Nobody launched them as weather sensors. Now thousands of satellites can help measure the atmosphere

Satellites are sent into orbit to communicate, observe, navigate or do some other carefully planned job. Nobody puts them up there to act as a weather station. Yet a large fleet of satellites can reveal something about the very thin atmosphere they travel through.

In a September 2026 analysis, astronomer Tony Phillips at SpaceWeather.com used publicly available orbital information from Starlink and other satellite constellations to calculate a daily indicator he calls a planetary barometer. It is a striking name, and a useful story, but the measurement is more specific than the nickname suggests.

The atmosphere does not stop at the edge of space

Low Earth orbit is not a perfect vacuum. At around 480 kilometres above Earth, the atmosphere is extraordinarily thin, but there is still enough gas to exert a tiny drag force on a moving satellite. That drag removes a little orbital energy over time and changes the satellite’s path.

Solar activity can heat the upper atmosphere, causing it to expand. At a satellite’s altitude, that can mean more particles and greater density than before. More density generally means more drag. Geomagnetic storms can also produce shorter-lived changes. NOAA describes these effects as an important source of uncertainty when calculating satellite orbits.

NOAA: how solar activity changes satellite drag →

How a satellite becomes an indirect sensor

The satellites do not carry a little air-density meter. The analysis starts with their changing orbital data. Researchers can compare where a satellite is expected to be with how its orbit behaves, then use a model of drag and the spacecraft’s motion to estimate what the surrounding atmosphere must be like.

Phillips expresses his result as a “sink rate”: how far a representative satellite at 480 kilometres would lose altitude in a day if its thrusters were switched off. That is a calculated indicator, not a measurement from a satellite that actually had its propulsion disabled. The reference altitude and the assumptions behind the calculation matter.

The method takes advantage of the fact that large satellite fleets are tracked over time. Changes across many objects can help show whether drag is increasing or easing. But spacecraft are not identical, operators make manoeuvres, and orbital data are not the same thing as a purpose-built scientific instrument. The result is an inference from motion, not a direct reading from an atmospheric probe.

SpaceWeather.com: Tony Phillips’s September 2026 analysis →

A barometer? Not quite

SpaceWeather.com itself makes the important distinction: satellite drag is most directly related to atmospheric density, not the pressure reading from a barometer on the ground. The analysis concerns the upper atmosphere along a satellite’s orbit, hundreds of kilometres above us. It does not tell a café in Cardiff whether to put the tables outside, and it will not replace a local weather forecast.

“Planetary barometer” is a memorable shorthand. “A model-based indicator of upper-atmosphere density inferred from orbital drag” is more accurate, but less likely to fit in a headline. The nickname is fine as long as we remember what it leaves out.

The science is established; this particular chart is an estimate

The underlying physics is not new: the upper atmosphere affects satellite motion, and orbital changes can be used to estimate atmospheric conditions. Peer-reviewed studies have used Starlink ephemerides and other orbital data to investigate thermospheric density. Those studies support the general research approach; they do not independently validate Phillips’s particular daily chart or every number in it.

That distinction is useful. This is an interesting example of data being reused, not a claim that a commercial satellite network has become a calibrated global weather instrument. Public orbital data and modelling create a new view of the atmosphere, with limits that depend on coverage, timing, manoeuvres and the model used.

Peer-reviewed research: thermospheric density estimates from Starlink ephemeris →

Peer-reviewed research: density estimates from Starlink orbital decay data →

The more interesting idea is what else infrastructure can tell us

A network designed for one job can quietly become useful for another. Smartphones built for communication help reveal traffic patterns. Modern vehicles can report road conditions. Wi-Fi measurements can help estimate whether a room is occupied. Satellites designed to deliver connectivity can also provide orbital data that researchers use to study the atmosphere.

These are not all the same kind of sensor, and they do not all produce measurements with the same accuracy or privacy implications. The common thread is that useful information can emerge from systems that were not originally designed to collect it. Sometimes the innovation is not a new device; it is recognising a second use for the infrastructure already in place.

There is no immediate action for an SME to take here. The takeaway is simply to notice where a network’s routine by-products—measurements, patterns or operating data—might answer a question that used to require new equipment. Satellite communications have other, more direct business uses too:

No Mobile Signal? That Excuse May Be Disappearing →

Sources and further reading

SpaceWeather.com’s “planetary barometer” is a practitioner analysis; NOAA and the research papers below support the general relationship between upper-atmosphere density, drag and satellite orbits, not the exact output of that chart.

SpaceWeather.com: “We’ve Turned Starlink into a Planetary Barometer” →

NOAA Space Weather Prediction Center: Satellite Drag →

Ou et al., Remote Sensing: Near-Real-Time Global Thermospheric Density Variations Unveiled by Starlink Ephemeris →

Earth, Planets and Space: Thermospheric density from Starlink TLE →

Found this useful? Forward it to someone who might too.