# The Well: Rayleigh-Taylor Instability Turbulence (3D)

## Facts

- Provider: Polymathic AI (The Well); Stefan Nixon and Stuart Dalziel (University of Cambridge), Romain Watteaux (CEA DAM)
- License: CC-BY-4.0 (https://arxiv.org/html/2412.00568)
- Size: 24 files
- Format: HDF5 (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/polymathic-ai/rayleigh_taylor_instability
- Access: public, readable without an account
- Category: Physics
- Source: https://arxiv.org/abs/2412.00568
- Page: https://scigantic.com/archive/f4ce6ef1-9ba7-5b76-afd9-68df245b7460
- Keywords: Physics, spectroscopy, HDF5, Hugging Face

## About

3D simulations of the Rayleigh-Taylor instability, where a heavy fluid above a light one mixes into turbulence, studying how the spectral shape and phase of the initial perturbation affect the transition to turbulence and its statistics. Solved with TurMix3D under the Boussinesq approximation on a unit cube, periodic on the side walls with slip at top and bottom; each trajectory has 60 time steps of 128 by 128 by 128 cubes. Fields are density and velocity. The datasheet lists 45 trajectories over five Atwood numbers (3/4, 1/2, 1/4, 1/8, 1/16) with log-normal initial spectra of varied mean, width and phase. Listing data/ (measured 2026-11-13) shows train/ (5 files, 180 GB, one per Atwood number), valid/ and test/ (5 files, 40 GB each); about 260 GB in the repository. Files are HDF5 in the common The Well format; open them with the_well package (WellDataModule with base path hf://datasets/polymathic-ai/) or directly with h5py. Part of The Well (Ohana et al., NeurIPS 2024 Datasets and Benchmarks, arXiv:2412.00568); released under CC-BY-4.0, cite The Well and the simulation's source papers.

## Load the data

```python
from huggingface_hub import hf_hub_download
path = hf_hub_download(repo_id="polymathic-ai/rayleigh_taylor_instability", filename="data/test/rayleigh_taylor_instability_At_0625.hdf5", repo_type="dataset")
print(path)
```

## Data structure

File types in sampled files: HDF5

Example file: `data/test/rayleigh_taylor_instability_At_0625.hdf5`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/f4ce6ef1-9ba7-5b76-afd9-68df245b7460
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/f4ce6ef1-9ba7-5b76-afd9-68df245b7460
- Over MCP: `get_schema_card` with id `f4ce6ef1-9ba7-5b76-afd9-68df245b7460` at https://api.scigantic.com/mcp

## Related topics

- [Spectroscopy datasets](https://scigantic.com/topics/spectroscopy)
- [HDF5 datasets](https://scigantic.com/topics/hdf5)
- [Physics datasets](https://scigantic.com/topics/physics)
