SuryaBench
Collection
Benchmark Dataset for Advancing Machine Learning in Heliophysics and Space Weather Prediction
•
7 items
•
Updated
Datasets:
Dataset Viewer (First 5GB)
The dataset viewer is not available for this split.
Server error while post-processing the split rows. Please report the issue.
Error code: RowsPostProcessingError
Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
Surya-Bench Coronal Field Extrapolation Task
This foundation model task uses the foundation model to emulate the physics-based ADAPT-WSA PFSS code, driven by SDO/HMI observations. The parameters to predict are spherical harmonic coefficients which represent the magnetic potential for a domain between the photosphere and a source surface (set to 2.51 Rs). Included in these FITS files are other variables relating the WSA output that might be useful, such as photospheric polarity and open/closed field line masks. The ADAPT-WSA code is an ensemble model. Data is included here for all 12 realizations (identified by R000 - R011 in the file names).
Spherical harmonics can be found in fits HDU index 3 (zero-based), as a (2, 91, 91) array. Each (91, 91) array is a triangular matrix with unused elements left as zero.
See also the associated code on GitHub, which includes a pytorch data loader for use with the foundation model.
Dataset Details
| Field | Description |
|---|---|
| Temporal Coverage | 2010-05-01 to 2021-12-31 |
| File Format | FITS |
| Data Shape | (2, 91, 91) |
| Number of Time Steps | 51156 |
| Cadence | daily |
| Spherical Harmonic Order | 90 |
| Spherical Harmonic Normalization | Schmidt |
| Total Size | 37 GB |
| Size per File | 746 KB |
Authors
Daniel da Silva, [email protected]
- Downloads last month
- 31
Size of the auto-converted Parquet files (First 5GB per split):
8.51 GB
Number of rows (First 5GB per split):
13,200
Estimated number of rows:
51,147