodology for auditing reproducibility. All code, prompts, and agent logs are publicly available.

2026-02-27T15:42:05Z Knut-Andreas Lie Olav Møyner Elling Svee Jakob Torben http://arxiv.org/abs/2603.03351v1 2026-02-26T17:17:03Z

Accurate subgrid-scale closures are essential for weather/climate models, where predicting extreme events is critical. Traditional closures have structural errors, e.g., producing excessive diffusion that dampens extremes. Artificial intelligence has gained attention for closure modeling, but the prediction of extreme events remains challenging. Supervised offline learning needs abundant high-fidelity training data and can lead to instabilities. Online learning algorithms are emerging as an alternative, but reliance on differentiable numerical solvers or scalable optimizers hinders broad use. Here, we introduce SMARL to develop closures for canonical prototypes of atmospheric/oceanic turbulence, using only the enstrophy spectrum, estimated from a few high-fidelity samples, as reward. This reward ensures that the model captures the cascades of scales in these simulations. These online-learned closures enable stable simulations, with up to five orders of magnitude fewer degrees of freedom, that reproduce high-fidelity simulation statistics and capture in particular extremes. We interpret the closures by analyzing the SMARL policy and demonstrate generalization to other flows. The results highlight SMARL as a potent tool for developing closures capable of capturing extremes in atmospheric/oceanic flows, opening new capabilities for effective climate modeling.

2026-02-26T17:17:03Z 21 pages; 8 figures Yifei Guan Lucas Amoudruz Sergey Litvinov Karan Jakhar Rambod Mojgani Petros Koumoutsakos Pedram Hassanzadeh http://arxiv.org/abs/2511.01536v2 2026-02-26T07:03:18Z

Seismic data restoration is a fundamental task in seismic exploration, yet remains challenging under complex and unknown degradations. Traditional model-driven or task-specific learning methods often require retraining for each degradation type and fail to generalize effectively to unseen field data. In this work, we introduce an unsupervised Posterior Sampling Framework (PSF) built upon Score-based Generative Models (SGMs) for unified seismic data restoration. PSF leverages the pre-trained unconditional SGMs as a seismic-aware generative prior and derives a generalized conditional score function associated with the forward operator of each inverse problem. This enables posterior sampling across different seismic restoration tasks without retraining or supervision. Additionally, an adaptive noise-level estimation mechanism is incorporated to dynamically regulate the noise suppression strength during sampling, enhancing flexibility under varying signal-to-noise ratios and degradation conditions.Extensive experiments on seismic denoising, interpolation, compressed sensing, and deconvolution demonstrate that PSF delivers high-quality samples and exhibits robust generalization to out-of-distribution data. These results highlight the potential of SGMs as a universal prior for seismic inverse problems and establish PSF as a flexible framework for unsupervised posterior inference across diverse degradation scenarios.

2025-11-03T12:58:13Z 17 Pages, 13 figures Chuangji Meng Jinghuai Gao Zongben Xu http://arxiv.org/abs/2602.22460v1 2026-02-25T22:49:56Z

Ultrafast lasers create extreme, non-equilibrium thermodynamic conditions that can transiently reach pressures and temperatures comparable to interior core of the earth. Here we show that femtosecond excitation of amorphous silica-hafnia multilayer dielectrics drives the formation of high-pressure crystalline phases of silica including stishovite, seifertite, and the pyrite-type high density structure, within confined subsurface regions.Using TEM, SAED, and 4D-STEM, we directly map nanoscale phase evolution and identify crystalline motifs embedded inside laser generated blisters.Complementary molecular dynamics simualtions reveal the thermodynamic pathway underlying these transformations, where rapid electronic pressure initiates densification and octahedral coordination, followed by temperature driven crystallization and displacive transitions during ultrafast quenching. The resulting polymorphs reflects a dual-stage pathway inaccessible under equilibrium processing. Our results establish femtosecond laser excitation as a viable route to synthesize and stabilize ultrahigh-density high pressure silica phases under ambient conditions, without a diamond anvil cell, with implications for laser-damage mechanisms, high-energy-density materials, and planetary physics.

2026-02-25T22:49:56Z Mohamed Yaseen Noor Aram Yedigaryan Gabriel Calderon Arshak Tsaturyan Elena Kanchan Jinwoo Hwang Carmen S. Menoni Jean-Philippe Colombier Enam Chowdhury http://arxiv.org/abs/2602.22293v1 2026-02-25T15:23:01Z

River systems operate as inherently interconnected continuous networks, meaning river hydrodynamic simulation ought to be a systemic process. However, widespread hydrology data scarcity often restricts data-driven forecasting to isolated predictions. To achieve systemic simulation and reduce reliance on river observations, we present GraphRiverCast (GRC), a topology-informed AI foundation model designed to simulate multivariate river hydrodynamics in global river systems. GRC is capable of operating in a "ColdStart" mode, generating predictions without relying on historical river states for initialization. In 7-day global pseudo-hindcasts, GRC-ColdStart functions as a robust standalone simulator, achieving a Nash-Sutcliffe Efficiency (NSE) of approximately 0.82 without exhibiting the significant error accumulation typical of autoregressive paradigms. Ablation studies reveal that topological encoding serves as indispensable structural information in the absence of historical states, explicitly guiding hydraulic connectivity and network-scale mass redistribution to reconstruct flow dynamics. Furthermore, when adapted locally via a pre-training and fine-tuning strategy, GRC consistently outperforms physics-based and locally-trained AI baselines. Crucially, this superiority extends from gauged reaches to full river networks, underscoring the necessity of topology encoding and physics-based pre-training. Built on a physics-aligned neural operator architecture, GRC enables rapid and cross-scale adaptive simulation, establishing a collaborative paradigm bridging global hydrodynamic knowledge with local hydrological reality.

2026-02-25T15:23:01Z 26 pages, 5 figures, 3 extended data tables, 3 extended data figures Hancheng Ren Gang Zhao Shuo Wang Louise Slater Dai Yamazaki Shu Liu Jingfang Fan Shibo Cui Ziming Yu Shengyu Kang Depeng Zuo Dingzhi Peng Zongxue Xu Bo Pang http://arxiv.org/abs/2602.20067v2 2026-02-25T09:33:15Z

Some turbulent flows self-organize into large-scale structures, rather than breaking up into ever-smaller scales. Underpinning this phenomenon is the existence of two sign-definite quantities which are conserved by the dynamics. Two-dimensional turbulence is a prime example, where large-scale mean flows, termed condensates, spontaneously emerge. We review a perturbative theoretical framework for the statistical description of such inhomogeneous turbulence, offering new perspectives on the role of the two conserved quantities. We illustrate the universal properties of the theory, comparing results from two-dimensional Navier-Stokes to those from the large-scale-quasi-geostrophic equation. These two models are limiting cases of the shallow water quasi-geostrophic equation, the former exhibiting long-range fluid element interactions, while the latter has local interactions. We then demonstrate these theoretical ideas in two new settings: first, in rotating three-dimensional turbulence, where two-dimensional condensates are known to form. Considering jet-type condensates, we derive the mean-flow profile and discuss a surprising symmetry breaking. Second, we vary the Rossby deformation radius in the shallow water quasi-geostrophic equation. We obtain novel domain-spanning condensates in all tested regimes and show that they follow two-dimensional Navier-Stokes for deformation radii above the forcing scale, and the large-scale quasi-geostrophic equation for those below, demonstrating the power of these asymptotic models.

2026-02-23T17:22:54Z Anna Frishman Sébastien Gomé Anton Svirsky http://arxiv.org/abs/2603.03346v1 2026-02-24T18:48:15Z

Modeling the unsaturated behavior of porous materials with multimodal pore size distributions presents significant challenges, as standard hydraulic models often fail to capture their complex, multi-scale characteristics. A common workaround involves superposing unimodal retention functions, each tailored to a specific pore size range; however, this approach requires separate parameter identification for each mode, which limits interpretability and generalizability, especially in data-sparse scenarios. In this work, we introduce a fundamentally different approach: a physics-constrained machine learning framework designed for meta-modeling, enabling the automatic discovery of closed-form mathematical expressions for multimodal water retention curves directly from experimental data. Mathematical expressions are represented as binary trees and evolved via genetic programming, while physical constraints are embedded into the loss function to guide the symbolic regressor toward solutions that are physically consistent and mathematically robust. Our results demonstrate that the proposed framework can discover closed-form equations that effectively represent the water retention characteristics of porous materials with varying pore structures. To support third-party validation, application, and extension, we make the full implementation publicly available in an open-source repository.

2026-02-24T18:48:15Z Yejin Kim Hyoung Suk Suh http://arxiv.org/abs/2602.20984v1 2026-02-24T15:05:41Z

Scientific exploration of the UK and Ireland's subsurface has made important contributions to scholarship and prosperity for people and the planet, including economic growth, sustainable use of natural resources, storage of greenhouse gases, and inspiring curiosity about the Earth beneath our feet. This article outlines a vision for an array of seismological instruments spanning the UK and Ireland, UKI Array, augmented by other types of geophysical sensors, to maximise the value offered by existing equipment pools. The mission is to research natural phenomena and structure in the deep and shallow Earth, to solve problems concerning hazards and resources, to connect scientists to schools and the broader public, and thus to inspire a new generation to learn about geophysics. The vision was created through a community driven process of engagement and participation. This paper describes the concept and design of the UKI-Array; a companion paper discusses related opportunities and potential applications.

2026-02-24T15:05:41Z Andrew Curtis Karen Lythgoe Stephen P. Hicks Lidong Bie Dominik Strutz Emma Chambers Brian Baptie Dave Cornwell Juliane Huebert Jessica Irving Glenn Jones Sergei Lebedev Walid Ben Mansour Aideliz Montiel Álvarez Stuart Nippress Koen Van Noten Tim Pharaoh Romesh Palamakumbura Nick Rawlinson Pablo Rodriguez Salgado James Verdon Chuanbin Zhu Wen Zhou Jelle Assink Ian Bastow Dino Bindi Tom Blenkinsop Raffaele Bonadio Michael Braim Tim Craig Elizabeth Day Giovanni Diaferia Stuart Dunning Ben Edwards Ake Fagereng Stewart Fishwick Amy Gilligan David Green David Healy Anna Horleston Mark Ireland Jenny Jenkins Jessica Johnson Mike Kendall Tom Kettlety Duygu Kiyan Paula Koelemeijer Rita Kounoudis Victoria Lane Chuanchuan Lu Alan MacDonald Fabrizio Magrini Auggie Marignier Carl Martin Martin Möllhoff Iain Neill Andy Nowacki Bob Paap Simone Pilia Sjoerd de Ridder Elmer Ruigrok Peidong Shi Anna Stork Alice Turner Jim Whiteley Anton Ziolkowski http://arxiv.org/abs/2512.09844v2 2026-02-24T12:55:10Z

The Great Oxidation Event (GOE) was a $200$ Myr transition circa 2.4 billion years ago that converted the Earth's anoxic atmosphere to one where molecular oxygen (O$_2$) was abundant (volume mixing ratio $>10^{-4}$). This significant rise in O$_2$ is thought to have substantially throttled hydrogen (H) escape and the associated water (H$_2$O) loss. Atmospheric estimations from the GOE onward place O$_2$ concentrations ranging between 0.1% to 150% PAL, where PAL is the present atmospheric level of 21% by volume. In this study we use WACCM6, a three-dimensional Earth System Model to simulate Earth's atmosphere and predict the diffusion-limited escape rate of hydrogen due to varying O$_2$ post-GOE. We find that O$_2$ indirectly acts as a control valve on the amount of hydrogen atoms reaching the homopause in the simulations: less O$_2$ leads to decreased O$_3$ densities that reduce local tropical tropopause temperatures by up to 17 K, which increases H$_2$O freeze-drying and thus reduces the primary source of hydrogen in the considered scenarios. The maximum differences between all simulations in the total H mixing ratio at the homopause and the associated diffusion-limited escape rates are a factor of 3.2 and 4.7, respectively. The prescribed CH$_4$ mixing ratio (0.8 ppmv) sets a minimum diffusion escape rate of $\approx 2 \times 10^{10}$ mol H yr$^{-1}$, effectively a negligible rate when compared to pre-GOE estimates ($\sim10^{12}-10^{13}$ mol H yr$^{-1}$). Because the changes in our predicted escape rates are comparatively minor, our numerical predictions support geological evidence that the majority of Earth's hydrogen escape occurred prior to the GOE. Our work demonstrates that estimations of how the hydrogen escape rate evolved through Earth's history requires 3D chemistry-climate models which include a global treatment of water vapour microphysics.

2025-12-10T17:23:09Z 35 pages, 7 figures, 1 table. Accepted for publication in Climate of the Past Gregory J. Cooke Dan R. Marsh Catherine Walsh Felix Sainsbury-Martinez Marrick Braam http://arxiv.org/abs/2602.20806v1 2026-02-24T11:46:42Z

Particles collected from the asteroid Ryugu by the Hayabusa2 spacecraft offer a unique opportunity to investigate the magnetic record of the primitive solar system, as any terrestrial magnetic contamination is minimal and can be accounted for. In previous studies, stepwise alternating field demagnetization (AFD) measurements of natural remanent magnetization (NRM) records have been conducted on seven Ryugu particles. However, due to the limited number of samples, there is no consensus regarding the interpretation of the results of these measurements. To address this problem, we performed stepwise AFD measurements of the NRM on 28 Ryugu particles. Twenty-three of the particles exhibited one or two stable NRM components, whereas the remaining five did not. Isothermal remanent magnetization-based paleointensity values derived from stable NRM components varied by more than one order of magnitude. These NRM characteristics were consistent with those observed in previous studies. Therefore, as a reflection of the original nature of the NRM record, some Ryugu particles exhibited stable NRM components, whereas others did not. The Ryugu particles investigated in this study and those from a previous study exhibited spatially inhomogeneous NRM directions within individual particles, constraining the NRM acquisition time to before the final solidification of the current Ryugu particles. A mechanism of remanence acquisition that can explain the observed NRM characteristics is a chemical remanent magnetization associated with the growth of framboidal magnetite during aqueous alteration in Ryugu's parent body.

2026-02-24T11:46:42Z Published in JGR Planets Journal of Geophysical Research: Planets 131 (2026) e2025JE009265 Masahiko Sato Yuki Kimura Tadahiro Hatakeyama Tomoki Nakamura Satoshi Okuzumi Sei-ichiro Watanabe Seiji Sugita Satoshi Tanaka Shogo Tachibana Hisayoshi Yurimoto Takaaki Noguchi Ryuji Okazaki Hikaru Yabuta Hiroshi Naraoka Kanako Sakamoto Toru Yada Masahiro Nishimura Aiko Nakato Akiko Miyazaki Kasumi Yogata Masanao Abe Tatsuaki Okada Tomohiro Usui Makoto Yoshikawa Takanao Saiki Fuyuto Terui Satoru Nakazawa Yuichi Tsuda 10.1029/2025JE009265 http://arxiv.org/abs/2506.19039v3 2026-02-23T16:37:37Z

Geological data show that, early in its history, the Earth had a large-scale magnetic field with an amplitude comparable to the one of the present geomagnetic field. However, its origin remains enigmatic and various mechanisms have been proposed to explain the Earth's field over geological time scales. Here, we critically evaluate whether tidal forcing could explain the ancient geodynamo, by combining constraints from geophysical models of the Earth-Moon system and predictions from turbulence studies. Our analysis shows that lunar tidal forcing could have been sufficiently strong before $-3.25$ Gy to trigger turbulence within the Earth's core, and potentially to sustain dynamo action during that interval. Then, we propose new scaling laws for the magnetic field amplitude $B$. We expect the latter to scale as $B \propto β^{4/3}$, where $β$ is the equatorial ellipticity of the liquid core, if the turbulence involves weak interactions of three-dimensional inertial waves. Alternatively, in the regime of strong tidal forcing, the expected scaling becomes $B \propto β$. When extrapolated to the Earth's core, it suggests that tidal forcing alone was too weak to possibly explain the ancient geomagnetic field. Therefore, our study indirectly favours another origin for the early Earth's dynamo on long time scales (e.g. exsolution of light elements atop the core, or thermal convection due to secular cooling).

2025-06-23T18:57:30Z 28 pages, 14 figures. Published on 28 January 2026 in C. R. Geosci Comptes Rendus. Geoscience, 2026, 358 (G1), 1-26 Jérémie Vidal David Cébron 10.5802/crgeos.324 http://arxiv.org/abs/2603.03344v1 2026-02-23T05:07:06Z

Earthquake detection and seismic phase picking are fundamental yet challenging tasks in seismology due to low signal-to-noise ratios, waveform variability, and overlapping events. Recent deep-learning models achieve strong results but rely on large datasets and heavy backpropagation training, raising concerns over efficiency, interpretability, and sustainability. We propose GreenPhase, a multi-resolution, feed-forward, and mathematically interpretable model based on the Green Learning framework. GreenPhase comprises three resolution levels, each integrating unsupervised representation learning, supervised feature learning, and decision learning. Its feed-forward design eliminates backpropagation, enabling independent module optimization with stable training and clear interpretability. Predictions are refined from coarse to fine resolutions while computation is restricted to candidate regions. On the Stanford Earthquake Dataset (STEAD), GreenPhase achieves excellent performance with F1 scores of 1.0 for detection, 0.98 for P-wave picking, and 0.96 for S-wave picking. This is accomplished while reducing the computational cost (FLOPs) for inference by approximately 83% compared to state-of-the-art models. These results demonstrate that the proposed model provides an efficient, interpretable, and sustainable alternative for large-scale seismic monitoring.

2026-02-23T05:07:06Z Yixing Wu Shiou-Ya Wang Dingyi Nie Sanket Kumbhar Yun-Tung Hsieh Yun-Cheng Wang Po-Chyi Su C. -C. Jay Kuo http://arxiv.org/abs/2510.14407v3 2026-02-22T16:42:49Z

Forecast models in statistical seismology are commonly evaluated with log-likelihood scores of the full distribution P(n) of earthquake numbers, yet heavy tails and out-of-range observations can bias model ranking. We develop a tail-aware evaluation framework that estimates cell-wise P(n) using adaptive Gaussian kernel density estimation and tests three strategies for handling out-of-range counts. Using the AoyuX platform, we perform a ~25-year month-by-month pseudo-prospective forecast experiment in the China Seismic Experimental Site (CSES), comparing Epidemic-Type Aftershock Sequence (ETAS) model with a homogeneous background (ETASμ) to a spatially heterogeneous variant (ETASμ(x,y)) across six spatial resolutions and five magnitude thresholds. Empirical probability density functions (PDFs) of counts per cell are well described by power laws with exponents a = 1.40 +- 0.21 across all settings. Using previous theoretical results, this provides a robust estimate of the productivity exponent, α = 0.57 +- 0.08 using a b-value equal to 0.8, providing a valuable quantification of this key parameter in aftershock modeling. Model ranking is sensitive to how the tail of the full distribution P(n) of earthquake counts is treated: power law extrapolation is both theoretically justified and empirically the most robust. Cumulative information gain (CIG) shows that ETASμ(x,y) outperforms ETASμ in data-rich configurations, whereas in data-poor settings stochastic fluctuations dominate. A coefficient-of-variation analysis of per-window log-likelihood differences distinguishes genuine upward trends in CIG from noise-dominated fluctuations. By aligning a fat-tail-aware scoring methodology with an open testing platform, our work advances fair and statistically grounded assessment of earthquake forecasting models for the CSES and beyond.

2025-10-16T08:06:40Z 30 pages, 7 figures, 1 table Jiawei Li Qingyuan Zhang Didier Sornette