The use of seismic noise interferometry to retrieve Green's functions and the analysis of volcanic tremor are both useful in studying volcano dynamics. Whereas seismic noise interferometry allows long-range extraction of interpretable signals from a relatively weak noise wavefield, the characterization of volcanic tremor often requires a dense seismic array close to the source. We here show that standard processing of seismic noise interferometry yields volcanic tremor signals observable over large distances exceeding 50 km. Our study comprises 2.5 yr of data from the U.S. Geological Survey Hawaiian Volcano Observatory short period seismic network. Examining more than 700 station pairs, we find anomalous and temporally coherent signals that obscure the Green's functions. The time windows and frequency bands of these anomalous signals correspond well with the characteristics of previously studied volcanic tremor sources at Pu'u 'Ō'ō and Halema'uma'u craters. We use the derived noise cross-correlation functions to perform a grid-search for source location, confirming that these signals are surface waves originating from the known tremor sources. A grid-search with only distant stations verifies that useful tremor signals can indeed be recovered far from the source. Our results suggest that the specific data processing in seismic noise interferometry—typically used for Green's function retrieval—can aid in the study of both the wavefield and source location of volcanic tremor over large distances. In view of using the derived Green's functions to image heterogeneity and study temporal velocity changes at volcanic regions, however, our results illustrate how care should be taken when contamination by tremor may be present.