The low-frequency radio sky is one of astronomy's most heavily used reference maps and, it turns out, one of its least accurately calibrated. A new measurement published in Nature Astronomy on June 29, 2026, reports that the sky between 60 and 350 megahertz is systematically brighter than the community's standard model has been telling us — by a factor of 1.2 across the 60–200 MHz range, climbing to 1.5 at 350 MHz. That is not a rounding error. It is a wholesale rescaling of a quantity that underpins everything from foreground subtraction in cosmic-dawn experiments to constraints on decaying dark matter.
The result comes from a deceptively simple instrument. Rather than a sprawling interferometer, the team used a single, carefully modelled log-periodic SKALA4.1 antenna — the same element type deployed across the Square Kilometre Array's low-frequency stations — placed at the centre of a 40-metre-diameter SKA-Low station ground mesh. The observations were carried out at Inyarrimanha Ilgari Bundara, the CSIRO Murchison Radio-astronomy Observatory in Western Australia, one of the radio-quietest sites on the planet. Over the campaign, the antenna swept up roughly half the celestial sphere.
Why a single antenna beats an army of them
Measuring the absolute brightness of the sky — the total power arriving at a given frequency, not just the contrast between bright and dim patches — is notoriously hard. Interferometers, which correlate signals from many antennas, are excellent at mapping structure but are blind to the smooth, sky-filling component that dominates at low frequencies. A single total-power antenna can see that diffuse glow, but only if you know, with brutal precision, how much of the recorded signal is the sky and how much is the instrument itself: the antenna's response pattern, the noise the receiver adds, and the frequency-dependent gain of the electronics.
That last problem is where this experiment innovates. The antenna was connected to a self-calibrating receiver called GINAN — short for Global Imprints from Nascent Atoms to Now. The receiver uses a new architecture that characterises its own noise contribution and bandpass in situ, while it is connected to the antenna, rather than relying on separate lab calibrations that can drift once the hardware is in the field. Combined with an accurate electromagnetic model of the SKALA4.1 element and its ground-mesh environment, that self-calibration is what lets the team quote an absolute brightness with enough confidence to challenge the reigning model.
What the reigning model got wrong
The benchmark here is the 2016 Global Sky Model (GSM2016), a widely used interpolation that stitches together radio surveys to predict the sky's brightness at any frequency and position. The new data say GSM2016 must be scaled up by a factor of 1.2 over 60–200 MHz, and the discrepancy grows toward higher frequencies, reaching 1.5 at 350 MHz. In other words, the model has been systematically underestimating the diffuse radio sky, and doing so more severely as you tune up the dial.
One immediate consequence lands squarely on a long-standing anomaly. Earlier work had already flagged an unexplained surplus of radio emission beyond what known source populations account for, the so-called excess radio background. Scaling the sky brighter does not make that excess go away; it makes it larger. The measurement significantly increases the previously inferred excess, sharpening rather than softening one of low-frequency radio astronomy's more stubborn puzzles.
Why It Matters
Accurate knowledge of the low-frequency sky is not a niche concern — it is load-bearing infrastructure for several frontiers at once. Experiments hunting for the cosmic dawn and the epoch of reionization, when the first stars and galaxies reshaped the intergalactic medium, must subtract Galactic and extragalactic foregrounds that outshine the target signal by orders of magnitude. If the foreground map is 20 to 50 percent too dim, those subtractions inherit the error.
The same measurement feeds into the Galactic cosmic-ray electron spectrum, into nanojansky-level source-population counts, and into constraints on the origins of the diffuse radio background — including the possibility that some of it comes from decaying dark matter. A brighter, deeper excess forces those models to be reconsidered: faint radio-source populations may need to be more numerous or more luminous than assumed, and dark-matter decay scenarios that were tuned to the old background now have a larger target to explain. Long-wavelength telescopes, which lean on sky models for their own calibration, stand to inherit the correction directly.
The broader lesson is a familiar one in precision astronomy: a widely trusted reference is only as good as the absolute calibration behind it, and here a single well-understood antenna has done what a forest of them could not. As SKA-Low itself ramps up, having a trustworthy absolute yardstick for the radio sky is exactly the kind of unglamorous groundwork that determines whether its headline science holds up.
