<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>NicheSTEM Open-Access Full-Text Archive</title>
    <link>https://nvkcnbsnkhie.org/</link>
    <description>Peer-Reviewed Pure STEM Research Repository (3035 Papers)</description>
    <language>en-us</language>
    <lastBuildDate>Mon, 16 Nov 2026 05:00:01 +0800</lastBuildDate>
    <atom:link href="https://nvkcnbsnkhie.org/feed.xml" rel="self" type="application/rss+xml"/>
    <atom:link href="https://pubsubhubbub.appspot.com/" rel="hub"/>
    <item>
      <title>Spatiotemporal learning improves quantification of pore-space evolution and fluid displacement</title>
      <link>https://nvkcnbsnkhie.org/?paper=3035</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3035</guid>
      <category>Rheology & Microfluidics</category>
      <description>Pore-space evolution and fluid displacement influence how rocks transmit and store fluids, affecting geological carbon storage, underground hydrogen storage and groundwater contaminant transport. A single time-resolved micro-CT experiment may produce tens to hundreds of three-dimensional images, and segmentation is required to quantify physical properties and run image-based simulations. Traditional methods such as thresholding perform poorly on noisy, low-contrast images and need manual adjustment per image; 3D deep learning networks segment each image without temporal information, so their errors vary between images and could be mistaken for physical change. Here we introduce a transformer-based 4D segmentation framework combining spatial and temporal information. It uses sequential images to resolve ambiguous phase assignments and separate physical change from segmentation error. The method is tested on images of mineral dissolution: it measures pore-volume growth and rock boundaries more accurately than an equivalent 3D model that ignores temporal information. It also captures the increase in permeability verified against reference segmentations and flow simulations. The advantage transfers to another dissolution sample without retraining. On two other datasets of fluid displacement, it reduces spurious changes in predicted rock and porosity fluctuations and accurately identifies filling events. Spatiotemporal segmentation turns sequences of hundreds of three-dimensional images into more reliable quantitative measurements of pore-space evolution and fluid displacement without per-image parameter adjustment to support the analysis of time-resolved CT imaging. (DOI: 10.48550/arXiv.2611.17111)</description>
    </item>
    <item>
      <title>Sub-Ice: semi-automated mapping of basal channels from ice shelf surface elevation data</title>
      <link>https://nvkcnbsnkhie.org/?paper=3034</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3034</guid>
      <category>Glaciology & Paleoclimatology</category>
      <description>Basal channels, troughs carved into ice shelf bases by buoyant plumes, modulate basal melt and can weaken ice shelves. Through hydrostatic adjustment they have a surface expression, yet their geometry is largely mapped manually, limiting the number of channels studied and the consistency between mappings. We present Sub-Ice, a semi-automated tool that maps basal channels from surface digital elevation models. From user-defined start and end points, Sub-Ice traces the channel centerline by iteratively locating elevation minima, extracts cross-sectional profiles and delineates channel edges using one of three methods (slope threshold, profile knee point or peak prominence), producing georeferenced centerlines and edges and along-channel width and depth. Applied to 34 channels on Roi Baudouin Ice Shelf, East Antarctica, Sub-Ice centerlines differ from manually digitized ones by a median of 132 m, at the upper end of the differences between manual records from different years (74-132 m), while following the channel trough more closely. Absolute channel widths depend on the edge method and the width of the edge search window, which we calibrate against a slope-based criterion. Rather than seeking a single "true" edge, Sub-Ice applies explicit, reproducible criteria, enabling consistent comparison of channel geometry across channels, ice shelves and time. (DOI: 10.48550/arXiv.2611.14219)</description>
    </item>
    <item>
      <title>Tidal potential measured with distant optical clocks</title>
      <link>https://nvkcnbsnkhie.org/?paper=3033</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3033</guid>
      <category>Glaciology & Paleoclimatology</category>
      <description>Fibre-linked optical clocks sense tidal changes of the gravitational potential, but separating overlapping tidal components and quantifying their precision remain challenging. Using strontium and ytterbium clocks 692 km apart, we measure over 353 days a response of 0.95 (0.11) to the combined solid-Earth and ocean-loading tide and 1.01 (0.13) for the lunar semidiurnal group, and separate the lunar and solar-frequency semidiurnal contributions. A signal-to-noise model based on the actual sampling accounts for the lunar precision and guides future clock-network observations. (DOI: 10.48550/arXiv.2611.14293)</description>
    </item>
    <item>
      <title>Differentiable Inverse Rendering Recovers Intrinsic Properties from Intravascular Optical Coherence Tomography</title>
      <link>https://nvkcnbsnkhie.org/?paper=3032</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3032</guid>
      <category>Microdosimetry & Hemodynamics</category>
      <description>Medical images encode intrinsic tissue properties that indicate disease progression, yet these properties remain entangled with the system-dependent physics of image formation. Differentiable inverse rendering asks whether they are recoverable by inverting the imaging system's forward physics rather than through post-hoc correction. We address this in optical coherence tomography (OCT), focusing on the optical attenuation coefficient, a per-pixel marker of pathology and cellular composition. Conventional recovery relies on closed-form inversions limited to simplified light-tissue models, or learned methods coupled to specific imaging configurations. We instead recover per-pixel attenuation from a single OCT image by gradient descent through a differentiable physics-based forward model. We verified the method against analytical solutions, validated it on five in vitro phantoms, and applied it without modification to 29{,}269 clinical intravascular OCT images from 128 patients. The inversion needs no training data or segmentations and is portable across platforms and retrospective archives. Replacing the forward model or recovering another optical parameter leaves the optimization procedure unchanged. We demonstrate per-pixel recovery is tractable because gradient information scales to this high-dimensional problem, where gradient-free methods become intractable. Our work supports quantitative tissue characterization from OCT images and provides a framework for extension across imaging modalities and organ systems. (DOI: 10.48550/arXiv.2611.12686)</description>
    </item>
    <item>
      <title>Quantum-walk transport oracles and amplitude estimation of a scalar range statistic in proton therapy</title>
      <link>https://nvkcnbsnkhie.org/?paper=3031</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3031</guid>
      <category>Microdosimetry & Hemodynamics</category>
      <description>Quantum amplitude estimation (QAE) promises a quadratic reduction in the sampling cost of Monte Carlo (MC) particle transport, but two obstacles have kept it from radiotherapy physics: the lack of an efficient unitary encoding of charged-particle transport, and the mismatch between QAE's single-scalar output and the per-voxel structure of dose calculation. We address both. We construct a discrete-time quantum-walk transport oracle for protons in which an energy register is driven by an energy-dependent coin encoding a continuous-slowing-down surrogate of the stopping power, so that stochastic energy loss and range straggling follow from the circuit dynamics. The oracle is unitary, with one fresh coin qubit per step, and its gate count grows linearly in the number of energy levels. On it we formulate a scalar range statistic, the probability that a proton penetrates beyond a distal depth threshold, as a directly marked amplitude, and estimate it by maximum-likelihood amplitude estimation (MLQAE) with Grover iterations and finite shots. The oracle reproduces a Bragg-shaped depth-deposit profile agreeing with a classical MC of identical rules to within 0.27%, and Grover amplification matches $\sin^2((2k+1)θ)$ to machine precision up to $k=32$. At matched budget the MLQAE error falls as $N^{-0.93}$ against $N^{-0.49}$ for sampling, 4.8 times smaller at the largest budget. A three-gate resource analysis (output structure, amortization, error-corrected wall-clock) finds full dose maps structurally unfavourable irrespective of hardware, and the scalar statistic short of a wall-clock crossover at clinical scale: the oracle's derived cost of $1.0 \times 10^6$ logical gates per call ($3.5 \times 10^7$ with rotation synthesis), dominated by the multiplexed coin, is 160 to 5400 times the permitted budget. We report the reduction that would be required and identify the term responsible for it. (DOI: 10.48550/arXiv.2611.13475)</description>
    </item>
    <item>
      <title>ISLA: A Coordinate-Free, Measure-Weighted Transformer for Boundary-Driven PDEs</title>
      <link>https://nvkcnbsnkhie.org/?paper=3030</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3030</guid>
      <category>Rheology & Microfluidics</category>
      <description>A foundational principle of physics is that all coordinate systems are equally valid: the behavior of a PDE-governed system does not depend on the origin or orientation of the frame, nor on the units or mesh used to write it down, and neither should a surrogate model. Yet the neural surrogates with the lowest errors on today's aerodynamics benchmarks lack these invariances and fail under simple transformations of the same problem: for models trained on PDEs of cars in one pose (i.e., translation and rotation) and from one mesher, rotating the full PDE raises the surface-pressure error from 4% to more than 100%, and re-sampling its boundary raises it to about 30%. We introduce ISLA (Invariant Slice Attention), a slice-attention transformer for steady, boundary-driven PDEs in which no learned operation receives a coordinate and every sum over the boundary is a quadrature. Positions enter only as distances and relative angles to learned slice anchors, lengths are measured in a unit computed from the boundary, and interior points query the boundary through a tree-accelerated learned kernel without influencing one another. With every boundary cell a token and an exact boundary read, ISLA is equivariant to rigid motions and reflections, invariant to the units of its declared data, and consistent under mesh refinement, while at every setting each query is independent of all other queries. On AeroZoo, 2,816 flows that vary in shape, size, speed, and boundary conditions, its mean reference-scaled velocity and pressure errors are at least 3.1 and 3.3 times lower than any baseline's. On AirfRANS its errors are second only to those of a boundary-integral model (GLOBE) and at least 1.6 times lower than those of Transolver and GeoTransolver. (DOI: 10.48550/arXiv.2611.17141)</description>
    </item>
    <item>
      <title>Discovering Equations for the Thermodynamics of South Asian Monsoon Precipitation</title>
      <link>https://nvkcnbsnkhie.org/?paper=3029</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3029</guid>
      <category>Glaciology & Paleoclimatology</category>
      <description>South Asian monsoon precipitation depends on the thermodynamic structure of the lower troposphere, but how its ingredients combine to shape precipitation remains unclear, partly because this dependence is nonlocal in the vertical. We use symbolic distillation, which combines neural networks and symbolic regression, to discover interpretable precipitation equations from 21 years of June-August reanalysis data. Neural networks first learn the vertical thermodynamic structure most relevant to precipitation, condensing each variable into a single interpretable feature not tied to a particular vertical grid. Then, symbolic regression derives closed-form equations from these features and the surface state. Our best equation explains 1.5 times as much precipitation variance as a widely used buoyancy-based diagnostic and recovers about 85% of neural network skill while using about a thousand times fewer parameters. Unlike the buoyancy diagnostic, which adds moisture and instability, the equations treat them as competing controls, with the heaviest precipitation occurring when the lower free troposphere is humid. Over ocean, sensible heating is associated with additional precipitation mainly when near-surface air is warm and moist, improving predicted precipitation and its diurnal cycle. Although the learned features were optimized only for precipitation prediction, the equations largely respond to each feature as expected from convective physics, consistent with the reanalysis. These responses reduce the dependence of monsoon precipitation on the entire vertical column to a few simple, testable constraints that could guide data-driven precipitation models to make them more physically consistent. (DOI: 10.48550/arXiv.2611.14401)</description>
    </item>
    <item>
      <title>Towards Distributed Quantum Computing: Quantum Gate Teleportation via Satellite-Distributed Entanglement under Atmospheric Depolarization</title>
      <link>https://nvkcnbsnkhie.org/?paper=3028</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3028</guid>
      <category>Glaciology & Paleoclimatology</category>
      <description>We investigate quantum gate teleportation as a primitive for satellite-enabled distributed quantum computing. In the considered architecture, a satellite distributes entangled photon pairs to two optical ground stations, which transfer the shared quantum resource to remote quantum processing units. This entanglement is then used to implement a nonlocal controlled-NOT gate between two circuit qubits located in different processors. We formulate the minimal four-qubit protocol, consisting of two network qubits that provide the entangled resource and two circuit qubits on which the distributed gate is performed. Atmospheric propagation effects in the gravitational field are modeled through an effective two-qubit depolarizing channel whose parameters depend on atmospheric and optical quantities, including the Fried parameter and receiver aperture. The performance of the protocol is quantified by the fidelity between the ideal CNOT output state and the state produced by noisy quantum gate teleportation. Our results show that the fidelity is governed by the quality of the distributed entanglement and depends jointly on atmospheric coherence, aperture size, and effective depolarization strength. These findings connect physical-layer modeling of satellite optical channels with gate-level performance in distributed quantum computing. (DOI: 10.48550/arXiv.2611.14623)</description>
    </item>
    <item>
      <title>TMAL: Topology-preserving Multi-offset Affinity Learning for Curvilinear Structure Segmentation</title>
      <link>https://nvkcnbsnkhie.org/?paper=3027</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3027</guid>
      <category>Microdosimetry & Hemodynamics</category>
      <description>Curvilinear structures such as retinal vessels and surface cracks need both pixel-level accuracy and structural continuity. Pixel-wise supervision does not model the connections between pixels, and topology-based losses still produce ordinary masks. Connectivity-based methods predict these connections directly, but they use binary connectivity targets over a fixed local range. Such labels give the same value to thin segments and to thick regions, and they cannot describe connections at larger offsets. To address these limitations, we propose Topology-preserving Multi-offset Affinity Learning (TMAL). TMAL uses soft affinities that emphasize thin regions near the boundary, and it predicts directional affinities at several offsets to reach further. It then combines these affinities with weights that come from the decoder features, so local continuity and wider context are used together. An auxiliary segmentation branch adds region-level supervision and keeps the overlap accuracy high. On the retinal vessel dataset DRIVE and the crack inspection dataset DeepCrack, TMAL reaches the highest Dice, IoU, and clDice scores, and it greatly reduces the connected-component error $Δβ_0$. Code is available at https://github.com/suho-han/TMAL. (DOI: 10.48550/arXiv.2611.13511)</description>
    </item>
    <item>
      <title>CURRENT: A Recurrent Neural Network for Concurrent Spatial and Temporal Super-Resolution of 4D Flow MRI</title>
      <link>https://nvkcnbsnkhie.org/?paper=3026</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3026</guid>
      <category>Microdosimetry & Hemodynamics</category>
      <description>Objective: To improve the spatial and temporal resolution of aortic 4D flow MRI through a unified deep learning framework for post-acquisition super-resolution (SR) of reconstructed velocity vector fields. Methods: We propose CURRENT, a recurrent neural network that performs spatial and temporal SR with a single network architecture. Supervised training was conducted using paired low- and high-resolution datasets generated by embedding computational fluid dynamics simulations of aortic flow into in vivo thoracic MRI backgrounds, followed by controlled spatial and temporal downsampling. CURRENT was evaluated for two-fold spatial and temporal SR against 4DFlowNet and for two-fold spatiotemporal SR against sequential spatial and temporal SR pipelines based on 4DFlowNet. Image-based and k-space-based quantitative metrics were considered and conventional interpolation were used as baselines. In addition, the impact of CURRENT's spatiotemporal SR on physics-informed relative pressure estimation was assessed. Finally, the ability of CURRENT to generalize from synthetic training data to in vivo acquisitions was investigated. Results: CURRENT achieved superior spatial SR performance and comparable temporal SR performance to dedicated 4DFlowNet models while providing a unified solution for both tasks. Simultaneous spatiotemporal SR using CURRENT yielded more accurate relative pressure estimates than either spatial or temporal SR alone and outperformed sequential 4DFlowNet-based spatial and temporal SR approaches. Furthermore, the proposed model demonstrated promising generalization to in vivo 4D flow MRI data despite being trained exclusively on synthetic datasets. Conclusion: CURRENT provides a unified framework for spatial and temporal SR of 4D flow MRI. The proposed approach has the potential to enhance physics-based analysis and broaden the applicability of accelerated 4D flow MRI. (DOI: 10.48550/arXiv.2611.14058)</description>
    </item>
    <item>
      <title>AeroZoo: A Benchmark of Well-Posed Flow Problems for Foundation Models of Aerodynamics</title>
      <link>https://nvkcnbsnkhie.org/?paper=3025</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3025</guid>
      <category>Rheology & Microfluidics</category>
      <description>Neural surrogates for computational fluid dynamics (CFD) are trained and ranked on datasets that each vary one design family in one flow domain. Such datasets cannot test the least that an aerodynamic foundation model must do: predict CFD-quality incompressible flow in seconds for new shapes in new enclosures. On DrivAerML, larger models and longer training barely lower Transolver's error; doubling the training cases cuts it by about 12% up to 400 cases, exhausting the dataset. We introduce AeroZoo, a benchmark and training corpus of 3,815 time-mean solutions to the unsteady Reynolds-averaged Navier--Stokes (RANS) equations with a transition model, across seven geometry families, eight enclosure configurations and Reynolds numbers from $10^3$ to $10^8$. Each sample poses a well-posed PDE problem in the form a solver receives it: the boundary meshes grouped by condition type, the prescribed values and the fluid properties. The 2,816 samples with a stable time mean form the canonical benchmark: two tasks, engineering metrics and 22 frozen splits, disjoint in geometry, that hold out geometry families, enclosures, boundary-condition layouts and Reynolds ranges. On the random split, boundary-condition tokens cut GeoTransolver's median velocity error from 76.9% to 8.4% of the inlet speed in single-seed runs (uniform inlet flow: 11.0%). ISLA, a transformer for boundary-driven PDEs, reaches 5.3% over four seeds, about 4 times the change from a finer mesh. On the same test samples, a held-out geometry family barely changes ISLA's error relative to uniform flow, and a held-out enclosure or boundary-condition layout raises it 2.4 to 9.2 times. The data and the scoring code will be released. (DOI: 10.48550/arXiv.2611.17152)</description>
    </item>
    <item>
      <title>Fragility of Aeolian Ripples</title>
      <link>https://nvkcnbsnkhie.org/?paper=3024</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3024</guid>
      <category>Glaciology & Paleoclimatology</category>
      <description>Sand surfaces exposed to the wind are covered by aeolian sand waves of diverse wavelengths, from centimetres for small ripples to decameters for dunes. Sand dunes can be said to be antifragile, since their growth is not prohibited by storms, whereas ripples and megaripples are fragile since they get flattened. The fluid-entrainment hypothesis attributes their fragility to a common cause: the spontaneous aerodynamic entrainment of grains. Here, we challenge this conventional view by revealing its major shortcomings and inconsistencies. An alternative framework, routed in the granular bed rheology, posits that impact ripples are erased when the two-phase flow of ballistically hopping grains on a solid sand surface gives way to a three-phase flow, featuring an intermediate, collision-dominated, fluid-like transport layer. Notably, both martian wind-drag ripples and terrestrial aerodynamic ripples, which emerge via a hydrodynamic instability, are insensitive to such "surface-melting". Their destruction is predicated on the transition from aerodynamically smooth to rough conditions, instead. Our theoretical arguments are well supported by a stability diagram in terms of Shields number and grain diameter, assembled from a carefully reassessed large body of literature, and backed up by dedicated control measurements. (DOI: 10.48550/arXiv.2611.14739)</description>
    </item>
    <item>
      <title>Data-driven Reconstruction of the Phase Dynamics for the El Niño-Southern Oscillation Reveals Seasonality and Statistical Synchronization with the Annual Cycle</title>
      <link>https://nvkcnbsnkhie.org/?paper=3023</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3023</guid>
      <category>Glaciology & Paleoclimatology</category>
      <description>In this study, we treat the El Niño-Southern Oscillation (ENSO) as a rhythmic system periodically forced by the annual cycle. We examine ENSO event seasonality and synchronization between the annual and ENSO cycles. We compute phase time series for both cycles from sea surface temperature data using a data-driven Koopman operator approach. We then use Bayesian inference to reconstruct phase equations describing how the phases of the two oscillators evolve. The reconstructed two-variable phase equations show that the annual cycle modulates the instantaneous frequency of the fundamental ENSO mode. Stochastic simulations of these equations incorporating the estimated stationary noise indicate that El Niño and La Niña events tend to last longer when they start in boreal summer and that the variability of their durations is greater when they start in boreal spring. Furthermore, the averaged form of the phase equations indicates statistical $3:1$ phase locking between the annual and ENSO cycles, whereas no statistical phase locking is found for the other $m:n$ ratios. (DOI: 10.48550/arXiv.2611.14868)</description>
    </item>
    <item>
      <title>Improved Myelin Water Fraction Estimation and Uncertainty Mapping in Multicomponent T2 Relaxometry via NNLS Bootstrap Aggregation</title>
      <link>https://nvkcnbsnkhie.org/?paper=3022</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3022</guid>
      <category>Microdosimetry & Hemodynamics</category>
      <description>Estimating the myelin water fraction (MWF) from multicomponent T2 relaxometry requires solving an ill-conditioned inverse problem, making estimates sensitive to noise. Regularization can improve stability but may introduce bias. Conventional NNLS-based methods also provide point estimates without directly quantifying voxelwise variability. To address these limitations, we propose the non-negative least squares Bootstrap Barycenter (NNLS-BB), combining residual resampling with repeated NNLS reconstruction. We average spectra and derived parameters across bootstrap replicates and use their standard deviations to characterize sensitivity to signal perturbations. The method requires neither an explicit spectral regularization penalty nor a training dataset. NNLS-BB was evaluated on 40,000 simulated voxels with signal-to-noise ratios of 50-1000, against conventional NNLS, two regularized NNLS methods, and a model-informed machine-learning (MIML) method. Scan-rescan repeatability was assessed across 20 white-matter regions in two independent cohorts comprising 13 healthy participants. NNLS-BB achieved the lowest MWF mean absolute error and root mean squared error in all simulated noise regimes. Mean absolute error decreased by 36-41% relative to conventional NNLS, alongside lower error dispersion. Compared with Bayes-Reg-$D^{-1}$ and MIML, NNLS-BB achieved comparable dispersion but substantially lower bias, although MIML reconstructed the full spectrum more accurately. In vivo, NNLS-BB yielded the lowest regional scan-rescan dispersion and highest mean correlation in both cohorts. Bootstrap standard-deviation maps provided complementary information about parameter variability. These results support residual-bootstrap aggregation for improving MWF estimation while providing voxelwise variability maps from a single acquisition. (DOI: 10.48550/arXiv.2611.14395)</description>
    </item>
    <item>
      <title>A first application of split conformal prediction for dose prediction in knowledge-based radiotherapy planning</title>
      <link>https://nvkcnbsnkhie.org/?paper=3021</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3021</guid>
      <category>Microdosimetry & Hemodynamics</category>
      <description>Radiation therapy treatment planning is a labour-intensive process prone to inter-expert variability. Deep learning (DL) dose prediction can accelerate this process, but most DL models do not quantify their reliability. In this study, we construct prediction intervals for DL-based radiotherapy dose prediction using split conformal prediction (CP), a model-agnostic and distribution-free method for uncertainty quantification that provides finite-sample coverage guarantees. We propose patient-level nonconformity scores obtained by aggregating voxelwise scores. Specifically, we formulate a maximum aggregation construction, which guarantees simultaneous coverage of the true dose in all voxels of a future patient with at least a user-defined probability, and a quantile-based construction, which guarantees, with the same probability lower bound, that a future patient attains a prescribed minimum level of voxel coverage. We also investigate groupwise calibration for region-specific guarantees and predictive uncertainty scaling for locally adaptive prediction intervals. The methods are evaluated on head-and-neck cases using a deep ensemble model. The proposed constructions achieved empirical coverage consistent with the nominal levels. Maximum aggregation led to conservative prediction intervals, whereas quantile aggregation produced narrower intervals. Groupwise calibration and uncertainty scaling reduced interval widths in some anatomical regions, while providing region-specific guarantees and local adaptivity. This is the first application of split CP to radiotherapy dose prediction. The explicit patient-level coverage guarantees offer a more interpretable characterisation of prediction reliability than raw uncertainty estimates. Split CP could help clinicians assess the reliability of model predictions and identify cases requiring further review. (DOI: 10.48550/arXiv.2611.14438)</description>
    </item>
    <item>
      <title>A Transferable Pretrained 3D Model for Impedance Inversion with Sparse-Well Prior Conditioning</title>
      <link>https://nvkcnbsnkhie.org/?paper=3020</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3020</guid>
      <category>Glaciology & Paleoclimatology</category>
      <description>Deep-learning-based seismic impedance inversion commonly requires retraining for each target survey, which may generalize poorly beyond the conditioning locations when well control is extremely sparse. We propose a transferable three-dimensional inversion framework that separates offline learning from target-survey deployment. A masked autoencoder first learns seismic structural representations from unlabeled multi-survey data, after which the frozen encoder is incorporated into an inversion network trained on synthetic geological models with sparse-well-derived low-frequency priors. During deployment, the complete inversion network remains fixed; the available conditioning well is used only to construct the low-frequency prior and align the global impedance range. Experiments on the Overthrust model show that the proposed method achieves the lowest full-volume NRMSE and reduces the P95 trace-wise NRMSE from 0.1310 to 0.1074 relative to the best target-trained baseline, indicating fewer localized large errors under different well locations. In a field survey with one conditioning well and three blind wells, it reduces the mean blind-well NRMSE by 20.95% relative to the target-trained U-Net while maintaining the highest correlation. These results demonstrate that the proposed framework transfers useful impedance information beyond the conditioning location with low deployment cost and without retraining the inversion network. (DOI: 10.48550/arXiv.2611.14922)</description>
    </item>
    <item>
      <title>A Diffusive T-Matrix Unifying the Electric and Magnetic Polarizability Tensors in Conducting Layered Media</title>
      <link>https://nvkcnbsnkhie.org/?paper=3019</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3019</guid>
      <category>Glaciology & Paleoclimatology</category>
      <description>Electromagnetic induction is used to detect and classify buried and submerged objects, such as unexploded ordnance in soil and on the seabed. A small object is commonly described by its magnetic polarizability tensor (MPT), which is independent of its position and of the sensor. The MPT assumes that the surrounding medium does not conduct. In seawater at extremely low frequency, or in clay at kilohertz frequencies, this assumption fails at depths comparable to the skin depth. The object then also diverts the current induced in the medium, a galvanic response that the MPT omits. We describe the object instead by its transition matrix (T-matrix), which keeps the independence from position and sensor and includes both responses. It is computed once per medium and frequency from a single finite element problem, for any shape and materials. Its inductive block reduces to the MPT when the medium conducts weakly, and its galvanic block reduces to the electric polarizability tensor in the static limit. We test the model against full finite element solutions for a conducting box, non-magnetic and permeable, in a layered sea and in clay over bedrock. Two orders beyond the dipole, it is accurate to within 1.4% over two decades of frequency and for sizes up to 0.7 of the depth, whereas the error of the MPT carried into the conducting medium reaches 91%. Blocks coupling the two responses, contained in neither tensor, contribute up to 13%. (DOI: 10.48550/arXiv.2611.15090)</description>
    </item>
    <item>
      <title>Improvements in Tropical Cyclone Representation in JRA-3Q over ERA5</title>
      <link>https://nvkcnbsnkhie.org/?paper=3018</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3018</guid>
      <category>Glaciology & Paleoclimatology</category>
      <description>Reanalysis products are widely used for long-term tropical cyclone (TC) studies, yet the fidelity of their TC representation is not fully established. We evaluate TC representation in the recently released JRA-3Q reanalysis against the higher-resolution ERA5 using the SyCLoPS tracker. JRA-3Q achieves substantially higher detection skill, with a hit rate of 90% and a false alarm rate of 6% during 1979-2024. Unlike ERA5, which exhibits markedly degraded detection skill during the pre-satellite era, JRA-3Q maintains stable performance over time. We analyze false alarms and missed TCs in the North Indian basin, TC vortex evolution following TC bogussing, and the spatial variance of column-integrated moist static energy, showing that JRA-3Q's improvements stem from both reduced model biases and a more effective and consistently applied TC bogus scheme. A decomposition of TC representation errors into contributions from tracking algorithm, reanalysis, and best-track uncertainty helps explain the sources of these biases. (DOI: 10.48550/arXiv.2611.15222)</description>
    </item>
    <item>
      <title>DaYu-TC: a global-regional nested AI system for tropical cyclone forecasting</title>
      <link>https://nvkcnbsnkhie.org/?paper=3017</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3017</guid>
      <category>Glaciology & Paleoclimatology</category>
      <description>Accurate tropical cyclone intensity forecasts are essential for disaster preparedness, yet persistent intensity underestimation limits the value of artificial intelligence (AI) weather models for predicting these destructive storms. Here we present DaYu-TC, a global-regional nested system that links the preservation of strong-wind and low-pressure signals with high-resolution regional prediction. Its global and regional models share an architecture and extreme-aware training objectives. The regional model combines evolving local states with global forecasts valid at each target time, connecting large-scale environmental information with finer-scale cyclone evolution. In an independent evaluation of 16 western North Pacific tropical cyclones in 2023, the global model retains competitive conventional forecast skill while reducing five-day maximum-wind root-mean-square error (RMSE) by 6-22% relative to other evaluated global AI models. Relative to the global model, regional nesting reduces maximum-wind RMSE by 22.0-33.1% and minimum-pressure RMSE by 28.6-39.8% across 12-120 h. The complete system outperforms concurrent operational high-resolution forecasts from the European Centre for Medium-Range Weather Forecasts for minimum pressure at 12-120 h and maximum wind speed at 48-96 h, with respective RMSE reductions of 26% and 12% at 72 h. Without retraining, the same regional model enhances three other global AI models, reducing overall maximum-wind RMSE by 22.6-28.6%. Five-day forecasts take under one minute on a single GPU. These results establish a transferable multiscale approach to reducing cyclone intensity underestimation, extending advances in AI weather forecasting towards more accurate prediction of high-impact weather events. (DOI: 10.48550/arXiv.2611.15266)</description>
    </item>
    <item>
      <title>Advancing global water cycle science with the upcoming MAGIC satellite mission through data assimilation</title>
      <link>https://nvkcnbsnkhie.org/?paper=3016</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3016</guid>
      <category>Glaciology & Paleoclimatology</category>
      <description>The GRACE and GRACE-FO missions provide unique global measurements of terrestrial water storage anomalies (TWSA) from space since 2003. However, in order to spatially downscale the GRACE/-FO resolution of 300 km further, and to vertically disaggregate TWSA into soil moisture, groundwater and other contributions, the TWSA observations were assimilated into hydrological and land surface models in the past while improving the realism of the models. With the upcoming MAGIC (Mass Change and Geosciences International Constellation) constellation consisting of two satellite pairs (NASA/DLR GRACE-C and ESA NGGM), significant advancements are expected for the monitoring of water mass changes, with improved temporal and spatial resolution, reduced latency, and with lower noise levels. In this study, we reviewed the capabilities and limitations of existing GRACE/-FO data assimilation systems. We then investigated the benefit of including data from the upcoming MAGIC constellation into existing assimilation frameworks to improve not only data sets for water and carbon cycle science but also in view of operational applications, within a closed-loop simulation. For example, we extended a global 17 years GRACE/-FO assimilation era to a long time series of 33 years with synthetic MAGIC data assimilated and show reductions in groundwater trends errors and errors in identifying hydrological extreme events. We further showed for Europe that both storage and river runoff simulations benefit from more frequent assimilation, which the integration of MAGIC data will enable due to the foreseen 5-daily (instead of monthly) TWSA product. Finally, we demonstrated for the Iberian Peninsula and for Eastern Europe that MAGIC TWSA data assimilation also improved simulated carbon cycle variables via coupling in the land surface model, especially in water-limited regions. (DOI: 10.48550/arXiv.2611.15421)</description>
    </item>
    <item>
      <title>Imputing Missing Near-Surface Air Temperature using Graph Attention Networks with Distance-Aware Edge Features</title>
      <link>https://nvkcnbsnkhie.org/?paper=3015</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3015</guid>
      <category>Glaciology & Paleoclimatology</category>
      <description>We introduce a graph attention network for reconstructing near-surface air temperature fields from incomplete observations. The model represents each grid cell as a node connected to its $k$ nearest neighbors, with attention weights that depend on node features, distance, and land-ocean transitions. This graph representation avoids the polar distortions associated with treating the globe as a flat latitude-longitude image, as most deep learning reconstruction methods do. This is particularly relevant for historical temperature records, which extend back to the mid-nineteenth century but remain incomplete, with the largest gaps early in the record and at high latitudes. We train our model on CMIP6 CESM2 historical simulations masked with the time-varying HadCRUT5 observational coverage from 1850 to 2014. We then compare our approach with ordinary kriging, GraphEM, and the LaMa inpainting network on a held-out realization. The graph model achieves the lowest spatial RMSE (0.997 K), 10.1% lower than LaMa and approximately 31% lower than kriging and GraphEM. The largest gains occur in the Arctic and Antarctic, while LaMa remains slightly more accurate in the densely observed tropics. The graph model's error also falls faster as coverage increases, so its advantage over LaMa grows by more than a third from the 1850s to the 2000s. An ablation study shows that the model's skill comes mainly from the geometric graph construction rather than from auxiliary physical covariates. (DOI: 10.48550/arXiv.2611.15598)</description>
    </item>
    <item>
      <title>Tendex Lines and Short-Time Deformation: The Geometry of Lunar Tides</title>
      <link>https://nvkcnbsnkhie.org/?paper=3014</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3014</guid>
      <category>Glaciology & Paleoclimatology</category>
      <description>Tendex lines and the stretchlines and squeezelines of Lagrangian coherent structure theory share a geometric construction: they integrate eigenvector fields of symmetric tensors. We connect their dynamical interpretations through the short-time deformation of particles released with zero initial velocity relative to a freely falling reference in an autonomous gravitational field. The leading departure of the Cauchy--Green tensor from the identity is quadratic in time and equals the tidal tensor multiplied by the squared observation time. The tidal tensorlines therefore give the limiting principal stretching and squeezing directions. For the planar lunar field, these tensorlines are radial lines and concentric circles centered on the Moon. The same tidal tensor defines a Lorentzian metric whose null directions have no quadratic change in line-element length. Its null geodesics are logarithmic spirals, describing the short-time limit of finite-time unit-stretch geometry. Near Earth, the principal tensorlines clarify the two outward tidal tendencies, while the null directions distinguish vanishing leading extension from vanishing relative acceleration. (DOI: 10.48550/arXiv.2611.16083)</description>
    </item>
    <item>
      <title>HybridCT: ML Accelerated Tomography for Climate Relevant 3D Cloud Observations</title>
      <link>https://nvkcnbsnkhie.org/?paper=3013</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3013</guid>
      <category>Glaciology & Paleoclimatology</category>
      <description>Low-level clouds are one of the largest sources of uncertainty in climate predictions, partly because we lack detailed three-dimensional (3D) measurements of their internal structure. Here, we present HybridCT, a cloud tomographic reconstruction scheme accelerated by machine learning (ML) that recovers 3D cloud extinction directly from passive multi-angle observations. An ML emulator maps radiances from multiple viewing angles to cloud optical thickness (COT), producing path-integrated projections, and a Simultaneous Algebraic Reconstruction Technique (SART) explicitly inverts these projections to retrieve 3D extinction. Training observations are generated from synthetic 3D cloud fields using radiative transfer simulations for nine viewing angles between $\pm$70$^{\circ}$. HybridCT reconstructs extinction with RMSE 13.3 km$^{-1}$ in 3.5 s, compared to 22.7 km$^{-1}$ in approx. 30 min for the state-of-the-art physics-based tomography framework AT3D, a speedup exceeding 500$\times$. HybridCT thus provides a scalable route toward large-scale observations of 3D cloud structure relevant to radiative-budget interpretation and evaluation of high-resolution cloud simulations. (DOI: 10.48550/arXiv.2611.16255)</description>
    </item>
    <item>
      <title>Resolving Radiogenic Heterogeneity in the Earth's Mantle with Multi-Site Geo-neutrino Measurements</title>
      <link>https://nvkcnbsnkhie.org/?paper=3012</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3012</guid>
      <category>Glaciology & Paleoclimatology</category>
      <description>We present a global fit to geo-neutrino measurements at four underground sites (Kamioka, Gran Sasso, Sudbury, Jiangmen), combining signal rates with site-specific predictions of the lithospheric contribution to constrain the quantity and distribution of radioactivity in the Earth's mantle. Assuming a uniform distribution of radioactivity in the mantle, we find a geo-neutrino rate of $S_\text{mantle} = 14.8 \pm 4.5$ NIU when the lithospheric predictions are allowed to vary within their uncertainties. This uniform mantle signal rate corresponds to radiogenic heating of $18.5 \pm 5.6$ TW. Although the uniform-mantle signal is resolved at $\sim3.2σ$, there is noticeable tension between the inferred mantle signals at Jiangmen and Kamioka, which may indicate non-uniform geo-neutrino production in the mantle. We explore six families of models to describe the heterogeneous spatial distribution of radiogenic elements in the Earth's mantle. Each model contains anomalous bodies or reservoirs carrying an excess of heat-producing elements over a depleted mantle (DM) background, with total inventory set by bulk-silicate-Earth mass balance and by a geochemical Th/U constraint. For each model, we compute the geo-response at each detector site and fit the resulting signals to the four measurements. A total of 968,274 individual model configurations are defined and tested. The results of $\sim$3 million distinct fits show that no plausible mantle or lithosphere model applied uniformly to all four sites reproduces the difference between the inferred mantle signals at the Jiangmen and Kamioka sites. Only local, site-specific prescriptions are able to reach the observed difference, suggesting that the Kamioka-Jiangmen tension is due to a local effect or statistical scatter, rather than large-scale mantle structure. (DOI: 10.48550/arXiv.2611.16715)</description>
    </item>
    <item>
      <title>Upscale energy transfer from mesoscale convective systems to African easterly waves</title>
      <link>https://nvkcnbsnkhie.org/?paper=3011</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3011</guid>
      <category>Glaciology & Paleoclimatology</category>
      <description>African easterly waves (AEWs) drive the synoptic variability of West African monsoon rainfall via intricate two-way interactions with mesoscale convective systems (MCSs). The composite MCS-AEW system is a canonical example of multiscale interactions in the tropical atmosphere, with upscale feedbacks from deep convection driving the evolution of a larger-scale circulation. Here we support this perspective through application of techniques from turbulence theory, calculating for the first time patterns of interscale kinetic energy transfer, $\mathcal{D}_L$, across different spatial scales $L$ around MCSs and AEW lifecycle phases. We study AEW and MCS dynamics in two pan-tropical 40 day Met Office Unified Model simulations with $\sim$5km grid spacings, for which $\mathcal{D}_L$ is calculated with the LoSSETT software package. We find explicitly-simulated MCSs show a distinctive vertical dipole structure of $\mathcal{D}_L$ at scales below 1000km, with strong upscale energy transfer located in the upper troposphere within storm anvils. Regionally, upscale energy transfer occurs most strongly where organised convection is active. We find 200hPa upscale transfer consistently leads AEW troughs, and that growing continental AEWs in the explicit simulation are characterised by strengthened upscale transfer of energy from MCS to AEW scales. Case-study analysis shows this transfer can contribute to the eddy kinetic energy growth of strengthening AEW troughs. These signals are weak in a simulation with parameterised convection: here, zonal mean energetics show enhanced barotropic conversion, caused by a stronger African Easterly Jet, provides a stronger AEW population. Our analysis of upscale energy transfers complements previous analysis of AEW potential vorticity dynamics, providing a route towards building new theoretical understanding of tropical dynamics from km-scale models. (DOI: 10.48550/arXiv.2611.16941)</description>
    </item>
    <item>
      <title>Spontaneous mechanical patterning in active elastic tissues with stress feedback</title>
      <link>https://nvkcnbsnkhie.org/?paper=3010</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3010</guid>
      <category>Microdosimetry & Hemodynamics</category>
      <description>The active mechanics of biological tissues is often governed by embedded active contractile fibers responding to mechanical feedback. We introduce a minimal model of a solid-like tissue where the mechanics is controlled by feedback of local active and elastic stresses on fiber organization. The resulting tissue generically exhibits a bistable response to external strain and spontaneously segregates into domains of orthogonally aligned fibers. Our theory thus provides a minimal mechanical feedback mechanism generating orthogonal structural alignment ubiquitous in biological tissues. (DOI: 10.48550/arXiv.2611.14943)</description>
    </item>
    <item>
      <title>Feedback-induced mechanical bistability and perpendicular-domain formation in active elastic tissues</title>
      <link>https://nvkcnbsnkhie.org/?paper=3009</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3009</guid>
      <category>Microdosimetry & Hemodynamics</category>
      <description>Contractile fibers in muscle tissues are highly aligned over the scale of many cells. Muscle layers surrounding organs with an inner lumen are often arranged in orthogonal directions. Here, we study a minimal mechanism of how structural order and its patterning into orthogonally aligned domains can emerge from mechanical feedback alone. The central idea is that contractile fibers align in the direction of the local active and elastic stresses, a feedback motivated by the maturation of muscle fibers under tension. We thus model muscle tissue as an active nematic solid with stress feedback: an elastic medium with embedded contractile fibers that are reinforced along the directions of the local tissue stresses. We find that the feedback from the contractile fiber stress is self-amplifying and drives spontaneous fiber alignment. Feedback from the elastic stress couples the alignment to the mechanical boundary conditions and tissue geometry. Depending on the mechanical boundary conditions, the elastic-stress feedback stabilizes two orthogonally aligned tissue states, resulting in a structurally and mechanically bistable tissue that shows a hysteretic stress-strain response. In spatially extended tissues, this bistability gives rise to orthogonally aligned domains separated by domain walls. In an annulus surrounding a pressurized lumen, a radially aligned inner domain and a circumferentially aligned outer domain stably coexist because the domain wall is pinned at a position that results in the fibers carrying the lumen pressure. Our results show that stress feedback can pattern structural order by coupling spontaneous alignment to the tissue geometry. Our mechanism thus provides a purely mechanical route to orthogonal fiber patterning in muscle tissues. (DOI: 10.48550/arXiv.2611.14964)</description>
    </item>
    <item>
      <title>A High-Fidelity Cardiac Digital Twin for Assessing Rate-Dependent Responses: A Comparative Multi-scale Study of TT2 and Tomek Models in Stratified 3D Ventricles</title>
      <link>https://nvkcnbsnkhie.org/?paper=3008</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3008</guid>
      <category>Microdosimetry & Hemodynamics</category>
      <description>Cardiac digital twins (CDTs) provide a computational platform for personalized cardiac modeling, but their predictive behavior depends strongly on the selected ionic model and the representation of transmural electrophysiological heterogeneity. In this study, we developed a multiscale CDT validation framework that integrates physiologically optimized hierarchical configurations. Using single-cell (0D) models and three-dimensional (3D) patient-specific ventricular models, we implemented the ten Tusscher-Panfilov model (TT2) and Tomek ionic model and compared their electrophysiological responses across a range of pacing frequencies, with basic cycle lengths (BCLs) from 600 to 4000 ms. The results show that tissue-level electrical coupling substantially reshapes the intrinsic dynamical properties observed at the single-cell level. By benchmarking model outputs against clinical data, this study quantifies fidelity differences between the two ionic models across multiple scales and provides a scientific basis for ionic model selection and CDT construction. (DOI: 10.48550/arXiv.2611.14995)</description>
    </item>
    <item>
      <title>A novel model-based equation for size dependent mean recovery coefficients for spheres and other shapes</title>
      <link>https://nvkcnbsnkhie.org/?paper=3007</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3007</guid>
      <category>Microdosimetry & Hemodynamics</category>
      <description>Background: In NM-imaging, theoretical curves for the recovery coefficient (RC) of the signal maximum and mean are known for spheres and cubes, if a 3D Gaussian PSF is assumed. The RC of the maximum is also known for cylinders. For these and other shapes empirical equations with one or two fit-parameters have been utilized. Methods: An equation for the RC for large objects of arbitrary shape is derived and generalized into an empirical equation for smaller objects, which is verified by numerical simulations. The proposed equation is compared to published results on SPECT kidney phantom measurements and to PET measurements on the NEMA IEC PET body phantom with six spheres. Results: The signal loss (1-RC) for large spheres is inversely proportional to the radius, where the slope is proportional to the FWHM of the spatial resolution. For non-spherical shapes the generalized instead of the volume equivalent radius should be utilized. For smaller objects, an equation with one added empirical fit-parameter is presented. It is demonstrated that the EANM-guidelines' two-parameter logistic function results in a poor fit if the theoretical slope and inverse proportionality are forced and it gives a suboptimal fit when both parameters are fitted. Conclusions: A novel model-based equation for the mean RC-curve is derived. It can be used for arbitrary shapes as long as the sphericity is taken into account and it is accurate down to RC=10%. One parameter is directly related to the spatial resolution, while the other is a shape depending fit-parameter. (DOI: 10.48550/arXiv.2611.15375)</description>
    </item>
    <item>
      <title>EQUIP-RPT: A framework for objective evaluation of quantitative imaging methods in radiopharmaceutical therapies</title>
      <link>https://nvkcnbsnkhie.org/?paper=3006</link>
      <guid isPermaLink="true">https://nvkcnbsnkhie.org/?paper=3006</guid>
      <category>Microdosimetry & Hemodynamics</category>
      <description>Quantitative imaging (QI) provides a non-invasive approach for dosimetry and quantitative response assessment in radiopharmaceutical therapies (RPTs). QI methods are being developed for instrumentation, acquisition, reconstruction, and image analysis. However, a key barrier to routine adoption of these methods is that evidence demonstrating reliable performance of QI methods on clinically relevant quantification tasks in RPTs remains limited. Our goal is to outline the Evaluation of Quantitative Imaging Performance in Radiopharmaceutical Therapy (EQUIP-RPT) framework, a structured approach for objective, task-based evaluation of QI methods for RPTs. We outline the EQUIP-RPT framework based on image-science and task-based image-quality evaluation methodology, practical considerations, and recent literature. The framework is organized around four linked design steps: (i) specification of the quantification task, (ii) design of the study population and imaging protocol, (iii) task-specific information extraction, and (iv) selection of figures of merit (FoMs). The EQUIP-RPT framework is presented in the context of four quantification tasks in RPT: activity uptake, absorbed dose, volumetric tumor burden, and lesion-level longitudinal quantifications. The framework distinguishes two paradigms for task-specific information extraction and organizes FoMs according to the performance attributes they characterize. We describe key elements to apply the EQUIP-RPT framework in simulation, physical phantom, and clinical studies. Finally, we show how EQUIP-RPT can be used to generate a quantitative claim based on the evaluation study. The EQUIP-RPT framework is designed to provide a structured approach for objective task-based evaluation of QI methods in RPTs, thereby facilitating clinical translation of these methods. (DOI: 10.48550/arXiv.2611.15411)</description>
    </item>
  </channel>
</rss>
