Disclaimer: “Beyond astro-ph” articles are not necessarily intended to be representative of the views of the entire Astrobites collaboration, nor do they represent the views of the AAS or all astronomers. While AAS supports Astrobites, Astrobites is editorially independent and content that appears on Astrobites is not reviewed or approved by the AAS.
Title: The Growing Impact of Unintended Starlink Broadband Emission on Radio Astronomy in the SKA-Low Frequency Range
Authors: Dylan Grigg et al.
First Author’s Institution: International Centre for Radio Astronomy Research, Curtin University, Australia
Status: Published 2025 July 17, Astronomy & Astrophysics [open access]
There has been an explosion in the number of communication satellites in low earth orbit (LEO) since 2019, facilitated by reusable space launch technologies and spurred on by the promise of low-cost, low-latency internet access. Starlink, a mega-constellation of satellites operated by SpaceX, currently dominates LEO, with batches of Starlink satellites being launched on average every 3 days and approximately 8000 in orbit as of this article’s publication.
These mega-constellation satellites have grave, well-documented effects on optical astronomy, as well as potential impacts on the composition of the atmosphere, including ozone depletion. The effects on radio astronomy is a developing field of research, but as previously documented in this Astrobite, Starlink satellites have already been consistently observed at low radio frequencies and notably at frequencies that are supposed to be protected for radio astronomy under international guidelines. Unlike terrestrial sources of radio interference, like television or cell phones, you can’t avoid satellites by building telescopes in remote, low-population areas on Earth.
This radio emission at lower frequencies from Starlink isn’t their downlink frequency, but instead unintentional electromagnetic radiation (UEMR), thought to be caused by the onboard electronics of the satellite. This UEMR is not currently regulated by the International Telecommunications Union (ITU), the organization responsible for managing and allocating the radio spectrum for various uses. At low radio frequencies (<5 GHz), only 5% of the radio spectrum is reserved for radio astronomy, so any interference within these frequency channels is doubly disruptive.
As the world’s largest radio observatory, the Square Kilometer Array Observatory (SKAO), comes online, astronomers are especially worried about effects mega-constellation satellites could have on the science goals of the multi-billion dollar project. The authors of today’s paper use an SKA pathfinder telescope, the Engineering Development Array 2 (EDA2), to investigate what impacts we can expect for SKA-Low, the low frequency half of the SKAO, from Starlink.
EDA2 is well situated for this task, as it is located in Western Australia at Inyarrimanha Ilgari Bundara (CSIRO Murchison Radio-astronomy Observatory), the same location as SKA-Low, and it uses the same number of antennas as an SKA-Low station. Previous studies using EDA2 used full sky radio images, detecting Starlink satellites within both 20 hours of data at 137.5 MHz and 23 hours at 159.4 MHz. A separate team, using the Low-Frequency Array (LOFAR) telescope, LOFAR, detected 68 satellites from 110 to 188 MHz, 29 satellites at 10 to 88 MHz, and 97 satellites at 100 to 188 MHz, each in separate 1-hour observations. The satellite emission was found to be both broadband (wide spread in frequency) and narrowband (low spread in frequency),
Today’s paper builds upon the work of these previous surveys by significantly increasing the number of observations, the area of sky searched, and the number of frequency channels, resulting in 29 separate, 24-hour observations at different frequencies. The authors chose frequencies related to key science goals of SKA-Low, for example frequencies used to study the Epoch of Reionization (EoR). They also chose frequencies that overlapped with protected ITU frequency bands.
For each timestep during one of the 24-hour observations, 31 radio images were created, corresponding to 31 different frequency channels. For each radio image, the predicted location of the satellite was fit to a 2D Gaussian, and if it passed a set of criteria, was labelled a detection. If five or more detections matched the trajectory of a single satellite across the sky, it was labelled an identification. A 1D Gaussian was then fit to the pixel amplitude of each individual identification and if it was less than a certain standard deviation value, was identified as narrowband. Otherwise it was assumed to be broadband.
In total, these observations resulted in 112, 534 individual identifications, corresponding to 1806 unique Starlink satellites. This is 28% of Starlink satellites in orbit at the time of observation. At some frequencies, almost 30% of images contained at least one identification, though the authors note the real value is likely higher, since identification criteria were very strict to avoid misidentifications.
The most common detection was the v2-mini Direct-to-Cell (DTC) model of Starlink, with 175 detected, corresponding to 71% of the total population. DTC satellites are designed to downlink directly with your cellphone, instead of to an intermediary, like a cellphone tower, meaning they are more powerful and larger than previous generations of Starlink satellites.
Not all the detected radio emission seems to be due to UEMR. Four satellites had narrowband emission at 99.7 MHz narrowband emission, which could be reflected FM radio from a transmitter in Geraldton, Australia, a city 300 km from EDA2. Two of these satellites were v2-mini Ku-band Starlink satellites and two were v2-mini DTC satellites. Additionally, there was a v2-mini Ku with exactly 100.00 MHz emission, only visible in a single channel, which does not match FM radio transmission frequencies, leaving its source currently unknown.
The authors estimate a lower limit of 93 Jy per beam in the frequency averaged images containing Starlink emission. Considering just 1 mJy of radio frequency interference could mess up an EoR power spectrum integration, this could severely affect SKA-Low EoR science. Of course the satellites will not always be visible but this is hugely over the limit of the survey.
This survey indicates the need for a serious increase in regulation for UEMR, as no regulation for it currently exists in the ITU radio standards, especially since the UEMR is visible in protected frequencies. The exact source of UEMR remains unknown but it could be related to propulsion or avionics since previous studies indicate radio emission corresponded to a change in altitude.
Astronomers are already working on mitigation efforts, while waiting for policy to catch up. For example, in 2024, the National Radio Astronomy Observatory (NRAO) in the USA and SpaceX announced an agreement in which Starlink satellites will turn off their downlink temporarily when in the region of sky NRAO telescopes are pointing to. Internationally, in 2022, the International Astronomical Union, the National Science Foundations NOIRLab, and the SKAO came together to create the Centre for the Protection of the Dark and Quiet Sky (CPS) to coordinate efforts on the issue of satellite interference. The CPS continues to work on this issue, including coordinating observations, developing mitigation techniques, and working with policy makers. The latest newsletter from the CPS is available here.
Astrobite edited by Mckenzie Ferrari
Featured image credit: Michael Goh/ICRAR/Curtin
