astro-legacy-archive/lwa1-lfsm

LWA1 Low Frequency Sky Survey

This dataset contains nine LWA1 Low Frequency Sky Survey maps at 35, 38, 40, 45, 50, 60, 70, 74, and 80 MHz. Each map and each uncertainty file is a separate configuration, preserving the source-file boundary. The survey covers the sky north of declination -40 degrees. Float32 NaNs occur in each paired map and uncertainty product; their measured masks coincide.

Intended uses

These maps support studies of low-frequency Galactic radio emission, foreground modelling, and comparisons across frequency. The uncertainty column measures variation across the snapshots contributing to each pixel and "capture the effects of uncertainties in the various calibrations and corrections applied to the data". The paper separately estimates the survey's flux-density calibration uncertainty at 20%.

Telescope	LWA1
Product	Low Frequency Sky Survey Stokes I maps and per-pixel uncertainties
Frequencies	35, 38, 40, 45, 50, 60, 70, 74, and 80 MHz
Bandwidth	957 kHz per map
Coverage	North of declination -40 degrees; measured span -39.84 to +89.82 degrees
Pixelisation	HEALPix, NSIDE 256 (786,432 pixels), RING ordering
Coordinate frame	Equatorial, equinox 2000.0
Units	K

How to use

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

from datasets import load_dataset
repo = "astro-legacy-archive/lwa1-lfsm"
sky = load_dataset(repo, "healpix-all-sky-rav-rsclean-map-74", split="train", columns=["I_STOKES"], streaming=True)
uncertainty = load_dataset(repo, "healpix-all-sky-rav-rsclean-err-74", split="train", columns=["I_STOKES"], streaming=True)
print(next(iter(sky))["I_STOKES"][0], next(iter(uncertainty))["I_STOKES"][0])

1915.99072265625 306.41888427734375

import json
import pyarrow.parquet as pq
from huggingface_hub import hf_hub_download
path = hf_hub_download("astro-legacy-archive/lwa1-lfsm", "healpix-all-sky-rav-rsclean-map-74.parquet", repo_type="dataset")
print(json.loads(pq.read_metadata(path).metadata[b"healpix_map"])["healpix"]["ordering"])

RING K NAN

The data example looks up equatorial (RA, declination) = (0, 0) degrees, RING pixel 391,680, in the separate 74 MHz map and uncertainty configurations. The metadata example reads the ordering, unit, and measured NaN code needed to interpret the source vectors.

Configs

There are two configurations per frequency, addressed by each source filename stem. Each contains 768 rows of 1,024-element vectors and is about 2.7 MiB. Measured NaN counts vary with frequency.

frequency (MHz)	measured NaN pixels	beam FWHM (deg)	temperature range (K)	uncertainty maximum (K)
35	141,809	4.65	0.0-97,571.7	40,615.9
38	141,356	4.28	5,207.7-89,864.5	38,413.2
40	141,358	4.07	4,539.1-84,662.8	36,716.9
45	142,430	3.62	3,021.5-78,356.3	32,334.3
50	142,536	3.26	1,829.8-71,190.5	27,850.9
60	142,383	2.71	1,356.8-60,386.7	23,768.8
70	141,356	2.33	913.5-56,400.1	24,608.3
74	143,374	2.20	518.4-54,980.4	25,143.0
80	142,380	2.04	860.6-52,155.6	25,148.9

Dataset structure

Every configuration has the source's single column:

column	type	unit	description
I_STOKES	fixed-size list[1024] of float32	K, from primary-header BUNIT	Stokes I sky brightness temperature in a map file, or per-pixel standard deviation of the contributing snapshots in an error file, as specified by the attached field description

The uncertainty maps give the "per-pixel standard deviation of all snapshots that contribute to that particular pixel". They are temperatures in K, not fractional uncertainties, and include calibration and correction effects that vary among snapshots.

Element j of row i is RING pixel i * 1024 + j. The map and uncertainty files contain NaNs with exactly coincident masks. The embedded coverage record identifies the measured NaN code, including its float32 bits and count. Neither file contains -1.0. The 35 MHz map has exactly seven 0.0 K pixels on the southern coverage edge; its next-lowest value is 3,359.82 K. Exactly-zero uncertainty values occur at every frequency, including 612 pixels at 35 MHz; the smallest non-zero 35 MHz uncertainty is 3.36 K. These measured zeros are data, not missing-data codes.

The maps are in equatorial coordinates. The product page describes "HEALpix RING data with 256 sides ... and equatorial coordinates", and Dowell et al. 2017 Section 3 gives the same frame. The headers contain EQUINOX = 2000.0.

Provenance

• Archive: LWA1 Low Frequency Sky Survey
• Source files and pinned SHA-256 hashes:

source file	SHA-256
healpix-all-sky-rav-rsclean-map-35.fits	ace77d09a54f4e4cb898c838f2d837c952e48614095893988a7b347827436fd8
healpix-all-sky-rav-rsclean-err-35.fits	5473ead978064f591edb1c142940beb7bee29026467f902a52665d76f4c965a4
healpix-all-sky-rav-rsclean-map-38.fits	c24bf9e8b910ae5b602d465bf8c395a1298f00156c7d6d5c3937a7b0b69f3648
healpix-all-sky-rav-rsclean-err-38.fits	3d4b1abd1fc63877e6ea910f34ed8f14f02a2882864cf642646182392968c652
healpix-all-sky-rav-rsclean-map-40.fits	701c3e8d51bc0dee240327ff4c48dd35d831d122587445abe96c29a87ee191de
healpix-all-sky-rav-rsclean-err-40.fits	13acfa9919a2e35fd4015711482307eacd2fadee279995551cf0e25396cbe125
healpix-all-sky-rav-rsclean-map-45.fits	0678f044a2a2fce54a3415cccf1e175f55e5e345f81d95fdbb4713b77cf3a908
healpix-all-sky-rav-rsclean-err-45.fits	908a867c6cdb9b65c23f9bed10323d0ee1560443407df1a2285050011fb64ba2
healpix-all-sky-rav-rsclean-map-50.fits	596b6074384114352c7e367c90a3be90d878f14ca5c6a9ca6086c8a7003af224
healpix-all-sky-rav-rsclean-err-50.fits	54eaa5d7ae891908fc958db93a45bb398e74a9285915f08803cf8ae1cbbcfb3a
healpix-all-sky-rav-rsclean-map-60.fits	e2fe42b2f55a6dfbf4bfcccdb8be172f2a48e034249db03b16c3308bfa57ed29
healpix-all-sky-rav-rsclean-err-60.fits	311f6dfc906ce103814d5904ee115b99e8ded6bc53f86dd1b404106a6dc5132e
healpix-all-sky-rav-rsclean-map-70.fits	cdb99ea336c7c3c64ae685652738642dd56410f0f9e7919f791adc8b9c0bc778
healpix-all-sky-rav-rsclean-err-70.fits	5cd56936a118d2a4e4fdb18a68151f4b46faa3d8e77cb1d96bc6c997ee0ebd8c
healpix-all-sky-rav-rsclean-map-74.fits	233dcb98ff244a539857ee909925e96b543a0e79d9650ed2c722b44f2be23597
healpix-all-sky-rav-rsclean-err-74.fits	2fad3f722a7a8fbe0cc8276d3b58062029b615a357954aa2fcfeda39e4f8f819
healpix-all-sky-rav-rsclean-map-80.fits	a5e55f59efabbc9b643589080c4bcd1572b2c242652e3e1a32d83defbc070945
healpix-all-sky-rav-rsclean-err-80.fits	a82638a69405e101a153d105d56b7426ef1149c24a7f0ff7f9b9a3ea24a91bdd

The archive records this notice verbatim: "Updated 2018 April 23 : The maps and the LFSM have been updated with improved astrometry. We have also made a small modification to the LFSM to remove an inflection in the spectrum below ~50 MHz." These are the files served by the archive at conversion, which its page identifies as the 2018 April 23 astrometry update; the FITS files carry no version identifier. The notice attributes the below-50-MHz inflection change specifically to the LFSM model.

Provenance (the original headers, source URL and hash, retrieval date, and the regeneration command) is embedded in each Parquet file. Rebuild the dataset with python tools/make_lwa1_lfsm.py.

The archive also serves Mollweide-projection images and the Low Frequency Sky Model (lfsm_20180424.tar.gz). Those companion products are not included here.

Considerations for use

• NaN masks vary by config; use each config's own measured mask.
• The overall flux-calibration budget is approximately 20%. The paper's concluding assessment is that the delivered maps are accurate to 15-25%, both for resolved structure and sky-averaged temperature.
• Dowell et al. report systematically low temperatures near the south Galactic pole in all nine maps, likely caused by limitations in the low-elevation dipole-response model coupled with the missing-spacing correction.
• The beam FWHM ranges from 4.65 degrees at 35 MHz to 2.04 degrees at 80 MHz; comparisons across frequency should account for this change.
• Exactly-zero map and uncertainty values are data, not missing-data codes.

Licensing and credit

The survey page carries "(c) 2018 The Long Wavelength Array Consortium | All Right Reserved" and publishes no permissive data licence. Its stated terms are represented as other. Credit is "Dowell et al. 2017, MNRAS 469, 4537".

NASA's LAMBDA archive independently redistributes the nine-frequency survey as "LWA1 Radio Maps" with citation-only framing, crediting "The LWA1 Low Frequency Sky Survey", Dowell, J., et al. 2017, MNRAS, 469, 4537.

Citations

Please cite Dowell et al. (2017), "The LWA1 Low Frequency Sky Survey", MNRAS 469, 4537-4550, DOI 10.1093/mnras/stx1136.

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 Community tab of any dataset. Confirmed corrections are made in the source conversion tools and published by rebuilding the dataset.