Abstract: Layered 1-D velocity models are widely used in seismic network routine locations and in seismological studies, such as earthquake relocation, focal mechanism inversion, synthetic seismogram calculation and geodynamics simulation. It’s also used as a reference model for 2-D and 3-D tomographic inversions. Therefore, obtaining a more reliable 1-D velocity model is an extremely important work for the study of earthquake source parameters and seismic tomography. The onshore-offshore area of the Pearl River Estuary is located at the transition zone between South China block and South China Sea bock, the special ocean-land transitional crustal type and the littoral fault zone, which is the regional seismic control structure passing through it, makes it a potential seismic source. Meanwhile, the Pearl River Delta has the most developed economy and dense population in South China. However, the 1-D velocity model used in seismic network routine operations has not been updated since 1990. To investigate the seismic structure and potential strong earthquake risk in this area, we conducted a 3-D active source seismic experiment in 2015, which incorporated sea-based airgun sources and land-based dynamite sources, and seismic recorders both at the onshore and offshore area in the Pearl River Estuary. A high quality subset of the data was used to derive an improved 1-D seismic V_P model for seismological studies. The model is constructed using the VELEST program with first arrival P-wave travel time data, together with station corrections, which account for shallow velocity anomalies from the true velocity model. The reliability of our new model is assessed by good fitting of the travel time data of airguns and dynamites and better earthquake relocation results.
The final 1-D model provides a good fit for travel time data. After iterative inversion, the root-mean-square travel-time error is 0.07s in the onshore area and 0.21s in the offshore area. Within 6km top of the model, the P-wave velocity of onshore area is 5.22~5.99km/s, and the offshore area is obviously lower, which is 2.11~6.03km/s. The retrieved values are in agreement with the thick sedimentary basins in the Pearl River Estuary Basin whose velocity is obvious lower. Then the velocity smoothly increases with depth, within the depth range of 6~15km, the P-wave velocity of onshore area is slightly lower than the offshore area, which may be due to the wide-spread low velocity layer at the middle crust depth in South China Block. Below the depth of 15km, the P-wave velocity of offshore area is greater than that of the onshore area, which is consistent with the high velocity layer in the base of the thinned continental crust and the gradually uplifting of Moho depth seaward as reported in the previous studies.
The spatial distribution of station corrections correlates well with the near-surface structure and geological features. In the area onshore of the Pearl River Estuary, positive values of station corrections are mostly observed in correspondence with the Pearl River Delta sedimentary basins due to its lower velocity values, such as Sanshui Sag, Shunde Sag and Dongguan Sag, etc. While stations located in granite, limestone and metamorphic rocks outcropping area show early P-wave arrivals(negative station corrections). In the area offshore of the Pearl River Estuary, the spatial distribution of station corrections shows a significant lateral variation and 80%larger than the onshore area. It has a good spatial correlation with the buried depth of the sedimentary basement inverted by reflection seismic survey, where the deposits are thicker, the station corrections are positive, the underground medium presents a low velocity, and vice versa. Negative values of station corrections are observed northwest of the NE-trending littoral fault zone, while positive values correspond to the thick sedimentary basins in the Pearl River Estuary Basin southeast of the littoral fault zone.
At last, we relocated 425 earthquakes in the onshore area and 234 earthquakes in the offshore area with M_L≥0.0 using simul2000 algorithm. The result shows that our new model is better than the South China model, the seismic travel time residual after relocation is greatly reduced, the land P wave residual is reduced by 22.6%, and the S wave is reduced by 21.2%. The sea P wave residual is reduced by 25.7%, and the S wave is reduced by 15.6%. The new model is better for regional earthquake location.
We provide a more reliable V_P velocity model, which can be used to earthquake location, earthquake source parameter inversion and 3-D velocity model studies in the Pearl River Estuary.

Key words: Pearl River Estuary, onshore-offshore seismic experiment, minimum 1-D velocity model, station corrections

摘要： 珠江口地区位于南海北部大陆的边缘, 具有洋陆过渡型地壳特征, 且NE向滨海断裂带从其中穿过, 强震风险不可忽视。 文中基于2015年珠江口海陆联合三维人工地震探测数据, 人工进行初至P波震相拾取, 并使用VELEST程序分别反演了陆域和海域的最小一维P波速度模型(走时残差均方根最小)和台站校正结果。 台站校正结果的空间分布与区域地形、 地质构造和沉积厚度相关较好, 正值多分布在珠江三角洲沉积盆地和珠江口盆地内, 而负值多分布在花岗岩等基岩出露地区以及滨海断裂带北侧和北部断阶带内的部分隆起地区。 新模型对人工地震走时的拟合精度较高, 陆域走时残差均方根为0.07s, 海域为0.21s。 与华南模型相比, 新模型对区域地震定位的效果更好, 重定位后, 陆域的P波地震走时残差降低了22.6%、 S波降低了21.2%; 海域的P波地震走时残差降低了25.7%、 S波降低了15.6%。 新模型可为区域地震定位、 地震参数和三维成像研究提供参考。

关键词: 珠江口, 海陆地震联测, 最小一维速度模型, 台站校正