Abstract

In 2023, several international collaborations comprising the International Pulsar Timing Array achieved a major scientific milestone with the first detection of a signal in pulsar timing observations consistent with the signature expected from a stochastic gravitational wave background, created by an ensemble of unresolved supermassive black hole binaries in the early Universe. The next breakthrough in the field is expected to be the first detection of a continuous wave from a single such binary, which would allow us to better understand their evolution and that of our Universe. However, this feat will require unprecedented precision in our timing measurements. To that end, in this thesis we conduct a detailed exploration of different noise budgeting techniques to improve the timing precision achieved by the North American Nanohertz Observatory for Gravitational Waves (NANOGrav). We present three distinct approaches: 1) Quantifying previously unaccounted-for error sources in pulsar observations. In particular, we study the biases introduced in the modeling of frequency-dependent effects due to incomplete frequency sampling. 2) Developing novel machine-learning-based algorithms to characterize single-pulse jitter and mitigate the timing uncertainty introduced by single-pulse variability, which causes deviations of the integrated pulse profile from the long-term average. 3) Incorporating high-precision pulsar astrometric estimates obtained using Very Long Baseline Interferometry (VLBI) measurements into NANOGrav’s timing models to eliminate the need to fit for those parameters in the timing solution, thereby mitigating potential power absorption from gravitational wave signals. We find that, for selected sources and after appropriate data processing, these techniques can enhance the quality of NANOGrav’s current datasets. In particular, we obtain consistent improvements as large as approximately 0.3 microseconds in timing precision when clustering single pulses of PSR J2145−0750 by fluence. We quantify that the error in time-of-arrival measurements from narrowband observations of PSR J1643−1224 may be underestimated by as much as approximately 22 microseconds. We also find a VLBI-consistent astrometric solution for PSR J0030+0451 that could prevent red-noise processes with residual amplitudes as large as approximately 0.8 microseconds from being absorbed into the pulsar astrometric fit. Applications of these techniques include refining current data processing pipelines and providing more robust error estimates for legacy observations, thereby increasing the effective time baseline using already available data.

Publication Date

1-2026

Document Type

Dissertation

Student Type

Graduate

Degree Name

Astrophysical Sciences and Technology (Ph.D.)

Department, Program, or Center

Physics and Astronomy, School of

College

College of Science

Advisor

Michael T. Lam

Advisor/Committee Member

Andrew Robinson

Advisor/Committee Member

Seth Hubbard

Campus

RIT – Main Campus

Download

Share

COinS