Joint Inversion Of Surface Wave Dispersions,Receiver Functions And PmP Times For Crustal Structures And Its Application In The Western Hubei Province

One of the fundamental problems in geophysical inversion is non-uniqueness.How to resolve this problem is an active researching topic.Since 1990s,it becomes a trend to incorporate multiple geophysical data sets to resolve the non-uniqueness problem.Often-used geophysical data include surface wave dispersions,receiver functions,travel times of seismic waves,gravities,and magnetotelluric data.One of the first and most widely used one is the combination of surface wave dispersions and receiver functions.This is due to that velocity and interface are the two most important parameters of Earth's structure.Moreover,the surface wave dispersions are sensitive to average shear wave velocity,while the receiver functions are sensitive to seismic interfaces.Besides,the receiver function can constrain the average Vp/Vs ratio when they contain clear multiple converted phases.Hence the information provided by surface wave dispersions and receiver functions is complementary.However,in practice,the information contained in receiver functions,especially those of multiple phases which have a long ray paths in the crust,have weak energy due to the attenuation and scattering in the crust.This makes the receiver functions unable to reduce the trade-off between the crustal thickness and Vp/Vs ratio and the non-uniqueness of solution becomes serious.This study is aimed at resolving the non-uniqueness by incorporating P-wave velocity information to determine earth structure.First,a deconvolution technique is developed to measure the time differences TPmP between the direct teleseismic S wave and its post-critical reflection phase SsPp.The traditional H-?method is then extended by jointly using TPmP and P-wave receiver functions.A joint inversion method based on the Neighborbood Algorithm is also developed to jointly invert T_PmP,P wave receiver functions and surface wave dispersions.Constrains from TPmP are studied.Finally,using two year records of 63 seismic stations of China national network and half-year records of 29 portable seismic stations,I applied the joint inversion and obtained crustal and uppermost mantle structure in the western Hubei Province and its adjacent regions.Following conclusions are obtained from this research:?1?Numerical tests confirm the correctness of our deconvolution technique to measure the time differences TPmP between the direct S waves and post-critical reflection phase SsPp.By applying this technique to two-year records of 63 China national seismic stations,we selected 23 high quality S wave seismic events.The good matching waveforms between observations and predictions further improved the feasibility of the deconvolution technique.A map of TPmP variation in the study area is obtained.This map shows that the variation of T_PmP has a high correlation with geology in the study area.The Jianghan and Nanxiang basins in the study area have low values of TPmP?5-7 s?,with the lowest value appeared in the Jianghan basins?5 s?.The Daba mountains,Huangling massif and Wuling uplift have high values of TPmP?7-11 s?,especially in the Wuling uplift where the value is the highest?9-11 s?.?2?The relations among TPmP and Ps and its behind multiple phases in P wave receiver functions show that:traditional H-?method brings large uncertainties to the estimated crustal thickness H in areas with a rapid varying Vp,take the numerical tests for example,in which Vp varied from 5.5 to 7.0 km/s,and it makes the H changed from35 to 45 km;the P wave information contained in TPmP is different from those of Ps and its behind multiple waves of P wave receiver functions,especially in H-Vp domain,where the Ps and its behind multiple waves exhibit a strong trade off of each other,the H and Vp can be better resolved by incorporating T_PmP;in the case that the multiple phases have a weak energy or not exist,H,?and Vp can still be uniquely determined by combine Ps phase and T_PmP.By applying to three stations with typical P wave receiver functions in the study area,the results show that:the extended H-?method have a same result with the traditional H-?method when there's enough energy of multiple waves;incorporating of TPmP can resolve H,?and Vp better than those of traditional H-?method;in the case that strong interface existed in the crust,the extended H-?method has the chances to recognize the Moho related Ps phase.By applying the extended H-?method to 63 China national seismic stations,this study obtained the variation of H,?and Vp in the study area.The results show that:the crust thickness is thin in eastern part of the study area,and the average crust thickness is about 30 km,with the thinnest area is in Jianghan and Nanxiang basins,where the crust thickness is about 27 km,in the western part of the study area,the average crust thickness is about 45 km,with the thickest area is about 55 km in Daba mountain.The distribution of?in the study area shows,the eastern basins and the southwestern part of Wuling uplift have a low value of??1.73?,while the Daba mountain and the eastern edge of Sichuang basin have a high value of??1.85?.The distribution of Vp in the study area shows,the southern part of Wuling uplift,Jianghan and Nanxiang basins have a low value of Vp,while the Daba mountain has a high value of Vp.?3?In the method of joint inversion of T_PmP,receiver functions and surface wave dispersions,the one dimensional numerical example shows that,joint inversion of surface wave dispersions and receiver function without multiple phases cannot recover the true crustal structure,especially for those of crustal thickness and Vp/Vs,after incorporating T_PmP,the inverted H and Vp/Vs can be resolved much better.The two dimensional numerical example show that,in the case that recever functions is disturbed by sediment layers and also without multiple phases,incorporating T_PmP can recover the crustal structure with a thick crustal thickness,while in the area with a thin crustal thickness the joint inversion cannot recover the true model since the interference between the sediment related phases and Moho related Ps phase.The inverted results can be recovered better when simultaneously incorporating T_PmP and T_Ps,especially for those of crustal thickness and Vp/Vs.TPs in the inversion actually plays a role of weighted factor to the Moho related Ps phase.A real world test to a seismic station HB_ZUX in the study area further testified the constrains of TPmP in the inversion,especially for crustal thickness and Vp/Vs.?4?The crustal and uppermost mantle Vs structure of the study area show that:the anomaly distribution of Vs for geology units is the same as the former studies,in which the Jianghan,Nanxiang,Sichuang basins and Daba mountain are low velocity anomaly in the upper crust,while the Huangling massif,Shenlong massif,Hannan massif and Wuling uplift exhibit high velocity anomaly in the upper crust.The low velocity anomalous body?Vs<3.4 km/s?of Daba mountain has a shape of a reverse trapezoid in the direction from northwest to sourtheast and it extrudes to the southwest.The most southwestern part is coupled with the low velocity anomaly of Sichuang basin,the depth of low velocity anomalous increases from 12-15 km in the southwest to the maximum depth of 18 km,and then it decreases to 5 km dramatically when reaching the Ankang fault.The reverse trapezoid anomalous low velocity body probabaly closely related to the clamping between Hannan massif in the northwest and Huangling-Shenlong massif in the southeast.The crustal thickness and Vp/Vs from the joint inversion show that,the crustal thickness of the eastern region is shallow than those of the western region,and the Vp/Vs of the eastern basins is low,while those of western mountains is high.These features are almost the same as the results obtained by the extended H-?method.The feature of media to high values of Vp/Vs and thick crustal thickness in Daba mountain probabaly indicates a mafic or a patially eclogitized lower crust is existed in this area.