arXane › Physics › Astrophysics
Wednesday, 4 November 2026: 10 new papers, 11 cross-lists and 6 replacements in Instrumentation and Methods for Astrophysics (astro-ph.IM), in arXiv's order, with full abstracts. The same papers as arXiv's own listing, set as a newspaper; every title links to arXiv.
Modern imaging surveys generate data at rates that exceed the throughput of photometry pipelines built on iterative, per-source fitting, a construct that does not map onto parallel hardware. We present a set of algorithms for point-source detection and photometry reformulated so that the entire reduction runs natively on a Graphics Processing Unit (GPU), implemented in GPUPhot. Their design rests on a single principle: every stage is cast as a convolution with a fixed kernel or an element-wise operation on full-frame maps. The algorithms are designed to operate on wide-field images with a spatially variable Point Spread Function (PSF), where the field-dependent PSF is described through a small set of fixed, flux-conserving principal-component eigenmodes with position-dependent coefficients, so that the spatially varying matched filter decomposes exactly into a few genuine convolutions while retaining its maximum-likelihood foundation. We show that, even in fields with strong background gradients and markedly inhomogeneous PSFs, the framework achieves sub-pixel astrometric accuracy and few-millimagnitude photometric precision for bright sources. The pipeline is already in production at the telescopes of the Teide Observatory, where it has processed 3.7 million images since 2023, amounting to 1.2 PB of data, at effective end-to-end rates below 15 s per 150-Mpx image, from ingestion of the raw frame to a source catalog with astrometric and photometric solutions. Because the same primitives underlie every stage, the individual algorithms are self-contained and applicable well beyond photometry. They form the basis of a complete, openly developed software package, released publicly with this paper.
In applied statistics, likelihoods involving high-dimensional integrals are often approximated via Monte Carlo summation. Controlling the resulting numerical variance is crucial for reliable inference. We provide a practical introduction to Latin hypercube sampling (LHS) and its extension based on orthogonal arrays (OA-LHS) as strategies to reduce this variance by stratifying the sampling distribution. LHS asymptotically removes variance contributions from individual coordinates, while strength-2 OA-LHS also removes those from pairwise interactions, yielding orders-of-magnitude improvements over plain Monte Carlo estimators for suitable integrands. As a case study, we investigate their application to the selection function entering hierarchical inference of gravitational-wave populations. At approximately one million injections, we find only a marginal reduction in variance, with low statistical significance. Nevertheless, the computational cost of constructing these designs is negligible relative to evaluating gravitational-wave injections, making even marginal gains worth considering for future applications. This practical introduction to LHS and OA-LHS offers a starting point for other applications in physics and astronomy involving computationally expensive Monte Carlo calculations.
Large-scale astronomical surveys define a heterogeneous, partially observed multi-survey learning setting: each object is observed by only a subset of surveys, while survey observations differ in resolutions, wavelength coverage, noise characteristics, and footprints. This setting gives rise to a natural machine learning problem: given observations of an object in some surveys, how to infer its appearance in other surveys. We formalize this as the problem of cross-survey modality imputation and introduce COSMIO, a multi-survey benchmark of aligned observations obtained by forward-rendered shared astrophysical scenes across LSST, Euclid, and Roman under survey-specific instrument characteristics. The benchmark enables systematic evaluation of imputation across source-target survey combinations and missing-survey patterns. We evaluate a broad range of machine learning methods for imputation, which reveal a clear domain gap: methods developed for existing imputation settings transfer poorly to the survey setting. We identify rare-event imputation as a central challenge. This regime arises in the early stages of new surveys, when rare but scientifically valuable phenomena may be available only from prior surveys. We use strong gravitational lensing, where light from a background galaxy is gravitationally deflected by a foreground galaxy, as a representative rare-event setting. We evaluate model performance on zero-shot imputation of gravitational lenses and on downstream scientific tasks using the imputed lens observations. On our constructed evaluation set of galaxy-galaxy strong lenses, performance degrades under rare-event distribution shift, and this degradation further propagates to the downstream lens-detection task. Code: https://github.com/fw-ic/COSMIO.
The International Liquid Mirror Telescope (ILMT) is a 4-meter survey telescope continuously observing towards the zenith in the SDSS g', r', and i' bands. This survey telescope is designed to detect various astrophysical transients (for example, supernovae) and very faint objects like multiply-imaged quasars and low surface brightness galaxies. Ensuring accurate and stable calibration of supernova (SN) and quasar measurements is essential for studying their intrinsic variability and long-term brightness evolution. In this work, we present a dedicated calibration approach tailored for ILMT point-like object observations obtained in Time Delayed Integration (TDI) mode. Using ZTF reference sources cross-matched with the Pan-STARRS1 catalog, we derive robust photometric zero-points along the ILMT scan. This calibration procedure enables the production of consistent, high-precision SN and quasar light curves, suitable for variability studies, structure-function analysis, and future reverberation-mapping applications.
This paper presents an automated end-to-end pipeline to detect and characterize artificial satellites and space debris streaks. As the exponential rise of orbital debris poses an escalating risk to global space assets, providing scalable, real-time tools is a critical contribution to the global Space Situational Awareness (SSA) network. The architecture utilizes a multi-stage geometric approach for raw image segmentation and high-precision morphological characterization. Experimental results demonstrate robust performance in real-world survey environments, successfully recovering streak profiles with an operational detection limit for features with integrated $SNR \ge 86$ and lengths $\ge 98$ arcseconds. This unsupervised framework strengthens global space sustainability efforts by monitoring RSO populations.
Fundamental physics faces several intrinsic limitations, such as the finite speed of light in vacuum, the absolute zero temperature, Heisenberg's uncertainty principle and so on. At the same time, the material resources necessary to conduct research in astrophysics, such as energy, fresh water, and rareearth elements, are also limited in both quantity and extraction capabilities, a concept often referred to as planetary boundaries. In the context of scientific research growing at a frenetic pace, following the greedy quenchless extractivist model of our era, and which is giving rise to an ever-increasing number of largescale, increasingly ambitious projects: what limits us, individually and collectively, in the practice of our occupation\,? In this conference proceeding, we discuss certain limiting factors we encounter in our daily work, which we need to reflect on collectively in order to move forward, toward a more sustainable and effective research that preserves the well-being of researchers, regardless of their career stage.
Athena++ is a finite-volume astrophysical magnetohydrodynamics (MHD) code. As of writing, isotropic thermal conduction and kinematic viscosity, Ohmic resistivity, and ambipolar diffusion are present in the main branch. Here we present a fork of Athena++, ``Athena-Aniso''. It implements fully anisotropic thermal conduction (parallel, perpendicular, and the Righi--Leduc effect); anisotropic viscosity (all five Braginskii viscosity tensors); the Hall effect; and several thermoelectric effects (Peltier, Seebeck, Nernst, and Ettingshausen). These effects can be weighted by ionisation fraction in a partially ionised gas, with the neutral part diffusing isotropically. We have also added isotropic bulk viscosity. The Biermann battery is not implemented here. The dissipative terms have super-time-stepping support, while the non-dissipative terms are combined with the ideal MHD flux. A CFL condition for all terms is derived and implemented using von Neumann stability analysis. The dissipative terms are found to be stable under first-order time integration. The non-dissipative terms are found to be stable under third-order time integration. We additionally derive new signal speeds for a modified HLLE Riemann solver. These suppress grid-scale non-linear instabilities still present in the non-dissipative terms. The Nernst/Ettingshausen system is found to be ill-posed unless it is stabilised by both perpendicular thermal conduction and perpendicular magnetic diffusion. We derive a well-posedness criterion for this. We provide numerical tests of each implemented effect against derived analytical solutions.
Imaging atmospheric Cherenkov telescopes (IACTs) cameras are constantly subject to signals from the night sky background and electronic noise during event recordings. Image cleaning methods are typically employed to eliminate pixels that are mostly noise-contaminated and lack sufficient Cherenkov light. In this paper, we present the implementation of the optimised next neighbour (ONN) technique to the Very Energetic Radiation Imaging Telescope Array System (VERITAS) reconstruction pipeline. With this method, cleaning thresholds are derived in a two-parameter space, composed of the minimum total integrated charge of a cluster of pixels, $q_{\rm min}$, and by the maximum difference in the pulse arrival time within the cluster, $\delta_t$. The cleaning contours in $q_{\rm min} - \delta_t$ are adjusted to set a high probability of noise suppression while optimising the retention of pixels containing Cherenkov light. In comparison to the VERITAS double-pass image cleaning method, effective areas are increased at the lower energy range ($\sim$ 60 - 90 GeV) by a factor of up to $\sim$ 25. Energy thresholds are reduced by up to $\sim$ 22\%. Testing the method on Crab Nebula data resulted in an increased number of reconstructed excess events at energies < 1 TeV by a factor of up to $\sim$ 82. Finally, an increase in the differential sensitivity of VERITAS by a factor of up to $\sim$ 8 was observed at $\sim$ 20 GeV. With the addition of BDT-based $\gamma$/hadron separation cuts, the method improves the detection significance of a 20 hour Crab Nebula dataset, while excess event rates are enhanced by $\sim$ 16\%. In summary, the ONN image cleaning provides a reduced energy threshold and a higher rate of event reconstruction. The technique is especially advantageous for the analysis of gamma-ray sources presenting transient and soft spectra.
Large language model (LLM) agents are now applied to astronomical data analysis. They retrieve the literature, reduce observations, and draft conclusions. Under pressure to complete a task, such agents can fabricate results, and the numbers they report may not come from the analysis they ran. We present auto_research, a semi-autonomous multi-agent pipeline that retrieves and analyses raw data from the High Energy Astrophysics Science Archive Research Center (HEASARC) and writes the analysis tools that it needs. We check the work of the agents at two points. A new tool enters the toolkit after it passes automatic checks, including repeated test runs and a scan for phrases that accompany invented values, and a code review by a model from a different provider; a human decides whether to promote it to the trusted branch. A number in a report passes only if a recorded output of a toolkit tool or the request contains it, or if the report gives its location in a cited paper. We test the pipeline with data from the Imaging X-ray Polarimetry Explorer (IXPE). Of 65 defects that we planted in the toolkit, 47 pass the automatic checks, and one passes the code review as well; the review also rejects 40 of 69 correct changes. The pipeline reproduces the published IXPE results for Cyg X-1: all 14 compared quantities agree within 3 sigma, and the polarization degrees are lower than the published values. With weaker models in the main roles, the pipeline recovers 12 of the 14 quantities and reports the other two as not measured. We also used the pipeline for a polarization-stability analysis of six observations, a survey of polarization against hardness in 29 observations, and a search for a time lag between the Stokes parameters q and u.
Prime-Cam is one of the primary instruments being developed for use by the CCAT collaboration with the Fred Young Submillimeter Telescope (FYST), which is undergoing construction at a height of 5600 m in Cerro Chajnantor. Prime-Cam can host up to seven instrument modules with kinetic inductance detectors (KIDs), including broadband modules ranging 280-850 GHz and imaging spectrometers for line-intensity mapping. Prime-Cam requires a number of hardware devices for standard operation and data collection, and so a distributed control system must be developed for software integration of these devices and user-friendliness when controlling Prime-Cam in the field, especially during remote operation that will be typical during observations. The Prime-Cam Control System (PCS) is being developed using the Observatory Control System (OCS) framework originally developed for Simons Observatory (SO) and is adapted for Prime-Cam and CCAT-specific requirements. Within OCS/PCS, hardware devices are controlled individually by long-running software processes known as agents, which are typically deployed in Docker containers running in the background of one of the Prime-Cam network computers. These agents are controlled by client scripts through the command line or frontend tools like Nextline and PCS-web. PCS also makes use of several external open source software tools for remote operations, data management, and logging. This system is being tested in the lab in preparation for deployment in 2026 for use with Prime-Cam on FYST.
The Laser Interferometer Space Antenna (LISA) is expected to observe gravitational-wave signals from massive black hole binaries, extreme mass ratio inspirals, and more. Despite the significant scientific potential, accurately extracting these signals from the data will only be possible if LISA's instrumental noise is understood: poor noise modelling will result in lost signals, biased parameter estimates, and incorrect astrophysical insights. Modelling LISA's instrumental noise is more challenging than its ground-based counterpart since the data is highly correlated, meaning that the noise covariance matrix can no longer be assumed diagonal in frequency and in the different data streams. In this work, we present an initial approach to perform non-parametric noise inference for LISA that requires no prior knowledge of the LISA observatory. By fitting the full noise covariance matrix with a combination of splines and inverse Lorentzians, we show that our model can estimate the power and cross spectral densities for LISA from 7 days of simulated noise-only data. We find that the median estimate from our non-parametric model whitens the data according to the expected unit normal distribution. We envision that our non-parametric model can be used as part of the global fit, as well as in combination with modelled approaches to identify, quantify, and constrain the frequency dependence of unknown noise artifacts.
The detection of organics and potential biomolecules is an important goal of planetary science and astrobiology spacecraft missions, and many types of analytical chemistry for studying organics depend upon liquid solvents as working fluids. However, most liquids are difficult to maintain in the vacuum of space, complicating their use and increasing spacecraft mass and cost. Here we present a study using Direct Analysis in Real Time mass spectrometry (DART-MS) as a rapid screen for the detectability of histidine, glycine, arginine, lysine, and aspartic acid at mole fractions of about $5\times10^{-4}$ to $10^{-3}$ in five vacuum-stable liquids: EMI-BF$_4$, EMI-acetate, DEMA-MsO, DEMA-OTf, and glycerol. In all of the ionic liquids, the solvent cation dominated the spectra and no amino acid was detected, apart from a weak histidine signal in EMI-BF$_4$ at 350$^\circ$ C (median 5\% of the normalized peak intensity). The protic DEMA ionic liquids did not enhance detection, and their ionic components instead dominated the mass spectra. Aspartic acid was not detected in any liquid. In glycerol solutions, histidine, glycine, arginine, and lysine were all detected, and histidine was frequently a dominant line. We thus find glycerol to be the most promising of the liquids tested for organic detection with an ideal operating DART helium temperature of 350--400 $^\circ$C. We also found no clear evidence of water uptake or aqueous alteration in the ionic liquids, which was a secondary goal relevant to ionic liquid spacecraft electric propulsion.
Carbon monoxide (CO) isotopologues are widely used to estimate protoplanetary disk gas masses, but their reliability is often compromised by CO freeze-out and subsequent chemical processing. The disk around the luminous Herbig star HD~179218 ($L_* \sim 100\,L_\odot$, $M_* \sim 3.0\,M_\odot$) offers a rare test case of CO-based mass measurements. Its CO snowline lies beyond the outer disk edge, and thus CO remains in the gas phase throughout. We present ALMA and NOEMA observations of six CO isotopologue lines ($^{13}$CO, C$^{18}$O, and C$^{17}$O, $J$=2--1 and 1--0). Fitting the spectral energy distribution yields $M_{dust} \approx 3 \times 10^{-5}\, M_\odot$. Modeling CO isotopologue emission with the thermochemical code DALI yields the gas-disk characteristic radius ($R_{c} \approx 90$~au) and the gas mass ($M_{gas} \approx 8 \times 10^{-3}\, M_\odot$). These results yield a gas-to-dust ratio of $\sim$270 for our fiducial dust opacity ($a_{\rm max} = 1$~mm), above the canonical interstellar value of $\sim$100. The thermal structure of our best-fitting model confirms that CO never freezes out in this disk, so the elevated ratio is not an artifact of CO depletion; together with the compact millimeter dust disk relative to the gas disk, it is consistent with substantial inward radial drift of millimeter dust. In addition, we find a gas cavity inside 50\,au that encloses the compact millimeter dust disk---the inverse of a typical transition disk---which we attribute to photoevaporative clearing of the inner gas after the pebbles drifted inward. HD~179218 thus provides a clean benchmark for studying the gas mass distribution and gas--dust interaction in protoplanetary disks, free from the CO-depletion biases that affect colder disks.
Classifying small exoplanets as rocky, volatile-rich, or iron-enriched is a prerequisite for efficient allocation of atmospheric characterization resources. The Nominally Rocky Planet Zone (NRPZ) provides a framework for this classification by comparing observed bulk densities against the expected density range for volatile-free rocky planets, calibrated from stellar abundance distributions. However, the efficacy of classifying a planet using the NRPZ is critically dependent on the precision of the underlying mass and radius measurements. Here, we perform an analytical propagation of observational uncertainties through the NRPZ classification framework for 37 candidate rocky exoplanets with radii below 2 Earth radii orbiting stars with visual magnitudes <11. Among the 37 planets, we find 12 already possess sufficiently precise mass and radius measurements to be confidently classified as Mini-Neptunes or Super-Mercuries. The remaining planets require approximately a ~60% reduction in the combined mass and radius uncertainties before they can be reliably classified. We find that 1-1,000 new observations are necessary to achieve these lower uncertainties, and these are likely lower limits. We argue that, while possible, this level of intense observation is likely not feasible for many of these planets and that mass-radius models alone are not likely to allow us to classify all but those planets with outlier densities. We therefore recommend a staged strategy in which improved mass and radius measurements are used to identify and prioritize ambiguous systems, followed (where feasible) by atmospheric observations with JWST or future facilities as a complementary test of whether selected planets retain substantial atmospheres.
The next generation of an international network of gravitational-wave observatories is now in the advanced planning stages. With Australia's substantial history of contributions in this field, its history of scientific mega-projects like the Square Kilometer Array, and its unique geographic location and conditions, it is ideally placed to host an observatory in the global network. It has already been demonstrated that a gravitational wave detector in Australia plays an important role in minimizing the sky localization uncertainty of gravitational wave events. We study a particular science case for an Australian observatory, focusing on its potential contributions to multimessenger astronomy. We demonstrate that an observatory in Australia can enable the detection of up to eight multimessenger gravitational-wave plus kilonova events per year, with localisations better than $1-10\deg^2$ in both second- and third-generation networks. We show that, in a global network of XG observatories, an Australian observatory with 2G sensitivity would still enable significant advances in multimessenger science. The full case for an Australian observatory must be considered in a much broader context; however, this work indicates it could make substantial contributions to multimessenger astrophysics.
Reprocessed X-ray emission from cold material surrounding compact objects carries critical information about the geometry and physical state of the system. In particular, the Fe K$\alpha$ fluorescence line provides a powerful diagnostic through its spectral and timing properties relative to the direct emission. Interpreting these observables requires self-consistent modelling of radiative transfer (RT) in three-dimensional (3D) space. We aim to implement photon arrival time lags in SKIRT, a Monte Carlo RT code, thereby enabling self-consistent modelling of both the spectral and timing properties of reprocessed emission in 3D space. We modified the code to accumulate photon path lengths over their lifetimes and convert them into arrival time lags relative to the direct emission. The implementation is validated using simple geometries and then applied to a realistic model of an active galactic nucleus (AGN), for which we compute the time-lag and energy distribution and the cross-correlation functions (CCFs) between the fluorescence line and broad bands. The simulations reproduce the expected time-lag distributions and their energy dependence for both free-electron scattering and the photoelectric absorption $+$ fluorescence. In the AGN application, the model naturally yields distinct lag components from the inner and outer structures, as well as energy-dependent time lags associated with the fluorescence line. The derived CCFs successfully recover both the peak time lag and its finite width, which arise from the extended geometry and intrinsic variability. In X-rays, this capability provides a robust framework for interpreting high-resolution spectroscopic data in conjunction with timing information from current and future missions such as XRISM and NewAthena. A wide range of applications is foreseen, including reverberation mapping studies of accreting systems.
Intact chains of mean-motion resonances may preserve information about the formation and early dynamical evolution of multiplanet systems. Although most mature multiplanet systems are non-resonant, the resonant fraction may have been substantially higher at young ages. Establishing whether exoplanet data are consistent with a resonant configuration is often challenging because the planetary orbits must be sufficiently well constrained. Moreover, while resonance configurations promote stability, standard orbital fitting methods do not impose such a dynamical constraint, with stability tested only a posteriori. As the first step in a series of papers aimed at favoring stable orbital solutions during the fitting process, we develop a method that restricts solutions to stable fixed-point orbits (fPOs)---coplanar, exactly periodic configurations associated with resonant dynamics. We show that dissipative evolution drive resonant systems toward an fPO state, motivating for fPO-constrained fitting for those planet systems that have avoided large-scale dynamical rearrangements since formation. The fPO model greatly reduces the number of free parameters and yields stable solutions by construction. We determine the best-fitting fPos through $\chi^2$ minimization and apply the method to astrometric data for the directly imaged PDS 70 and HR 8799 systems and to radial-velocity data for GJ 876. For HR 8799 we independently recover a previously reported resonant solution, illustrating the fidelity of our method. For PDS 70 we find solutions fall in two groups, with the group where the outer planet c mass exceeds planet b more aligned with previous measurements. Finally, For GJ 876 our fPO fit is on par with previous unconstrained fitting, which strengthens the case that its planets reside deep in resonance.
Many fields of physical interest on the sphere, such as wind, the polarisation of the cosmic microwave background and the shear of weak gravitational lensing, are spin fields, whose values change phase when the local frame rotates. We extend spherical Fourier neural operators (SFNOs) to fields of any spin, through lifted spherical convolutions that are exactly rotationally equivariant. Lifting spin fields to Mackey functions on the rotation group provides the conceptual framework, while computation is performed exactly and efficiently through their harmonic representation and by appealing to sampling theorems on the sphere and rotation group. We construct spin SFNOs (S$^2$FNOs) from these building blocks. We demonstrate empirically that rotational equivariance holds to machine precision for these layers. We achieve state-of-the-art performance on existing benchmarks, and introduce new benchmarks on which spin-aware models outperform their scalar counterparts.
We compare feature-based machine learning and end-to-end deep learning for variable star classification using the Zwicky Transient Facility (ZTF) StarEmbed dataset, which contains about 40,000 objects across seven imbalanced classes. Classical models trained on physically motivated features consistently outperform deep networks trained directly on raw light curves. We find that limited labeled data contributes to the performance gap, while the multivariate, irregularly sampled, and morphologically diverse nature of light curves poses additional challenges for deep learning models. While deep learning holds promise for learning representations directly from observations, our results indicate that current models fail to generalize with the available training set size. These findings motivate further investigation of the benefits of scaling labeled datasets and self-supervised pretraining on large unlabeled surveys. This work establishes a quantitative baseline for future research and highlights the need for further innovations to unlock the potential of deep learning in astronomical time series analysis.
The astrophysical discovery potential of future gravitational wave detectors will hinge on accurate and rapid waveform generation. Third generation detectors are expected to observe many long duration compact binary coalescence signals that can probe binary formation and evolution as well as the nature of matter under extreme conditions. We develop a unified framework that combines physical modeling with deep learning. The framework uses an analytic residual spectral representation, shared learning, and decoders for each binary type. It generates waveforms for binary black holes, binary neutron stars, and neutron star black hole binaries, providing a single unified generator for all compact binary coalescences. The generated waveforms include the physical effects described by each reference model, such as spin precession, higher order modes, matter effects, and postmerger radiation for binary neutron stars. The waveforms span hundreds to thousands of cycles, with mean mismatches ranging from $5.89\times10^{-4}$ to $5.34\times10^{-2}$ relative to the reference waveforms. Batch generation achieves a maximum speedup of about 1800 over reference waveform generation.
We investigate the potential of the Ariel space telescope to detect and characterise exocomet transits through their wavelength-dependent photometric signatures. Building on a radiative-transfer framework, we model exocomet transits for three representative dust morphologies, three grain-size distributions, and four homogeneous dust compositions using synthetic photometric bands covering the Ariel wavelength range. The modelled events typically last 1-2 days and produce photometric variations of a few millimagnitudes. We find that the wavelength dependence of the transit signal provides stronger diagnostics of the dust properties than transit morphology alone. In particular, colour variations relative to the VISPhot band and their wavelength-dependent slopes separate the considered compositions into two broad groups: carbonaceous dust and graphite, and olivine and pyroxene. While the compositions within each group show similar behaviour, the two groups remain distinguishable across the dust morphologies and grain sizes considered. We further assess the detectability of these signals using the $\beta$~Pictoris spectral energy distribution and ArielRad simulations. Integrating over broad synthetic photometric bands substantially improves the signal-to-noise ratio, allowing millimagnitude-level signals to be detected with short integration times for sufficiently bright targets. Ariel's simultaneous optical-to-infrared coverage therefore offers the possibility not only of detecting exocomet transits, but also of characterising their chromatic signatures and constraining broad classes of dust properties. We also discuss the limitations of the modelling assumptions and the potential of coordinated observations with long-baseline photometric missions.