Reflexionsseismiska mätningar över Burträsk-förkastningen: Migrated cross-dip stack

In 2008, a 22 km long high resolution reflection seismic profile was acquired over the Burträsk fault in Northern Sweden. The Burträsk fault formed at the end of the last glacial, triggered by the rapid melting of the Scandinavian ice sheet and the following crustal rebound. The aim of the survey was to image the structure of the fault with depth. The profile followed existing roads resulting in a very crooked acquisition geometry. A list of the key acquisition parameters is given below. The dataset was first processed and published in 2011. In 2018, the dataset was reprocessed using an additional cross-dip correction step to improve the imaging of reflection with a dip component in the cross-profile direction. The following data files are available for downloading: 1) raw shot gathers (decoded and quality controlled, with geometry information in the headers) 2) original stacked section from Juhlin and Lund (2011) 3) cross-dip corrected stacked section from Beckel and Juhlin (2019) 4) cross-dip corrected migrated section from Beckel and Juhlin (2019) A short summary of the main processing steps for each dataset is given below. For a complete description, please refer to the above mentioned publications. Key acquisition parameters: Number of channels: 280 Near offset: 0 m Geophone spacing: 20 m Geophone type : 28 Hz single Nominal source spacing : 20 m Source type : VIBSIST Hit interval between hammer blows : 100-200 ms Sweeps per source point :3-4 Nominal fold : 140 Recording instrument : SERCEL 408L Sample rate : 1ms Field low cut : Out Field high cut : 400 Hz Record length : 23 s Profile length : 22 km Source points : 799 Dates acquired: 2008/08/05-2008/08/16 Short summary of the main processing steps: 1) BUR_shots_geom.sgy The raw shot records are pre-processed with VIBSIST decoding, quality control. Information about the shot and receiver geometry has been added to the trace headers. 2) BUR_original_stack.sgy The processing flow of the original data stack includes trace editing and balancing, spectral equalization, time-variant bandpass filtering, refraction and residual statics, horizontal median filtering, NMO correction, DMO correction, stacking and coherency filtering (for a complete description see Juhlin & Lund, 2011) 3) BUR_crossdip_stack.sgy The main processing steps of the crossdip stack are very similar to those of the original stack, including trace editing and balancing, spectral equalization, time-variant bandpass filtering, refraction and residual statics, horizontal median filtering, NMO correction, DMO correction, stacking and coherency filtering. Before DMO correction, an additional cross-dip correction step has been added. For a complete description see Beckel & Juhlin, 2019. 4) BUR_crossdip_mig.sgy The processing flow of the migrated section includes trace editing and balancing, spectral equalization, time-variant bandpass filtering, refraction and residual statics, horizontal median filtering, NMO correction, cross-dip correction, DMO correction, stacking, coherency filtering and migration (for a complete description see Beckel & Juhlin, 2019). This dataset contains the migrated and depth converted crossdip stack of the Burträsk profile (Beckel & Juhlin, 2019). Processing steps for this data set: 1. Read decoded VIBSIST data 2. Bulk static shift to zero time 3. Apply geometry 4. Manual first-break picking 5. Trace balance: 0–3000 ms 6. Ground-roll and first-break muting in local time– frequency domain: 25 % threshold 7. Spectral equalization: 30 Hz window, 25–40–120–150 Hz bandpass 8. Time-variant bandpass filtering: 0–200 ms: 35–60–120–180 Hz 250–500 ms: 30–50–120–180 Hz 600–900 ms: 25–40–110–165 Hz 1100–3000 ms: 20–35–100–150 Hz 9. Refraction statics: floating datum, replacement velocity from model 10. Trace editing 11. Horizontal median filter: 11 traces, 5300 m s-1 and 3000 m s-1 12. Butterworth filter: 20–40–90–120 Hz 13. Spherical divergence correction: 0.8 tpower, 2.0 vpower 14. Velocity analysis 15. NMO correction: 40 % stretch mute 16. Residual statics 17. Cross-dip correction: 5400 m s-1 , 20 % taper; Velocity analysis 18. DMO correction; Velocity analysis 19. Stacking 20. FX deconvolution: 19 trace window 21. Trace balance 22. Stolt migration: 5400 m s-1 , 0.6 stretch factor 23. Zero mute 24. Approximate depth conversion: 5400 m s-1 Processed seismic data stored as one file for the entire seismic profile, according to SEG technical standard SEG-Y revision 1 (SEG-Y_r1.0, 2002); https://seg.org/Publications/SEG-Technical-Standards2008年，科研团队在瑞典北部的布尔特拉斯克断层（Burträsk fault）上方采集了一条总长22千米的高分辨率反射地震剖面（reflection seismic profile）。布尔特拉斯克断层形成于末次冰期结束之际，由斯堪的纳维亚冰盖（Scandinavian ice sheet）快速消融及随后的地壳反弹所触发。本次勘探的核心目标是获取该断层的深部结构成像。由于剖面沿现有道路布设，采集几何（acquisition geometry）形态存在显著弯曲。下文列出了本次勘探的关键采集参数。 该数据集最早于2011年完成处理并发表；2018年，研究团队通过新增跨倾角校正（cross-dip correction）步骤对数据集进行了二次处理，以提升跨剖面方向倾角分量反射波的成像质量。 可供下载的数据文件包括： 1) 原始炮集数据（raw shot gathers，已解码并经过质量控制，道头中包含采集几何信息）； 2) Juhlin与Lund（2011）发表的原始叠加剖面（stacked section）； 3) Beckel与Juhlin（2019）发表的跨倾角校正叠加剖面； 4) Beckel与Juhlin（2019）发表的跨倾角校正偏移剖面（migrated section）。 下文将简要说明各数据集的主要处理步骤，完整技术细节请参阅上述参考文献。 ### 关键采集参数 - 道数：280道 - 近偏移距：0米 - 检波器（geophone）间距：20米 - 检波器类型：28 Hz单分量检波器 - 标称震源间距：20米 - 震源类型：VIBSIST（可控震源系统） - 锤击间隔：100-200毫秒 - 单震源点扫频次数：3-4次 - 标称覆盖次数：140次 - 记录仪器：SERCEL 408L - 采样率：1毫秒 - 野外低截止滤波：未启用 - 野外高截止滤波：400 Hz - 记录时长：23秒 - 剖面总长：22千米 - 震源点数量：799个 - 采集日期：2008/08/05-2008/08/16 ### 各数据集主要处理步骤简要说明 1. **BUR_shots_geom.sgy**：原始炮记录首先通过VIBSIST解码完成预处理与质量控制，随后将震源与检波器的几何信息写入道头文件。 2. **BUR_original_stack.sgy**：原始叠加数据的处理流程包括道编辑与道平衡、频谱均衡、时变带通滤波、折射静校正与剩余静校正、水平中值滤波、正常时差校正（Normal Moveout, NMO）、倾角时差校正（Dip Moveout, DMO）、叠加以及相干滤波（完整处理细节请参阅Juhlin & Lund, 2011）。 3. **BUR_crossdip_stack.sgy**：跨倾角校正叠加剖面的主要处理流程与原始叠加剖面基本一致，同样包含道编辑与道平衡、频谱均衡、时变带通滤波、折射静校正与剩余静校正、水平中值滤波、NMO校正、DMO校正、叠加以及相干滤波，仅在DMO校正前新增了跨倾角校正步骤（完整处理细节请参阅Beckel & Juhlin, 2019）。 4. **BUR_crossdip_mig.sgy**：偏移剖面的处理流程包括道编辑与道平衡、频谱均衡、时变带通滤波、折射静校正与剩余静校正、水平中值滤波、NMO校正、跨倾角校正、DMO校正、叠加、相干滤波以及偏移成像（完整处理细节请参阅Beckel & Juhlin, 2019）。该数据集包含布尔特拉斯克剖面经跨倾角校正、偏移成像及近似深度转换后的结果（Beckel & Juhlin, 2019）。 ### 该数据集详细处理步骤 1. 读取已解码的VIBSIST原始数据 2. 整体静校正移位至零时刻 3. 加载采集几何信息 4. 手动拾取初至波 5. 0–3000 ms时间段内的道平衡处理 6. 在局域时频域内压制面波与初至波：采用25%阈值 7. 频谱均衡：使用30 Hz滑动窗口，带通范围为25–40–120–150 Hz 8. 时变带通滤波： - 0–200 ms：35–60–120–180 Hz - 250–500 ms：30–50–120–180 Hz - 600–900 ms：25–40–110–165 Hz - 1100–3000 ms：20–35–100–150 Hz 9. 折射静校正：采用浮动基准面，替换速度由速度模型提供 10. 道编辑处理 11. 水平中值滤波：采用11道滑动窗口，速度阈值为5300 m·s⁻¹与3000 m·s⁻¹ 12. 巴特沃斯滤波（Butterworth filter）：20–40–90–120 Hz 13. 球面发散校正：t幂次0.8，v幂次2.0 14. 速度分析 15. NMO校正：拉伸压制阈值为40% 16. 剩余静校正 17. 跨倾角校正：采用5400 m·s⁻¹速度，20%锥度截断；随后再次进行速度分析 18. DMO校正；随后再次进行速度分析 19. 叠加处理 20. FX反褶积（deconvolution）：采用19道滑动窗口 21. 道平衡处理 22. Stolt偏移成像（Stolt migration）：采用5400 m·s⁻¹速度，拉伸因子为0.6 23. 零时刻压制处理 24. 近似深度转换：采用5400 m·s⁻¹速度 所有处理后的地震数据均按照SEG技术标准SEG-Y格式（SEG-Y）修订版1（SEG-Y_r1.0, 2002）存储为单文件，覆盖整条地震剖面；相关标准可参阅：https://seg.org/Publications/SEG-Technical-Standards