5:15 PM - 6:30 PM
[SCG53-P03] Evaluation of simple source time function in rapid source parameter determination system (AQUA)
Keywords:the 2011 off the Pacific coast of Tohoku earthquake, rapid source parameter determination system, Centroid moment tensor
In the rapid source parameter determination system called Accurate and QUick Analysis System for Source Parameters (AQUA), source parameters including location, size, and moment tensor (MT) are automatically determined with good accuracy. By incorporating the velocity type strong motion seismometer in the NIED full-range seismograph network (F-net) and changing the processing flow, it became possible to analyze the 2011 off the Pacific coast of Tohoku earthquake (M9.0)(Kimura et al., 2020). However, its size was underestimated as Mw8.6. Since appropriate estimation of the earthquake size is important for disaster mitigation including tsunami warning and initial damage assessment, the simple source time function (STF) in AQUA was evaluated, as suggested by Asano and Kimura (2012).
The isosceles triangle type STF is incorporated into the current processing flow after determining the MT solution with fixed horizontal location (called as ´AQUA-MT´) and its source duration is estimated by a grid search. Here, the centroid time and depth are fixed to those of AQUA-MT to shorten the processing time. When the variance reduction (VR) decreases, the source duration with the highest VR is used hereafter, including the horizontal location estimation (called as ´AQUA-CMT´). Based on this flow, we reanalyzed the 2011 Off Tohoku earthquake and its major aftershocks and earthquakes with Mw6.8 or greater. In addition, the 2003 Tokachi-oki Earthquake (Mjma8.0) and the 2004 Kii Peninsula and Tokaido-oki earthquake (Mjma7.4) were reanalyzed as past major earthquakes. For these two earthquakes, waveform data recorded by the F-net broadband seismometer with epicentral distance of 400 km or more is used.
In the 2011 Off Tohoku earthquake, selecting a 70 sec source duration increased VR from 67.0% to 81.4% and Mw from 8.6 to 8.8. The centroid time, depth, and the focal mechanism are almost the same. The additional processing time was estimated to be about 1 sec. The source duration of major aftershocks of the 2011 Off Tohoku earthquake (15:08 and 15:15, 11 March 2011 (JST); Mjma7.4 and 7.6; note that the latter is the largest aftershock) and two past major earthquakes were estimated as 50, 30, 40, and 40 sec, respectively. Mw increased from 7.4 to 7.6, 7.8 to 7.9, 7.8 to 8.0, and 7.4 to 7.5, respectively.
The incorporation of the STF reduced the underestimation of Mw for the 2011 Off Tohoku earthquake while its size is still underestimated. The source duration of 70 sec is comparable with analysis of rupture process (e.g., Suzuki et al., 2012) and can be considered to represent a major part of the slip history. This duration is not negligible compared to the bandwidth of the analysis (50-200 sec). Hence, we can consider that the incorporation of the STF improved size underestimation. Mw of GlobalCMT for the aftershock on 15:15 and the 2003 and 2004 earthquakes are 7.9, 8.3, and 7.4, respectively. Incorporation of the STF reduced difference from GlobalCMT in most cases.
A grid search of the source duration with a fixed centroid time and depth suppresses a significant increase in processing time. Since STF is introduced only when VR increases after determination of AQUA-MT, the determination ratio would be maintained. By incorporating STF, we can expect to be able to better estimate earthquake size while maintaining process speed and determination ratio.
The isosceles triangle type STF is incorporated into the current processing flow after determining the MT solution with fixed horizontal location (called as ´AQUA-MT´) and its source duration is estimated by a grid search. Here, the centroid time and depth are fixed to those of AQUA-MT to shorten the processing time. When the variance reduction (VR) decreases, the source duration with the highest VR is used hereafter, including the horizontal location estimation (called as ´AQUA-CMT´). Based on this flow, we reanalyzed the 2011 Off Tohoku earthquake and its major aftershocks and earthquakes with Mw6.8 or greater. In addition, the 2003 Tokachi-oki Earthquake (Mjma8.0) and the 2004 Kii Peninsula and Tokaido-oki earthquake (Mjma7.4) were reanalyzed as past major earthquakes. For these two earthquakes, waveform data recorded by the F-net broadband seismometer with epicentral distance of 400 km or more is used.
In the 2011 Off Tohoku earthquake, selecting a 70 sec source duration increased VR from 67.0% to 81.4% and Mw from 8.6 to 8.8. The centroid time, depth, and the focal mechanism are almost the same. The additional processing time was estimated to be about 1 sec. The source duration of major aftershocks of the 2011 Off Tohoku earthquake (15:08 and 15:15, 11 March 2011 (JST); Mjma7.4 and 7.6; note that the latter is the largest aftershock) and two past major earthquakes were estimated as 50, 30, 40, and 40 sec, respectively. Mw increased from 7.4 to 7.6, 7.8 to 7.9, 7.8 to 8.0, and 7.4 to 7.5, respectively.
The incorporation of the STF reduced the underestimation of Mw for the 2011 Off Tohoku earthquake while its size is still underestimated. The source duration of 70 sec is comparable with analysis of rupture process (e.g., Suzuki et al., 2012) and can be considered to represent a major part of the slip history. This duration is not negligible compared to the bandwidth of the analysis (50-200 sec). Hence, we can consider that the incorporation of the STF improved size underestimation. Mw of GlobalCMT for the aftershock on 15:15 and the 2003 and 2004 earthquakes are 7.9, 8.3, and 7.4, respectively. Incorporation of the STF reduced difference from GlobalCMT in most cases.
A grid search of the source duration with a fixed centroid time and depth suppresses a significant increase in processing time. Since STF is introduced only when VR increases after determination of AQUA-MT, the determination ratio would be maintained. By incorporating STF, we can expect to be able to better estimate earthquake size while maintaining process speed and determination ratio.