5. Discussion
The seismicity of the NTF and NMF is investigated from documented historical earthquakes to November 2019 TKC earthquake. Many historical earthquakes are referred to NTF (Fig. 1a). The last two historical earthquakes of 1721 AD and 1780 AD cover all NTF segments. EHB catalog shows most of the seismic activity on the western termination of NTF. However, the GCMT catalog does not have any earthquake on the NTF.
The IRSC network earthquake catalog has improved from 2006 in terms of completeness and location accuracy. There are 512 earthquakes (45 of which have Ml>=2.5; Table S2) in the distance of 5 km from NTF and NMF for a period from 2006 until before AVD. These seismic activities are distributed along all segments of NTF and NMF unless the central segment of NTF that is situated North of SND and shows much less seismicity compared to its neighbor segments. Two remarkable peaks are observable in the cumulative scalar seismic moments of these earthquakes on both lobes of the central segment of NTF near SND (Figure 2b). A probable explanation for such behaviour is that the segment of NTF near SND is partially creeping. Djamour et al., (2011) and Rizza et al., (2013) reported a decrease in right-lateral surface deformation rate from West of BA (Longitude 47°) to the East from 7mmy-1 to 5mmy-1. Su et al., (2017) remarked that the region near BA is affected by deep magmatic activities of SND. This segment is close to the thermal areas (hot springs) near BA reported by Ghalamghash et al. (2019). Tomography studies by Rezaeifar et al., (2016) and Bavali et al., (2016) revealed a heterogeneous structure in this region with high and low-velocity anomalies. A low-velocity region has obtained beneath SND at depths deeper than 8 km that extends until the NTF by Bavali et al., (2016) (Fig. 5). However, at shallower depths, a relatively high-velocity anomaly obtained by Rezaeifar et al., (2016), and interpreted as cooled magmatic rocks of SND. The observed thermal activities near BA area are probably due to the existence of some dyke-like branches of the SND deep magma chamber in that area that was also suggested by Ghalamghash et al. (2019) as many young craters with dacitic to rhyolitic parasitic cones of magma of neo-Sahand were observed toward NNE of SND near BA (see Fig. 3). The other explanation will be the possible aid of NTF fractured area which is extended down to the depth of 20 km, in bringing heat to the surface. The existing heat increases the pore-fluid pressure in the fault area and unclamps this segment of NTF, facilitating its creep.
However, this segment of NTF was ruptured as a part of the 1721 AD M7.6 earthquake. Harris, (2017) mentioned that the creeping segments are also potential to rupture in M~6.8 earthquakes, and they usually rupture together with their nearby segments (i.e. Van den Ende et al., 2020). Dynamic weakening is the probable mechanism for rupture of such fault segments (i.e. Noda & Lapusta, 2013). The same mechanism may have happened during the 1721 AD earthquake, and that is most likely the reason for segmentation of NTF during 1721 AD and 1780 AD historical earthquakes.
The effect of raise of pore-fluid pressure in facilitating the fault creep/slip is widely observed and reported mostly for Strike-slip faulting mechanism (e.g. Avouac, 2015, Floyd et al., 2016, Goebel et al., 2017, Scuderi et al., 2017, Michel et al., 2018, Johann et al., 2018, Eaton & Schultz, 2018, Zhu et al., 2020, Momeni & Madariaga, 2020).
A slip model is estimated for the creeping segment of NTF from 1721 AD until before the 2012 AVD considering that half of the 7 mmy-1 right-lateral deformation rate obtained by Djamour et al., (2011) and Rizza et al., (2013) was happening in creep mode. This creep has occurred at the longitudes between 46.55° E to 46.85° E and with a locking depth of 20 km. Having a maximum cumulative slip of 1.02m, the obtained cumulative scalar seismic moment is 2.0 * e+19 Nm equal to Mw6.8 (for the fully creep mode, this value raises to 4.0 * e+19 Nm equal to Mw7.0). This also remarks that the other segments of NTF have a considerable amount of accumulated tectonic stress.
The creeping segment of NTF transferred positive Coulomb stress field of >4 bar on the neighbor segments, and brought them closer to failure (Fig. 3b). That is confirmed by the observation of two peaks of cumulative scalar seismic moments on both lobes of the creeping segment. These earthquakes can be considered as aftershocks of the creep event. Aftershocks surrounding a slip model is a consistent feature of large earthquakes (see Henry & Das, 2002).
The 3D stress field produced by this creep source on AVD is computed. The estimated slip model for the creeping segment can transfer positive Coulomb stress of >1 bar on the AVD and trigger them (Fig. 6). After the AVD until November 2019, one peak of the cumulative scalar seismic moment is observed for earthquakes occurred in 5 km distance from NTF, and that is in the middle of the central NTF. There is also one peak on the NMF. Observation of seismic activity on the previously creeping segment of NTF and absence of two peaks of cumulative scalar seismic moments on both lobes of that segment suggest a change in its rheology from creep to stick-slip after AVD. This change is probably due to the positive static normal stress field of >0.1 bar that was transferred from AVD on half of the creeping segment of NTF. Also, Momeni et al., (2019) compute the stress field by AVD on NTF and NMF and reported transfer of positive Coulomb stress of >0.1 bar on the central segment of NTF as well as NMF. For the NMF, two relatively small peaks of cumulative scalar seismic moment release are observed before the AVD (Fig. 2b). After AVD, one big peak is observed in between the two previous peaks suggesting that this part of NMF was partially locked, and triggered by the 2012 AVD.
Seismic quiescence of the creeping segment of NTF near SND from 2006 together with the observed magmatic activities in that area proposes a strong relation between the volcanic activity of SND and frictional properties of that segment of NTF. Compared to the central NTF, the western segment that is closer to the Tabriz city shows higher seismic activity. Also, the Eastern segment shows considerable seismic activity which highlights its importance as another potential segment of the NTF for future large earthquakes.
The smooth geometry of the central and western segments of NTF may facilitate the rupture expansion on them. However, low seismic coupling in the creeping central NTF may act as a barrier and stop ruptures from expansion toward the western segment.
The suggested 20km thick seismogenic layer for NTF (Djamour et al., 2011; Moradi et al. 2011), highlights its potential for the production of large earthquakes and with low-frequency seismic energy contents that can reach to Tabriz city with less damped energy and affect the tall buildings.