{"success":true,"data":{"author":"Polymathic AI (The Well); Drummond Fielding, Flatiron Institute CCA and Cornell University","bucketAccessType":"public","bucketName":"polymathic-ai/turbulent_radiative_layer_2D","bucketPrefix":"","category":"Physics","cloudProvider":"huggingface","createdAt":"2026-11-14T01:01:19.805Z","description":"2D simulations of a turbulent radiative mixing layer: cold dense gas below hot dilute gas, moving past each other at subsonic speed, unstable to the Kelvin-Helmholtz instability. Mixing creates intermediate-temperature gas that cools rapidly, a process found throughout astrophysical systems such as galactic winds and the circumgalactic medium. Computed with Athena++ on a 384 by 128 grid (x in [-0.5, 0.5], y in [-1, 2]), periodic in x and zero-gradient in y; each trajectory has 101 time steps. Fields are density, pressure and velocity. The datasheet lists 90 trajectories: 10 noise seeds for each of 9 cooling times t_cool from 0.03 to 3.16. Listing data/ (measured 2026-11-13) shows train/ (9 files, 5.9 GB, one per cooling time), valid/ and test/ (9 files, about 1 GB each); about 7.9 GB in the repository. A 3D companion set exists. 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.","displayTitle":"The Well: Turbulent Radiative Mixing Layer (2D)","id":"acd2cd0e-2940-59a1-9dda-d43e9e6dd44a","license":"CC-BY-4.0","licenseEvidenceUrl":"https://arxiv.org/html/2412.00568","objectCountEstimate":34,"probeStatus":"ok","referenceUrl":"https://arxiv.org/abs/2412.00568","region":"us","schemaCard":{"accessHints":[{"language":"python-datasets","snippet":"from datasets import load_dataset\nds = load_dataset(\"polymathic-ai/turbulent_radiative_layer_2D\")"},{"language":"python-hub","snippet":"from huggingface_hub import snapshot_download\nsnapshot_download(repo_id=\"polymathic-ai/turbulent_radiative_layer_2D\", repo_type=\"dataset\")"}],"caveats":["HuggingFace metadata: author=polymathic-ai, license:cc-by-4.0, downloads=7524"],"format":"mixed","generatedAt":"2026-11-14T01:01:24.780Z","layout":{"example":"data/test/turbulent_radiative_layer_tcool_0.03.hdf5","pattern":"{repoId}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"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","encoding":"hex","key":"data/test/turbulent_radiative_layer_tcool_0.03.hdf5","truncatedAt":512}],"sampleObjects":["data/test/turbulent_radiative_layer_tcool_0.03.hdf5","data/test/turbulent_radiative_layer_tcool_0.06.hdf5","data/test/turbulent_radiative_layer_tcool_0.10.hdf5","data/test/turbulent_radiative_layer_tcool_0.18.hdf5","data/test/turbulent_radiative_layer_tcool_0.32.hdf5"],"sidecars":[{"bytes":"---\nlanguage:\n- en\nlicense: cc-by-4.0\ntags:\n- physics\n\ntask_categories:\n- time-series-forecasting\n- other\ntask_ids:\n- multivariate-time-series-forecasting\n---\n\n# How To Load from HuggingFace Hub\n\n1. Be sure to have `the_well` installed (`pip install the_well`)\n2. Use the `WellDataModule` to retrieve data as follows:\n\n```python\nfrom the_well.data import WellDataModule\n\n# The following line may take a couple of minutes to instantiate the datamodule\ndatamodule = WellDataModule(\n    \"hf://datasets/polymathic-ai/\",\n    \"turbulent_radiative_layer_2D\",\n)\ntrain_dataloader = datamodule.train_dataloader()\n\nfor batch in dataloader:\n    # Process training batch\n    ...\n```\n\n# Turbulent Radiative Layer - 2D\n\n**One line description of the data:** Everywhere in astrophysical systems hot gas moves relative to cold gas, which leads to mixing, and mixing populates intermediate temperature gas that is highly reactive—in this case it is rapidly cooling.\n\n**Longer description of the data:** In this simulation, there is cold, dense gas on the bottom and hot dilute gas on the top. They are moving relative to each other at highly subsonic velocities. This set up is unstable to the Kelvin Helmholtz instability, which is seeded with small scale noise that is varied between the simulations. The hot gas and cold gas are both in thermal equilibrium in the sense that the heating and cooling are exactly balanced. However, once mixing occurs as a result of the turbulence induced by the Kelvin Helmholtz instability the intermediate temperatures become populated. This intermediate temperature gas is not in thermal equilibrium and cooling beats heating. This leads to a net mass flux from the hot phase to the cold phase. This process occurs in the interstellar medium, and in the Circum-Galactic medium when cold clouds move through the ambient, hot medium. By understanding how the total cooling and mass transfer scale with the cooling rate we are able to constrain how this process controls the overall phase structure, energetics and dynamics of the gas in and around galaxies.\n\n**Associated paper**: [Paper](https://iopscience.iop.org/article/10.3847/2041-8213/ab8d2c/pdf).\n\n**Domain expert**: [Drummond Fielding](https://dfielding14.github.io/), CCA, Flatiron Institute & Cornell University.\n\n**Code or software used to generate the data**: [Athena++](https://www.athena-astro.app/).\n\n**Equation**:\n\n$$\n\\begin{align*}\n\\frac{ \\partial \\rho}{\\partial t} + \\nabla \\cdot \\left( \\rho \\vec{v} \\right) &= 0 \\\\\n\\frac{ \\partial \\rho \\vec{v} }{\\partial t} + \\nabla \\cdot \\left( \\rho \\vec{v}\\vec{v} + P \\right) &= 0 \\\\\n\\frac{ \\partial E }{\\partial t} + \\nabla \\cdot \\left( (E + P) \\vec{v} \\right) &= - \\frac{E}{t_{\\rm cool}} \\\\\nE = P / (\\gamma -1) \\, \\, \\gamma &= 5/3\n\\end{align*}\n$$\n\nwith \\\\(\\rho\\\\) the density, \\\\(\\vec{v}\\\\) the 2D velocity, \\\\(P\\\\) the pressure, \\\\(E\\\\) the total energy, and \\\\(t_{\\rm cool}\\\\) the cooling time.\n\n![Gif](https://users.flatironinstitute.org/~polymathic/data/the_well/datasets/turbulent_radiative_layer_2D/gif/density_normalized.gif)\n\n| Dataset    | FNO | TFNO  | Unet | CNextU-net\n|:-:|:-:|:-:|:-:|:-:|\n| `turbulent_radiative_layer_2D`  | 0.5001| 0.5016 |0.2418| \\\\(\\mathbf{0.1956}\\\\)|\n\nTable: VRMSE metrics on test sets (lower is better). Best results are shown in bold. VRMSE is scaled such that predicting the mean value of the target field results in a score of 1.\n\n\n## About the data\n\n**Dimension of discretized data:** 101 timesteps of 384x128 images.\n\n**Fields available in the data:** Density (scalar field), pressure (scalar field), velocity (vector field).\n\n**Number of trajectories:** 90 (10 different seeds for each of the 9 \\\\(t_{cool}\\\\) values).\n\n**Estimated size of the ensemble of all simulations:** 6.9 GB.\n\n**Grid type:** uniform, cartesian coordinates.\n\n**Initial conditions:** Analytic, described in the [paper](https://ui.adsabs.harvard.edu/abs/2020ApJ...894L..24F/abstract).\n\n**Boundary conditions:** Periodic in the x-direction, zero-gradient for the y-direction.\n\n**Simulation time-step ( \\\\(\\Delta t\\\\)):** varies with \\\\(t_{cool}\\\\). Smalles","encoding":"utf8","key":"README.md","name":"README.md","size":5804,"truncated":true}],"sourceObjectKey":"data/test/turbulent_radiative_layer_tcool_0.03.hdf5","starterCell":"from huggingface_hub import hf_hub_download\npath = hf_hub_download(repo_id=\"polymathic-ai/turbulent_radiative_layer_2D\", filename=\"data/test/turbulent_radiative_layer_tcool_0.03.hdf5\", repo_type=\"dataset\")\nprint(path)","starterSamplePath":"data/test/turbulent_radiative_layer_tcool_0.03.hdf5","totalObjects":34},"summary":"2D simulations of a turbulent radiative mixing layer: cold dense gas below hot dilute gas, moving past each other at subsonic speed, unstable to the Kelvin-Helmholtz instability.","title":"The Well: Turbulent Radiative Mixing Layer (2D)","updatedAt":"2026-11-14T02:36:57.949Z","visibility":"public","websiteUrl":"https://polymathic-ai.org/the_well/"}}