Supercomputers uncover a new class of cosmic explosions hidden in plain sight
Supercomputers uncover a new class of cosmic explosions hidden in plain sight
previous arrow
previous arrow
next arrow
next arrow
Shadow
Featured

Supercomputers uncover a new class of cosmic explosions hidden in plain sight

High-performance simulations reveal that the mysterious transient AT2019ijn may be powered by an off-axis relativistic jet from an intermediate-mass black hole, opening a new frontier in time-domain astrophysics.

Modern astronomy has entered an era where telescopes no longer make discoveries in isolation. Increasingly, the most profound scientific breakthroughs arise from a powerful synergy between observational surveys and high-performance computing. While modern instruments can detect extraordinary events billions of light-years away, deciphering their nature often relies on sophisticated numerical modeling to reconstruct the underlying physics.
 
A compelling example is the transient AT2019ijn, an unusual optical and radio outburst that defies conventional classification. Characterized by a rapid rise in brightness, a prolonged blue phase, and an exceptionally bright, long-lasting radio afterglow, the event fits poorly into established categories like supernovae or fast blue optical transients (LFBOTs), suggesting it represents an entirely new phenomenon.
 
This discovery is particularly significant for the high-performance computing community because it could not have been understood through observation alone. Successfully reconstructing one of the most energetic explosions ever observed in a dwarf galaxy required a complex suite of computational tools, including large-scale Bayesian inference, relativistic jet simulations, Markov chain Monte Carlo (MCMC) optimization, synchrotron emission modeling, and tidal disruption event (TDE) fitting.

An explosion that refused to fit the rules

AT2019ijn was discovered in the nucleus of a dwarf galaxy approximately 3.4 billion light-years away (redshift 0.2729). It reached an optical luminosity of about –21 magnitude in just over five days before fading over more than a month while maintaining a remarkably high blackbody temperature of roughly 15,000–16,000 K. These characteristics resemble fast blue optical transients, yet its slow decay is far more typical of tidal disruption events or superluminous supernovae. The real surprise came hundreds of days later.
 
Radio observations revealed emission that continued to rise long after the optical flash had faded, peaking 641 days after discovery at a luminosity of around 2 × 10³¹ erg s⁻¹ Hz⁻¹, more than an order of magnitude brighter than previously known radio-bright LFBOTs and comparable to relativistic jetted tidal disruption events. Such behavior immediately suggested that conventional explosion models were insufficient.

Turning observations into physics

Understanding the source required far more than comparing observations with previous events. The research team combined observational astronomy with advanced computational astrophysics to determine which physical scenario best reproduced every aspect of the transient.
 
Their first step involved fitting the optical spectral energy distribution using an MCMC framework with 64 walkers and 2,000 sampling steps to estimate the evolving temperature, luminosity, and emitting radius of the transient. These calculations established the unusually persistent thermal properties that distinguish AT2019ijn from known fast optical transients. The radio observations presented an even greater computational challenge.

Modeling a relativistic jet

To explain the delayed radio brightening, the researchers investigated whether AT2019ijn launched a relativistic jet pointed away from Earth. They employed VegasAfterglow, a high-performance numerical framework designed for multiwavelength afterglow simulations and Bayesian parameter estimation. The software models how relativistic jets propagate through the interstellar medium while accounting for synchrotron radiation, relativistic beaming, jet geometry, and energy transport.
 
The parameter space explored was enormous. The simulations considered initial Lorentz factors between 5 and 1,000, isotropic-equivalent jet energies spanning six orders of magnitude, interstellar medium densities covering five orders of magnitude, jet opening angles from 0° to 30°, and viewing angles ranging from directly on-axis to completely off-axis. Each candidate solution was evaluated through MCMC optimization using 16 walkers and one million sampling steps.
 
Such large Bayesian searches are precisely the kind of workload that benefits from leadership-class supercomputing systems, where thousands of parameter combinations can be evaluated simultaneously.

The best-fitting universe

The simulations converged on a remarkably energetic solution. The preferred model indicates that AT2019ijn produced a narrow relativistic jet with an opening angle of roughly 7°–10°, viewed from approximately 40° off-axis. The inferred isotropic-equivalent kinetic energy approaches 10⁵⁴ erg—comparable to the most energetic relativistic explosions known.
 
Because the jet was not pointed directly toward Earth, relativistic beaming initially suppressed the radio signal. As the jet slowed while interacting with surrounding gas, its emission gradually entered our line of sight, naturally producing the observed radio peak more than 600 days after the optical outburst. Without computational modeling, this delayed evolution would have remained difficult to interpret.

Testing competing physical models

The study did not stop with jet modeling. Researchers also examined whether the optical emission could originate from a newly born magnetar, a rapidly rotating neutron star with an extremely strong magnetic field. Bayesian fitting reproduced several optical properties, suggesting a millisecond spin period and magnetic field near 10¹⁴ gauss. However, the enormous radio energy proved difficult to reconcile with a magnetar unless highly specialized conditions were invoked.
 
The team then modeled the event using MOSFiT, a widely used computational framework for tidal disruption events. The best-fitting solution involved an intermediate-mass black hole of approximately 10⁵ solar masses disrupting a low-mass star. Bayesian model evaluation using the Widely Applicable Information Criterion (WAIC) indicated that this scenario is consistent with known tidal disruption events while naturally explaining the unusually rapid rise of the transient. Combining the optical fits, radio simulations, and host galaxy properties led the researchers to favor a jetted tidal disruption event involving an intermediate-mass black hole.

Supercomputing changes time-domain astronomy

The broader significance extends well beyond a single transient. Future observatories, including the Vera C. Rubin Observatory, the Square Kilometre Array, and the Nancy Grace Roman Space Telescope, will discover millions of transient events every year. Finding them will no longer be the limiting factor. Interpreting them will.
 
Each newly detected transient may require thousands or millions of numerical realizations spanning relativistic hydrodynamics, radiation transport, Bayesian inference, jet evolution, and statistical model comparison before astronomers can identify its physical origin. The bottleneck is rapidly shifting from telescope sensitivity to computational capability.

The next generation of discovery

AT2019ijn may ultimately represent the first recognized member of a previously unknown family of relativistic optical transients. The authors conclude that combining wide-field optical surveys with deep radio monitoring will be essential for discovering additional examples and determining how frequently intermediate-mass black holes launch relativistic jets. For the supercomputing community, the message is equally compelling.
 
The future of transient astronomy will not be defined solely by larger telescopes or more sensitive detectors. It will be shaped by the computational power needed to recreate extreme astrophysical environments, evaluate millions of possible universes, and identify the one that best matches reality. In that sense, every new supercomputer becomes more than a scientific instrument. It becomes a machine capable of revealing the hidden engines powering the most extraordinary explosions in the cosmos.
Antarctic ice meets the rocky coastline. Researchers traced landscape features from the two-kilometre-high coastal escarpment of Dronning Maud Land to the subglacial Gamburtsev Mountains, buried beneath 1–3 km of ice Credit Matt Palmer
Antarctic ice meets the rocky coastline. Researchers traced landscape features from the two-kilometre-high coastal escarpment of Dronning Maud Land to the subglacial Gamburtsev Mountains, buried beneath 1–3 km of ice Credit Matt Palmer
Featured

Rebuilding a lost continent: Supercomputers reveal Antarctica before the ice

Advanced landscape evolution models uncover how the rise of Antarctica’s mountains may have primed the continent for its first great ice sheet

Reconstructing a continent’s geological evolution over hundreds of millions of years is among the most computationally demanding challenges in science. As mountains rise, rivers carve landscapes, and continents drift, these processes interact with shifting atmospheric and oceanic conditions to create an incredibly complex multiphysics puzzle.
 
By integrating geological observations, thermochronology, and paleoclimate data with large-scale numerical simulations, an international research team has successfully peered back through 160 million years of Antarctic history. Their findings suggest that the continent's dramatic topography predates the formation of its first major ice sheet, fundamentally reshaping our understanding of Antarctica’s transition from a temperate region to a frozen wilderness. Beyond these geological insights, the study underscores the indispensable role of high-performance computing in unlocking our planet's deep history.

Turning deep time into a computational problem

The evolution of Antarctica cannot be observed directly.
 
Instead, scientists must solve an immense inverse problem.
 
Starting with sparse geological evidence, including thermochronology measurements, erosion histories, present-day topography, geophysical observations, and paleoclimate records, they seek to reconstruct landscapes that disappeared tens of millions of years ago.
 
Accomplishing that requires coupling numerical models spanning tectonics, erosion, river incision, surface processes, paleoclimate, and ice-sheet evolution.
 
Rather than relying on static geological reconstructions, the researchers employed forward landscape evolution simulations that began approximately 160 million years ago and advanced to the Eocene–Oligocene transition roughly 34 million years ago. The simulations used one-kilometer spatial resolution, a computational timestep of 1,000 years, and generated topographic outputs every two million years, producing a detailed digital reconstruction of Antarctic landscape evolution across more than 126 million years.
 
That temporal scale alone illustrates the extraordinary computational challenge.

Simulating continental evolution

At the heart of the study lies a sophisticated landscape evolution model built upon the open-source FastScape framework, one of computational geoscience’s leading numerical platforms for simulating long-term erosion and tectonic evolution. The supplemental methods describe extensive parameter optimization, erosion modeling, thermochronology predictions, and uncertainty analysis used to reproduce the continent’s ancient topography.
 
The model incorporated numerous interacting physical processes, including:
  • River incision and drainage evolution
  • Hillslope diffusion
  • Surface erosion
  • Lithospheric flexure
  • Escarpment formation
  • Mountain uplift
  • Sediment transport
Unlike simplified geological reconstructions, these simulations allowed the Antarctic landscape to evolve naturally according to the governing physical equations.
 
Every simulated timestep updated elevation, erosion, drainage patterns, and surface morphology, gradually transforming an initial continental configuration into the reconstructed Antarctica observed at the onset of major glaciation.

Optimizing millions of years of history

Running a landscape model is only the beginning.
 
Determining whether that simulation accurately represents reality requires comparing model output against multiple independent geological datasets.
 
The researchers therefore performed extensive parameter optimization using an automated misfit-minimization approach that simultaneously evaluated escarpment position, plateau elevations, Gamburtsev Mountain heights, erosion histories, and thermochronological age constraints.
 
Rather than producing a single deterministic solution, the study explored numerous parameter combinations to quantify uncertainty and identify the highest-quality reconstructions.
 
This type of computational optimization exemplifies modern Earth-system modeling.
 
Instead of asking, “Can this model reproduce Antarctica?”
 
Scientists ask, “Among thousands of physically plausible models, which best matches every available observation?”
 
Answering that question requires substantial computational resources.

Bridging geology and climate

The reconstructed landscapes became inputs for additional climate simulations.
 
The researchers coupled paleotopographic reconstructions with an energy-balance climate model to investigate how evolving mountain ranges altered Antarctic temperatures and snowfall.
 
The supplemental analyses demonstrate that the model successfully reproduces expected polar amplification behavior across varying global mean temperatures while remaining consistent with independent paleoclimate reconstructions.
 
This coupling between landscape evolution and climate modeling is particularly significant.
 
Mountain building influences atmospheric circulation.
 
Atmospheric circulation affects snowfall.
 
Snowfall determines where glaciers can form.
 
Those glaciers subsequently reshape the landscape through erosion.
 
Capturing these feedbacks requires solving tightly coupled numerical systems spanning multiple scientific disciplines.

The hidden role of the Gamburtsev Mountains

Among the study’s most intriguing conclusions is the importance of the Gamburtsev Subglacial Mountains.
 
Buried beneath kilometers of ice in East Antarctica, these mountains have long puzzled geologists because they rival major alpine ranges despite lying deep within an ancient continental craton.
 
The simulations indicate that elevated interior topography existed well before continent-wide glaciation, providing favorable conditions for early ice accumulation once global temperatures cooled sufficiently. References throughout the study connect this interpretation with decades of geophysical, thermochronological, and landscape investigations of the Gamburtsev Mountains and surrounding East Antarctica.
 
Rather than mountains simply surviving beneath the ice sheet, the research suggests they may have actively helped initiate Antarctica’s transition into an ice-covered continent.

Supercomputers as geological time machines

Perhaps the greatest achievement of the project lies not in a single scientific conclusion but in its computational methodology.
 
The simulations reconstruct processes occurring over geological timescales that dwarf the duration of human civilization.
 
No laboratory experiment can reproduce 160 million years of erosion.
 
No field expedition can observe mountain formation across tens of millions of years.
 
Only numerical simulation allows scientists to investigate such questions quantitatively.
 
By integrating geological observations, thermochronology, paleoclimate constraints, landscape evolution algorithms, and uncertainty quantification into a unified computational workflow, researchers transformed fragments of Earth’s history into a coherent digital narrative.

Why HPC matters for climate science

The broader implications extend far beyond Antarctica.
 
Modern climate science increasingly depends on understanding Earth’s long-term geological evolution.
 
Continental topography influences atmospheric circulation.
 
Ocean basin geometry governs heat transport.
 
Mountain ranges alter precipitation.
 
Landscape evolution affects carbon cycling through weathering and erosion.
 
Each process operates over millions of years yet continues influencing climate today.
 
High-performance computing enables researchers to couple these processes into comprehensive Earth-system models capable of exploring interactions that would otherwise remain inaccessible.
 
As computing power continues to increase, future models will incorporate finer spatial resolution, more sophisticated physical parameterizations, and increasingly realistic coupling between tectonics, climate, ice sheets, and ocean circulation.

Reconstructing the future by understanding the past

Although this research examines events tens of millions of years old, its relevance is thoroughly modern. Understanding how Antarctica first became glaciated provides critical context for predicting how its ice sheets may respond to future climate change. The study demonstrates that today’s Antarctic landscape is the product of an extraordinarily long geological evolution, one that can now be explored with unprecedented fidelity through advanced numerical simulation.
 
The message is equally compelling for the supercomputing community: the world’s fastest machines are no longer limited to forecasting tomorrow’s weather or simulating future technologies. They are now being used to reconstruct worlds that vanished millions of years ago, turning every processor core into a window into Earth’s deep past. Each simulation deepens our understanding of how landscapes, climates, and ice sheets evolved in tandem; as computing power grows, so does our ability to explore not just where our planet is headed, but the complex geological journey that shaped the world we inhabit today.
Jill Mesirov, PhD
Jill Mesirov, PhD
Featured

The future of cancer research runs on supercomputers

Sanford Burnham Prebys recruits one of the world’s leading computational biologists to accelerate AI-driven biomedical discovery through advanced computing

For decades, supercomputers have reshaped our understanding of the universe through feats like simulating stellar explosions, modeling global climate patterns, and engineering next-generation aircraft. Today, these immense computational capabilities are being directed inward to address one of science’s greatest challenges: deciphering the complex language of human biology.
 
This shift was underscored this week by the appointment of Dr. Jill P. Mesirov, a pioneer in computational biology and cancer genomics, as Distinguished Professor and Senior Vice President for Computational Science at Sanford Burnham Prebys. Her recruitment marks a strategic effort to integrate advanced computing, artificial intelligence, and data science into the core of the institute’s biomedical research. For the high-performance computing (HPC) community, this move is more than just a key hire; it signals the definitive convergence of supercomputing, AI, and modern medicine.

Biology has become an HPC problem

Biological research has undergone a remarkable transformation during the past twenty years. Sequencing a human genome once required years of effort and billions of dollars. Today, thousands of genomes can be sequenced in days. Single-cell sequencing technologies now generate millions of individual cellular measurements from a single experiment, while spatial transcriptomics and advanced imaging systems produce multidimensional datasets measured in petabytes.
 
Extracting meaningful biological insight from these data is no longer primarily an experimental challenge.
 
It is a computational one.
 
Modern cancer research depends upon algorithms capable of integrating genomic, transcriptomic, proteomic, metabolomic, and clinical datasets simultaneously. These analyses involve billions of variables and demand computational infrastructures that resemble those found in national supercomputing centers.
 
Mesirov has spent her career developing precisely these kinds of computational approaches, helping establish data science as a central pillar of biomedical research. Her work has contributed to widely used computational tools and analytical frameworks that enable researchers to interpret complex genomic information and identify molecular mechanisms underlying disease.

Beyond bioinformatics

The title “computational biology” scarcely captures the breadth of modern biomedical computing.
 
Today’s computational scientists build machine-learning models capable of identifying previously unknown cancer subtypes, predicting patient responses to therapy, reconstructing cellular signaling networks, and discovering molecular biomarkers hidden within enormous genomic datasets.
 
Each of these workflows requires sophisticated numerical methods executed across large-scale computing systems.
 
Genome-wide association studies routinely analyze millions of genetic variants across hundreds of thousands of individuals.
 
Single-cell RNA sequencing experiments may profile millions of cells simultaneously.
 
Deep-learning pathology systems process gigapixel microscope images using thousands of GPU cores.
 
Drug discovery platforms evaluate billions of molecular interactions through simulation and AI-guided optimization.
 
Collectively, these workloads have made biomedical research one of the fastest-growing consumers of high-performance computing resources worldwide.

Supercomputers as biomedical instruments

Traditional scientific instruments observe nature.
 
Supercomputers increasingly function as instruments themselves.
 
Rather than collecting photons or particles, they construct mathematical representations of biological systems, allowing scientists to investigate processes that cannot be observed directly.
 
Large GPU clusters now train foundation models on genomic sequences, protein structures, electronic health records, and multimodal imaging data. These models can identify relationships that would be impossible for human investigators to recognize manually.
 
Increasingly, biological discovery begins not in the laboratory but inside large-scale computational infrastructure.
 
This shift explains why research institutions are investing heavily in computational leadership alongside experimental expertise.
 
By recruiting Mesirov, Sanford Burnham Prebys is reinforcing the idea that future biomedical breakthroughs will emerge from close integration between laboratory science and advanced computing. The institute has emphasized expanding capabilities in data science and AI as part of its broader research strategy.

AI changes the scale of discovery

Artificial intelligence is rapidly changing every stage of biomedical research.
 
Deep neural networks now predict protein structures with remarkable accuracy.
 
Generative AI models assist researchers in designing new therapeutic molecules.
 
Machine learning accelerates image segmentation, genomic classification, biomarker discovery, and clinical decision support.
 
Yet AI itself depends upon extraordinary computational infrastructure.
 
Training state-of-the-art biomedical foundation models requires clusters containing thousands of GPUs connected through high-bandwidth interconnects, supported by distributed storage systems capable of delivering terabytes of data every second.
 
The resulting computational demands rival those of traditional scientific supercomputing applications.
 
As AI becomes embedded within biomedical research, institutions capable of combining biological expertise with leadership-class computing infrastructure will possess a growing competitive advantage.

The rise of computational medicine

Medicine is steadily becoming a predictive science.
 
Instead of reacting after disease develops, researchers increasingly seek to model disease progression before symptoms appear.
 
Digital representations of tumors can simulate therapeutic response.
 
Network models identify previously unknown disease pathways.
 
Multiomic analyses reveal subtle molecular signatures long before conventional diagnostics detect abnormalities.
 
These capabilities depend upon sophisticated computational pipelines integrating simulation, statistical inference, machine learning, uncertainty quantification, and large-scale data management.
 
Each represents a mature discipline within high-performance computing.
 
Rather than replacing laboratory experiments, supercomputers now guide them.
 
Scientists can prioritize promising therapeutic targets computationally before committing years of experimental effort.
 
This dramatically shortens the path from hypothesis to discovery.

Converging scientific disciplines

The significance of Mesirov’s appointment extends beyond cancer research.
 
It reflects a broader transformation occurring across scientific computing.
 
Historically, computational biology evolved separately from traditional HPC disciplines such as computational fluid dynamics, astrophysics, and climate modeling.
 
Today, those boundaries are dissolving.
 
Shared technologies, including GPU acceleration, distributed computing, cloud-native workflows, AI frameworks, high-performance storage, and advanced visualization, are becoming universal scientific tools.
 
The same accelerator architectures used to simulate galaxy formation now train genomic foundation models.
 
Parallel computing techniques originally developed for physics increasingly drive precision medicine.
 
The future of supercomputing is no longer defined by scientific discipline.
 
It is defined by computational capability.

Building the next generation of discovery

Perhaps the most inspiring aspect of this appointment is what it represents for the future of biomedical research.
 
Scientific progress has always depended upon better instruments.
 
Microscopes revealed cells.
 
DNA sequencers revealed genomes.
 
Today, supercomputers reveal patterns hidden within biological complexity.
 
Every additional GPU, every faster interconnect, every more efficient algorithm expands researchers’ ability to understand disease at unprecedented resolution.
 
By bringing one of computational biology’s most influential leaders to Sanford Burnham Prebys, the institute is making a clear statement about where biomedical science is headed.
 
The laboratories of the future will still contain microscopes, sequencers, and imaging systems.
 
But they will also rely upon leadership-class computing clusters, artificial intelligence, and computational scientists capable of translating massive datasets into actionable biological knowledge.
 
For the supercomputing community, that evolution represents one of the most exciting frontiers in computational science.
 
The next life-saving medical breakthrough may not emerge solely from a laboratory bench.
 
It may first appear within the processors of a supercomputer, where mathematics, biology, and artificial intelligence converge to reveal discoveries that would otherwise remain invisible.
POPULAR RIGHT NOW
THIS YEAR'S MOST READ
MOST READ OF ALL-TIME