The NANOGrav 15-Year Data Set: A Case Study for Simplified Dispersion Measure Modeling for PSR J1455-3330 and the Impact on Gravitational Wave Sensitivity

Abhimanyu Susobhanan; Akash Anumarlapudi; Alexander McEwen; Anne M. Archibald; Ann Schmiedekamp; Benetge B. P. Perera; Bradley W. Meyers; Brent J. Shapiro-Albert; Carl Schmiedekamp; Cherry Ng

arxiv: 2506.03597 · v2 · submitted 2025-06-04 · 🌌 astro-ph.HE

The NANOGrav 15-Year Data Set: A Case Study for Simplified Dispersion Measure Modeling for PSR J1455-3330 and the Impact on Gravitational Wave Sensitivity

Michael T. Lam , David L. Kaplan , Gabriella Agazie , Akash Anumarlapudi , Anne M. Archibald , Zaven Arzoumanian , Paul T. Baker , Paul R. Brook

show 41 more authors

H. Thankful Cromartie Kathryn Crowter Megan E. DeCesar Paul B. Demorest Timothy Dolch Elizabeth C. Ferrara William Fiore Emmanuel Fonseca Gabriel E. Freedman Nate Garver-Daniels Peter A. Gentile Joseph Glaser Deborah C. Good Jeffrey S. Hazboun Ross J. Jennings Megan L. Jones Matthew Kerr Duncan R. Lorimer Jing Luo Ryan S. Lynch Alexander McEwen Maura A. McLaughlin Natasha McMann Bradley W. Meyers Cherry Ng David J. Nice Timothy T. Pennucci Benetge B. P. Perera Nihan S. Pol Henri A. Radovan Scott M. Ransom Paul S. Ray Ann Schmiedekamp Carl Schmiedekamp Brent J. Shapiro-Albert Joseph Simon Ingrid H. Stairs Kevin Stovall Abhimanyu Susobhanan Joseph K. Swiggum Haley M. Wahl

This is my paper

Pith reviewed 2026-05-19 11:43 UTC · model grok-4.3

classification 🌌 astro-ph.HE

keywords pulsar timing arraysdispersion measure modelinggravitational wave backgroundNANOGravPSR J1455-3330red noisesolar windtiming parallax

0 comments

The pith

Replacing the piecewise dispersion measure model for PSR J1455-3330 with a fixed solar wind density plus linear trend turns its parallax into a detection and makes its red noise match the common array signal.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper investigates how dispersion measure modeling choices affect the timing solution for one long-observed pulsar in the NANOGrav 15-year data set. It finds that a simpler physical model consisting of constant solar wind density and a linear dispersion measure trend produces different results than the standard piecewise-constant function. Under the new model the pulsar's timing parallax becomes a significant detection rather than an upper limit, red noise appears at significant levels, and that red noise matches the common signal seen across the rest of the array. These changes matter for gravitational wave searches because the parameters extracted from pulsar timing arrays depend on how individual pulsar noise is modeled even when overall detection significance stays similar. The authors therefore advocate for pulsar-by-pulsar physical modeling in precision timing experiments.

Core claim

When the original piecewise-constant dispersion measure model is replaced with a fixed solar wind density plus linear trend for PSR J1455-3330, the pulsar's timing parallax signal changes from an upper limit to a significant detection, red noise becomes significant, and the red noise is consistent with the common signal found for the other pulsars. Neither of these signals is radio-frequency dependent.

What carries the argument

The simplified dispersion measure model of fixed solar wind density plus linear trend in dispersion measure, used in place of the piecewise-constant function to capture the physical behavior specific to this pulsar.

If this is right

The pulsar's timing parallax changes from an upper limit to a significant detection.
Red noise becomes significant in the timing residuals.
The red noise is consistent with the common signal found for the other pulsars.
Neither the parallax nor the red noise signal is radio-frequency dependent.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

Applying similar physically motivated simplifications to other pulsars without detected intrinsic red noise could refine the overall gravitational wave sensitivity of the array.
Better-constrained parallax values would improve distance estimates used in interpreting the gravitational wave background.
The result highlights that model selection for individual pulsars can alter the inferred spectrum of the common red process even if the array-wide detection significance stays unchanged.

Load-bearing premise

The simpler physical model of fixed solar wind density plus linear dispersion measure trend is preferred for PSR J1455-3330 on the basis of physical motivation and the absence of long-timescale red noise in the original modeling, rather than a quantitative model comparison that accounts for the change in degrees of freedom.

What would settle it

A new data set or longer baseline that either fails to recover a significant parallax or shows that the red noise under the simpler model does not remain consistent with the common array signal.

read the original abstract

Evidence for a low-frequency gravitational-wave background using pulsar timing arrays has generated recent interest into its underlying contributing sources. However, multiple investigations have seen that the significance of the evidence does not change with choice of pulsar modeling techniques but the resulting parameters from the gravitational wave searches do. PSR J1455-3330 is one of the longest-observed pulsars in the array monitored by the North American Nanohertz Observatory for Gravitational Waves (NANOGrav) but showed no evidence for long-timescale red noise, either intrinsic or the common signal found among many pulsars in the array. In this work, we argue that NANOGrav's piecewise-constant function used to model variations in radio-frequency-dependent dispersive delay should not be used for this pulsar, and a much simpler physical model of a fixed solar wind density plus a linear trend in dispersion measure is preferred. When the original model is replaced, (i) the pulsar's timing parallax signal changes from an upper limit to a significant detection, (ii) red noise becomes significant, and (iii) the red noise is consistent with the common signal found for the other pulsars. Neither of these signals are radio-frequency dependent. While the same physical motivation will not apply to many of the pulsars currently used in pulsar timing arrays, we argue for careful physically-motivated timing and noise modeling of pulsars used in precision timing experiments.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

For PSR J1455-3330, replacing the piecewise DM model with fixed solar wind plus linear trend turns the parallax into a detection and makes the red noise match the common signal.

read the letter

The key takeaway is that for this one long-observed pulsar the choice of dispersion measure model changes two important timing results: parallax goes from an upper limit to a significant value, and red noise appears at a level consistent with the array-wide common process. The paper takes the existing NANOGrav 15-year residuals, swaps in a simpler physical model (fixed solar wind density plus one linear DM trend), and shows the before-and-after fits. That is the concrete new information they add; it is not a new technique but a targeted demonstration that the modeling decision matters for this pulsar’s contribution to the gravitational-wave search. They do a clear job laying out the physical motivation and reporting the shifts in parallax significance and red-noise amplitude and spectrum. The work is straightforward and the numbers are easy to follow. The main soft spot is the lack of a formal model-comparison number. The piecewise-constant model has many more free parameters than the two-parameter alternative, yet the paper relies on physical argument and the prior absence of red noise rather than a Bayes factor or information criterion that accounts for the difference in degrees of freedom. Without that check it is still possible that some unmodeled DM structure is being pushed into the parallax or red-noise terms. The rest of the analysis looks solid and the circularity burden is low because the consistency check is against the rest of the array. This paper is for people who actually run or review PTA noise models and timing solutions. Anyone thinking about how individual pulsar modeling choices feed into the common-signal parameters will find it useful. It is narrow but the point is practical and worth checking. I would send it to peer review; the data exist to test the model preference more rigorously, and the issue is real enough that referees should see it.

Referee Report

1 major / 1 minor

Summary. The paper argues that for PSR J1455-3330 in the NANOGrav 15-year data set, the standard piecewise-constant dispersion measure (DM) model should be replaced by a simpler physical model consisting of a fixed solar wind density plus a linear DM trend. This replacement changes the pulsar's timing parallax from an upper limit to a significant detection, makes red noise significant, and renders that red noise consistent with the common signal seen across the rest of the array. The work uses this case study to advocate for physically motivated timing and noise modeling in pulsar timing arrays.

Significance. If the central claim holds, the work provides a concrete demonstration that DM modeling choices for individual pulsars can alter the detection of parallax and red-noise signals and their consistency with a common process. The before-and-after comparisons for this long-observed pulsar illustrate how a physically motivated simplification can affect PTA gravitational-wave sensitivity, offering a useful example for the community even if the same modeling choice does not apply to most other pulsars.

major comments (1)

[Section on DM modeling and results for PSR J1455-3330] The manuscript motivates the switch from the piecewise-constant DM model to a fixed solar wind density plus linear trend by noting the absence of long-timescale red noise under the original model and by physical considerations, but does not supply a quantitative model-comparison statistic (e.g., Bayes factor, marginal likelihood ratio, or penalized information criterion) that accounts for the large difference in degrees of freedom. The piecewise-constant model has one free parameter per epoch or segment while the alternative has only two; without such a comparison the reported changes in parallax significance, red-noise detection, and consistency with the common signal remain vulnerable to the possibility that unmodeled DM structure is being absorbed elsewhere in the timing solution.

minor comments (1)

[Abstract] The abstract and introduction clearly state the before-and-after results; a brief parenthetical note on the number of free parameters in each DM model would help readers immediately appreciate the degrees-of-freedom contrast.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their constructive review and recommendation for major revision. The central concern regarding the lack of a quantitative model-comparison statistic for the DM modeling of PSR J1455-3330 is valid, and we will incorporate a Bayes factor analysis in the revised manuscript to address it directly while preserving the physical motivation for the simpler model.

read point-by-point responses

Referee: [Section on DM modeling and results for PSR J1455-3330] The manuscript motivates the switch from the piecewise-constant DM model to a fixed solar wind density plus linear trend by noting the absence of long-timescale red noise under the original model and by physical considerations, but does not supply a quantitative model-comparison statistic (e.g., Bayes factor, marginal likelihood ratio, or penalized information criterion) that accounts for the large difference in degrees of freedom. The piecewise-constant model has one free parameter per epoch or segment while the alternative has only two; without such a comparison the reported changes in parallax significance, red-noise detection, and consistency with the common signal remain vulnerable to the possibility that unmodeled DM structure is being absorbed elsewhere in the timing solution.

Authors: We agree that a formal quantitative comparison is needed to strengthen the argument and mitigate concerns about parameter absorption. Our preference for the fixed solar wind plus linear trend model is driven by the specific geometry and long observational baseline of PSR J1455-3330, where a linear DM trend is physically expected rather than epoch-by-epoch variations. The piecewise-constant approach, while flexible, risks absorbing other signals into the many DM parameters, which is consistent with the observed shifts in parallax significance and red-noise properties. In the revised manuscript we will add a Bayes factor (or equivalent marginal likelihood ratio) computed within the same PTA analysis framework to explicitly account for the differing degrees of freedom. This will provide a direct statistical measure supporting the simpler model and address the vulnerability noted by the referee. revision: yes

Circularity Check

0 steps flagged

No significant circularity; model swap yields independent consistency check

full rationale

The paper's central chain re-models existing timing residuals for PSR J1455-3330 with a fixed solar-wind plus linear DM trend instead of the piecewise-constant DMX model. The reported parallax detection, red-noise significance, and consistency with the array-wide common signal are direct numerical outcomes of that re-fit and a subsequent comparison to the independent common-process parameters derived from the rest of the NANOGrav array; none of these quantities are defined by construction from the input residuals or from a self-referential fit. Model preference rests on physical motivation and the observed lack of long-timescale red noise under the original model rather than on a self-citation chain or an ansatz smuggled from prior work by the same authors. The analysis therefore remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central claim rests on the physical preference for a fixed solar-wind plus linear DM model over the piecewise function; no new free parameters are introduced beyond the linear trend coefficient, which is fitted to the data.

free parameters (1)

linear DM trend coefficient
Slope of the dispersion measure change over time, fitted to the timing residuals under the simpler model.

axioms (1)

domain assumption The piecewise-constant DM model is not physically appropriate for PSR J1455-3330 because the pulsar lacks long-timescale red noise under the original analysis.
Invoked in the abstract to justify replacing the standard NANOGrav DM modeling choice.

pith-pipeline@v0.9.0 · 6054 in / 1478 out tokens · 27367 ms · 2026-05-19T11:43:16.749493+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

We can estimate the scintillation timescale... σDM(15.7 yr) = 5.9×10^{-4} pc cm^{-3}... using Eq. 3... DDM(τ) = 1.47×10^{-15} (pc cm^{-3})^2 (ν/GHz)^2 (τ/Δtd)^{5/3}
IndisputableMonolith/Foundation/BranchSelection.lean branch_selection unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

ΔAIC = 6.4... Model B is 24 times more preferred than Model C... BIC values

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.