README.md

---
language:
- en
tags:
- inversion
- seismic
- imaging
- subsurface
size_categories:
- 10B<n<100B
viewer: false
---

<h1 align="center">
  <span style="font-family: 'Papyrus', sans-serif; color: red; font-weight: bold;">OpenSWI</span>:
  <span style="font-family: 'Papyrus', sans-serif; color:rgb(14, 126, 146);">A Massive-Scale Benchmark Dataset for </span> <br> 
  <span style="font-family: 'Papyrus', sans-serif; color:rgb(14, 126, 146);">Surface Wave Dispersion Curve Inversion</span>
</h1>

<h5 align="center"><a href="https://liufeng2317.github.io/">Feng Liu</a>, Sijie Zhao, Xinyu Gu, Fenghua Ling, Peiqin Zhuang, Yaxing Li*, Rui Su*, Lihua Fang, Jianping Huang, Lei Bai</h5>

![](./Figure/OpenSWI.png)

## 📖 **Overview**

**OpenSWI** is a comprehensive 1D dataset for surface-wave dispersion curve inversion, specifically designed for both shallow subsurface exploration (\~3 km) and deep geological studies (\~300 km). The dataset contains synthetic data derived from a variety of geological models, as well as real-world observational data, providing an invaluable resource for assessing and enhancing the generalization capabilities of deep learning models. The dataset includes:

* [**OpenSWI-Shallow**](https://huggingface.co/datasets/LiuFeng2317/OpenSWI/tree/main/Datasets/OpenSWI-shallow/0.2-10s-Aug): 1D velocity profiles derived from 2D velocity models (OpenFWI dataset), paired with corresponding surface wave dispersion curves.
* [**OpenSWI-Deep**](https://huggingface.co/datasets/LiuFeng2317/OpenSWI/tree/main/Datasets/OpenSWI-deep/1s-100s-Aug): 1D velocity profiles generated from high-resolution 3D geological models, sourced globally and regionally, tailored for deep geological studies.
* [**OpenSWI-Real**](https://huggingface.co/datasets/LiuFeng2317/OpenSWI/tree/main/Datasets/OpenSWI-real): AI-ready observational data from Long Beach, USA, and the China Seismological Reference Model Project, with 1D velocity profiles and corresponding surface wave dispersion curves.

These datasets are ideal for training and evaluating deep learning models focused on surface-wave dispersion curve inversion tasks.

---

## 📊 **OpenSWI Datasets**

<table>
<thead>
<tr>
  <th>Group</th>
  <th>Datasets</th>
  <th>Period Range (s)</th>
  <th>Depth Range and Interval (km)</th>
  <th>Extracted 1D Velocity</th>
  <th>Augmented 1D Velocity</th>
</tr>
</thead>
<tbody>

<tr>
  <td rowspan="5"><a href="https://huggingface.co/datasets/LiuFeng2317/OpenSWI/tree/main/Datasets/OpenSWI-shallow/0.2-10s-Aug">OpenSWI-shallow</a></td>
  <td>OpenFWI-FlatVelA</td>
  <td>0.1-10 s</td>
  <td>0-2.8 km / 0.04 km</td>
  <td>30,000 x 4 x 70</td>
  <td>1,490,415 x 4 x 70</td>
</tr>
<tr>
  <td>OpenFWI-Flat-FaultA</td>
  <td>0.1-10 s</td>
  <td>0-2.8 km / 0.04 km</td>
  <td>292,941 x 4 x 70</td>
  <td>2,925,151 x 4 x 70</td>
</tr>
<tr>
  <td>OpenFWI-CurveVel</td>
  <td>0.1-10 s</td>
  <td>0-2.8 km / 0.04 km</td>
  <td>295,773 x 4 x 70</td>
  <td>2,952,975 x 4 x 70</td>
</tr>
<tr>
  <td>OpenFWI-Fold-Fault</td>
  <td>0.1-10 s</td>
  <td>0-2.8 km / 0.04 km</td>
  <td>537,774 x 4 x 70</td>
  <td>5,369,692 x 4 x 70</td>
</tr>
<tr>
  <td>OpenFWI-StyleA</td>
  <td>0.1-10 s</td>
  <td>0-2.8 km / 0.04 km</td>
  <td>2,344,958 x 4 x 70</td>
  <td>9,345,103 x 4 x 70</td>
</tr>

<tr>
  <td rowspan="14"><a href="https://huggingface.co/datasets/LiuFeng2317/OpenSWI/tree/main/Datasets/OpenSWI-deep/1s-100s-Aug">OpenSWI-deep</a></td>
  <td>LITHO1.0</td>
  <td>1-100 s</td>
  <td>0-300 km / 1.0 km</td>
  <td>40,959 x 4 x 300</td>
  <td>245,771 x 4 x 301</td>
</tr>
<tr>
  <td>USTClitho1.0</td>
  <td>1-100 s</td>
  <td>0-300 km / 1.0 km</td>
  <td>9,125 x 4 x 300</td>
  <td>54,750 x 4 x 301</td>
</tr>
<tr>
  <td>Central-and-Western US</td>
  <td>1-100 s</td>
  <td>0-300 km / 1.0 km</td>
  <td>6,803 x 4 x 300</td>
  <td>40,818 x 4 x 301</td>
</tr>
<tr>
  <td>Continental China</td>
  <td>1-100 s</td>
  <td>0-300 km / 1.0 km</td>
  <td>4,516 x 4 x 300</td>
  <td>27,096 x 4 x 301</td>
</tr>
<tr>
  <td>US Upper-Mantle</td>
  <td>1-100 s</td>
  <td>0-300 km / 1.0 km</td>
  <td>3,678 x 4 x 300</td>
  <td>22,061 x 4 x 301</td>
</tr>
<tr>
  <td>EUcrust</td>
  <td>1-100 s</td>
  <td>0-300 km / 1.0 km</td>
  <td>43,520 x 4 x 300</td>
  <td>261,155 x 4 x 301</td>
</tr>
<tr>
  <td>Alaska</td>
  <td>1-100 s</td>
  <td>0-300 km / 1.0 km</td>
  <td>19,408 x 4 x 300</td>
  <td>116,448 x 4 x 301</td>
</tr>
<tr>
  <td>CSEM-Europe</td>
  <td>1-100 s</td>
  <td>0-300 km / 1.0 km</td>
  <td>21,931 x 4 x 300</td>
  <td>131,586 x 4 x 301</td>
</tr>
<tr>
  <td>CSEM-Eastmed</td>
  <td>1-100 s</td>
  <td>0-300 km / 1.0 km</td>
  <td>12,782 x 4 x 300</td>
  <td>76,692 x 4 x 301</td>
</tr>
<tr>
  <td>CSEM-Iberian</td>
  <td>1-100 s</td>
  <td>0-300 km / 1.0 km</td>
  <td>9,102 x 4 x 300</td>
  <td>54,612 x 4 x 301</td>
</tr>
<tr>
  <td>CSEM-South Atlantic</td>
  <td>1-100 s</td>
  <td>0-300 km / 1.0 km</td>
  <td>7,371 x 4 x 300</td>
  <td>44,226 x 4 x 301</td>
</tr>
<tr>
  <td>CSEM-North Atlantic</td>
  <td>1-100 s</td>
  <td>0-300 km / 1.0 km</td>
  <td>14,541 x 4 x 300</td>
  <td>87,246 x 4 x 301</td>
</tr>
<tr>
  <td>CSEM-Japan</td>
  <td>1-100 s</td>
  <td>0-300 km / 1.0 km</td>
  <td>14,641 x 4 x 300</td>
  <td>87,846 x 4 x 301</td>
</tr>
<tr>
  <td>CSEM-Astralasia</td>
  <td>1-100 s</td>
  <td>0-300 km / 1.0 km</td>
  <td>4,131 x 4 x 300</td>
  <td>24,786 x 4 x 301</td>
</tr>

<tr>
  <td rowspan="2"><a href="https://huggingface.co/datasets/LiuFeng2317/OpenSWI/tree/main/Datasets/OpenSWI-real/">OpenSWI-real</a></td>
  <td><a href="https://huggingface.co/datasets/LiuFeng2317/OpenSWI/tree/main/Datasets/OpenSWI-real/LongBeach">LongBeach</a></td>
  <td>0.263 - 1.666 s</td>
  <td>0-1.4 km / 0.04 km</td>
  <td>5,297 x 4 x 35</td>
  <td>-</td>
</tr>
<tr>
  <td><a href="https://huggingface.co/datasets/LiuFeng2317/OpenSWI/tree/main/Datasets/OpenSWI-real/CSRM">CSRM</a></td>
  <td>8 - 70 s</td>
  <td>0-120 km / 1.0 km</td>
  <td>12,901 x 4 x 120</td>
  <td>-</td>
</tr>

</tbody>
</table>

> Details of the source of datasets can be found at [OpenSWI](./Datasets/README.md)

### 🏞️ **OpenSWI-Shallow**

* **Resource Models**: Includes diverse geological features such as flat layers, faults, folds, and their combinations, representing typical shallow subsurface structures.
* **Profiles**:

  * **~22 million 1D velocity profiles**, each paired with corresponding **Rayleigh wave dispersion curves**.
  * The dataset spans a **period range from 0.2 to 10 seconds** and covers **100 sampling points** (including uniform, random, and logarithmic distributions) for each dispersion curve. This variety ensures robust training and evaluation across different geological scenarios.
* **Geological Diversity**: The models include a broad spectrum of real-world shallow subsurface structures, such as:

  * **Flat Layers**
  * **Faulted Layers (Flat-Fault)**
  * **Folds**
  * **Folds with Faults (Fold-Fault)**
  * **Real Style (Field)**

These diverse models make the dataset highly applicable for both synthetic and real-world seismic data inversion tasks.

---

### 🌍 **OpenSWI-Deep**

* **Resource Models**: Includes over **14 global and regional 3D geological models**, derived from high-resolution seismic data. These models represent deep geological structures, from the crust to the mantle, spanning a variety of tectonic settings and geological environments.
* **Profiles**:

  * **~1.26 million 1D velocity profiles**, derived from the 3D models.
  * The profiles span a **period range from 1 to 100 seconds**, covering **300 sampling points** (including uniform, random, and logarithmic distributions) for each dispersion curve.
  * These profiles offer high-resolution data suitable for deep geological studies and support advanced seismic inversion techniques.

* **Geological Diversity**: The 3D models come from various sources, including well-established models such as:
  * **LITHO1.0** (Pasyanos et al., 2014)
  * **USTClitho1.0** (Xin et al., 2018)
  * **Central and Western US Models** (Shen et al., 2013)
  * **Continental China** (Shen et al., 2016)
  * **US Upper-Mantle Model** (Xie et al., 2018)
  * **EUcrust Model** (Lu et al., 2018)
  * **Alaska Model** (Berg et al., 2020)
  * **CSEM-Europe Model** (Blom et al., 2020; Fichtner et al., 2018; Çubuk-Sabuncu et al., 2017)
  * **CSEM Eastern Mediterranean Model** (Blom et al., 2020; Fichtner et al., 2018)
  * **CSEM Western Mediterranean Model** (Fichtner et al., 2018; Fichtner et al., 2015)
  * **CSEM South Atlantic Model** (Fichtner et al., 2018; Colli et al., 2013)
  * **CSEM North Atlantic Model** (Fichtner et al., 2018; Krischer et al., 2018)
  * **CSEM Japanese Island Model** (Fichtner et al., 2018； Simutė et al., 2016)
  * **CSEM Australasian Model** (Fichtner et al., 2018; Fichtner et al., 2010)

These models provide a comprehensive representation of both regional and global deep geological structures, enhancing the dataset’s value for training deep learning models on complex inversion tasks.

---

### 🏔️ **OpenSWI-Real**

* **Long Beach, USA (Fu et al., 2022)**:

  * Contains **5,297 stations**, each with phase velocity dispersion curves.
  * The **period range** for the curves is **0.263 to 1.666 seconds**, focusing on shallow subsurface structures.
  * This dataset provides real-world observational data for evaluating model performance and generalization in seismic inversion tasks.

* **China Seismological Reference Model (Xiao et al., 2024)**:

  * Includes **12,901 grid points**, each with phase and group velocity dispersion curves.
  * The **period range** for these curves is **8 to 70 seconds**, ideal for studying deeper geological structures.
  * Data is sourced from a dense network of seismic stations across mainland China, offering comprehensive coverage for advanced inversion tasks.

----

## 📧 **Contact**

**Principal Developer**

[Feng Liu](https://liufeng2317.github.io/)  

🏛️ Affiliations:  
- Shanghai Artificial Intelligence Laboratory
- Shanghai Jiao Tong University

📧 **Contact Information**:  
- [![Email](https://img.shields.io/badge/[email protected]?style=flat&logo=gmail)](mailto:[email protected])  
- [![Alt Email](https://img.shields.io/badge/[email protected]?style=flat&logo=gmail)](mailto:[email protected])


----

## 📚 **Citation**

If you use this dataset in your research, please cite:

```bibtex
@software{liufeng_2025_openswi,
  title={OpenSWI: A Massive-Scale Benchmark Dataset for Surface Wave Dispersion Curve Inversion},
  author={Feng Liu, Sijie Zhao, Xinyu Gu, Fenghua Ling, Peiqin Zhuang, Yaxing Li, Rui Su, Lihua Fang, Jianping Huang, Lei Bai},
  year={2025},
}
```

----


## 📝 **License**

Released under the **CC BY 4.0 License**. See the full license in the repository.


---
## 🔍 **Reference**
<details>
<summary>Click to expand related works</summary>

1. **X. Xiao et al.**, "CSRM‐1.0: A China Seismological Reference Model," *JGR Solid Earth*, vol. 129, no. 9, p. e2024JB029520, Sept. 2024, [doi: 10.1029/2024JB029520](https://doi.org/10.1029/2024JB029520).
   
2. **L. Fu et al.**, "Improved high‐resolution 3D vs model of Long Beach, CA: Inversion of multimodal dispersion curves from ambient noise of a dense array," *Geophys. Res. Lett.*, vol. 49, no. 4, p. e2021GL097619, Feb. 2022, [doi: 10.1029/2021GL097619](https://doi.org/10.1029/2021GL097619).
  
3. **C. Deng et al.**, “OpenFWI: Large-scale multi-structural benchmark datasets for full waveform inversion,” *Neural Information Processing Systems*, Nov. 2021, Accessed: Feb. 29, 2024. [Online]. Available: [https://www.semanticscholar.org/paper/2d13799dcbd0aefb08c379f58cd6004b1376ca33](https://www.semanticscholar.org/paper/2d13799dcbd0aefb08c379f58cd6004b1376ca33)

4. **N. Blom et al.**, "Seismic waveform tomography of the central and eastern Mediterranean upper mantle," *Solid Earth*, vol. 11, no. 2, pp. 669–690, Apr. 2020, [doi: 10.5194/se-11-669-2020](https://doi.org/10.5194/se-11-669-2020).
   
5. **E. M. Berg et al.**, "Shear velocity model of Alaska via joint inversion of Rayleigh wave ellipticity, phase velocities, and receiver functions across the Alaska transportable array," *J. Geophys. Res.: Solid Earth*, vol. 125, no. 2, Feb. 2020, [doi: 10.1029/2019jb018582](https://doi.org/10.1029/2019jb018582).

6. **H. Xin et al.**, "High‐resolution lithospheric velocity structure of continental China by double‐difference seismic travel‐time tomography," *Seismol. Res. Lett.*, vol. 90, no. 1, pp. 229–241, Jan. 2019, [doi: 10.1785/0220180209](https://doi.org/10.1785/0220180209).

7. **J. Xie et al.**, "3-D upper-mantle shear velocity model beneath the contiguous United States based on broadband surface wave from ambient seismic noise," *Pure Appl. Geophys.*, vol. 175, no. 10, pp. 3403–3418, Oct. 2018, [doi: 10.1007/s00024-018-1881-2](https://doi.org/10.1007/s00024-018-1881-2).

8. **Y. Lu et al.**, "High-resolution surface wave tomography of the European crust and uppermost mantle from ambient seismic noise," *Geophys. J. Int.*, vol. 214, no. 2, pp. 1136–1150, Aug. 2018, [doi: 10.1093/gji/ggy188](https://doi.org/10.1093/gji/ggy188).

9. **L. Krischer et al.**, "Automated large‐scale full seismic waveform inversion for North America and the North Atlantic," *J. Geophys. Res.: Solid Earth*, vol. 123, no. 7, pp. 5902–5928, July 2018, [doi: 10.1029/2017JB015289](https://doi.org/10.1029/2017JB015289).

10. **A. Fichtner et al.**, "The collaborative seismic earth model: generation 1," *Geophys. Res. Lett.*, vol. 45, no. 9, pp. 4007–4016, May 2018, [doi: 10.1029/2018gl077338](https://doi.org/10.1029/2018gl077338).

11. **Y. Çubuk-Sabuncu et al.**, "3-D crustal velocity structure of western Turkey: constraints from full-waveform tomography," *Phys. Earth Planet. Inter.*, vol. 270, pp. 90–112, Sept. 2017, [doi: 10.1016/j.pepi.2017.06.014](https://doi.org/10.1016/j.pepi.2017.06.014).

12. **S. Simutė et al.**, "Full‐waveform inversion of the Japanese islands region," *J. Geophys. Res.: Solid Earth*, vol. 121, no. 5, pp. 3722–3741, May 2016, [doi: 10.1002/2016jb012802](https://doi.org/10.1002/2016jb012802).

13. **W. Shen et al.**, "A seismic reference model for the crust and uppermost mantle beneath China from surface wave dispersion," *Geophys. J. Int.*, vol. 206, no. 2, pp. 954–979, Aug. 2016, [doi: 10.1093/gji/ggw175](https://doi.org/10.1093/gji/ggw175).

14. **A. Fichtner and A. Villaseñor**, "Crust and upper mantle of the western Mediterranean – constraints from full-waveform inversion," *Earth Planet. Sci. Lett.*, vol. 428, pp. 52–62, Oct. 2015, [doi: 10.1016/j.epsl.2015.07.038](https://doi.org/10.1016/j.epsl.2015.07.038).

15. **M. E. Pasyanos et al.**, "LITHO1.0: An updated crust and lithospheric model of the Earth," *JGR Solid Earth*, vol. 119, no. 3, pp. 2153–2173, Mar. 2014, [doi: 10.1002/2013JB010626](https://doi.org/10.1002/2013JB010626).

16. **W. Shen et al.**, "A 3‐D model of the crust and uppermost mantle beneath the Central and Western US by joint inversion of receiver functions and surface wave dispersion," *JGR Solid Earth*, vol. 118, no. 1, pp. 262–276, Jan. 2013, [doi: 10.1029/2012JB009602](https://doi.org/10.1029/2012JB009602).

17. **F. Rickers et al.**, "The Iceland–Jan Mayen plume system and its impact on mantle dynamics in the North Atlantic region: evidence from full-waveform inversion," *Earth Planet. Sci. Lett.*, vol. 367, pp. 39–51, Apr. 2013, [doi: 10.1016/j.epsl.2013.02.022](https://doi.org/10.1016/j.epsl.2013.02.022).

18. **L. Colli et al.**, "Full waveform tomography of the upper mantle in the South Atlantic region: imaging a westward fluxing shallow asthenosphere?," *Tectonophysics*, vol. 604, pp. 26–40, Sept. 2013, [doi: 10.1016/j.tecto.2013.06.015](https://doi.org/10.1016/j.tecto.2013.06.015).

19. **C. Trabant et al.**, "Data products at the IRIS DMC: stepping stones for research and other applications," *Seismol. Res. Lett.*, vol. 83, no. 5, pp. 846–854, Sept. 2012, [doi: 10.1785/0220120032](https://doi.org/10.1785/0220120032).

20. **A. Fichtner et al.**, "Full waveform tomography for radially anisotropic structure: new insights into present and past states of the Australasian upper mantle," *Earth Planet. Sci. Lett.*, vol. 290, no. 3–4, pp. 270–280, Feb. 2010, [doi: 10.1016/j.epsl.2009.12.003](https://doi.org/10.1016/j.epsl.2009.12.003).

21. **A. Fichtner et al.**, "Full seismic waveform tomography for upper-mantle structure in the Australasian region using adjoint methods," *Geophys. J. Int.*, vol. 179, no. 3, pp. 1703–1725, Dec. 2009, [doi: 10.1111/j.1365-246x.2009.04368.x](https://doi.org/10.1111/j.1365-246x.2009.04368.x).


</details>