[Paper Review] Timing Noise Analysis of NANOGrav Pulsars

D. Perrodin, Fredrick Jenet+8 more|arXiv (Cornell University)|Nov 14, 2013

Pulsars and Gravitational Waves Research3 citations

TL;DR

This study analyzes five years of timing data from 17 NANOGrav millisecond pulsars to assess the presence of red timing noise, which could limit gravitational wave detection sensitivity. Using autocovariance and nearest-neighbor correlation analyses, the authors find that only two pulsars exhibit weak red noise, while the rest are well described as white noise, implying that increasing observation cadence remains an effective strategy for improving timing precision and enhancing gravitational wave detection sensitivity.

ABSTRACT

We analyze timing noise from five years of Arecibo and Green Bank observations of the seventeen millisecond pulsars of the North-American Nanohertz Observatory for Gravitational Waves (NANOGrav) pulsar timing array. The weighted autocovariance of the timing residuals was computed for each pulsar and compared against two possible models for the underlying noise process. The first model includes red noise and predicts the autocovariance to be a decaying exponential as a function of time lag. The second model is Gaussian white noise whose autocovariance would be a delta function. We also perform a ``nearest-neighbor" correlation analysis. We find that the exponential process does not accurately describe the data. Two pulsars, J1643-1224 and J1910+1256, exhibit weak red noise, but the rest are well described as white noise. The overall lack of evidence for red noise implies that sensitivity to a (red) gravitational wave background signal is limited by statistical rather than systematic uncertainty. In all pulsars, the ratio of non-white noise to white noise is low, so that we can increase the cadence or integration times of our observations and still expect the root-mean-square of timing residual averages to decrease by the square-root of observation time, which is key to improving the sensitivity of the pulsar timing array.

Motivation & Objective

• To determine whether red timing noise limits the sensitivity of pulsar timing arrays (PTAs) to nanohertz gravitational wave backgrounds.
• To assess whether the timing residuals of NANOGrav pulsars are consistent with white noise or exhibit correlated, red-noise-like behavior.
• To evaluate whether increasing observation cadence or integration time would still reduce timing residual RMS as √N, a key requirement for improving PTA sensitivity.
• To compare multiple noise characterization methods—autocovariance, nearest-neighbor correlation, and exponential fitting—on real PTA data.
• To inform future PTA observing strategies by quantifying the dominance of white versus non-white noise components in individual pulsars.

Proposed method

• Computed the weighted autocovariance of timing residuals for each of the 17 NANOGrav pulsars over a five-year baseline using Arecibo and Green Bank Telescope data.
• Compared the autocovariance to two models: a decaying exponential (characteristic of red noise) and a delta function (characteristic of white noise).
• Performed a nearest-neighbor correlation analysis to test the statistical consistency of residuals with white noise, reporting the probability (p-value) of white noise compatibility.
• Fitted a decaying exponential function to the autocovariance to quantify red noise amplitude and assess model fit quality.
• Calculated a 'time factor' to estimate how many additional observations could be made before red noise dominates, preserving the √N scaling of RMS reduction.
• Used χ² statistics to evaluate the goodness of fit of the autocovariance to a delta-function model, indicating whiteness of residuals.

Experimental results

Research questions

• RQ1 Do the timing residuals of NANOGrav pulsars exhibit significant red noise that would limit the sensitivity of pulsar timing arrays to gravitational wave backgrounds?
• RQ2 Which pulsars show evidence of non-white noise, and how strong is this effect compared to white noise?
• RQ3 Can the standard assumption of white noise hold for most pulsars, allowing RMS to decrease as 1/√N with increased observation cadence?
• RQ4 Which noise characterization method—nearest-neighbor correlation, autocovariance fitting, or delta-function χ² test—provides the most reliable assessment of residual whiteness?
• RQ5 What is the practical impact of non-white noise on future PTA observing strategies, particularly in terms of cadence and integration time optimization?

Key findings

• Only two pulsars, J1643-1224 and J1910+1256, show weak evidence of red noise, with red noise amplitudes of 2.2 and 5.3 μs respectively, indicating minimal deviation from white noise.
• The remaining 15 pulsars are well described as white noise, with the nearest-neighbor correlation test showing p-values consistent with white noise (e.g., p > 0.05 for most), and χ² tests rejecting the exponential model with high confidence.
• The autocovariance of residuals does not follow a decaying exponential, indicating that the red noise model is a poor fit for the observed data across the NANOGrav pulsar sample.
• The 'time factor'—indicating how many more observations can be made before red noise dominates—was high for all pulsars (e.g., 14.9 for J1909-3744), confirming that RMS reduction as 1/√N remains valid.
• The overall lack of strong red noise implies that the sensitivity of the NANOGrav PTA to a stochastic gravitational wave background is limited by statistical uncertainty rather than systematic noise.
• The results support increasing observation cadence across all NANOGrav pulsars, as white noise remains dominant and RMS is expected to scale as 1/√N, enhancing GW detection prospects.