Insights into the Pulsar Timing Array hypothesis space

Abstract

We present novel insights into the pulsar-noise model space in pulsar timing array (PTA) experiments. Through a comparative analysis of the same Parkes PTA second data release observations processed with the 2020 and 2023 pipelines, we show that data processing materially changes the distribution of posterior support across competing noise hypotheses and increases sensitivity to weak contributions such as the solar wind. A solar-wind component appears in the highest-posterior hypothesis for ten pulsars under the 2023 pipeline, against four under the 2020 pipeline, while the corresponding mean electron density estimates are consistent with PPTA DR3 results. Furthermore, hypothesis-space analysis exposes model competition and degeneracies that are hidden by single-model summaries. Approximately 75% of the pulsars retain substantial posterior support for alternative noise descriptions. Therefore, understanding the nature of the pulsar noise hypothesis space is crucial for robust inference and nanohertz gravitational-wave background searches.