Bjorn Larsen, J. G. Baier, Daniel J Oliver, K. E. Wayt, Yu-Ting Chang, Jeffrey S. Hazboun, Chiara M. F. Mingarelli, Joseph Simon, Matthew T. Miles, Gabriella Agazie, Akash Anumarlapudi, Anne M. Archibald, Zaven Arzoumanian, P. T. Baker, Paul R. Brook, H. Thankful Cromartie, Kathryn Crowter, Megan E. DeCesar, Paul B. Demorest, Timothy Dolch, E. C. Ferrara, William Fiore, Emmanuel Fonseca, Gabriel E. Freedman, N. Garver-Daniels, Peter A. Gentile, Joseph Glaser, Deborah C. Good, Ross J. Jennings, Megan L. Jones, D. L. Kaplan, M. Kerr, Michael T. Lam, D. R. Lorimer, Jing Luo, Ryan S. Lynch, Alexander McEwen, M. A. McLaughlin, Natasha McMann, Bradley W. Meyers, Cherry Ng, David J. Nice, Timothy T. Pennucci, Benetge B. P. Perera, Nihan S. Pol, H. A. Radovan, S. M. Ransom, Paul S. Ray, Ann Schmiedekamp, Carl Schmiedekamp, Brent J. Shapiro-Albert, I. H. Stairs, Kevin Stovall, Abhimanyu Susobhanan, Joseph K. Swiggum, Haley M. Wahl
Abstract Pulsar timing arrays (PTAs) conduct low-frequency gravitational-wave (GW) searches, which require comprehensive accounting of various noise sources to achieve robust results. Interstellar propagation effects (e.g., dispersion and scattering) are especially complex noise sources, introducing chromatic delays that can reduce sensitivity to GWs and bias their inference if left unmodeled. These delays also strongly depend on the line-of-sight properties to each individual pulsar. To address this, we present customized chromatic noise models for 67 pulsars in the NANOGrav 15 yr dataset. These models are selected from an expanded suite of Gaussian processes to simultaneously characterize multiple types of chromatic delays and are tailored to each pulsar’s dataset. Alongside probing the interstellar medium, we use these models to infer the solar wind electron density over the course of ∼1.5 solar cycles. We also find evidence for nondispersive chromatic delays in 21 out of 67 NANOGrav pulsars. After applying our chromatic models, we observe significant impacts on the inference of achromatic noise in 19 out of 67 pulsars, finding in several cases that a previously significant achromatic noise process can be partially or entirely described as chromatic. These results demonstrate that refined noise modeling is essential to enhance the sensitivity and accuracy of low-frequency GW searches with PTAs.