The M_w = 5.8 2019 Silivri earthquake, NW Türkiye: is it a warning beacon for a big one?

Abstract

The September 26, 2019 Silivri earthquake (M_W = 5.8) occurred along the North Anatolian Fault beneath the Marmara Sea and its epicenter was in an identified seismic gap. Coseismic stress calculations demonstrate that the 1999 İzmit earthquake (M_W = 7.4) caused stress increase from 0.057 to 0.114 bars at its hypocenter, depending on the various reported rupture parameters. In addition, over 20 years following the 1999 earthquake, and constituting the main difference from previous studies, viscoelastic postseismic stress computations indicate stress increase from 0.081 to 0.135 bars at the hypocenter. In spite of the positive stress transfer, the 2019 earthquake occurred long after the end of the computed aftershock time span (~ 16 years) of the 1999 earthquake. Plots of the seismicity around selected points within the gap also show that the background seismicity level following the 1999 earthquake was reached in 2003. Therefore, it is suggested that the 2019 earthquake was not an aftershock but rather an independent event, and its occurrence was hastened about 4 years due to stress loading. Further analysis of the seismicity between 1978 and 2020 indicates that the b value increased from a range of 1.0–1.1 to 1.6–1.8 till 2002, then progressively decreased to 0.9–1.0, which is consistent with positive stress transfer. The stress increase ranging from 0.19 to 2.52 bars on the segments within the gap brought forward their seismic cycles about 33 and 2 years from east to west, respectively. These additional clock advances in the seismic cycles due to stress load urgently require risk mitigating actions.

Graphical abstract

Appendix A: analysis of the seismicity catalogue

The seismicity beneath the Marmara Sea along the NNAF is analysed to interpret the earthquake hazard using seismicity parameter b value. The seismicity data for the 1978–2018 period is taken from the homogenized catalogue of Tan (2021) and is based on moment magnitude. The study of Tan (2021) contains new conversion equations between moment magnitude and other magnitude scales. The equivalent moment magnitudes for the all earthquakes that have no M_W are calculated using these magnitude relations to homogenize the catalogue. The reader is referred to Tan (2021) for further information. The catalogue is extended till July 2020 using KOERI catalogue. The same procedure as in Tan (2021) was applied to the Kandilli Observatory and Earthquake Research Institute (KOERI) catalogue after 2018 to estimate the equivalent moment magnitude for earthquakes with no given M_W. Therefore, the catalogue is homogeneous by means of magnitude. The epicentral distribution of M_W ≥ 2.5 seismicity from this combined raw catalogue for the Marmara Region after 1970 is demonstrated in Fig. 1b.

Figure 

Fig. 9

a Cumulative number of earthquakes for M_W ≥ 2.5 seismicity after 1970 in the raw catalogue. See Fig. 1b for the epicentral distribution; b Cumulative number of earthquakes for M_W ≥ 2.5 seismicity in the declustered catalogue after 1978; c Magnitude of completeness (M_C) of the declustered catalogue calculated over time after 1978. The computed time variations indicate that the catalogue is complete for M_W ≥ 2.8

9a indicates a dramatic change in the cumulative number of earthquakes after 1978, reflecting a main improvement in the local network. Therefore, it was decided to use seismicity after 1978 in the study, and declustering is carried out using the algorithm defined by Reasenberg (1985). The cumulative number of earthquakes and computed Mc time variations of the seismicity are demonstrated in Fig. 9b and c, respectively. Figure 9b suggests relatively homogeneous distribution compared to the raw catalogue (Fig. 9a) while Fig. 9c indicates that the seismicity is complete for M_W ≥ 2.8 earthquakes. To be more confident about the completeness of the catalogue, M_W ≥ 2.9 magnitude threshold is selected for the seismicity analysis. Epicentral distribution and cumulative numbers of earthquakes after this selection are shown in Fig. 

Fig. 10

a Epicentral distribution and b cumulative number of earthquakes for M_W ≥ 2.9 seismicity in the declustered catalogue after 1978. The red rectangle encloses the selected seismicity along the Northern Strand of the North Anatolian Fault beneath the Marmara Sea (see text for details and Fig. 7a for the cumulative number of earthquakes within the rectangle area). Red and yellow stars represent the epicenters of 17 August 1999 İzmit and 26 September 2019 Silivri earthquakes. MMF Main Marmara Fault, PI Princes Islands fault segment

10a and b, respectively. Cumulative numbers of earthquakes for declustered M_W ≥ 2.9 seismicity as shown in Fig. 10b suggest a relatively homogeneous catalogue. From the seismicity shown in Fig. 10a, the seismicity along the NNAF is selected for the seismicity analysis of the Marmara Sea (the seismicity enclosed by the red rectangle on the figure). The cumulative number of earthquakes after that last spatial selection of seismicity is shown in Fig. 7a.

To calculate the b value of the FMD of earthquakes (Gutenberg and Richter 1944) from the background seismicity, we have used the Weighted Least Squares Method (Menke 1989). A software package called ZMAP was used for the seismicity analysis (Wiemer and Wyss 1997; Wiemer 2001). The b value is an important seismotectonic parameter that has been shown to be inversely related to the stress in the crust and to vary mostly in the range 0.5–1.5 in finer scales, with a global mean of 1.0. (Scholz 1968; Wiemer and Wyss 1997).