Starlink and Astronomy: The Real Conflict

When the first sixty Starlink satellites launched in May 2019, astronomers noticed within days. They were far brighter than anticipated, and the prospect of thousands more prompted immediate alarm.

Seven years on, the situation is genuinely mixed. Optical brightness has been reduced substantially through deliberate engineering. Radio interference remains largely unaddressed. Both facts deserve stating plainly.

What a satellite does to an astronomical image

Astronomical imaging works by collecting photons over long exposures — minutes to hours — to build up signal from extremely faint objects. A satellite crossing the field during that exposure leaves a bright streak across the frame.

For a narrow-field telescope pointed at one galaxy, a streak is an annoyance: mask the affected pixels, or discard and repeat the exposure. Irritating, not fatal.

For wide-field survey telescopes the calculus changes completely. A survey instrument images a large patch of sky repeatedly to catch things that change — supernovae, asteroids, transient events. A wide field is far more likely to contain a satellite, and the whole point of the survey is that exposures cannot simply be repeated at leisure. Twilight observations, which is exactly when satellites are sunlit, are the worst affected — and twilight is also when asteroid searches near the Sun have to operate.

What SpaceX actually changed

To SpaceX's credit, the response was engineering rather than public relations, and it went through several iterations.

DarkSat, launched in early 2020, applied a darkening coating to reflective surfaces. It cut brightness meaningfully but caused thermal problems — a satellite that absorbs sunlight rather than reflecting it has to shed that heat somewhere.

VisorSat followed with a deployable sunshade physically blocking sunlight from the most reflective surfaces. It worked better and became standard.

The current approach, introduced with the V2 generation, uses dielectric mirror film: a reflective film that redirects sunlight away from the ground rather than absorbing it. It achieves better darkening than the visor without the thermal penalty of the coating.

Alongside the hardware, SpaceX adjusted satellite attitude during orbit-raising to reduce the reflective cross-section presented to the ground, which is why modern trains are less conspicuous than the earliest ones.

• 2020 — DarkSat coating trial: effective on brightness, poor thermally
• 2020 — VisorSat deployable sunshade: adopted as standard for v1.0 and v1.5
• 2023 — dielectric mirror film on V2 Mini: current approach, best results to date
• Ongoing — attitude management during orbit raising, reducing brightness of new batches

The magnitude 7 target

The astronomy community, working through the IAU's centre for dark and quiet sky protection, proposed a brightness target of apparent magnitude 7 — dimmer than the naked-eye limit of roughly 6.5, and chosen as the level at which impact on professional instruments becomes manageable.

Modern operational Starlink satellites at their normal attitude are close to this target, and a substantial improvement on the first generation. They are not uniformly there, and satellites during orbit raising remain much brighter.

It is worth being clear about what the target means. Magnitude 7 does not make the satellites harmless to astronomy. It makes them tractable — streaks that software can identify and mask rather than saturating detectors. It is a mitigation, not a solution, and the astronomy community has been consistent about that distinction.

Magnitude 7 is the threshold where satellite trails become a data-processing problem instead of a data-destroying one. It is progress, not resolution.

Radio astronomy is the harder problem

Optical brightness gets the attention because it is visible. The radio problem is more serious and much less tractable.

Radio telescopes are among the most sensitive instruments ever built, detecting signals from astronomical sources measured in almost unimaginably small power levels. A satellite transmitting overhead is, from that instrument's perspective, extraordinarily loud.

Frequency coordination handles part of this: Starlink's downlink bands are allocated, and radio astronomy has protected bands of its own. But two things leak past that arrangement. Out-of-band emission — signal spilling outside the assigned band — is unavoidable at some level in any transmitter. And unintended electromagnetic radiation from onboard electronics, entirely separate from the communications payload, has been detected from Starlink satellites by European radio observatories at frequencies nowhere near the downlink bands.

That second category is not covered by existing spectrum regulation at all, because regulation was written for deliberate transmissions. There is no established framework for limiting a satellite's incidental electronic noise, and there are now thousands of satellites producing it.

Where coordination has worked

There has been genuine cooperation, which is worth acknowledging in a debate that often lacks it.

SpaceX has worked with the US National Science Foundation on coordination arrangements — including steering beams away from radio astronomy sites, sharing operational data so observatories can plan around passes, and providing predicted positions with enough accuracy for scheduling.

Position data availability matters more than it sounds. If an observatory knows precisely where satellites will be, it can schedule exposures to avoid them, or at minimum know which frames to flag. That is the practical mitigation available today, and it depends on exactly the kind of open orbital data this site is built on.

The honest position

Starlink has real value. Connecting places that terrestrial infrastructure has never reached is not a trivial benefit, and dismissing it to protect astronomy is not a serious argument.

The costs are also real. Wide-field surveys are measurably affected, radio observatories face interference with no regulatory remedy, and the cumulative effect of every constellation now being built — Starlink is not alone — is not something any single operator's mitigations can solve.

The uncomfortable structural issue is that the night sky is a shared resource with no owner and no established mechanism for allocating access to it. National regulators license spectrum and launches; nobody licenses the visual and radio environment above everyone's heads. Constellation operators are making decisions with global consequences under frameworks that were never designed for the question.

That is not an argument against Starlink. It is an argument that the governance has not caught up with the engineering, which is a fair description of where things actually stand.