Posted by Nearly every solar system live view on the web is a simulation. Beautiful ones, ticking away in a browser tab, planets sliding around the Sun while you watch. Almost none of it is a picture of anything. It is arithmetic dressed up in pixels. Underneath that layer, though, sits a thinner and much stranger one that really is observation: a small set of spacecraft pointing cameras at the Sun and at Earth and pushing frames to the ground fast enough that you can look at them while they are still warm.
That layer is patchy, unevenly funded, and full of delays nobody advertises. Worth knowing before you trust a timestamp.
The Sun on a Twelve-Second Heartbeat
NASA’s Solar Dynamics Observatory launched in February 2010 into an inclined geosynchronous orbit, and the orbit was chosen for a boring reason: it keeps the spacecraft in view of a pair of dedicated dishes at White Sands, New Mexico, nearly all the time. No storing data for later. SDO’s Atmospheric Imaging Assembly photographs the full solar disc in ten ultraviolet and extreme-ultraviolet channels at 4096 pixels square, cycling through them every twelve seconds or so, while the Helioseismic and Magnetic Imager maps the surface magnetic field underneath. The result is well over a terabyte a day, coming down more or less continuously.
The public “latest images” pages hang off that stream and typically run a few minutes behind the Sun itself. As live as this business gets.
The gaps are instructive. Twice a year, near the equinoxes, Earth slides between SDO and the Sun for a few weeks — an eclipse season, with daily blackouts of up to an hour or so. The Moon photobombs occasionally too, drifting across the disc and taking a bite out of it. Anyone watching a solar system live view built on SDO frames will eventually hit a stretch where the star just disappears, and the cause is orbital geometry rather than a dead server.
A Telescope That Covers the Sun to See It
SOHO, a joint ESA and NASA mission, has been on station since 1996 at the first Lagrange point, roughly 1.5 million kilometres sunward of us, hanging permanently between Earth and the star. It was a two-year mission. Three decades later it is still working, having survived a spectacular near-loss in mid-1998 when a chain of bad ground commands sent it tumbling, silent and freezing, for weeks before controllers coaxed it back.
Its coronagraphs, LASCO C2 and C3, see by hiding things. A physical occulting disc blocks the blinding photosphere so the faint corona around it registers at all, along with whatever else wanders through the field: planets, background stars, and comets.
The comets are the good part. LASCO frames go public fast and in bulk, and for years amateurs have combed them looking for small bright streaks falling sunward. More than five thousand comets have turned up that way, the majority found by volunteers who never touched a telescope. That is what an open feed actually does — it hands the discovery to whoever bothers to look.
The Whole Daylit Earth, About a Day Late
DSCOVR launched in February 2015 and parked at that same L1 point, facing the other way. Its Earth Polychromatic Imaging Camera stares back at the planet and photographs the entire sunlit face in ten narrow colour bands, ten to twenty times a day. Nothing else does this. Every other “whole Earth” picture you have seen is either a mosaic stitched from low orbit or an Apollo photograph from the early 1970s. EPIC has caught things no mosaic could: in July 2015 it recorded the Moon crossing in front of Earth, showing the lunar far side lit up against the Pacific.
The pictures are also stale by the time you see them. NASA posts them roughly twelve to thirty-six hours after capture, once frames have been downlinked, sorted and colour-processed.
Here is the odd thing. The same spacecraft carries the most time-critical instruments in the space weather system — a Faraday cup and a magnetometer sampling the solar wind before it reaches us. Wind measured at L1 arrives at Earth somewhere between fifteen minutes and an hour later, depending on its speed, and that gap is the entire practical warning time for a geomagnetic storm. One bus, two data streams, two completely different clocks: one measured in hours, the other in minutes.
It fails, too. DSCOVR sat in safe mode for most of nine months across 2019 and early 2020 with an attitude control problem, and the Earth pictures simply stopped.
Alerts That Arrive After the Event
The X-ray sensors aboard NOAA’s GOES satellites watch the Sun continuously in a fixed band, and that flux number is what defines a flare’s class, from A through M and X. When an X-class flare goes off, its X-rays and ultraviolet light cross 150 million kilometres in eight minutes and twenty seconds — which is also how long the detection takes, because the detection is the light. Shortwave radio on the daylit side of the planet is already blacking out when the alert publishes. Calling that a warning is generous.
The forecastable part is slower and messier. A coronal mass ejection, the billion-tonne cloud of magnetised plasma a big flare often throws off, takes one to three days to cross the same distance, which is why the Space Weather Prediction Center can put out geomagnetic storm watches with real lead time. In May 2024 conditions hit G5 on the NOAA scale, the first time since October 2003, and aurora showed up over latitudes that essentially never see them.
Even then, arrival estimates were off by hours. The detail that decides whether a storm is severe — whether the cloud’s magnetic field points south and couples to Earth’s own — is not reliably known until the plasma sweeps past the L1 monitors, under an hour out. Nobody has solved this. It is a measurement problem, not a modelling one, and it is the honest reason space weather forecasting still carries error bars that would embarrass a meteorologist.
The Window That Keeps Going Blank
Then there is the International Space Station, whose external cameras produce the one part of any solar system live view that a viewer would recognise as television. It is live within a few seconds of encoding and buffering. It also breaks constantly, and the breakages have good explanations. The station laps the planet every ninety minutes, which means sixteen sunrises a day and roughly half of every orbit spent in darkness where the camera sees nothing. Video comes down over Ku-band through a network of relay satellites in much higher orbits, and when a handover between them slips you get the grey card and the line about signal acquisition. The earlier experiment, HDEV, was four commercial off-the-shelf cameras sealed in a box and bolted to the European lab module in 2014; it ran until August 2019, when it stopped returning data for good.
Where the Public Feeds Actually Come From
Almost everything described here is public, and most of it flows through a handful of channels. Helioviewer serves compressed browsable versions of SDO, SOHO and STEREO imagery. NOAA’s forecast office publishes its alerts and solar wind readings as plain machine-readable files that anyone can poll. NASA’s community coordinated modelling centre keeps a catalogue of flares, CMEs and their modelled arrival times. The apps and dashboards you have seen are, with very few exceptions, front ends over these same feeds.
Latency comes from three places, and they stack. Geometry and the speed of light set a floor. Contact time adds more — a spacecraft can only send when an antenna is listening, and the Deep Space Network is oversubscribed. Processing adds the rest, since quick-look frames get replaced weeks or months later by calibrated science versions with the instrument’s own quirks removed. Plenty of missions never stream at all: Parker Solar Probe records and dumps, so its record-setting approach in December 2024, which took it within about six million kilometres of the solar surface, was confirmed first by a bare beacon tone with the data following much later. Solar Orbiter sends in bursts. Voyager 1’s signals now take close to a day each way.
So when a solar system live view labels a panel “now”, the useful question is which now it means.
The Same Problem, Scaled Up
Everything above matters for the search for life too, because that field runs on exactly the same tension between fast and checked. Breakthrough Listen has released petabytes of raw observations from Green Bank and Parkes into public archives, which is roughly the SOHO comet strategy applied to radio astronomy. Its most famous candidate, BLC1, was recorded at Parkes in 2019 while the telescope was pointed at Proxima Centauri, noticed in the archive a year later, and finally traced to human-made interference in 2021.
None of that was live. It could not have been, and pretending otherwise is how bad claims get made.
If you enjoy that kind of caution — real feeds, real delays, and a clear line between what an instrument saw and what it means — you will find plenty to dig into at SETIworld, where we follow solar observatories, planetary missions and the long search for company in the universe. Come and read alongside us.
