astro-legacy-archive/planck-frequency-maps-pr4

Planck PR4 NPIPE Frequency Maps

This dataset contains the nine Planck Public Release 4 frequency maps produced by the NPIPE joint LFI/HFI processing pipeline. It serves the full-mission, full-channel source tables exactly as released: source column names, units, fixed-vector shapes, row order, dtypes, and value bits are preserved. Metadata makes the source representation self-describing without flattening, renaming, or normalising it.

The included maps are the R4.00 products at 30, 44, 70, 100, 143, 217, 353, 545, and 857 GHz. Field-only extracts, detector and detector-set maps, ring-half and year splits, low-resolution products, component maps, and simulations are outside this dataset.

Release	Planck PR4 / NPIPE, R4.00
Frequencies	30, 44, 70, 100, 143, 217, 353, 545, 857 GHz
Pixelisation	HEALPix; LFI NSIDE 1024, HFI NSIDE 2048; NESTED
Coordinate card	COORDSYS='G'
Polarisation convention	POLCCONV='COSMO'
Processing version	PROCVER='npipe6v20'
Signal unit, all bands	K_CMB

How to use

python -m venv .venv && .venv/bin/pip install datasets huggingface_hub pyarrow

from datasets import load_dataset
dataset = load_dataset("astro-legacy-archive/planck-frequency-maps-pr4", "HFI_SkyMap_545_2048_R4.00_full", split="train[:1]", columns=["I_STOKES"])
row = dataset[0]
print(row["I_STOKES"][:3])

[1.1965783834457397, 1.2781219482421875, 1.2653427124023438]

import json
from huggingface_hub import hf_hub_download
import pyarrow.parquet as pq
path = hf_hub_download("astro-legacy-archive/planck-frequency-maps-pr4", "545-ghz.parquet", repo_type="dataset")
print(json.loads(pq.read_schema(path).metadata[b"healpix_map"])["schema"])

astro-legacy-archive/healpix-map

The embedded document records the 545 GHz pixelisation, ordering, coordinate-card value, unit, coverage, and vector pixel_rule. Consumers apply that rule to the source’s 1,024-element cells; other configurations retain their source-defined field spellings.

Dataset structure

Each LFI config retains 12,288 source rows and each HFI config retains 49,152. Every column is a fixed-size list of 1,024 source values. The embedded pixel_rule states the mapping: element j of row i is HEALPix pixel i * 1024 + j in the ordering named by ORDERING. Thus the maps contain 12,582,912 LFI pixels or 50,331,648 HFI pixels without recasting the source table into one row per pixel.

The 30--353 GHz configs contain Stokes I, Q, and U, HITS, and all six elements of the per-pixel IQU covariance matrix. The 545 and 857 GHz configs contain I_STOKES, HITS, and II_COV only. All field names are uppercase exactly as supplied.

Every PR4 full-channel file, including 545 and 857 GHz, labels its sky signal K_CMB and its variance (K_CMB)^2; these must not be replaced with the PR3 high-frequency MJy/sr units. HITS has the source unit NONE. The source table value COORDSYS='G' and its keyword basis are embedded, as are NESTED, NSIDE, POLCCONV='COSMO', and each source TUNIT value. Structured physics records the nominal channel frequency and effective-beam FWHM with Planck release-paper bases.

There are no nulls, NaNs, infinities, exact zeros, or occurrences of the header's BAD_DATA value in any source array. Every HITS value is a positive integer. Coverage therefore contains no asserted source codes and reports zero matching pixels.

Unlike preceding Planck releases, NPIPE retains the CMB Solar dipole in its frequency maps for component separation and high-precision relative calibration. Its calibration at 44--217 GHz acts as a constrained filter that suppresses large-scale CMB E-mode power. The NPIPE paper measures this effect with simulations and supplies the E-mode transfer functions; quantitative polarization work below about ell=20 must use those transfer functions and matching simulations. The paper finds no such suppression at the foreground-calibrated 30 and 353 GHz channels.

Provenance

The files are the PR4/NPIPE products indexed by the Planck Legacy Archive and served from the IRSA Planck map directory. IRSA exposes PR4 beneath a release_3 path; product identity instead comes from the explicit NPIPE listing, the R4.00 filenames, and the source headers.

Source URLs, exact byte counts and SHA-256 hashes, original headers, retrieval date, config-to-file mapping, and regeneration command are embedded in each Parquet. Every selected binary table reconstructs in source column and row order with identical TFORM, shape, dtype, and value bits. FITS container serialisation is outside that promise.

The canonical 545 GHz FITS file is 603,988,416 bytes, 2,304 bytes short of the next FITS block boundary. Its declared binary-table data are complete, and all three arrays reconstruct bit-exactly; the source is preserved as delivered rather than padded.

Licensing and credit

No PR4-specific DOI or distinct PR4 licence statement was found on the Planck Legacy Archive, IRSA, or LAMBDA pages. ESA's PR3 DOI record says, verbatim, "Data hosted in the ESA Space Science Archives are distributed under the CC BY-NC 3.0 IGO license," but this card does not silently assign that PR3 release-specific record to PR4.

The embedded source credit is: "Planck Collaboration, Planck intermediate results LVII, Astronomy & Astrophysics 643, A42 (2020), doi:10.1051/0004-6361/202038073. Data provided by the Planck Collaboration and hosted by IRSA."

When research benefits from LAMBDA, LAMBDA requests:

We acknowledge the use of the Legacy Archive for Microwave Background Data Analysis (LAMBDA), part of the High Energy Astrophysics Science Archive Center (HEASARC). HEASARC/LAMBDA is a service of the Astrophysics Science Division at the NASA Goddard Space Flight Center.

These files were retrieved from IRSA rather than LAMBDA, so that text is recorded here as the archive's conditional request, not stamped as the source credit.

Citation

Please cite Planck Collaboration, "Planck intermediate results. LVII. Joint Planck LFI and HFI data processing", Astronomy & Astrophysics 643, A42 (2020), doi:10.1051/0004-6361/202038073.

Errata and requests

To report an erratum, propose a correction, or ask a question about this dataset, open a Discussion or Pull Request in its Community tab; posting requires a free Hugging Face account. Dataset requests and archive-wide questions may be posted in the README space's Community tab or the Community tab of any dataset. Confirmed corrections are made in the source conversion tools and published by rebuilding the dataset.