Wine and Cheese Spring 2025

This page records the schedule, titles and abstracts of the JHU/STScI CAS Astrophysics Wine & Cheese Series in Spring 2025.

Wine and Cheese sessions with one speaker will have a 50 minute talk with 10 minutes for questions. Sessions with two speakers will have two 25 minute talks, each with 5 minutes for questions.

03 February

Cameron Trapp (JHU)

Torques and Radial Flows of Gas in Simulated Milky-Way Mass Galaxies
Observations indicate that a continuous gas supply is needed to maintain observed star formation rates in large, disky galaxies. To fuel star formation, gas must reach the inner regions of such galaxies. Despite its crucial importance for galaxy evolution, how and where gas joins galaxies is poorly constrained observationally and rarely explored in fully cosmological simulations. I will discuss the results of our initial study investigating gas accretion and transport in the FIRE-2 cosmological zoom-in simulations for 4 Milky Way mass galaxies. We generally found that gas joins just interior to the disk edge before radially transporting through the disk at average speeds of 1-5 km/s. This corresponds to radial mass fluxes of a few solar masses per year, comparable to the galaxies’ star formation rates. I will also discuss more recent work focused on understanding the angular momentum transfer required for these gas flows, including torques arising from various gravitational and magnetohydrodynamical forces. Finally, I will give a quick introduction to the work I will be doing as a new postdoc here at JHU, working with the FOGGIE simulations.

10 February

George Wong (IAS)

Precision Black Hole Astrophysics in the Era of Event-Horizon-Scale Observation
Black holes are ubiquitous and essential to our understanding of the universe, shaping galaxy evolution, driving magnetized outflows, and providing a natural laboratory for theories of gravity, high-energy plasma physics, and relativistic accretion. Their extreme environments also offer insights into broader astrophysical processes, from planetary accretion disks to pulsar magnetospheres and beyond. Over the past decade, cutting-edge interferometric experiments like the Event Horizon Telescope (EHT) have produced exquisite, transformative horizon-scale observations. These data provide an unprecedented opportunity to probe relativistic plasma physics, test general relativity in the strong-field regime, and constrain the mechanisms of accretion and jet formation. I will discuss the state-of-the-art in supermassive black hole accretion modeling and highlight how these methods have been applied to EHT data, producing quantitative constraints on near-horizon physics. I will then describe recent advances in identifying robust observational signatures of black hole spin and spacetime geometry from semi-analytic arguments. The next generation of black hole experiments promises to revolutionize our understanding of accretion and jet physics through a combination of space-based interferometry, expanded VLBI arrays, and high-energy multi-wavelength observatories. I will conclude with a discussion of how we will bridge the gap between modeling and observations, paving the way for precision black hole astrophysics in the coming decade.

17 February

Viraj Pandya (Columbia)

Decoding the Complexity of Galaxy Formation with Physics-Informed, AI-Accelerated Dynamical Models
Galaxies are complex dynamical systems evolving against the backdrop of cosmology. One of the grand challenges of modern astrophysics is to build a fully predictive theory for galaxy formation so that we can reliably use them to understand the fundamental physics of dark matter, dark energy and inflation. However, the relevant physical processes are poorly understood and non-linearly coupled over an enormous range of spatiotemporal scales. This, combined with our inability to observe the time evolution of individual systems, necessitates phenomenological approaches. I will describe my interdisciplinary efforts to use population studies to decode the ordered complexity of galaxies. First, I will show that my JWST discovery of a significant excess of linear, elongated galaxies at high-redshift has deep implications for the origin of the Hubble Sequence. Galaxies may not start out as axisymmetric disks as commonly assumed, but rather as prolate (cigar-shaped) or triaxial (surfboard-shaped) ellipsoids. This preferential elongation is naturally expected from the tidal field of the filamentary cosmic web, and it motivates an exciting search for intrinsic alignments among this dominant population of "early blue ellipticals" with JWST, Roman and Euclid. Although this puzzle may seem niche, it bridges together many different subfields (including astrostatistics, stellar dynamics, Galactic archaeology and dark matter phenomenology) and unlocks fresh science cases for HWO and GMT. Second, I will introduce a novel automatically differentiable, GPU-capable framework to help definitively answer ~50-year-old questions about how galaxies self-regulate and why they follow remarkably tight scaling relations. In my new model, supernovae and black holes over-pressurize galactic atmospheres, which naturally explains why galaxies are so inefficient at forming stars. But robustly testing this paradigm requires physics-informed, data-driven, AI-accelerated techniques for parameter inference and causal discovery. I will highlight some promising early applications of gradient descent optimization, Hamiltonian Monte Carlo and population-level implicit likelihood inference. These reveal that observations of the circumgalactic medium may hold the key to breaking numerous modeling degeneracies. I will conclude by outlining how my future research will enable transformative cosmology-style precision astrophysics for galaxy formation, provide a much-needed interpretable emulator for expensive hydrodynamical simulations, and rigorously connect observations with theory to maximize the impact of multiple JHU-led missions (e.g., SDSS-V, Rubin, PFS and NASA's next Great Observatories).

24 February

Ting Li (Toronto)

From Stellar Streams to Near Field Cosmology: Insights from Large-Scale Spectroscopic Surveys
Stellar streams serve as exceptional tracers in near-field cosmology, providing critical insights into galaxy formation and evolution, as well as the fundamental nature of dark matter. My talk will feature two major ongoing spectroscopic programs targeting the Milky Way’s streams. The Southern Stellar Stream Spectroscopic Survey (S5) is the first systematic effort to map known streams in the Southern Hemisphere, utilizing the AAOmega spectrograph on the Anglo-Australian Telescope. Complementing this, the Milky Way Survey of the Dark Energy Spectroscopic Instrument (DESI) is a recently launched 5-year initiative targeting the Northern Hemisphere. Together, these surveys are delivering unprecedented 6D kinematic and chemical data on dozens of streams, transforming our understanding of the Milky Way’s chemo-dynamical evolution, including the tidal disruption of satellite galaxies and globular clusters. I will also discuss the broader implications of these findings for dark matter and near-field cosmology, concluding with a perspective on next-generation spectroscopic surveys.

03 March

Sumit Sarbadhicary (JHU)

Where do stars explode in the interstellar medium?
The physics of stellar feedback is arguably our largest uncertainty in the understanding of galaxy evolution. Particularly contentious is the question of where stars explode with respect to gas clouds in the interstellar medium (ISM), which plays a major role in how efficiently supernovae (SNe) drive turbulence in the ISM vs outflows in order to regulate star-formation. Addressing this from modern observations is urgently needed by the field. The nearest galaxies (1-20 Mpc) are the most exciting laboratories for this purpose, since one gets the most detailed observations of the multi-phase ISM, the stellar populations powering feedback, and supernova sites from historical supernovae or their progenitors (e.g. red supergiants, Wolf-Rayet stars). In this talk, I will present the first set of direct observational constraints on where stars explode from ongoing high-resolution surveys of nearby galaxies with JWST, HST, VLA, ALMA and VLT/MUSE. We find that at least 50% of core-collapse supernovae occurred outside molecular clouds, implying that their environments were cleared by pre-SN/prior-SN explosions, and their feedback is occurring in the lower density atomic ISM. With a survey of red-supergiants and Wolf-Rayet stars in M33, we find that this fraction exploding outside molecular clouds has a mass-dependence, with 44% for stars > 30 Msun, and 72% for stars 8-30 Msun. Independently with JWST, we find that about 20-40% of supernova remnants show direct evidence of interaction with dense gas/molecular clouds in the infrared. I will discuss how these empirical measurements of SN fractions interacting with dense vs diffuse gas can be a novel observational anchor for stellar feedback models that underpin hydrodynamical simulations of galaxy formation, and ongoing expansion of this work to a sample of nearly 80 star-forming galaxies.

Marco Chiaberge (STScI/JHU)

On Human Spaceflight and Mice: Exercise Countermeasures to Joint Cartilage Degradation in Long-Duration Spaceflight (to the Moon and Mars)
Astronauts aboard the International Space Station exercise regularly to counter the damaging effects of microgravity. However, one aspect that has been largely overlooked—yet is becoming an increasing focus of research—is the degeneration and thinning of articular cartilage, which may lead to osteoarthritis as a consequence of long duration spaceflight. For extended missions such as a prolonged human presence on the Moon, or a mission to Mars, this may lead to functional impairments for the crew. In this talk, I will take you through our 5+ year journey, during which our multidisciplinary team from JHU, Carnegie and other institutions designed, built and utilized an experimental apparatus for plyometric (i.e. jump) training in mice to investigate the potential benefits of jumping for cartilage health. I will share the unexpected results we obtained—originally from what was meant to be a simple test experiment—and discuss their implications for long-term human space missions, as well as potentially for osteoarthritis treatments on Earth.

10 March

Sihao Cheng (IAS)

From Data to Discovery: Mining Hidden Insights from Astronomical Surveys
Astronomical surveys hold immense discovery potential, but unlocking their secrets often requires creative mining. I will discuss efficient strategies to extract insights from data. First, I will present my discovery of buoyant "icebergs" in white dwarf stars and hidden resonance structures in the Milky Way, both of which are enabled by novel data analyses with guidance from physics. Next, I will discuss the frontier of decoding complex datasets in astrophysics, with new techniques that achieve neural-net performance while preserving interpretability. These cases highlight a critical lesson: discovery is driven not only by new data, but also by innovative ideas that transcend disciplinary boundaries. I will conclude with how these strategies can be applied to next-generation sky surveys to open new discovery spaces.

24 March

Eric Murphy (NRAO)

The next-generation Very Large Array: The Next Great Ground-Based Observatory
Inspired by dramatic discoveries from the Jansky VLA, VLBA, and ALMA, a large collecting area radio interferometer that will open new discovery space from proto-planetary disks to distant galaxies is being designed by the U.S. National Radio Astronomy Observatory (NRAO) and the broad scientific and technical communities. The next-generation VLA (ngVLA), which was strongly endorsed by the Astro2020 Decadal Survey as an essential research facility whose construction should begin this decade, is envisaged as an interferometric array with ten times greater sensitivity and spatial resolution than the current VLA and ALMA, operating in the frequency range of 1.2 - 116 GHz. Replacing both the VLA and VLBA, the ngVLA will be optimized for observations in the spectral region between the superb performance of ALMA at mm and sub-mm wavelengths, and the future Phase I Square Kilometer Array (SKA-1) at decimeter-scale and longer wavelengths. As such, the ngVLA will uniquely tackle a broad range of outstanding scientific questions in modern astronomy by simultaneously delivering the capability to: unveil the formation of Solar System analogues on terrestrial scales; probe the initial conditions for planetary systems and life with astrochemistry; characterize the assembly, structure, and evolution of galaxies from the first billion years to the present; use pulsars in the Galactic center as fundamental tests of gravity; and understand the formation and evolution of stellar and supermassive blackholes in the era of multi-messenger astronomy. In this presentation, I will provide a project update and discuss the overall science case that led to the current technical design of the ngVLA, as well as provide an update the progress of the project.

31 March

Digvijay Wadekar (JHU)

Uncovering New Heavy Black Hole Mergers in Public Gravitational Wave Data
To date, approximately 100 binary black hole (BH) mergers have been identified in the LIGO-Virgo-KAGRA gravitational wave data. However, nearly all previous searches have relied on waveform templates that include only the dominant quadrupole mode, neglecting higher-order harmonics predicted by general relativity. In this talk, I will present ~10 new BH merger detections in the LIGO-Virgo O3 data discovered using a novel search pipeline that incorporates these higher-order modes. Several of these events exhibit astrophysically intriguing features, including black hole masses in the intermediate-mass and pair-instability mass gap ranges, high-redshifts (1<z<2), significant mass asymmetries, and positive effective spins. I will also highlight recent upgrades to our IAS pipeline that leverage AI/ML techniques to enhance sensitivity. Finally, I will shift focus to share an overview of my broader work applying interpretable AI/ML methods to problems in galaxy formation and cosmology.

William Balmer (JHU)

Living on the Wedge: Novel JWST Coronagraphy of Giant Exoplanets
What are planets made of, and how do they form? Only a handful of planets outside the solar system have been "driectly imaged" - that is, resolved apart from their bright hosts - yet direct imaging provides some of the most complete information about exoplanetary atmospheres and dynamics, and will be crucial in the quest to observe the occurrence of life in the galaxy in the next century. The planets we can resolve with current telescopes are young, widely separated, and massive compared to our own, and pose challenges to our understanding of planet taxonomy and formation. While these objects might not be massive enough burn deuterium, and are therefore commonly called "exoplanets," did they form from the top down through molecular cloud collapse or disk fragmentation, like stars and brown dwarfs do, or bottom up through the accretion of solids and gas, like most planets are expected to? High contrast observations from JWST in the mid-infrared can reveal key compositional clues to this mystery. I will present our GTO team's novel coronagraphic images of the emblematic HR 8799 and 51 Eri planetary systems using the NIRCam Long Wavelength Bar in an offset ``narrow" position. These observations reveal the known planets in each system at high throughput "in a new light," even at spatial separations challenging for a 6.5m telescope in the mid-IR. The chosen filters constrain the strength of CO, CH4, and importantly, for the first time, CO2 absorption, largely independent of cloud and mixing degeneracies that have hampered ground-based studies. The HR 8799 planets display a sequence of colors correlated with their orbital separation. They also show significantly stronger CO2 absorption than expected from solar metallicity models, indicating that they are metal enriched by at least 0.5 dex compared to their host. We detected 51 Eri b 4.1 micron and not at longer wavelengths, which, given the planet's temperature, is indicative of significantly out-of-equilibrium carbon chemistry, and a moderately enhanced metallicity. This appears to indicate these massive, widely separated giants formed via core accretion. These results present an opportunity to trace the metallicity and formation histories of directly imaged planets unambiguously for the first time, and imply efficient core assembly in the massive disks surrounding A-type stars. The work I'll present was published this month in AJ, and received a joint JHU/STScI/NASA press release, which you can read ahead of time here: https://webbtelescope.org/contents/news-releases/2025/news-2025-114. I'll also briefly discuss a few on-going efforts to image the coldest planets possible with JWST, and my team's successful JWST Cycle 4 programs in both these arenas.

7 April

Elena Rossi (Leiden)

Galactic Centres and their High-Energy Multimessenger Phenomena
Galactic Nuclei are the densest stellar systems in the Universe, and they host massive black holes (MBHs). A host of energetic and unique phenomena, --such as tidal disruption events, hypervelocity stars and extreme mass ratio inspirals--, arise from these outstanding features of galactic nuclei. In this talk, | I will present my group work on state-of-the-art dynamical calculations that link the rate and properties of these different phenomena and show how they can be used to investigate MBH environments at scales hardly accessible (or not yet accessible) by direct observations.

14 April

Vida Saeedzadeh (JHU)

Cool and gusty, with a chance of rain: dynamics of multiphase circumgalactic medium around massive galaxies
The circumgalactic medium (CGM) is a dynamic and multiphase medium that plays a key role in galaxy evolution. In this talk, I will present results from my PhD research using the high-resolution Romulus simulations to investigate the origin and evolution of the CGM around massive central galaxies in group-scale halos. I identified seven evolutionary patterns of CGM and explored the cooling pathways. The gas cools via two modes: (1) filamentary inflows and (2) condensations forming from rapidly cooling density perturbations. I showed that in cosmological simulations, these perturbations are primarily seeded by orbiting substructures. The condensations can form even when the median t_cool/t_ff of the X-ray emitting gas is above 10 or 20. Strong perturbations can trigger runaway cooling regardless of the background gas conditions. Conversely, I also find perturbations whose local t_cool/t_ff ratios dip below the threshold but do not condense — instead, they bounce. These are weak perturbations temporarily swept up in satellite wakes and carried to larger radii, where t_cool/t_ff decreases because t_ff is increasing, not because t_cool is decreasing. This study highlights the challenge of using a simple threshold argument to infer the CGM's evolution. It also highlights that the median hot gas properties are suboptimal determinants of the CGM's state and dynamics. Realistic CGM models must incorporate the impact of mergers and orbiting satellites, along with the CGM's heating and cooling cycles.

Finally, I’ll close with a brief look at my current research with the FOGGIE collaboration, where I’m studying the CGM in UV emission. This work explores how spatial and spectral resolution, as well as instrumental sensitivity, affect which CGM structures are detectable and how much of the total observable mass and gas kinematics are influenced by these criteria.

Laura Herold (JHU)

Cosmological neutrino mass bounds with DESI from a Bayesian & frequentist perspective
The DESI galaxy survey has recently placed the tightest constraint on the sum of neutrino masses to date. For such effects “below the detection limit”, where data can only infer upper bounds, Bayesian and frequentist methods can give important complimentary information. I will begin with an overview of the frequentist profile likelihood method, its advantages and limitations. Using a frequentist and Bayesian toolbox, I will discuss neutrino mass constraints from Planck and DESI data. In particular, I will focus on the impact of different assumptions about the neutrino mass hierarchy on the inferred mass bounds: while it has been shown for previous experiments that the normal and inverted neutrino mass hierarchies are well approximated by three degenerate-mass neutrinos, we scrutinise this approximation in light of DESI BAO data.

21 April

Guadalupe Tovar Mendoza (JHU)

Cuentos de las Estrellas: Understanding the Properties and Impacts of Stellar Flares Across Time and Wavelength
Stellar magnetic activity remains a limiting factor for both planet detection and characterization, particularly around active M-type stars. Here, I present the results of my PhD studies which focused on understanding of the morphology, energies, temperatures, and frequency of stellar flare events across various time and wavelength regimes. By leveraging existing datasets such as Spitzer, Kepler, and TESS I studied how stellar activity impacts our search for life on exoplanets. In part one of my thesis, I used Kepler data to study the morphology or shape of stellar flares. I derived a continuous and analytical model to describe the shape of flare events in photometric observations. I tested the flare template on datasets of various cadences and wavelengths and found the model is versatile enough to fit flare events in all of the tested cases. In the second part of my thesis, I studied the energetics of flare events at infrared wavelengths. I used archival Spitzer observations of the active, planet hosting star, TRAPPIST-1, to study how flare temperatures vary at infrared observations and compare our results to the latest JWST observations of TRAPPIST-1. The final part of my thesis considered what the upcoming flagship mission, Nancy Grace Roman Space Telescope will provide for flare science. Specifically, we simulated 10,000 Roman light curves with realistic noise properties and used a flare frequency distribution model to inject flares into the light curves and perform recovery tests. We found that Roman will need to consider cadence, filters, and read-out strategies in order to maximize and expand our knowledge of stellar magnetic activity.

Xiaosheng Zhao (JHU)

From 21 cm Astrophysics to Galactic Archaeology: Enriching Physics-Driven Analysis with Machine Learning
The era of machine learning (ML) has opened new opportunities for extracting physical insights from complex astronomical data. During my PhD, I explored challenges in extracting physical information from 3D 21 cm images during the epoch of reionization, where strong foregrounds, signal non-Gaussianity, and slow forward modeling limit conventional analysis. I applied implicit likelihood inference (simulation-based inference) and introduced new summary representations using convolutional neural networks and wavelet scattering transforms. I also developed diffusion models to accelerate simulation-based studies without sacrificing information content.

Now in my postdoctoral research, I focus on Galactic Archaeology using stellar spectra. I investigate how undetected secondary-component stars in local group galaxies affect the uncertainty of radial velocity measurements from low- and medium-resolution stellar spectra. Parallel to this, I am developing a contrastively trained foundation ML model that aligns stellar spectra across instruments to support cross-survey calibration, parameter estimation, and anomaly detection.

Together, these projects demonstrate how machine learning methods can enrich physics-driven analysis across different domains, offering scalable and informative tools for the next generation of astronomical datasets.