Identifying deep moonquake nests using machine learning model on single
lunar station on the far side of the Moon
Abstract
One of the future NASA space program includes the Farside Seismic Suite
(FSS) payload, a single station with two seismometers, on the far side
of the Moon. During FSS operations, the processing of the data will
provide us with new insight into the Moon’s seismic activity. One of
Apollo mission finding is the existence of deep moonquakes (DMQ), and
the nature of their temporal occurrence patterns as well as the
spatially clustering. It has been shown that DMQs reside in about 300
source regions. In this paper we tackle how we can associate new events
with these source regions using the single station data. We propose a
machine learning model that is trained to differentiate between DMQ
nests using only the lunar orbital parameters related to DMQ time
occurrences. We show that ML models perform well (with an accuracy
>70%) when they are trained to classify less than 4 nests.