Moon crust and upper mantle revealed by seismic interferometry methods applied to Apollo seismic data

The Moon has recently been put back into the spotlight for space exploration. As our closest astral neighbor, the Moon is an accessible differentiated planetary body which provides insight into the formation and differentiation of inner solar system bodies. Back in the seventies, NASA’s Apollo program conducted several exploration and sample return missions to better understand our natural satellite. Five of these missions included the deployment of seismometers for passive and active seismology. We process the passive seismology records by using seismic interferometry methods. Cross-correlations and auto-correlations between the different seismometers deployed by the Apollo missions are computed and analyzed to retrieve Moon’s internal structure.

The seismic interferometry method is applied to the coda of seismic events detected by the Apollo seismic network. The events are selected based on their quality and the presence or lack of significant gaps in the data. Auto-correlations are computed on time series whitened using 1-bit normalization. Each event is processed independently. The Signal-to-Noise Ratio (SNR) of correlations is calculated by comparing the envelope of the stacked correlations to the amplitude of the residual fluctuations between events and between different time windows in the coda. An additional data selection step is added in order to exclude data that do not show an enhancement of the auto-correlation spectrum at frequencies for which the seismometer is most sensitive.

The evolution of the auto-correlation SNR as a function of time in the coda is consistent with the SNR of the seismic data. Auto-correlations of vertical components show seismic phase arrivals in the 13-90s time range which are consistent between two Apollo stations.

These arrivals are interpreted in terms of seismic waves reflected on crust and mantle interfaces. A preliminary crust and mantle model of these interfaces is deduced and validated by comparing vertical and horizontal auto-correlation results. The analysis of the cross-correlations between Apollo stations is also presented. Additional validation tests are implemented to ensure that the detected arrivals are not due to data artefacts at a given periodicity.