首页

Galaxy-galaxy strong lensing is a powerful tool for studying galaxy structure and dark matter substructure. Detecting these subhaloes constrains the halo mass function and, through it, the nature of dark matter. Euclid, Roman and CSST will soon deliver tens of thousands of lenses, and JWST is already providing high-resolution imaging for hundreds of new lenses. Together these surveys will let us study the full population of strong lenses and search for substructure at lower masses than ever before. In this talk I will explain how we search for dark matter substructure in galaxy-scale lenses with extended sources. I will focus on the PyAutoLabs modelling tools, including how we add angular complexity to mass profiles through “multipole” perturbations, and discuss the application of these methods to the lens SPT2147. In this system, adding an NFW subhalo to an elliptical power-law plus external shear model gives a ln(Bayes factor) of up to 50, indicating a strong subhalo detection. However, when the model also includes angular complexity, the ln(Bayes factor) drops to around 10. This illustrates a central problem in lens modelling: is the angular complexity from the multipole model absorbing a real subhalo signal, is the subhalo model absorbing real angular complexity, or is it some of each? If it is both, how do we separate them?  This is not an isolated degeneracy – many quantities in strong lensing suffer similar problems. I will close by addressing how we are building better population priors from both simulations and observations. Such priors are needed to break the degeneracy between multipoles and subhaloes, and to give a solid foundation for modelling the thousands of lenses still to come.

Host: Dandan Xu