Effect of dust in circumgalactic haloes on the cosmic shear power spectrum

Weak gravitational lensing is a powerful statistical tool for probing the growth of cosmic structure and measuring cosmological parameters. However, as shown by studies such as M\'enard et al. (2010), dust in the circumgalactic region of haloes dims and reddens background sources. In a weak lensing analysis, this selects against sources behind overdense regions; since there is more structure in overdense regions, we will underestimate the amplitude of density perturbations $\sigma_8$ if we do not correct for the effects of circumgalactic dust. To model the dust distribution we employ the halo model. Assuming a fiducial dust mass profile based on measurements from M\'enard et al. (2010), we compute the ratio $Z$ of the systematic error to the statistical error for a survey similar to the Nancy Grace Roman Space Telescope reference survey (2000 deg$^2$ area, single-filter effective source density 30 galaxies arcmin$^{-2}$). For a waveband centered at $1580$ nm ($H$-band), we find that $Z_{H} = 0.37$. For a similar survey with waveband centered at $620$ nm ($r$-band), we also computed $Z_{r} = 2.8$. Within our fiducial dust model, since $Z_{r} > 1$, the systematic effect of dust will be significant on weak lensing image surveys. We also computed the dust bias on the amplitude of the power spectrum, $\sigma_{8}$, and found it to be for each waveband $\Delta \sigma_8/\sigma_8 = -3.1\times 10^{-4}$ ($H$ band) or $-2.2\times 10^{-3}$ ($r$ band) if all other parameters are held fixed (the forecast Roman statistical-only error $\sigma(\sigma_8)/\sigma_8$ is $9\times 10^{-4}$).