Title: Improving flux ratio anomaly precision by measuring gravitational lens multipole moments with extended arcs

Abstract: In a strong gravitational lens, perturbations by low-mass dark matter halos can be detected by differences between the measured image fluxes relative to the expectation from a smooth model for the mass distribution which contains only the gravitational effects of the main deflector. The abundance of these low-mass structures can be used to constrain the properties of dark matter. Traditionally only the lensed quasar positions have been to predict the smooth-model flux ratios. We demonstrate that significant additional information can be gained by using the lensed quasar host galaxy which appears as an extended arc and constrains the smooth-model over a much larger angular area. We simulate Hubble Space Telescope-quality mock observations based on the lensing system WGD2038-4008 and we compare the model-predicted flux ratio precision and accuracy for two cases; one of which the inference is based only on the lensed quasar image positions, and the other based on the extended arcs as well as lensed quasar image positions. For our mock lens systems we include both elliptical, and higher order m=3 and m=4 multipole terms in the smooth-mass distributions with amplitudes based on the optically measured shapes of massive elliptical galaxies. We find that the extended arcs improve the precision of the model-predicted flux ratios by a factor of 6-8, depending on the strength of the multipole terms. Furthermore, with the extended arcs, we are also able to accurately recover the m=3, 4 mass multipole strengths and angles $a_3/a$, $a_4/a$, $\phi_3-\phi_0$, and $\phi_4-\phi_0$ to a precision of 0.002, 0.002, $3^\circ$ and $3^\circ$, respectively. This work implies that lensed arcs can constrain deviations from ellipticity in strong lens systems, and potentially lead to more robust constraints on substructure properties from flux ratios.