Spatiotemporal gravity signals from gravity measurement have contributed to the understanding of the surface and subsurface mass redistribution, which reflects the crustal deformation and density changes (Van camp et al., 2017; Battaglia et al., 2018). In the past few decades, the satellite and terrestrial gravity data have been employed to investigate the geophysical process of earthquake (Tanaka et al., 2010; Okubo et al., 2020; Bouih et al., 2022). Recently, it is reported that the gravity changes derived from satellite gravity measurements may be connected to the pre-seismic signals potentially caused by the deep mass redistribution of the large subduction earthquakes, such as the 2010 Mw8.8 Maule earthquake (Bouih et al., 2022) and the 2011 Mw9.0 Tohuku-Oki earthquake (Panet et al., 2018, 2022). However, the large observation errors of the GRACE satellite gravity (Wang et al., 2019) and footprints of space gravity data (Tapley et al., 2004) may make it difficult to obtain the local gravity changes with spatial resolution less than several hundred kilometers.
Compared to satellite gravity observation, the Terrestrial Hybrid Repeated Gravity Observation (THRGO) system (Hinderer et al., 2016; Okubo et al., 2020) can solve the problem of spatial limitation. In this system, the microgal-level (1 microgal = 1 µGal = 10− 8 m/s2) absolute and relative gravity observations can be periodically performed in local areas. In the recent decades, with the rapid development of high-precision absolute gravimetric instruments, the time-varying gravity data obtained by the THRGO system have been employed to investigate the crustal deformation, density change, and deep crustal fluid movement more often in the tectonic active region. In the western Europe, low gravity change rates and slow intraplate vertical movements were obtained by Van Camp et al. (2011) using the repeated absolute gravimetry. With the absolute gravity observation data over 20 years in Japan, Tanaka et al. (2018) reported negative mass anomalies in two recent long-term slow slip events in the Tokai area. Sun et al. (2008) combined the absolute gravity observation data and the GPS measurement results to reveal the deep mass change and the increasing of crustal thickness beneath the Tibetan Plateau. In the western China, Zhu et al. (2010) found based on absolute and relative gravity measurement data with an accuracy of ~ 15 µGal, that the gravity field changed significantly before the 2008 Wenchuan MS8.0 earthquake in a broad region with the size of about hundreds of square kilometers. Chen et al. (2016) found the gravity increase at four absolute gravimetry stations in the south Tibet near the epicenter of the 2015 Nepal Mw7.8 earthquake, indicating the possible mass distribution changes in the broad source region of this earthquake. Xuan et al. (2019) used the repeated relative gravity observation data in the northeastern Tibetan Plateau to derive the density variations at different depths over different time before the 2016 MS6.4 Menyuan earthquake.
However, what causes the observed gravity changes in terrestrial gravity measurement and whether the gravity changes can be reliably used in the earthquake precursor research are still under debating. The arguments mainly focus on the factors affecting the measurement results of absolute and relative gravimetry, such as instrumental artifacts, hydrogeological effects, and topography deformations (Van camp et al., 2016, 2017; Yi et al., 2016). It is well known that the surface observable gravity changes related to seismic tectonic movement in deep crust are characterized by low signal-to-noise ratios (Crossley et al., 2013). Consequently, whether the observed gravity change has an enough signal noise ratio (SNR) for investigating the tectonic movement is still a great challenge. However, Zhang et al. (2020) revisited the observed gravity changes at Pixian absolute gravimetry station before the 2008 Wenchuan MS8.0 earthquake using absolute gravimetric with a good SNR. They found that the corrected residual gravity changes are much larger than the possible gravity effects due to vertical deformation and hydrological processes. Their results provided a new insight on interpreting the mechanisms of deep mass changes. In terms of geodynamic numerical simulation, Liu et al. (2022) proved that the presence of observable yearly transient change in gravity is related to the deep fluid movement. Therefore, by means of the THRGO, it is beneficial to isolate the gravity signals potentially related to the deep mass migration.
In this study, we investigate the gravity change derives from the absolute and relative gravity observation before and after the 2013 Lushan MS7.0 earthquake. The mainshock occurs on a blind revers-fault in the Longmenshan fault belt in the eastern Tibetan Plateau. The epicenter is located in the northern section of the Sichuan THRGO network. Several previous studies have reported that there are significant gravity changes before the MS7.0 Lushan earthquake based on the THRGO dataset (Zhu et al., 2013; Hao et al., 2015; Xuan et al., 2015). However, how the gravity change relates to the mechanism of the occurrence of Lushan earthquake remains debated. Therefore, we try to isolate the gravity signals that are potentially related to deep tectonic movement. We interpret the gravity changes caused by deep crustal sources to study the seismogenic environment of the Lushan earthquakes.
This study is arranged as follows. In section 2, we describe process and quality control of the gravity observation data. In section 3, the inversion method is introduced to obtain the deep source apparent density model. In section 4, we use the equivalent source model (ESM) to estimate the deep source mass change parameters in the significant gravity change region. In this section, we also give the geodynamic interpretation of the equivalent source inversion results as well as the related evidence from other geophysical observations. In section 5, we summarize our findings and draw some concluding remarks.