Title: Statistical Microlensing Toward Magnified High-Redshift Star Clusters

Abstract: We study light variability of gravitationally magnified high-redshift star clusters induced by a foreground population of microlenses. This arises as the incoherent superposition of light variations from a large number of source stars traversing the random magnification pattern on the source plane. The light curve resembles a scale-invariant, Gaussian process on timescales of years to decades, while exhibits rapid and frequent micro-caustic crossing flares of larger amplitudes on timescales of days to months. For a concrete example, we study a young Lyman-continuum-leaking star cluster recently discovered in the lensed Sunburst Arc at $z=2.37$. We show that one magnified image happens to be intervened by a faint foreground galaxy, and hence should exhibit a variable flux at the $1$--$2\%$ level, which is measurable in space with $\sim 1$--$3\,$ks exposures on the Hubble Space Telescope and more easily with the James Webb Space Telescope, or from the ground using a $\sim$4-meter telescope without adaptive optics. Detailed measurement of this variability will enable us to determine the absolute macro magnification and hence the intrinsic mass and length scales of the star cluster, test synthetic models of stellar population, and probe multiplicity of massive stars. We furthermore suggest that monitoring the other lensed images of the star cluster, which are free from significant intervention by foreground stellar microlenses, will allow us to probe planetary to stellar mass compact objects constituting as little as just a few percent of the dark matter. Given the typical surface density of intracluster stars, we expect this phenomenon to be relevant for many other gravitationally magnified star clusters at Cosmic Noon behind galaxy cluster lenses.