In the designs for a permanent lunar settlement, nearly everything is specified. There are drawings for pressure modules, for reactors, for rovers, for airlocks and for growth chambers. What barely appears in those drawings is the surface all of it is meant to stand on.
This is not an oversight by the engineers. It is a habit we carry with us from Earth. Here, ground is a precondition you investigate once and then tick off. A soil report is produced, a foundation is sized, concrete is poured. After that nobody thinks about what is underneath, as long as nothing unusual happens.
On the Moon that sequence does not hold. There the ground is not what gets settled before building. It is what gets consumed during operations, and consumed by exactly the activities that keep the settlement alive. Every landing that brings supplies alters the surface the supplies are set down on.
I call this the bearing term. It means the span of time for which a demonstration of load capacity still holds. On Earth that span is long enough that we are allowed to ignore it. On the Moon it is shorter than the interval between two approaches.
Ground That Does Not Recover
Lunar soil is not sand. It is shattered rock, ground down by impacts over billions of years, with no water and no wind to round the edges. The grains are sharp and they interlock. That interlocking is the reason the loose upper layer carries anything at all.
Load capacity therefore rests on a fabric that was assembled mechanically and can be destroyed mechanically. On Earth, processes go to work after every disturbance to consolidate soil again. Rain washes fines into voids, freeze and thaw rearrange the grain skeleton, roots run through the subsurface, traffic compacts it. A dirt track torn up by a heavy transport is a dirt track again a year later.
The Moon has none of those processes. What is loosened stays loosened. What is removed does not come back. The only force that redistributes material there over the long run is impact, and impacts work on geological timescales rather than operational ones.
Then there is the lower gravity. One sixth of Earth's pull means that any grain once freed travels much further before it settles again. It also means the overburden stress holding the soil together is smaller. The ground forgives less and it does not heal.
What an Engine Does to a Surface
How severe this loading is not a matter of speculation. We have a measurement, and it is more than fifty years old.
When Apollo 12 touched down in November 1969, the Surveyor 3 probe stood roughly one hundred and sixty metres away, having landed two and a half years earlier. The crew walked over and cut pieces out of it that came back to Earth. Examination showed that the probe's surfaces had been struck by material thrown up by the lander at touchdown. The descent engine produced thrust on the order of twenty kilonewtons. By today's standards that is a small device.
The reason this effect reaches so far lies in the absence of an atmosphere. On Earth, air slows every disturbed grain within a few metres. On the Moon nothing slows it. A grain accelerated by an engine follows a ballistic path until it hits something. Models for the landers now being planned assume ejection speeds of up to roughly 2,380 metres per second. That is precisely lunar escape velocity. Some of the material leaves the body altogether.
The ejection angle is almost always below ten degrees, and in the Apollo landings mostly below five. The stream does not go upward, it goes flat across the plain, like a shotgun fired at knee height. Anything standing in that plane is in the field of fire.
Since last year there has also been a direct measurement. When the Blue Ghost lander set down in Mare Crisium in March 2025, a camera array recorded the event. The imagery begins roughly twenty-eight metres above the ground. At around fifteen metres the interaction between the control thrusters and the surface becomes visible. From there so much material is thrown up that the view is lost for a time. Four further cameras surveyed the surface after touchdown to determine how much ground had been displaced.
So the erosion is now being recorded. What follows from the recording is open.
The landers now in question are considerably larger. A Starship-class vehicle stands around fifty metres tall. The manufacturer has provided high-mounted thrusters for the final phase of descent specifically so that the main plume does not strike the surface directly. NASA is separately running trials in a vacuum chamber roughly eighteen metres across, firing an engine into a bed of lunar soil simulant. The setup is described as the most complex of its kind ever run in a vacuum chamber.
You do not run tests like that when you already know the answer.
The Bearing Term
Something follows from these two findings that appears as a line item in no mission plan.
If the ground does not regenerate, and if every landing strips it across an area, then load capacity on the Moon is not a property a surface has. It is a demonstration made at a point in time, which then ages. Between that demonstration and the next landing it holds. After the next landing, whether it still holds is an open question.
This sounds like a technical nicety. It is the difference between a settlement and a depot. A depot can sit on a surface whose condition nobody tracks, because in the worst case it is allowed to be lost. An inhabited facility cannot. Its foundations, its lines, its sealing faces and its anchorages all assume that nothing beneath them moves.
A differential settlement of a few centimetres across the length of a pressure module is damage on Earth. In an environment where the inner shell is the only difference between habitation and death, it is something else.
That creates a task barely recognised as a profession here, because our terms are so long: someone has to demonstrate the state of the surface continuously. Not once before construction, but after every event that changed it. Whoever takes that on effectively decides when landings may resume and where. This is not an administrative question. It is an operational one that follows from the physics of the place.
The Test Is the Ordinary Day
Here it is worth taking a second look at a facility that appears in every settlement plan and is almost always sized wrongly.
A lunar clinic is usually conceived from the emergency backwards. You assume an accident, a certain number of casualties, a stock of materials. What comes out of that is a storeroom with a bed in it.
I regard this as the decisive design error. A clinic that exists only for emergencies does not work in an emergency. Medical equipment sitting unused for months near vacuum and through thermal cycling is of unknown condition at the moment it is needed. Personnel who do not rehearse the procedures perform them for the first time under time pressure. Only a station operated between emergencies is a station during one.
And that is exactly where the surface comes in. An operation that cannot be interrupted needs continuous power from a ridge others also draw on. It needs a driveable connection to the landing pads on which a casualty arrives in minutes rather than hours. It needs a standing rule on admission. And it needs the road to still be a road after the last landing.
This is why the clinic is the best test of any settlement design. It reveals immediately whether a design treats a common substructure as one or merely mentions it. Anyone who sizes the infirmary under the same flag they sized the habitat under has not grasped that the second casualty comes from wherever he comes from.
The Reserve Beneath the Same Surface
There is a second layer under the surface, and it is the reason the south pole is a candidate for settlement in the first place.
Permanently shadowed crater floors hold temperatures between roughly twenty and one hundred and twenty kelvin. That is cold enough that volatiles barely sublimate over billions of years. Those floors are cold traps, accumulating what comets and meteorites have delivered across the entire history of the Earth-Moon system. Water is part of it, and probably organic compounds as well.
For science this is the only undisturbed record of its kind in the inner solar system. For settlement operations it is the only place to obtain water without lifting it up out of Earth's gravity well. Archive and resource do not sit side by side. They are the same material.
Damage does not begin with extraction. A committee of the US national academies worked through the effect of lander exhaust years ago. Combustion products, water among them, disperse through the thin exosphere and find their way into the cold traps. A substantial share of the water released ends up in precisely those shadowed areas whose water content was to be measured. Measure after landing and you are measuring your own propulsion to an unknown degree.
Heated extraction is then the second step. Within the volume it processes, it erases the layering, the isotopic signature and the provenance. Afterwards the material is still there as water. As a record it is gone.
A reporting duty does exist. Under the most recent version of the international planetary protection rules, any mission approaching shadowed areas or the polar regions must submit a complete inventory of its organic materials. What that captures is what a vehicle brings with it. What it does not capture is the order in which things are measured and consumed. The same rules state plainly that they do not protect the Moon for its own sake.
So nobody decides the sequence. It falls out of who arrives first.
The Release That Happens Once
Three decisions are therefore in play that never present themselves as decisions.
Whether a surface is released for permanent operations, whether a facility is built as an operating installation or as a stockpile, and whether a crater floor is measured or consumed. None of the three is reversible the first time. And none of them is currently answered by anyone appointed to answer it.
Over recent weeks I have spoken about the Moon with specialists in geotechnics, architecture and astrobiology, and each of them pointed independently at the same gap from a different side. The geotechnical engineer asks who signs off the surface. The architect asks what a clinic lives on between emergencies. The astrobiologist asks what has been measured before anything is touched. It is the same question three times over: who owns the condition of the ground?
While nobody answers it, it still gets answered. A place where no one sets the order receives the order of arrival. This is not a governance problem to be supplied later. It is a physical outcome, fixed with the first heavy engine.
What Follows
The Moon is being discussed as a destination, but it is already a construction site. On a construction site one simple rule applies, familiar to anyone who has run one: you do not start with the building. You start with laying out the site, with the routes, with the sequence of trades, and with the question of who is answerable for the condition of the ground.
On Earth that role is so self-evident it never enters a debate. On the Moon it does not exist, and what separates the two cases is not a treaty. It is the absence of rain, wind and time, which here repair after the fact what nobody attended to.
What follows is unspectacular and therefore easy to miss. Before the first permanent structure stands, there has to be a body that accepts a surface and renews that acceptance after every approach. There have to be facilities that run in ordinary time rather than waiting for the emergency. And there has to be a measurement taken before the first extraction, because afterwards it can no longer be taken.
None of this calls for a new institution. All of it calls for someone to take responsibility before the surface decides the matter by being used up.
A question to @NASA, to @SpaceX, to @Blue Origin, to @Firefly and to the seventy states of the Artemis Accords: which of you says that nobody may land right now?
Website: https://planet-futures.org