An Emerging Risk to Radio Astronomy

A new study from scientists at the International Centre for Radio Astronomy Research (ICRAR) has delivered comprehensive evidence of the impact that satellite mega-constellations, particularly those generated by the Starlink network - will have on radio astronomy and the behemoth science project, the Square Kilometre Array (SKA).

Over the course of 29 days and 76 million full-sky images, astronomers using the Engineering Development Array 2 (EDA2 - a prototype instrument with similar design features to the SKA-Low telescope) located at the Inyarrimanha Ilgari Bundara (CSIRO Murchison Radio Astronomy Observatory) in Western Australia, detected more than 112,000 instances of radio-frequency emission from 1,806 unique Starlink satellites.

“We knew from our previous studies that some satellites transmit unintentionally across a wide range of frequencies,” said lead author and PhD candidate Dylan Grigg from Curtin University. “The difference with Starlink is that there are thousands of them in orbit compared to the handful of other satellites we saw doing this previously.” 

“With the increase in Starlink launches over the last few years since we first observed them, it makes sense that we would detect more now as the cause of this unintended emission still persists in the design of the satellite.”

“What perhaps surprised us the most was the extent at which we detected them over radio frequencies which are protected for radio astronomy by the International Telecommunication Union,” said Grigg. “We are hoping that continued discussion with the community and SpaceX could lead to improvements in future satellite design.”

Starlink satellites emit both intended and unintended signals across a broad range of frequencies, including those reserved by international agreement through the International Telecommunication Union (ITU) for radio astronomy. In the worst-affected datasets, a Starlink satellite appeared in nearly 30% of all-sky images, demonstrating the scale of the interference.

The SKA-Low telescope, a multi-billion-dollar, multi-national project, is poised to explore cosmic dawn and the Epoch of Reionisation (EoR) at unprecedented sensitivity. To achieve this sensitivity, the Australian Government has dedicated a 260 km radius radio quiet zone around the telescope. This assists with radio frequency interference (RFI) that is terrestrial-based, but is less effective when it comes to RFI from above.

Researchers warn that this level of radio contamination could overwhelm ultra-sensitive science goals, and that urgent global regulatory and technical mitigation is needed.

“The impact of the unintended emission will likely depend on the science case, but has the potential to result in increasing data loss for projects,” said co-author Professor Steven Tingay from Curtin University. “Knowledge of when a Starlink satellite is passing over the patch of sky that is being observed will become increasingly important for projects.”

“To put the signals in context, an unintended signal from a single Starlink satellite can often be as strong as the strongest astronomical radio sources in the sky.  So, the interference is comparable to the brightest signals we deal with, and we are interested in those signals that are far weaker.”

“So, the Starlink signals can drown out the weaker signals.”

This new study presents the largest radio survey to date monitoring satellite interference across the frequency range of the SKA-Low telescope (50-350 MHz). Unlike previous studies that focused on narrower bands or used targeted tracking, this survey leveraged EDA2’s all-sky, 24-hour imaging to detect satellites passively and autonomously.

Across 29 observing sessions at different frequencies, both narrowband and broadband emissions were detected. Critically, the emissions span primary and secondary frequency allocations that are protected under ITU rules for radio astronomy, including the bands 73.0-74.6 MHz and 150.05-153.0 MHz, which fall within the SKA-low observing bandwidth.

Even more concerning, the study found that the average flux density from unintended emissions was five orders of magnitude greater than the contamination threshold estimated to ruin EoR studies.

Once completed, the SKA will be the world’s largest radio telescope and is designed to answer fundamental questions about the universe: from the formation of the first stars to the nature of dark energy. SKA-Low, is also designed to explore the faint radio signals from the EoR, a period over 13 billion years ago when the first stars and galaxies ignited, reionising the neutral hydrogen that filled the early Universe.

“The Epoch of Reionisation experiment attempts to measure extremely faint signals generated from neutral hydrogen from billions of years ago, early in the Universe and soon after the Big Bang,” said Grigg. 

“Any bright foreground signals (e.g. satellites and anthropogenic RFI) can obscure these faint EoR signals by many orders of magnitude, meaning affected data need to be identified and removed from the analysis. An increasing amount of unintended emissions from satellites will result in a higher rate of data loss for experiments such as this, and when the interfering signals are weak, affected data can be hard to recognise.”

However, the sensitivity of SKA-Low - one of its scientific strengths - makes it susceptible to even faint RFI. The new findings suggest that Starlink’s unintended electromagnetic radiation (UEMR) alone could be enough to affect several key experiments if not mitigated.

Additionally, with this RFI occurring across a wide swath of the spectrum and originating from thousands of moving objects, filtering them out with software may be challenging. Physical, regulatory, and engineering-based solutions may be required at both the satellite and ground-based levels.

The Starlink network, operated by SpaceX, is the largest satellite constellation ever launched, currently numbering over 7,000 satellites. The company regularly launches batches of up to 50 satellites at a time and has approval for tens of thousands more. Other operators, such as Amazon, OneWeb, and future Chinese and Russian systems - are planning constellations of comparable scale, which will eventually compound the issues astronomers are facing.

Starlink satellites operate in low Earth orbit, typically 400–800 km above the surface, offering low-latency communication services to users across the globe. The v2-mini and v2-mini Direct to Cell satellites detected in this study dominate the fleet’s current makeup.

Satellite emissions can be classed under two categories: Intended emissions, which are used for communication and telemetry and are restricted to specific frequencies that the ITU has assigned, and Unintended emissions which are generated from onboard electronics, propulsion systems. The unintended emissions are broadband, which means they are spread across lots of frequencies.

Radio astronomers have long relied on small portions of the radio spectrum that are protected under ITU regulations - designated for scientific use. Yet this study confirms that Starlink satellites are emitting directly into those protected bands. These emissions are not part of the satellite’s official intended downlink frequencies but are likely due to incidental leakage from propulsion systems.

This is a growing issue: emissions are now being detected from Starlink satellites even after they’ve reached their operational orbits, with signatures at frequencies above 200 MHz. And while ITU rules regulate intentional transmissions, they currently have no specific provisions for UEMR - placing astronomers in a regulatory grey zone.

“In an ideal world, the best mitigation strategy would be engineering solutions on the Starlink satellites themselves that mitigate the generation of these signals,” said Prof. Tingay. “Identifying where on the satellites the unintended emission comes from, and engineering the component differently, would be the best mitigation.”

“For an observer, mitigation may be as simple as excluding data from the analysis, which comes at the loss of costly telescope time. This is due to the emission from the satellites being orders of magnitude brighter than the astrophysical signals which are trying to be observed.”

“Algorithmic mitigation or data cleaning will have to be developed to potentially recover less affected data products, but it is likely that the cost and complexity of doing this will exceed the already costly and complex task of basic data processing.  Even if possible, observer-side mitigation could easily double the already very high cost of data processing.” 

With mega-science projects like the SKA now heavily invested in, astronomers and policymakers are starting to navigate the complex regulations that will balance the drivers of global connectivity with preservation of the sky for astronomical purposes. This latest survey, with its open-access dataset, is a suitable tool for regulators, engineers and scientists to utilise in considering a sustainable future for astronomy.

As the Low-Earth Orbit becomes busier, radio quiet zones start to grow louder. 

Video Credit: Astro_Work (YouTube - Supplied)

We acknowledge the traditional owners of the lands in which our instruments are based. The SKA-Low site is located on the traditional lands of the Wajarri Yamatjii people.