How low do they go: satellites impacting low-frequency radio astronomy

Title: Bright unintended electromagnetic radiation from second-generation Starlink satellites

Authors: C. G. Bassa, F. Di Vruno, B. Winkel, G. I. G. Józsa, M. A. Brentjens, and X. Zhang

First Author’s Institution: ASTRON, Netherlands Institute for Radio Astronomy

Status: Published in Astronomy & Astrophysics [open access]

A man-made satellite does not emit much intrinsic radiation in visible light, or light from the part of the electromagnetic spectrum that humans can perceive. Instead, essentially all of the visible light from a satellite is reflected from the Sun. As the number of satellites has sky-rocketed with the introduction of ever-growing satellite “constellations” like SpaceX’s Starlink– which plans to be a fleet consisting on the order of 10000 satellites and has already vastly increased the number of satellites in orbit–  the effects of this form of light pollution has rapidly increased its negative impacts on both ground and space-based observations. An example of the effects of these reflections on astronomical observations is shown in Figure 1, where each of the numerous streaks represents an individual satellite.

Figure 1: The result of a 5.5 minute exposure from the Víctor M. Blanco 4-meter Telescope at the Cerro Tololo Inter-American Observatory. Long exposures like this are required to achieve high sensitivity, but Starlink satellites streak through the field of view during that time, producing the bright, straight lines in the image. Credit: NSF’s National Optical-Infrared Astronomy Research Laboratory / CTIO / AURA / DELVE.

However, these satellites aren’t just reflecting visible light. They have a suite of electronics on board to execute their intended functions; for constellations like Starlink, these intended functions require them to emit in a designated radio band so that they can transmit information. Although the electronics are only meant to transmit at frequencies in the 10s of gigahertz (GHz) range, they also produce unintended electromagnetic radiation (UEMR) at parts of the spectrum that were meant to be protected for radio astronomy and communications as outlined by the  Radiocommunication Sector of the International Telecommunication Union (ITU-R). Importantly, this emission isn’t reflected but is intrinsically produced by the satellites. The authors of today’s paper quantify this UEMR from the second generation of Starlink satellites and what that could mean for the future of radio astronomy at low frequencies.

In order to study the emissions from these constellations, the authors look at two, one-hour observations taken with the Low-frequency Array (LOFAR). This is one of the most sensitive radio arrays in the world operating at low frequencies (near the same frequencies used for FM radio stations) and the most sensitive in the Northern hemisphere. One hour was collected with the low-band array (LBA) operating between 10-88 Megahertz (MHz), and the other hour was collected with the high-band array (HBA) operating 110-188 MHz. As a low-frequency instrument, LOFAR takes advantage of interferometry to increase its sensitivity and spatial resolution beyond what a single antenna can achieve. Although LOFAR consists of antennas spanning much of Europe, the authors today use the six central collections (or “stations”) of antennas in the Netherlands.

The signals from these stations are combined to study 91 regions of the sky (or “beams”) that the array records the intensity of as a function of time and frequency. Several examples of this data product, called dynamic spectra, are shown in Figure 2. SpaceX does not provide information on the orbital parameters of their satellites. Instead, the authors used parameters based on observations of the satellites made by the United States Space Force in order to predict when a satellite might pass through at least one of the 91 LOFAR beams. 

Figure 2: Flux density (energy flux per frequency bin, given in units of Janskys) as a function of frequency (vertical axis) and time offset (horizontal axis) for various frequency ranges. There is bright emission spanning ~40-70 MHz, referred to as broadband emission, that originates from the Starlink v2-Mini satellite Starlink-31441. (Adapted from Figure A.1 in the paper)

Impressively, if not concerningly, they indeed detected all 97 satellites that they predicted to pass through an HBA beam. They detected emission from both the first- and second-generation satellites with the HBA, with the second-generation showing significantly brighter UEMR than the first-generation satellites. At the lower frequencies that the LBA observes, only the second-generation satellites were detected; signals from the first-generation satellites were nowhere to be found in the LBA dynamic spectra.

It is worth noting that the second-generation satellites are in lower orbital altitudes than the first-generation satellites; they may appear brighter or detectable simply because they are closer to us rather than by being intrinsically more luminous. To check if this is the case, the authors scale all of the satellites’ emissions as if they originated from a uniform altitude of 1000 kilometers (km). This scaling reveals that the second-generation satellites are intrinsically more luminous than their first-generation siblings in addition to being lower in orbit. The evaluation of this luminosity in terms of electric field strength is shown in Figure 3.

Figure 3: The normalized distribution of the number of satellites as a function of the electric field strength they emit with (in units of micro-volts per meter on a decibel scale). The different generations of satellites are indicated by different colors, with v1.0 and v1.5 referring to the first generation and v2-Mini and v2-Mini [DTC] referring to second generation satellites. This shows that the second-generation satellites have much stronger electric fields to produce UEMR than the first-generation satellites. (Figure 2 in paper).

The implications of this are dire. The electric fields that the first-generation satellites were producing were already in excess of the ITU-R threshold for the HBA band. The fact that these fields have not only gotten stronger with the latest generation of satellites but have also begun producing emission detectable at lower frequencies is troubling at best. This could be especially catastrophic to time-domain science (like searches for unpredictable stellar emission or other transient sources) as well as science that relies on high-sensitivity data of the entire sky (like attempts to observe the first signals of galaxies and stars reionizing the universe). 

Unfortunately, as the definitions of the ITU-R stand, the UEMR described in this paper may fall outside the scope of regulation; although there are various pathways and protocols for mitigating UEMR from the ground, a regulatory framework for space-based UEMR is missing. This is not the first article to detail the adverse effects of Starlink constellations on radio astronomy, but hopefully the growing body of literature detailing the extent of UEMR– and its worsening effects– will motivate engineers and corporations to consider their role and responsibility to our skies.

Edited by Will Golay

Featured image credit: A snapshot from the satellite map service

Author

  • Ivey Davis

    I’m a fourth year astrophysics grad student working on the radio and optical instrumentation and science for studying magnetic activity on stars. When I’m not crying over RFI, I’m usually baking, knitting, harassing my cat, or playing the banjo!

    View all posts

Submit a Comment

Your email address will not be published. Required fields are marked *