7. Discussion
The seismicity of the MSH is investigated from documented historical earthquakes previous to 22 June 2020. Three M>=6.5 historical earthquakes cover all the MSH segments. However, instrumental seismicity is relatively poor. The EHB catalog (Engdahl et al., 2006) shows three earthquakes near the Eastern and Central segments of MSH, close to DMV. The GCMT catalog shows two 5<M<5.3 earthquakes on the Central and Eastern segments of the MSH. The IRSC network earthquake catalog has improved from 2006 in terms of completeness. They show 67 M>=2.5 earthquakes within a distance of 5 km from the fault before the 2020 mainshock. Most of this seismicity concentrated on the Central segment of MSH, South of the DMV. Three peaks are observable in the cumulative Scalar Seismic Moment chart of these earthquakes (Fig. 5c). Interestingly, the central peaks that are mostly related to the 2006 and 2007 earthquakes, coincide with the estimated rupture areas during the Ms 5.2 1930 and Mb 4.0 1955 earthquakes. A possible explanation is that they are late aftershocks of these earthquakes. The Western one is close to thermal areas reported by Eskandari et al. (2018). A low-velocity region has obtained Southwest of DMV that extended until the MSH down to the depth of 15 km in a tomography study by Mostafanejad et al. (2011) (Fig. S1a). The observed thermal activities in the same area are probably due to the existence of some branches of the DMV magma chambers in that area which was also suggested by Eskandari et al. (2018).
The rupture process and the fault geometry of the 7 May 2020 M5.1 Damavand earthquake was investigated by inverting both the local broadband seismic data for the moment tensor and the near-field strong-motion displacement time series for its extended rupture model. The mainshock occurred on the central segment of the MSH: It nucleated ~15 km SSW of the DMV crest and at a depth of ~14 km. The rupture is estimated in an elliptical patch with a major-minor axis of 5 km-3.6 km. It evolves mostly toward the Northwest along strike and to the up-dip direction at a sub-shear speed of ~2.75 km/s for 2.8 s. The estimated geometry is ~WNW (292°) strike and ~60° dip to the North. The obtained scalar seismic moment by point-source moment tensor inversion is 4.8 e+16 Nm while using an extended rupture model, this value reduces to 4.04 e+16 Nm, suggesting the release of some of the scalar seismic moment at relatively lower frequencies between 0.03 Hz to 0.08 Hz.
The interpolated PGA from 33 recorded stations of the ISMN network suggest a west-northwestward directivity, which is to some extent consistent with our source model showing a westward directivity (Fig. 4a). For each station, peak values of the geometric average of the two horizontal components of strong motions are considered as horizontal PGA. The damping observed in PGA in the center of Tehran is interpreted as attenuation due to the deepest part of the sedimentary basin (see Majidnejad et al., 2017). The Fourier spectra of the strong motion data for stations FRK3 and LVS1 that have negligible site effect, show the low-frequency content of this event with a corner frequency of 1 Hz. Far-field Brune models for an M5.1 earthquake is estimated in Tehran region (Brune, 1970) in which source slip patch radius is roughly 4.3 km, S-wave velocity is 3.5 km /s, and \(\rho\) is equal to 1.5 (Figs. 6 and 7). The fmax is obtained from the smoothed spectra (see Konno & Ohmachi, 1998; Figures 6, 7, S12, S13) in the ranges between 6 Hz to 16 Hz in the Tehran region. Such difference is mainly related to the site attenuation (i.e., scattering and dissipation) (see Gomberg et al., 2012; Hanks, 1982).
The stress drop has obtained 2.6 bar from the extended rupture model, posing roughly a circular slip patch with a radius of 4.3 km, and a scalar seismic moment of 4.8 e+16 Nm (see Madariaga, 1977). An empirical relation between scalar moment and stress drop by Ide and Beroza, (2001) suggests a stress drop of ~10 bars for an M5.1 earthquake. The obtained relatively low stress drop is consistent with the relatively large rupture length. We note that the obtained rupture length (estimated between 7 km to 9 km) is relatively large for such a magnitude earthquake (for example see Momeni et al., 2019 for a rupture length of ~12 km estimated for an Mw6.5 earthquake; and Vicic et al., 2020 with rupture length of ~4 km for an Mw5.1 earthquake).
The mainshock exhibits a left-lateral strike-slip mechanism (Rake=14°) the same as the general mechanism of MSH proposed by Tatar et al., (2012), a geodetic study of Djamour et al., (2010), and geological-paleoseismological studies by Nazari et al., (2009) and Solaymani-Azad et al., (2011). A maximum slip of ~3 cm was estimated between depths of 12 km and 11 km. The rupture stopped at a depth of 8 km.
The mainshock rupture and the early aftershocks occurred between the two peaks of cumulative scalar seismic moments on the MSH, proposing that this part of the fault was somehow locked compared to two other neighbors that experienced the 1930 and 1955 earthquakes.
The aftershocks were distributed toward the West and up-dip, consistent with the main rupture direction and general orientation of the MSH. The largest aftershock with M4.1 occurred 20 days after the mainshock with a left-lateral strike-slip mechanism, the same as the mainshock. Aftershocks surrounding the mainshock slipped area (Figs. 4a, 5), is a consistent feature of large earthquakes (see Henry and Das, 2002).
The 2020 seismic activity occurred at a depth range between 15 km to 8 km, where Tatar et al. (2012) also detected most of the microearthquakes. This range is almost the same as the upper-crystalline layer of the velocity model obtained by Abbasi et al. (2010) for the region. This relatively thick and deep part of the seismogenic layer may have the potential for the production of large earthquakes with low-frequency contents that can reach Tehran with less damped seismic energy and affect the tall buildings, the same as the 7th May 2020 M5.1 mainshock.
The smooth geometry of the central segment of MSH may facilitate the rupture expansion on it. Occurrence of the 1930 (Ms 5.2), 1955 (Mb 4.0), 1983 (Mw 5.3), and 2020 (Mw 5.1) earthquakes in the South of the DMV, together with its seismic activity from 2006, suggest a strong relationship between the volcanic activity of DMV and relatively high seismicity rate of the central segment of the MSH. Also, most of the microseismic activity and larger microearthquakes were reported by Tatar et al. (2012) on the central segment of MSH, just to the South of DMV between longitudes from 51.75 E to 52.2 E, while their seismic network was well-distributed on the two other segments of MSH.
Previous studies suggested the existence of a hot young sill-like magma chamber of DMV in the Southwest of its current crater (i.e. Mostafanejad et al., 2011; Shomali and Shirzad, 2014; Yazdanparast and Vosooghi, 2014; Eskandari et al., 2018). While the old magma chamber of Damavand is detected toward the North-Northeast of the crater and is detected as a cooled high-velocity dike-like structure (Mostafanejad et al., 2011). The existing young magma chamber may increase the pore pressure on the left-lateral strike-slip MSH which consequently decreases the effective normal stress on it and facilitates the rupture nucleation-expansion (Fig. 8). Such phenomena have been widely observed and reported mostly for Strike-slip and Normal faulting mechanisms (i.e. Saar and Magna, 2003, Goebel et al., 2017, Scuderi et al., 2017, Johann et al., 2018, Eaton and Schultz, 2018, Benson et al., 2020). On the other hand, such a mechanism may not allow considerable accumulation of strain on this part of the MSH near DMV (i.e. Yagi et al., 2016).
The 2020 M5.1 earthquake is the largest well-recorded event on the MSH after the 1983 event. This segment of the MSH has experienced the 1830 IX 7.1 historical earthquake. All of the evidences indicate that the 2020 M5.1 mainshock and recent seismicity of the central segment of MSH are related to the existence/activity of the magma chamber of DMV. We also stress that 1930, 1955, and 1983 earthquakes on the South of DMV might have happened as a result of the same unclamping mechanism due to the existing high pore pressure.
Compared to the Central segment of Mosha, the Western segment that is closer to Tehran city is silent. However, GPS studies confirm its lower deformation rate (1mm/y, Djamour et al., 2010). The occurrence of earthquakes like the 2012 Ahar-Varzaghan doublet (Mw 6.5 and Mw 6.3) with almost no detected seismic activity in the IRSC network before the mainshock and low deformation rate (i.e., Momeni et al., 2019) highlights the importance of a detailed seismic-geodetic study on the Western segment of MSH that will affect the seismic hazard of that region, and especially Tehran city. Also, the Eastern segment of MSH shows seismic activity which highlights its importance as another potential segment of the MSH for future large earthquakes.