Pulsar Timing Residuals Exhibit Multiple Dynamical Regimes: An Empirical Analysis of NANOGrav Data Barry Kieran Radcliffe and ChatGPT March 2, 2026

This paper presents an empirical analysis of pulsar timing residuals from the NANOGrav 12.5-year dataset, with the objective of determining whether these residuals are consistent with uncorrelated noise or exhibit structured temporal behavior. Rather than introducing new theoretical assumptions, the study analyzes residual time series directly across multiple pulsars, focusing on variability, statistical distribution, and autocorrelation structure. The results show that pulsar timing residuals are not uniformly stochastic. Instead, they exhibit a range of behaviors, including noise-like behavior, moderate temporal correlation, and strong long-term coherence. Autocorrelation analysis reveals that some pulsars display near-zero correlation across time, while others exhibit persistent structure extending across many temporal lags. These findings indicate that pulsar timing residuals cannot be described by a single statistical model or universal organizing variable. Instead, they fall into distinct dynamical regimes, suggesting system-dependent behavior. This result provides an empirical constraint on interpretations of pulsar timing data and highlights the importance of distinguishing between shared qualitative features and heterogeneous system-level dynamics. This work builds on prior analysis (Paper 1), which identified similar statistical signatures across multiple astrophysical systems, by examining whether such structure is uniformly expressed within a single domain. The results show that while similar types of structure may be present across systems, their manifestation within pulsar timing data is not uniform.