Speaker
Description
Operating in the millihertz band, LISA will probe a rich superposition of GW signals. Astrophysical population models predict a sufficient number of signals in the LISA band to blend together and form an irresolvable Galactic foreground noise.
In addition, a stochastic gravitational wave background (SGWB) of cosmological origin could add an unknown component to the measured noise.
A central challenge is then to separate these unknown contributions from the Galactic foreground. Current analyses build parametric templates for each population independently. I will show that, in the presence of multiple unresolved components, these templates are not independent, leading to inconsistencies when interpreting the combined signal.
Using a representative SGWB model, I quantify the impact of this effect on parameter estimation.
Neglecting inter-template correlations can lead to significan biases, and, for several benchmark cases, those biases exceed the statistical reconstruction uncertainties. These findings expose a key limitation of existing template-based strategies and indicate that unbiased component separation will likely require additional information, such as constraints from resolved sources obtained through global-fit analyses.