Parameter extraction of the stochastic gravitational wave background with peak-like templates in millihertz

Abstract: We investigate a framework for extracting parameters of stochastic gravitational wave background (SGWB) with peak-like templates in the millihertz frequency band, and analyzing transient contamination effects on parameter reconstruction. We present the spectrum and spectrogram under different conditions and provide the results of parameter reconstruction. Using templates from the early universe, we demonstrate that the peak-like templates outperform the broken power law (BPL) templates in power-law exponents recovery and peak frequency localization. The reconstruction results obtained using data from Fast Fourier Transform (FFT) are better than those obtained using data from Short-Time Fourier Transform (STFT) which is based on the spectrogram. For the single-peak template, the estimation accuracy of the exponent and peak frequency surpasses that of the BPL template by an order of magnitude, but demonstrates less precision in amplitude estimation compared to BPL. Regarding the double-peak template, parameter estimation results derived from the FFT methodology consistently outperform those obtained using STFT. Nevertheless, transient signals exhibit a detrimental impact on parameter estimation precision, causing errors to increase by an order of magnitude, particularly in multi-peak scenarios. This framework provides an example for using templates to analyze data from space-based gravitational wave detectors.