2026ApJ...999..124K

The “near-far” approach to studying reionization leverages the star formation histories of the Milky Way (MW) or Local Group (LG) galaxies, derived from resolved photometry, to infer the low-mass/faint end of the stellar mass functions (SMFs) or the ultraviolet luminosity functions (UVLFs) of high-redshift galaxies (z ≳ 6), beyond the current James Webb Space Telescope detection limits (M_UV ≳ -15). Previous works considered only intact low-mass galaxies in the MW and LG, neglecting disrupted galaxies such as stellar streams and phase-mixed objects. Using the FIRE-2 simulations, we show that these disrupted galaxies contribute up to ∼50% of the total stellar-mass budget of the proto-MW/LG at z = 6-9. Including all the progenitors of these disrupted galaxies improves the normalization of the recovered SMFs/UVLFs by factors of ∼2-3 and reduces the halo-to-halo variation in the slope by ∼20%-40%. This enables robust constraints down to at least the resolution limit of the simulations, near M_⋆ ∼ 10⁵M_⊙or M_UV ∼ -10 at z ≳ 6. We also show that “fossil-record” reconstructions-which assume each present-day system descends from a single reionization-era progenitor-are sensitive to the stellar-mass/UV-magnitude thresholds, which introduces bias in the inferred slopes at the low-mass/faint end. Additionally, we demonstrate that neglecting disrupted systems underestimates the contribution of galaxies with M_UV ≲ -15 to the reionization-era UV luminosity density. Finally, we estimate that a significant fraction (∼50%) of streams with M_⋆ ≳ 10⁶M_⊙ at z = 0 should be detectable from upcoming Rubin Observatory and Roman Space Telescope observations.