Using Structure Functions to Characterize Protoplanetary Disk Velocity Residuals
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Astrophysics Radio. Generated commentary on the latest astrophysics papers.
Vera: We're looking at the paper: "Using Structure Functions to Characterize Protoplanetary Disk Velocity Residuals".
Jocelyn: Here's the gist of the paper. Protoplanetary disks are highly dynamic environments where high angular and spectral resolution observations reveal non-Keplerian motions indicative of (magneto-)hydrodynamical instabilities or planet-disk interactions The second-order structure function is proposed as a tool for recovering structural properties of velocity perturbations in the presence of significant noise The study introduces the second-order structure function, S2, to identify and characterize structure in apparently noise-dominated velocity maps This approach involves using morphological models to generate ensembles of realizations that mimic known (magneto-)hydrodynamical instabilities The researchers employ correlated, zero-mean Gaussian random fields (GRFs) drawn on the disk polar grid, allowing for the construction of a noiseaware pipeline that de-noises at the level of S2 and assesses significance against matched noise-only null ensembles Structural Heuristics and Characterization A set of six empirical heuristics is proposed to summarize information from the structure function plots, including magnitude, scale, stationarity indices, and anisotropy These heuristics are defined as: 1.
Vera: Magnitude and Scale: derived from the average amplitude of variations by rescaling the radius-averaged S2 plateau to 2σ2 2.
Jocelyn: Stationarity: encoded in indices αˆr and αˆϕ, which capture how correlation lengths vary with reference radius 3.
Vera: Anisotropy: defined by the ratio A = lϕ/lr, where A > 1 describes structure elongated azimuthally Detection and Differentiation of Instabilities The pipeline is applied to velocity fields from numerical simulations of the gravitational (GI), magnetorotational (MRI), and vertical-shear (VSI) instabilities The analysis detects excess variance above the noise model robustly across all three simulations, with significance ranging from z ≈ 7 to z ≈ 120 depending on the noise conditions The MRI simulation is cleanly distinguished from both hydrodynamic simulations based on these structural heuristics Application to Real Data and Limitations The framework is applied to 12CO observations of TW Hya, revealing azimuthally elongated velocity structure that persists after removing the large-scale m = 1 component The study shows that the choice of background model directly changes the recovered amplitude and azimuthal heuristics Limitations include the fact that a Gaussian description cannot reproduce anti-correlation features like those in the GI spiral arm, necessitating dedicated coherent-ridge models Furthermore, the parametric model struggles even where it describes the field, as none of its shape parameters are informative for these simulations The paper concludes that excess variance is detected robustly against a matched noise-only null across all tested noise conditions The azimuthal length ˆls is the most readily detected property, and only where the field is genuinely anisotropic While parameter recovery is parameterdependent, the radial correlation length l0,r is recovered most reliably The application to inclined disks shows that structure remains robustly detected to i = 60◦, but the stretched beam smooths the field and compresses Aˆ The null must be forward-modeled as formal uncertainties on the velocity map overestimate its noise The autocorrelation of interferometric image noise is the point spread function rather than its autocorrelation, suggesting a beam-convolved white-noise null is correlated on a scale √2 too long References Astropy Collaboration, Robitaille, T.
Paper summary: Vera: So wrapping up this paper, "Using Structure Functions to Characterize Protoplanetary Disk Velocity Residuals," the authors are essentially proving that you can detect excess variance above a noise model consistently across different simulations of instabilities Jocelyn They used the second-order structure function, S2, and developed empirical heuristics to characterize the resulting velocity fields Subrahmanyan
Jocelyn: The main point is that this method gives us a way to quantify which properties of the disk velocity field remain recoverable even when there's a lot of noise present in the measurements Vera They found that azimuthally elongated structure, defined by an anisotropy ratio A greater than one, was the most easily detected property Subrahmanyan
Subrahmanyan: The paper shows that this method works robustly against matched noise-only null ensembles even under different noise conditions, with significance levels ranging from seven to one hundred twenty depending on the specific noise characteristics Jocelyn They also showed that for real data like 12CO observations of TW Hya, they could detect azimuthally elongated velocity structure that persisted after accounting for the large scale one component Vera
Vera: It seems like this work provides a concrete set of tools—the heuristics and the noise-aware pipeline—for astronomers to test if a measured velocity field actually contains physical structure or just random noise Jocelyn It’s about establishing what can be recovered in the presence of noise Subrahmanyan
Subrahmanyan: The implication for us is that when we look at these dynamic environments, we should use this S2 approach to decide which features are real physical signals and which are just observational artifacts Vera This helps guide future high-resolution observations of protoplanetary disks Jocelyn
Conclusion: Vera: So, we're wrapping up this look at the paper "Using Structure Functions to Characterize Protoplanetary Disk Velocity Residuals." Basically, they’ve shown how to use these structure functions to figure out what's actually going on in the velocity field when you have a lot of noise.
Jocelyn: Right. They took some data, and instead of just looking at the big picture, they used this second-order structure function, S2, as a way to find hidden patterns in the velocity perturbations.
Subrahmanyan: From my side, it’s about moving away from just seeing a smooth average and starting to characterize the actual structure present in these dynamics.
Vera: Exactly. And they developed these six empirical heuristics—things like magnitude, scale, stationarity indices alpha r and alpha phi, and anisotropy—to summarize all that information from the plots.
Jocelyn: The anisotropy ratio A is especially interesting; if it’s greater than one, it tells you the structure is stretched out in a specific direction across the disk.
Subrahmanyan: That directional stretching relates directly to how instabilities like the magnetorotational instability or gravitational instabilities manifest in a physical disk.
Vera: They tested this pipeline on simulations of three different kinds of instabilities—gravitational, magnetorotational, and vertical-shear—and they found excess variance above the noise model in all three cases.
Jocelyn: And they were pretty robust about it; the significance levels varied a lot depending on how much noise was there, sometimes as high as one hundred twenty.
Subrahmanyan: The clean distinction between the magnetorotational instability and the other two hydrodynamic simulations based on these structural features is a solid result for connecting theory to simulation.
Vera: They then applied this entire framework to real observational data from 12CO observations of TW Hya, finding that they could detect an azimuthally elongated velocity structure even after stripping away the big one component.
Jocelyn: That means we can use this approach to look for these kinds of features in actual telescope data, not just simulations.
Subrahmanyan: But there are some caveats they point out; the choice of the background model actually changes what amplitude and azimuthal features you recover, which means we have to be careful about our assumptions.
Vera: Right, and they also mentioned that a simple Gaussian description isn't good enough for complex features like spiral arms in gravitational instabilities, requiring more specialized models.
Jocelyn: So the takeaway is that this method helps us decide what physical signals are strong enough to recover when the data is noisy, but it doesn't solve every problem on its own.
Subrahmanyan: It’s a tool for filtering and characterizing structure in these complex environments, which opens up new ways to interpret what we see in protoplanetary disks.
Vera: It definitely gives us a better way to assess if that signal we're seeing is real physics or just noise.
Jocelyn: And it sets the stage for follow-up observations that can actually test those structural predictions.
Massachusetts Institute of Technology Department of Earth, Atmospheric and Planetary Sciences · Center for Astrophysics Harvard & Smithsonian Department of Astronomy University of Florida Niels Bohr International Academy Max Planck Institute for Astronomy Dipartimento di Fisica Università degli Studi di Milano
astro-ph.EP
Submitted: 2026-11-05
Updated: 2026-11-05
Comments: Accepted to ApJ
License: http://creativecommons.org/licenses/by/4.0/
Importance score: 97/100
The gist: Protoplanetary disks are highly dynamic environments where high angular and spectral resolution observations reveal non-Keplerian motions indicative of (magneto-)hydrodynamical instabilities or…
Key concepts
- Second-order structure function (S2)
- S2 is a mathematical tool used to reveal patterns or structure within velocity maps that are otherwise dominated by noise. It helps identify correlated variations in the velocity field, allowing researchers to distinguish real physical structures from random noise by analyzing how these variations change across different spatial scales.
- Empirical Heuristics
- These are six specific rules used to summarize the information extracted from S2 plots. These heuristics include measures of magnitude, scale, stationarity (how correlation lengths change with radius), and anisotropy (the shape or elongation of the structure). They provide a standardized way to quantify and describe the characteristics of the detected velocity structures.
- Noise-aware Pipeline
- This is a computational method designed to clean up velocity data by assessing its significance against known noise patterns. The pipeline uses ensembles of random fields drawn from disk models to create a 'null' reference. By comparing the actual data's S2 signal against these null models, researchers can robustly determine if the observed variations are statistically significant physical structures or just random noise.
- Anisotropy (A = lϕ/lr)
- Anisotropy describes whether a detected velocity structure is elongated in one direction compared to another. The ratio A compares the correlation length in the azimuthal ($ ext{l} ext{ extphi}$) direction to the radial ($ ext{l} ext{r}$) direction. An anisotropy value greater than 1 indicates that the structure is stretched or elongated around the disk, which can be a sign of specific instabilities.
Terminology
Summary
Protoplanetary disks are highly dynamic environments where high angular and spectral resolution observations reveal non-Keplerian motions indicative of (magneto-)hydrodynamical instabilities or planet-disk interactions The second-order structure function is proposed as a tool for recovering structural properties of velocity perturbations in the presence of significant noise
Methodology and Modeling
The study introduces the second-order structure function, S2, to identify and characterize structure in apparently noise-dominated velocity maps This approach involves using morphological models to generate ensembles of realizations that mimic known (magneto-)hydrodynamical instabilities The researchers employ correlated, zero-mean Gaussian random fields (GRFs) drawn on the disk polar grid, allowing for the construction of a noiseaware pipeline that de-noises at the level of S2 and assesses significance against matched noise-only null ensembles
Structural Heuristics and Characterization
A set of six empirical heuristics is proposed to summarize information from the structure function plots, including magnitude, scale, stationarity indices, and anisotropy These heuristics are defined as:
-
Magnitude and Scale: derived from the average amplitude of variations by rescaling the radius-averaged S2 plateau to 2σ2
-
Stationarity: encoded in indices αˆr and αˆϕ, which capture how correlation lengths vary with reference radius
-
Anisotropy: defined by the ratio A = lϕ/lr, where A > 1 describes structure elongated azimuthally
Detection and Differentiation of Instabilities
The pipeline is applied to velocity fields from numerical simulations of the gravitational (GI), magnetorotational (MRI), and vertical-shear (VSI) instabilities The analysis detects excess variance above the noise model robustly across all three simulations, with significance ranging from z ≈ 7 to z ≈ 120 depending on the noise conditions The MRI simulation is cleanly distinguished from both hydrodynamic simulations based on these structural heuristics
Application to Real Data and Limitations
The framework is applied to 12CO observations of TW Hya, revealing azimuthally elongated velocity structure that persists after removing the large-scale m = 1 component The study shows that the choice of background model directly changes the recovered amplitude and azimuthal heuristics Limitations include the fact that a Gaussian description cannot reproduce anti-correlation features like those in the GI spiral arm, necessitating dedicated coherent-ridge models Furthermore, the parametric model struggles even where it describes the field, as none of its shape parameters are informative for these simulations
Conclusion and Future Directions
The paper concludes that excess variance is detected robustly against a matched noise-only null across all tested noise conditions The azimuthal length ˆls is the most readily detected property, and only where the field is genuinely anisotropic While parameter recovery is parameterdependent, the radial correlation length l0,r is recovered most reliably The application to inclined disks shows that structure remains robustly detected to i = 60◦, but the stretched beam smooths the field and compresses Aˆ The null must be forward-modeled as formal uncertainties on the velocity map overestimate its noise The autocorrelation of interferometric image noise is the point spread function rather than its autocorrelation, suggesting a beam-convolved white-noise null is correlated on a scale √2 too long
References
Astropy Collaboration, Robitaille, T. P., Tollerud, E. J., et al. 2013, A&A, 558, A33
Astropy Collaboration, Price-Whelan, A. M., Sipőcz, B. M., et al. 2018, AJ, 156, 123
Astropy Collaboration, Price-Whelan, A. M., Lim, P. L., et al. 2022, ApJ, 935, 167
Bae, J., Teague, R. p. 0 Draft version November 6, 2026
Typeset using LATEX twocolumn style in AASTeX7. 1
Using Structure Functions to Characterize Protoplanetary Disk Velocity Residuals
Richard Teague ,
1 Sean M. Draft version November 6, 2026
Typeset using LATEX twocolumn style in AASTeX7. 1
Using Structure Functions to Characterize Protoplanetary Disk Velocity Residuals
Richard Teague ,
1 Sean M. Draft version November 6, 2026
Typeset using LATEX twocolumn style in AASTeX7. 1
Using Structure Functions to Characterize Protoplanetary Disk Velocity Residuals
Richard Teague ,
1 Sean M. Draft version November 6, 2026
Typeset using LATEX twocolumn style in AASTeX7. 1
Using Structure Functions to Characterize Protoplanetary Disk Velocity Residuals
Richard Teague ,
1 Sean M.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Barraza-Alfaro, M., Flock, M., Béthune, W., et al. 2025, ApJL, 984, L21
Bruno, R., & Carbone, V. p. 1 Lazarian 2005) and in quantifying fluctuations in the solar
wind and magnetized astrophysical plasmas (R. Carbone 2013).Where the paper says it
Casassus, S., & Pérez, S. 2019, ApJL, 883, L41
Czekala, I., Loomis, R. A., Teague, R., et al. 2021, ApJS, 257, 2
Esquivel, A., & Lazarian, A. p. 1 Esquivel & A. Lazarian 2005) and in quantifying fluctuations in the solar
wind and magnetized astrophysical plasmas (R.Where the paper says it
Flock, M., Ruge, J. P., Dzyurkevich, N., et al. 2015, A&A, 574, A68
Fukagawa, M., Izquierdo, A. F., Stadler, J., et al. 2026, ApJL, 1000, L15
Hall, C., Dong, R., Teague, R., et al. 2020, ApJ, 904, 148
Heinemann, T., & Papaloizou, J. p. 0 Heinemann &
J. Papaloizou 2009; M.Where the paper says it
Heyer, M. H., & Brunt, C. p. 1 2
Low 2002; M. Brunt 2004) as well
as in recovering the underlying velocity statistics from
line-of-sight velocity centroids (A.Where the paper says it
Hunter, J. D. 2007, Computing in Science and Engineering, 9, 90
Izquierdo, A. F., Stadler, J., Galloway-Sprietsma, M., et al. 2025, ApJL, 984, L8
Izquierdo, A. F., Bae, J., Facchini, S., et al. 2026, ApJL, 1000, L13
Jorsater, S., & van Moorsel, G. A. 1995, AJ, 110, 2037
Kolmogorov, A. N. 1941, Doklady Akademii Nauk SSSR, 30, 301
Larson, R. B. 1981, MNRAS, 194, 809
Lesur, G., Flock, M., Ercolano, B., et al. 2023
Monin, A. S., & Yaglom, A. p. 0 Kolmogorov 1941), the structure function is a
standard descriptor of turbulent and geophysical flows
and is widely used in terrestrial applications such as
the study of atmospheric and oceanographic turbulence
(A. Yaglom 1975).Where the paper says it
Nelson, R. P., Gressel, O., & Umurhan, O. M. 2013, MNRAS, 435, 2610
Ossenkopf, V., & Mac Low, M. p. 0 Ossenkopf & M. Mac
arXiv:2611.Where the paper says it
Paciorek, C. J., & Schervish, M. p. 1 Paciorek & M. Schervish (2006), and validated against
the analytic structure function predicted by the kernel
parameters.Where the paper says it
Pinte, C., Teague, R., Flaherty, K., et al. 2023
Speedie, J., Dong, R., Hall, C., et al. 2024, Nature, 633, 58
Teague, R. 2019, The Journal of Open Source Software, 4, 1220
Teague, R., & Foreman-Mackey, D. 2018, Research Notes of the American Astronomical Society, 2, 173
Teague, R., Bae, J., Andrews, S. p. 0 1
Using Structure Functions to Characterize Protoplanetary Disk Velocity Residuals
Richard Teague ,
1 Sean M. Andrews ,
2 Jaehan Bae ,
3 Marcelo Barraza-Alfaro ,
1, 4
Abigail Bodner ,
1 Jonathan D.Where the paper says it
Virtanen, P., Gommers, R., Oliphant, T. E., et al. 2020, Nature Methods, 17, 261
Winter, A. J., Benisty, M., Izquierdo, A. F., et al. 2025, ApJL, 990, L10
Wölfer, L., Izquierdo, A. F., Booth, A., et al. 2026, A&A, 712, A177
Yen, H.-W., & Gu, P.-G.
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Improved structure function analysis for noisy data: The system can use a noiseaware pipeline to de-noise at the level of S2 and assess significance against matched null ensembles, recovering structural properties including correlation lengths, radial scaling, and anisotropy.
-
Enhanced model comparison capabilities: The system can distinguish between different (magneto-)hydrodynamical instabilities by comparing the recovered structural heuristics for each simulation in a vector space of shape parameters.
-
Robust parameter recovery under noise: The system can employ a parametric Gaussian description to fit the structure function, which is shown to be
decisively better
than empirical heuristics for recovering correlation lengths and amplitude, especially when compared against the scatter of the sampling ensemble.
Abstract
Protoplanetary disks, the birthplace of planetary systems, are highly dynamic environments. High angular and spectral resolution observations have revealed non-Keplerian motions in nearly every source studied, indicative of (magneto-)hydrodynamical instabilities or planet-disk interactions. However, these velocity perturbations rarely exhibit ordered structure, limiting our ability to attribute them to a specific mechanism. We propose the second-order structure function as a tool for recovering structural properties of velocity perturbations in the presence of significant noise. Using morphological models, we demonstrate how easy-to-calculate heuristics and parametric model fits together identify structure in noisy data and recover its properties, including correlation lengths, their radial scaling where genuinely present, and the anisotropy. We further outline a workflow to de-noise the data and to build a matched null model against which inferences are compared. Applying the framework to one numerical simulation each of the gravitational, magneto-rotational and vertical-shear instabilities, we recover their structure at high significance and the structural heuristics separate all three pairs, with the gravitational and vertical-shear instability simulations remaining the hardest to distinguish once observationally degraded. As a demonstration on real data, 12CO observations of TW Hya reveal azimuthally elongated velocity structure that persists once the large-scale m=1 component is removed. We also delineate where a Gaussian description breaks down for coherent features such as spiral arms, and how inclination affects the recovered heuristics. The methodology is implemented in the eddy package and provides a foundation for statistically characterizing turbulent velocity structure in high-resolution disk observations.
Sources
- Hydro-, Magnetohydro-, and Dust-Gas Dynamics of Protoplanetary Disks
- Kinematic Structures in Planet-Forming Disks
Related papers
- A Bright L4 Neptune Trojan Found by Stacking Archival ZTF Images
- PDS 70 c and SR 12 c: Observational Constraints on Giant-Planet and Satellite Formation
- JWST Rocky Worlds DDT Observations of LTT 1445 A b and c in Secondary Eclipse Using MIRI/F1500W
- A Joint Study of Epoch Astrometry and Radial Velocities of Gaia-4b
- An analytical framework for atmospheric tides on rocky planets. II. Semidiurnal thermotidal torque
- Signatures of CH 4 and NH 3 in the atmosphere of beta Pictoris b