Ask Someone You Trust To Check This

Someone is running an experiment on the sky. Nobody is writing down the results.

Page class: PUBLIC CASE. It carries no figure that is not on a Results page. Every claim is typed — see the ontology.

Imagine a factory near your town announced the following.

We are going to release 210 tonnes of vaporised aluminium into the air every year. That is about ten times what nature puts there. We will also release copper, lithium, lead, niobium, hafnium and about fifteen other metals. We do not know what any of it does. We are not going to measure it. And we would like permission to increase it fiftyfold.

That factory would be shut by the end of the week. Every parent within thirty miles would be at the council meeting.

This is happening. It is happening about twenty kilometres above your head, which is the only reason there is no council meeting.

---

What is actually up there

In February and March of 2023, a NASA aircraft flew out of Fairbanks, Alaska, into the stratosphere, carrying an instrument that vaporises individual specks of dust with a laser and reads what they were made of. One particle at a time. It read more than five hundred thousand of them.

About one in ten of the larger particles up there is now part satellite.

Aluminium, mostly. Also niobium and hafnium, which come from rocket nozzles, and silver, which the researchers think comes from electronics. Over twenty elements, in the exact proportions you find in spacecraft alloys. There is no natural process that makes those ratios. It is not ambiguous where it came from.

This is not a forecast. It is a measurement, published in the Proceedings of the National Academy of Sciences, and it describes air that exists now.

The same paper says that within a few decades, that one-in-ten could become one in two.

And then it says — in the authors' own words — that the effect of this on the atmosphere is unknown.

---

Why that last sentence should stop you

The stratosphere is not scenery. It is the layer that holds the ozone.

Ozone is the reason you can stand outside for an hour instead of twenty minutes. It is the reason wheat and rice yield what they yield. It is the reason plankton survive in the top few metres of the ocean, and plankton make roughly half the oxygen in the breath you are taking right now.

We spent thirty years and an international treaty repairing that layer after refrigerants damaged it. It worked. A paper in Nature last year confirmed the ozone layer is genuinely recovering. It is one of the few environmental things humanity has actually fixed.

And into that same layer, we are now putting a growing quantity of vaporised metal, and the honest scientific answer to what does it do is: nobody knows.

Five research groups have tried to model it. One found it makes the ozone hole slightly worse. One found it makes it better. One found a small increase in ozone. One found the damage comes from a completely different chemical. One group ran a single model with one set of inputs, and got opposite answers depending on which standard definition of "warming" it used.

That is not a scientific consensus being ignored. That is a scientific question that has not been answered, about the layer of air that stops you burning.

---

So what could actually happen

"Nobody knows" is where most articles stop. It is not where you have to stop, and it is not what a scientist would do.

When you cannot measure a thing directly, you take every published result, you run each one forward to the scale that is actually being proposed, and you look at the spread. You do not pick a winner. You lay out the envelope — and the size of the envelope tells you how much you are gambling.

Of the five groups, one published a number for global ozone loss across the whole planet: Vliex and colleagues, using GEOS-Chem, modelling the satellites coming down today. Their answer for the current rate is a loss of about 86 mDU. Total column ozone is around 300,000 mDU, so today that is three hundredths of one percent. Nothing. You would never find it in the noise.

Now run their own number forward to what has been filed for:

what is on the table	how much bigger than today	ozone loss it implies	as % of the ozone column
what is happening now	1×	86 mDU	0.03%
what is already approved (19,408 satellites)	9×	771 mDU	0.26%
the 100,000-satellite application	46×	3,973 mDU	1.3%
the million-satellite application	377×	32,302 mDU	10.8%

To read that last row you need one comparison. At the very worst of the CFC crisis — the aerosol cans, the refrigerants, the thing that got a global treaty and thirty years of repair — the global average ozone column dropped by about 5%.

The bottom row of that table is roughly twice that. Permanently, because it is a tap left running rather than a chemical being phased out.

That is not a prediction, and we will not dress it up as one. It is one group's published mechanism, multiplied. Three things could each make it wrong:

• The multiplication is almost certainly too simple. Chemistry like this saturates — as the sky fills with particles, each additional one does less. The real curve bends. It could bend a lot.
• The other four groups disagree, one of them in the opposite direction. Run Revell's result forward instead and you project ozone going up. Run Maloney's and you project a weaker ozone hole. Both are peer-reviewed. Neither is fringe.
• These are applications, not approvals. 19,408 satellites are authorised. The 100,000 and the million are requests that may be refused.

Which is exactly the point.

Look at the honest range for the million-satellite case: somewhere between ozone slightly improving and twice the worst damage in the history of the ozone layer.

That is the width of what we do not know. Not a small uncertainty around a comfortable answer — a spread that runs from "fine" to "twice the CFC crisis", on a question about the layer of air that stops ultraviolet light reaching your skin, your crops, and the plankton that makes half your oxygen.

You would not fly on an aircraft whose engineers disagreed that much about whether the wings stay on. You would ground it and make somebody measure.

---

It does not start when they fall. It starts in a mine.

Everything above counts satellites burning up. That is the easy half to count, and it is not the whole system.

A satellite's footprint begins long before it re-enters. Aluminium and silicon and copper and lithium have to be mined and refined. Wafers get fabricated. Kerosene gets refined and oxygen gets liquefied. Then a rocket burns most of that propellant on the way up, through the same stratosphere, and only then does anything reach orbit at all.

And here is the uncomfortable part for our own argument: the scientific literature currently blames the way up more than the way down.

Barker and colleagues, modelling all space activity in 2026, found ozone loss dominated by chlorine from solid rocket propellant — a launch emission. In their analysis satellite mega-constellations account for only about 9% of the damage, and specifically because they fly kerosene rockets, which emit no chlorine at all.

So: on that particular pathway, SpaceX's choice of fuel is better than the alternative. We are saying that plainly, in the middle of a page arguing they should be measured, because it is true and because a page that only ever produces bad news about one company is not evidence, it is advocacy.

What it means for the rest of this page is a limit on our own work. Our ledger counts the stage that is easiest to count and not the stage the science currently blames most. That is a gap in our accounting, and closing it needs a per-launch propellant and emissions inventory we have not built yet.

But the direction it points is not reassuring. It means the atmospheric bill for a constellation is larger than the re-entry figures on this page, not smaller. We have been under-counting.

---

What this actually looks like: a slope, not a cliff

Nobody should tell you there is a date.

There is no year when something snaps, no threshold that gets crossed, no moment when an alarm goes off — and that is precisely the problem, not a reason for comfort.

Here is the actual shape of it:

• The amount goes up — from nothing to nearly half of everything entering the atmosphere, in six years.
• What goes up stays for one to four years before settling out.
• The source is continuous, because satellites are replaced as they die. It is not a spill. It is a tap.
• And the effect is not measured, so nothing ever triggers.

A cliff would at least get noticed. Somebody would see the number cross the line.

There is no line. Nobody wrote one. That is the finding.

So it does not arrive as a catastrophe. It arrives the way real environmental damage almost always arrives — as a slow change in something you assumed was fixed, noticed late, in data somebody happened to still be collecting. Thinner ozone means more ultraviolet at the surface. More ultraviolet means measurably lower yields in wheat and rice, and it means damage to the plankton in the top few metres of the ocean, which is where roughly half the world's oxygen is made and where the marine food chain starts.

We are not telling you that is happening. We are telling you it is the thing at risk, that the published estimates of the risk range from slightly positive to twice the worst of the CFC era, and that no one is required to find out which.

Not in your lifetime, not in your children's. Not on any schedule at all. Because there is no schedule.

---

The part that should make you angry

Here is the entire legal requirement, in the world, for a company burning satellites in the atmosphere. This is the complete text of the rule:

"The number of satellites that re-entered the atmosphere."

Twice a year, they file a number. How many. That is all.

Not how much it weighed. Not what it was made of. Not what altitude it burned at. Not one measurement of the air.

It gets worse when you look at why. Under the law, a "re-entering vehicle" means something built to come back in one piece. A satellite designed to burn up is not one — so the environmental review that would normally apply simply never applies. The category doesn't fit, so nobody has to look.

When the American aviation regulator wrote its report to Congress about exactly this — satellites from large constellations burning up in the atmosphere — the words ozone, stratosphere, and air quality do not appear anywhere in it. Not once. The entire report is about whether a piece might land on somebody.

And right now, today, there are two live proposals: one to exempt space operations from environmental review altogether on the grounds that space is outside anyone's territory, and one to waive thirteen environmental laws for launch companies — including the Clean Air Act.

So: the quantity is rising, the science is unresolved, and the rules that might one day have required someone to check are being taken apart.

---

How fast this is moving

In 2019, one company accounted for none of the mass falling into Earth's atmosphere.

In 2025, it accounted for 45.8% of all catalogued human-made reentering mass. Nearly half — from zero, in six years. (An earlier version of this sentence said "natural and human-made combined," which was wrong by our own arithmetic: against the natural meteoric flux, all human injection together is currently 3 to 6 percent. The correction is dated 2026-09-01, and it is exactly the kind our validation harness exists to force.)

Six years.

And there is now an application in front of the American regulator for a constellation of one million satellites, to run data centres in orbit. By the company's own filing, about forty thousand of them would burn up in the atmosphere every year, permanently, as they are replaced by newer ones.

For context on that number: the entire natural stratospheric aerosol layer — everything, all of it, built up over all of history — weighs a few hundred thousand tonnes. A constellation at that scale would put a comparable quantity through the stratosphere and keep it there, continuously, forever.

Mount Pinatubo erupted in 1991 and put roughly thirty to sixty times the normal amount of material into the stratosphere. It cooled the entire planet by about half a degree for two years. Everyone noticed. It was in every newspaper.

That was one eruption, decaying away. This would be a tap left running.

---

What we are not telling you

We are not telling you the sky is falling. We think you have been lied to enough by people who needed you frightened.

We are not saying anyone will die. We are not saying the ozone layer is collapsing — we just told you it is recovering, and linked the paper. We are not saying the scientists who found "no effect" are corrupt or wrong. We do not know that they are wrong. That is the point.

We ran our own frightening calculation. It produced a specific year by which the sky would be, in a technical sense, full. We checked it, found that one of the numbers in it was wrong in a way that would have made the same maths say the sky was already full in 1850, and we threw it out. You can read the whole thing, including the version we deleted, on the evidence page.

We would rather hand you something a hostile expert cannot break than something that makes you afraid for an afternoon.

What we are saying is narrower and, we think, worse:

You cannot fix what you have decided not to measure. And every decision being made right now is a decision to keep not measuring, while the amount goes up.

---

Why this is about children, specifically

Not because anyone is in danger this year. Because of how the clocks line up.

Material put into the stratosphere stays there for one to four years before it drifts down. A satellite constellation is planned in decades. A regulatory exemption lasts until somebody repeals it, which nobody ever does. And if there is an effect, it will show up in the data long after the thing causing it has become normal, profitable, and employing thousands of people in someone's district.

There is a window where measuring is cheap and stopping is easy. We are in it. It does not close with an announcement. It closes because everybody assumed this was somebody else's department.

That is how almost every large irreversible thing has ever happened. Not by a decision. By nobody's decision.

Your children will not inherit a verdict on this. They will inherit whichever answer turned out to be true, plus the fact that nobody bothered to find out while it was still cheap to.

---

We did not just complain. We built the instrument.

Everything above says: nobody is measuring. It would be cheap to leave it there. Complaint is free, and the people making it are usually selling something.

So here is what we did instead. We took every axis of this problem — the mass coming down, the ozone column, the heat balance, the biosphere's own couplings — and put them in one account, computed in exact integers, running live on nine independent machines. Not a paper. Not a model you have to take on faith. A running court you can question directly, this minute, with no account and no key.

Ask it about the fleet that is already approved — not the speculative one, the one with licences — and it answers:

Three separate planetary accounts move at once. The metal falling from orbit reaches 13 to 26 percent of everything nature delivers to the stratosphere. Regional ozone moves 3 percent. The heat balance moves 1 percent of everything humanity's greenhouse gases have ever added. And no instrument on Earth is required to report any of the three.

Each of those alone looks survivable, and a reader shown any one of them would reasonably move on. That is precisely how this has stayed invisible: every published study covers one axis, and the biosphere column is empty in all five. No two of them even share enough axes to be added together. The joint number did not exist anywhere until it was computed here — not because it was hard, but because no funding line spans four disciplines.

And when we push the court past the filed million-satellite scale, it does something a model built to frighten you would never do: it disqualifies itself. The projection exceeds the physical ozone column, and instead of printing an impossible number, the court answers PROJECTION_DISQUALIFIED and refuses to print a value at all. Where the biology is unmeasured, it answers NOT_KNOWN and leaves it out of the sum — which means every total it gives you is a floor.

An instrument that can refuse is an instrument you can trust when it doesn't.

The researcher you send this to can replay the whole thing in one call — the address, the tool and the four admitted scenarios are on the study page, together with the raw data the baselines come from: 56,575 daily ozone readings from five stations between the Arctic and the South Pole, and 46 years of measured greenhouse forcing. Every value an exact integer. Every grade printed on the wire.

---

What to do, which takes one message

Send this to one person who reads scientific papers for a living. An atmospheric chemist. A physicist. A PhD student who would genuinely enjoy proving strangers wrong on the internet.

Ask them to check four things:

1. Murphy et al. 2023, PNAS 120(43) e2313374120. Is it really one in ten? Does it really say 210 tonnes of spacecraft aluminium against about 20 from meteors?
2. The five modelling papers. Do they really disagree about the direction of the effect? All five are cited on the evidence page.
3. The rule. Is the only requirement really just a count of satellites?
4. Our arithmetic. One command, no account, no permission:

git clone https://github.com/gaiaftcl-sudo/uum8dSolarResearch.git
cd uum8dSolarResearch && bash reproduce/validate.sh

That script re-derives every number on our pages from the original data and fails loudly if any of them is wrong. It has caught us repeatedly — including twice in one day: a headline quoted to nine significant figures when three quarters of the underlying mass carries the catalogue's own estimate flag, and a unit error that printed 437% where the true figure was 43.7%. Both corrections are published, dated, with the wrong number named.

5. The court itself. Ask the running system, not us — the live call and its four admitted scenarios are on the study page. Nine machines answer with the same integers, and the response tells you which parts are measured, which are projections, and which are simply not known.

If they find a mistake, tell us and we will publish it with our name on it. We have already done that several times, including once where we published a number five times too large.

And if it holds up, ask them the only question that matters:

Why is nobody required to measure this?

That question does not need the science settled. It gets stronger the more the scientists disagree — because disagreement is the reason to measure, not the reason to wait.

---

The whole ask, in one sentence

Nobody is asking anyone to stop launching. Weigh what comes down, write down what it was made of, and stop dismantling the only rules that could ever make someone do it.

That costs almost nothing. It requires no one to be proved right. It just requires somebody to look.

Right now, by law, nobody has to.

---

Every figure above, every citation, every caveat, and every number we withdrew: the evidence page.