Revised August 2024
Rubin Observatory is nearing completion, and its LSST will soon offer an unprecedented, detailed view of the changing sky. Starting in 2025, Rubin Observatory will capture about 1,000 images of the sky, every night, for ten years. Each image will cover a 9.6 square degree field of view, or about 40 times the area of the full Moon. Because of the telescope's large light-collecting area, each 30-second exposure will reveal distant objects that are about 20 million times fainter than those visible with the unaided eye from a dark location. This large combination of light-collecting area and field of view on the sky is unprecedented in the history of optical astronomy, which is one reason Rubin/LSST was the top-ranked ground-based project in the National Academy of Science 2010 Decadal Survey of Astronomy and Astrophysics.
LSST survey images will contain data for about 20 billion galaxies and a similar number of stars, and will be used for investigations ranging from cosmological studies of the Universe to searches for potentially dangerous Earth-impacting asteroids. However, the revolutionary discoveries anticipated from the LSST could be significantly degraded by the alarming pace of new deployments of LEO communications satellites. The same properties that make LSST uniquely well-suited for discovery across all areas of astrophysics — a wide field of view and a large light-collecting area — also render it vulnerable to this new source of light pollution.
LSST and all ground-based astronomy operates by looking through an unprecedentedly crowded LEO. This new commercial space race began in 2019 with the launch of the first SpaceX Starlink satellites, and other commercial satellite operators rapidly entered the scene, including but not limited to include Amazon Kuiper, OneWeb, AST SpaceMobile, Planet Pelican, and SSST Qianfan (Thousand Sails). Each operator makes independent decisions about hardware, orbits, launches, lifetimes, radio transmissions, maneuvering capabilities, and more.
In general, impacts to science can be lessened with fewer satellites, dimmer satellites, lower orbital altitudes, or a combination of these. While astronomers cannot control these variables or reliably predict the future satellite population distribution, we have consistently recommended operators implement brightness mitigations and choose orbital altitudes below 600 km. This is because lower satellites orbit faster and are more out of focus. As a result, they spend less time illuminated above the horizon and the resulting streak appears dimmer due to a lower peak surface brightness. In contrast, astronomers do have control over where we point the telescope when, characterizing and mitigating the camera detector artifacts, and our image processing pipelines.
During a nominal 30-second LSST visit, illuminated satellites in LEO will move about 15 degrees across the sky (about four times the diameter of the field of view). They are most numerous a few hours after sunset and before sunrise and closer to the horizon. Simulations of the LSST observing cadence and about 40,000 LEO satellites show that about 10% of all LSST images would contain at least one satellite trail, and the majority of images in twilight would contain streaks (Hu et al. 2022). Given that applications for over one million commercial LEO satellites have now been filed, there may well be 10 times this many.
Measurements of the brightness of some current LEO satellites indicate their trails will cause residual artifacts in reduced data from the LSST Camera. If these satellites can be darkened to about 6th–7th magnitude, then an instrument signature removal algorithm can remove some of these residual crosstalk artifacts (Tyson et al. 2020). However, the bright main satellite trail would still be present, potentially creating bogus alerts and systematic errors at low surface brightness (Hasan et al. 2022). While some darkening mitigations have been demonstrated to be effective, virtually all present-day commercial LEO satellites are brighter than 6th–7th magnitude at least some of the time. A typical satellite is designed to operate for five years and then burn up in Earth’s atmosphere, which raises other environmental concerns. Launch, orbit raise, various experiments, attitude adjustments, collision avoidance maneuvers, and deorbit are all common life cycle elements when a satellite tends to appear brighter than in its typical operational orbit. Even if these activities are a small fraction of a satellite’s life, it does not scale sustainably to a population of tens or hundreds of thousands of such satellites.
In the case of satellites that are significantly brighter than 6th–7th magnitude, such as larger Direct To Cell satellites, entire LSST camera detectors may be saturated, which would render that exposure scientifically useless. Proactive avoidance of certain regions of the sky at certain times via the LSST scheduler algorithm may be necessary (Hu et al. 2022). This is technically feasible with sufficiently accurate satellite pass and brightness predictions (Fankhauser et al. 2023). Simulations show it is possible to sacrifice about 10% of LSST observing time to reduce the fraction of LSST exposures with streaks by a factor of 2.
Because of the huge data volume, LSST science will typically be limited by systematics rather than by sample variance (area incompleteness). For example, the ability to detect asteroids approaching from directions interior to Earth's orbit will be disproportionately impacted because those directions are visible only during twilight when illuminated LEO satellites are most numerous. Precision cosmological studies are another example; they are very sensitive to small systematic effects, and might suffer from artifacts due to crosstalk correction or insufficient masking of satellite streaks. While it is LSST policy to not issue alerts for artificial satellites, it will be impossible to fully eliminate bogus alerts caused by glints or flares from satellites maneuvering to adjust orbits or for collision avoidance.
We do not anticipate significant impacts to LSST science from small LEO debris with radius less than 10 cm (Tyson et al. 2024). Most of this population is sufficiently faint to escape detection due to being out of focus and having a small reflective area. The exception is debris composed of mirror-like material, which can produce very bright glints. We do anticipate glint trains, or “strings of pearls,” due to tumbling debris, as has been seen in images from the Zwicky Transient Facility (ZTF), and are developing a glint detection algorithm to distinguish such signals from transients or Solar System objects. We also expect the overall sky background brightness level to increase over the ten years of LSST due to the proliferation of very small debris (Kocifaj et al. 2021).
Founded in 2022 following SATCON2 (see, e.g., Hall et al. 2021, Rawls et al. 2021), the IAU CPS is co-hosted by NSF NOIRLab and the SKA Observatory. It aims to coordinate efforts and unify voices across the global astronomical community with regard to the protection of the dark and quiet sky from satellite constellation interference, and is composed of four Hubs designed to address the problem from multiple angles: SatHub (for observations and astronomical data analysis), Policy, Industry and Technology, and Community Engagement. Anyone can apply to join the CPS and contribute to its mission.
In June 2024, the UN’s Committee on the Peaceful Uses of Outer Space (COPUOS) approved the inclusion of an item on one of its subcommittee’s agenda for the next five years to address the emerging issues and challenges posed by large satellite constellations. Several dozen nations have been supportive of efforts to address this topic within COPUOS, and have established a Group of Friends of the Dark and Quiet Sky to promote awareness of the issue.
In July 2024, the NSF awarded $750k to a three-year project led by the IAU CPS SatHub to minimize the science impact caused by satellite constellations on astronomical observatories, with a focus on Rubin. The grant includes development of tools and services that would provide sufficiently accurate, precise, and up-to-date satellite position information along with improved brightness modeling to enable LSST to implement a satellite avoidance scheme. The funded work will also more quantitatively assess impacts on Solar System discovery and characterization studies as well as transient science with LSST.
In addition, the IAU CPS has produced recommendations and set some helpful precedents in the US, such as requiring a coordination agreement between NSF and satellite operators as a condition of FCC licensure. The CPS is concerned with the entire electromagnetic spectrum, including unintended radio emissions (Di Vruno et al. 2023).
LSST Science Pipelines team members are working to implement and validate a streak detection algorithm, a glint detection algorithm, and other techniques to eliminate artificial satellites from alerts and minimize bogus sources in catalog-based data products. We do not plan to modify any image pixels. Instead, following difference imaging in both the Alert Production and Data Release Production pipelines, we will identify bright streaks, add a STREAK mask plane, and remove affected sources from catalogs where appropriate. Using methods developed to detect Solar System objects, we will also search for signatures of tumbling debris glints and remove them from affected source catalogs. While we will make every effort to accurately identify and label situations that are most likely caused by artificial satellites and debris, we cannot guarantee that LSST data products will be free of contamination. Overall, large numbers of bright satellites — and the necessary steps to avoid, identify, and otherwise mitigate them — will impact the ability of LSST to discover the unexpected.
Reports, Recommendations, and Non-Peer-Reviewed Pieces
ESA 2024 Space Environment Report https://sdup.esoc.esa.int/discosweb/statistics/
IAU CPS 2024 Recommendations Paper https://cps.iau.org/news/cps-urges-action-in-first-recommendations-paper/
Falle et al. 2024 https://www.science.org/doi/10.1126/science.adi4639
GAO 2022 Report on Large Constellations of Satellites https://www.gao.gov/products/gao-22-105166
Lawrence et al. 2022 https://www.nature.com/articles/s41550-022-01655-6
Mallama et al. 2023 https://arxiv.org/abs/2309.14152
SpaceX Darkening Mitigations 2022 https://api.starlink.com/public-files/BrightnessMitigationBestPracticesSatelliteOperators.pdf
Di Vruno et al. 2023 https://doi.org/10.1051/0004-6361/202346374
Fankhauser et al. 2023 https://iopscience.iop.org/article/10.3847/1538-3881/ace047
Hasan et al. 2022 https://www.sciencedirect.com/science/article/pii/S2213133722000245
Hu et al. 2022 https://iopscience.iop.org/article/10.3847/2041-8213/aca592
Kocifaj et al. 2021 https://academic.oup.com/mnrasl/article/504/1/L40/6188393
Murphy et al. 2023 https://www.pnas.org/doi/10.1073/pnas.2313374120
Nandakumar et al. 2023 https://www.nature.com/articles/s41586-023-06672-7
Tyson et al. 2024 https://iopscience.iop.org/article/10.3847/2041-8213/ad41e6
Tyson et al. 2020 https://iopscience.iop.org/article/10.3847/1538-3881/abba3e/meta
Walker & Hall 2020 https://aas.org/sites/default/files/2020-08/SATCON1-Report.pdf
Dark & Quiet Skies 1, 2020 https://www.iau.org/static/publications/dqskies-book-29-12-20.pdf
Hall et al. 2021 https://baas.aas.org/pub/2021i0205/release/1
Rawls et al. 2021 https://baas.aas.org/pub/004iuwwu/release/1
Dark & Quiet Skies 2, 2021 https://zenodo.org/record/5874725
Peel et al. 2024 https://arxiv.org/abs/2404.18742
IAU CPS https://cps.iau.org
AAS COMPASSE https://compasse.aas.org
https://www.zotero.org/groups/4501709/satcons/library
