Study On The Rupture Directivity Of Moderate Earthquakes

Moderate earthquakes(M5.5-7)occur in both interplate and intraplate regions with relative short recurrence interval,and may cause severe damage.Thus,rapid determination of source parameters of moderate earthquake is essential for earthquake disaster mitigation.Point approximation is routinely adopted for representing earthquake source model,however,the fault plane and auxiliary plane in point source model are equivalent and thus indistinguishable.Instead,rupture directivity parameters,including fault plane,rupture direction,rupture length and rupture velocity,can help resolve ruptured fault plane and should be adopted to describe the source finiteness feature.Earthquake rupture directivity affects the ground motion in near-fault regions significantly.For example,regions along the rupture direction usually suffer stronger ground motion with enhanced amplitude and shorter duration,while regions against the rupture direction experience weaker shaking and longer duration.Thus,rapid determination of rupture directivity helps improve more realistic shake map,which provides essential information for hazard evaluation and rapid rescue.Broadband three-component seismic waveforms contain abundant temporal and spatial information of source rupture process,and can be used to determine the fault plane and rupture length.For regions with accurate 3D velocity models and dense network,there are effective methods to estimate the rupture directivity.However,there are some regions with active seismicity and dense population yet without reliable 3D velocity models or dense seismic network.such as Yunnan.Sichuan,Italy,Turkey and the central U.S.In these regions,new methods are needed to determine earthquake rupture directivity.We develop a method based on Qin et al.(2004)to resolve rupture directivity of moderate strike-slip earthquakes.In this method,a reference earthquake is used to provide path calibration and suppress 3D structure effect.and the method is referred to as "reference earthquake method".We validate the method s effectiveness via forward tests,and find this method can determine the correct fault plane and rupture direction robustly even when there is partial bilateral rupture or relative location error(smaller than 25%rupture length).Then,we apply this method to the 2008 Mw5.2 Illinois earthquake,the 2014 Mw6.1 Ludian earthquake and the 2016 Mw5.4 Gyeongju earthquake.The Illinois earthquake is found to rupture towards southeast for about 2-3 km along the fault plane striking 297°.The Ludian earthquake is found to rupture towards southeast for at least 4 km along the fault plane striking 168°(Baogunao-Xiaohe Fault).The Gyeongju earthquake is found to rupture towards southwest for about 4 km along the fault plane striking 24°(Yangsan Fault).The derived fault plane,rupture direction and rupture length for these three earthquakes are consistent with other studies such as aftershock distribution and intensity map.Since the reference earthquake method usually requires a proper reference event,the method is less applicable to regions with low historical seismicity.We develop a reference station method based on ambient noise location method proposed by Zhan et al.(2011)and Zeng et al.(2014).This method explores the feasibility of using estimated Green's functions(EGFs)extracted from Noise Cross-correlation Functions(NCFs)as path calibration to resolve the rupture directivity.Two methods are developed,which are based on waveforms and travel time,respectively.For the reference station method based on waveforms,we investigate the stability of this technique against the choice of reference station and point source parameters(focal mechanism and focal depth).It is found that combination of reference stations could lead to smaller misfit and improve resolving capacity,and this method is modestly robust against error in source parameters.For the reference station method based on travel time,we investigate the source parameter influence on Rayleigh wave group travel time and how the spectral null affects group travel time measurement.We find that if focal mechanism and focal depth are known,source term correction can be applied to Rayleigh wave group travel time in periods outlying the spectral null.Then,we apply both methods to the 2011 Mw5.7 Oklahoma earthquake and find this earthquake ruptured to southwest for 6-11 km along fault plane striking 234°.If a library of EGFs is established,it will contribute to rapid rupture directivity determination for future strike-slip earthquakes.For dip-slip earthquakes that rupture along dip,we propose the reduced finite source method to resolve the rupture directivity rapidly with teleseismic P wave.The reduced finite source model combines the point source model and rupture directivity calibration,and we develop the telRup codes to compute teleseismic body waves with source finiteness feature.We verify the effectiveness of this method with forward tests.and find that the method is robust against different data preprocessing schema,uncertainties in point source parameters,uncertainties in source finiteness parameters and uncertainties in velocity model.We apply this method to the 1994 Mw6.7 Northridge earthquake,the 2008 Mw6.0 Nevada earthquake,the 2011 Mw5.8 Virginia earthquakeand the 2016 Mw5.9 Menyuan earthquake.The results indicate that the Northridge earthquake ruptured upwards along the southwest dipping fault plane for about 20 km.the Nevada earthquake ruptured downwards along the southeast dipping fault plane for about 13 km,the Virginia earthquake ruptured upwards along the southwest dipping fault plane for about 8 km,and the Menyuan earthquake ruptured downwards along the southwest dipping fault plane(secondary fault of Lenglongling Faults)for about 8 km.We also explore the feasibility of resolving the rupture directivity rapidly with teleseismic P wave time shift data,and investigate the stability of point source parameter inversion with the CAP technique and teleseismic P wave.The 2015 Mw7.9 Nepal earthquake is chosen as a case study,and the results show that inversion using stations within 40° could provide stable focal mechanism,focal depth,duration and moment magnitude in 10 minutes after the earthquake occurs.Moreover,the time shift data can provide good estimation of rupture length after P wave onset correction.Based on the results derived with different focal mechanism and stations at different epicentral distances,the Nepal earthquake is estimated to rupture towards southeast for 155±15 km with a velocity 3.1 ±0.3 km/s.In summary,this paper develops a few methods to resolve rupture directivity for moderate earthquakes with different rupture patterns.The effectiveness,stability and applicability of different methods are investigated with forward tests and real earthquake case studies,and we expect these methods help provide information for calculating realistic shake map.