In order to validate the accuracy of estimated Green’s functions (EGFs), which are widely used in ambient seismic noise tomography, a broadband seismograph was installed in the epicentral area of an M 5 earthquake, with ground‐truth location from Interferometric Synthetic Aperture Radar observations. EGFs between this station and permanent stations are compared with surface waves excited by the earthquake. The group velocity dispersion measured from EGFs at large interstation distances (∼1000 km or longer) are consistent with measurements from the earthquake, and the EGFs from symmetric noise correlation functions (NCFs) are usually more accurate. At shorter interstation distances (300–1000 km), the match between noise and earthquake dispersion is good for most stations, although we observed mismatch for a few stations. The mismatch is probably caused by low signal‐to‐noise ratio of NCFs or nondiffusive noise wavefield at short distances.
Online Material: Figure showing mismatch of group velocity dispersion between earthquake and noise data, and figure of phase velocity dispersion between earthquake data and noise correlation functions at station ENH.