Thermal Moonquake Characterization and Cataloging Using Frequency-Based Algorithms and Stochastic Gradient Descent

Abstract

Thermal moonquakes are recurring seismic signals detected on the lunar surface that are temporally correlated with the lunar day-night cycle, although their precise mechanism is not known. The Lunar Seismic Profiling Experiment (LSPE) was a deployment of four geophones during the Apollo 17 mission to characterize the near-surface structure of the landing site using explosive shots. The array was reactivated in passive recording mode several times after the mission was completed and recorded thousands of thermal moonquakes. In this study, we expand on an initial detection catalog and determine waveform parameters and source information for the seismic signals. We used a spectrogram-based approach to obtain fine-tuned arrival times, quantify the envelope length using emergence, and compute peak-ground-velocity. We obtain the incident azimuth direction by applying stochastic gradient descent to a travel-time misfit equation. We found that thermal moonquakes are split into two main classes: (a) impulsive, high-amplitude events that are produced by the lunar module descent vehicle located east of the LSPE array in response to rapid temperature transitions during sunrise and sunset and (b) emergent events, that are natural responses to incident sunlight whose duration is directly linked to the temperature of the regolith. We hypothesize that the correlation between temperature and emergence could be due to changes in regolith scattering or higher energy daytime events occurring further away.