✦ Kiri Gardens · A visitor brochure & science paper · Welcome in

Welcome to Kiri Gardens.

Shorts and Lumen — the orange-haired and pink-haired guides in their white suits, space eyes bright — standing in front of the amber cover with the grove inside, the collector lens standing apart on the rim, black mushrooms and pines outside. The art reads: Welcome to Moon Base!
Shorts & Lumen, your guides. That's them — space eyes, white suits, the whole base behind them: the amber cover with the grove inside, the collector lens standing apart on the rim, black mushrooms and pines outside. They live here. They'll keep this quick.

We live in a Moon crater, under a golden cover, among trees. Our place is made from mined sulphur. Ours has plants. That's our personality — and honestly, that's most of what you need to know. The rest of this brochure is how it works, the house rules, some pretty pictures, and — hanging quietly overhead, not hogging the page — the science.

⛏️ Made of mined sulphur🌿 Full of plants🧑‍🚀 Space Kids & schools at the centre🌈 Everybody belongs here✨ Glazed crystal, not grey stone
Your guide · The edge & the works

Shorts

“I work the sulphur plant and the sorted bays, out on the dusty side. Matter of fact: we took dust that traps wheels and turned it into a fortress. The clever temperatures? Leave those to the professionals — my job is that it works, every day.”

Your guide · Light & fibre

Lumen

“I tend the light — collected on the rim, carried down in fibre, glowing in the gardens. Matter of fact: sunlight, delivered. You don't need the maths to enjoy breakfast under it. If you do want the maths, it's all further down, sourced.”

The Kiri Gardens way   Clean and proud, and we don't pretend we reinvented the wheel. You already know what recycling is. You already know hydroponics and light systems. So we won't lecture you on them. Ours just works — because we appreciated every little element the challenge threw at us: the dust, the ice, the sulphur, the dark. Eigenvectors? Like rocks in the creek. They sit there, they shape the flow, you step on them to cross. That's the level of fuss they get here.

And because this is also the science paper — the one we present to NASA — everything below the welcome is checked against real papers and NASA material, in story order from the start: a polar crater chosen for rim light and floor ice, a base built from its own sorted trash, water pulled from dark ice into a loop that reuses ~98% of it, Lewa in the middle — home, school and the jar libraries — gardens, fishponds, the air people and trees trade, and then the newest chapters: the glazed cover, the rim lens and glass eye, the living sheath, and the eagle suit. What works, we keep. What doesn't, we abandon or redesign — never soften.

In this brochure — in story order

Welcome · How it works · The rules · The art · The science

  1. Founding — trash becomes treasure; water, one train from dark ice to the closed loop; and how the cover is built: fibre-pole skeleton first, every bag stacked, one melt.
  2. The base in cross-section — the crater, Lewa, houses, terraces, the ice vault, and the pressure numbers that decide the cover.
  3. Light & the cover's Lewa — the rim collector, fibre-delivered light, the glass eye and its radiation garden.
  4. The people — the Space Kids, teachers, Kiri, Ellie, Lumen and Shorts, each in their place.
  5. The living sheath — the tree skin tested part by part, and mature trees moved outside.
  6. The ahupuaʻa — rim to Lewa to ice basement: how air, water and waste flow downhill through the zones.
  7. Tanks, chemicals & waste — sorted — every stream labelled, thresholded, and kept at the edge to be mined later.
  8. Ahupuaʻa inside ahupuaʻa — every room and garden indexed, with its own quarantine boundary.
  9. Life under the cover — the eagle suit and the libraries of biology.
  10. Verdicts in full — each founding idea judged on its own evidence: kept, redesigned, or abandoned.
Verdict · The cover chapter · Keep — and keep studying
The cap itself: pressure skin + fused regolith sheet

“Paraterraforming” / the worldhouse — proposed in print since 1992, structurally modelled for a crater lid in 2025. A fused regolith sheet over a regolith-fibre pole skeleton — shell action, the embedded ribs and the anchor ring hold the lens shape against the pressure; the sheet shields radiation and its mass helps hold the pressure down — everywhere except one small glass eye at the apex, the cover's Lewa. Unproven at our scale, but the physics is honest and the problems are named, not hidden.

Verdict · The sheath chapter · Keep — with one load-bearing redesign
The sheath: real parts, one part contradicted as stated

Pressure skin, root pouch, channel networks and self-repair all exist and are tested. Feeding leaves gas through a contact material does not — leaves breathe through stomata into air. Give the sheath a thin gas gap inside and it becomes a legitimate programme: a conformal inflatable greenhouse, not a membrane feeder. Full verdicts below.

Here's how it works

Four things. That's the whole tour. Shorts and Lumen will handle the rest if you ask them nicely.

🌞

1 · Light comes in by fibre

A lens on the sunny rim gathers the low polar light, sorts it, and sends it down fibre trunks. Inside, the gardens glow amber. No raw sky falls on anyone — except in one tiny glass eye, on purpose, for an experiment.

💧

2 · Water goes round

Ice from the dark floor, cleaned up at the edge, then round and round the base. We reuse almost all of it. The fishponds — the loko iʻa — are the living seal that checks it on the way back in.

♻️

3 · Nothing is “away”

You know recycling. Ours is just thorough: everything is sorted at the edge into five tidy bays, and most of it becomes something else — building, fertiliser, fuel, treasure for later. Dust included. Especially dust.

🌳

4 · Plants do half the work

People breathe out, trees breathe in, and they swap. Gardens, forest terraces and school plots aren't decoration here — they're crew. Botany is a founding principle, not a hobby corner.

~1 km
across — one crater, one neighbourhood
~98%
of our water, reused (demonstrated 2023)
10.3 tf/m²
of push on the cover at Earth pressure — yes, we planned for it
80–96%
of the time, the rim is in sunlight
35–100 K
on the dark floor — where the ice keeps

Those are the megalithic numbers. They hang over the place the way the cover does — holding everything up, not getting in anyone's way. Every one of them is sourced in the science sections below.

Here are the rules

Visitor rules, which are just the base's rules said friendly. They keep the middle clean, the air sweet, and everybody — kids first — safe.

🧑‍🚀 Suits outside, smiles inside

Outside the cover it's vacuum — suit on, no exceptions. Inside, you're indoors even when you're outdoors. That's the treat of the place.

🚪 Everything through the lock

People, plants and parcels cross at the locks, and dusty things get cleaned there. Nobody carries dust, brine or gas through a clean corridor. Ever.

🏫 The middle is for living

Lewa — home, school, the jar libraries — stores nothing hazardous. Dirty work lives at the edge with Shorts. Clean in the centre, busy at the rim: that's the map and the rule in one.

➡️ Waste flows out, only out

Follow the arrows downhill, rim to pond. If you're carrying something the base is done with, you're walking outward. Ellie at the ponds decides what may come back in.

🌱 Ask the garden first

Every garden and room is indexed and keeps its own boundaries. Want something from a bed? Ask — the Teachers log it, the channels deliver it to the right address. No picking at random.

🌳 Trees graduate, they don't get thrown out

A mature tree gets fitted with its living sheath indoors, tested, then walked out by Kiri to stand in the sun. It's a ceremony. Visitors are welcome to watch — from inside.

Here's some pretty art — the scenes themselves

The actual pictures: the grove, the crater before and after its cover, the film stills, the plan, the region, and the tree the sheath was designed for — shown big, because that's what they're for. Notice the glaze: furnishings, fittings, the cover itself — molten sulphur glass, warm and glossy, like the orange juice in the kids' glasses catching the light.

Inside the cover: glass domes among gardens, a waterfall, koi ponds and a little bridge, with kids in flight suits carrying jars on their rounds
The grove, from insideDomes, gardens, the waterfall and the koi ponds under the great cover — kids mid-rounds with their jars. This is what “a garden with a town in it” looks like.
The crater before dressing: a bare amber bowl with sulphur piles, rim works and the first green trial plots
Before — the bare craterAmber walls, sulphur piles, rim works just starting. The crater does the work; this is it, undressed.
The same crater dressed: green grove, river and ponds under the golden cover, rim beams firing light down from the collector
After — the dressed craterSame bowl, green under its golden cover, the river running, rim beams firing the collected light down. Matter of fact: it greens up.
Reference-01: the sulphur-glass breakfast room — amber glazed tables, spiral stair and plants everywhere — a girl eating cereal while a boy sips orange juice
Reference-01 — breakfast in the glazeThe sulphur-glass room, the boy mid-sip of orange juice. This picture is the finish standard for the whole base: everything glazed like this.
Three young astronauts on the crater rim at night, one holding a small glowing Earth, the real Earth huge in the starry sky above
The rim trioOur guides and a friend on the rim with the little glowing Earth. Home is right there — you can see it from work.
The trio from behind, looking up at Earth as an eagle crosses in front of it above the crater rim
The eagle crossingEarth overhead, the eagle on its glide. Under the cover there's air to fly in — that's the whole trick of the place.
Four young astronauts in suits around a paper map of Tycho Crater spread on the rocks, one looking through a brass telescope
SurveyorsThe map comes out, the telescope goes up, somebody points. Kiri Base Station, Earthview Post, Observatory Ridge — all walkable from here.
The compound from above: five golden domes linked by ring corridors, terraced gardens and ponds across the crater floor
The compound from aboveGolden domes on their ring, terraces round the walls, ponds on the floor. Loved things in the middle — you can read the old map right off it.
Annotated radial plan of the base: Lewa the wild rim, Māla the gardens, light houses, the Light Hub bag dome at the centre, the lens apart, black mushrooms outside
The plan — the map of the baseLewa on the wild rim, Māla the gardens, the Light Hub at the centre, the lens standing apart, black mushrooms outside. Everything has an address.
The whole region from above: the covered crater at the centre, reservoir, ice mine, a second covered crater, giant sheath trees outside, light lines stitching it together
The region, wideReservoir, ice mine, the second crater under its own lens, the giants standing outside — and the light lines stitching it all together.
Species plate: a tall ponderosa pine in pixel detail, plated orange bark, needle clusters and cones
The ponderosa — species plateThe sheath tree, drawn properly: plated bark, long needles, cones. Raised inside, graduated outside in its living sheath.

✨ The look is canon: glazed crystal, everywhere

Say it once, so every drawing, chart, film and model from here on agrees: this base does not look like stone. It looks like glazed crystal. The cover, the furnishings, the collector, the fittings — everything the sulphur touches carries the same sleek sulphur glaze: honey-amber, glossy, light caught in it. Dust went in dull and came out shining. That is the genesis look, the science look and the art style, in one sentence — the molten slag-glass moon base, presented to NASA exactly as it is: bright, warm, and full of the most charming crew in their suits, walking around in it like they own the place. They do.

Diversity is canon too, not decoration: different kids, different gardeners, different jobs, one watershed. The schools and the botany aren't a chapter of the base. They are the base.

Founding — what the base is made of, and how it was raised

Shorts, matter of fact: “Three facts, then we're done explaining and you can go look at the gardens.” Three facts come first — before every verdict — and each is told once, here. Terms, fixed for the whole page: the cover is the whole closure — pressure skin + fused sheet + glass eye; the sheet is the fused regolith layer alone; the skeleton is the pole armature inside it; the collector is the separate lens on the rim. Nothing arrives from Earth except people, seed stock and first tools — the base is built from what the Moon and its own waste already hold.

Founding I · Trash becomes treasure — the base is built from waste

Doctrine: real practice

There is no “away” on the Moon. Every output of every process is sorted at the edge, labelled, and kept as feedstock — the waste stream is the supply chain. The full stream-by-stream table lives once, in the inventory; the doctrine is this:

  • Sorted, never mixed. Five bays at the edge — oxidiser, fuel, toxic, inert, biological — because a mixed pile destroys both safety and value. Reference art: the five bays and the plants around them are drawn in the plan view under The people.
  • The big piles are building material. MRE slag and glass → fibre, aggregate and shielding blocks. Bulk tailings → radiation shielding by the metre, roadbed, landing pads. Sulphur fines → remelted directly into the next batch of sulphur concrete. Demolished sulphur-concrete blocks are feedstock again — the material remelts at ~120 °C.
  • The small piles are the scarce elements. Phosphate sludge — phosphorus has no lunar source yet demonstrated, so per kilogram it is arguably the most valuable pile on the base. Ice-plant ammonia → the greenhouse's nitrogen. End-of-life batteries, segregated by chemistry → cobalt, nickel, lithium, copper. Brine salts → N, P, K — only if the pretreat is non-chromium (the law in Founding II); with Cr(VI) pretreat the same pile is hazardous waste, not ore.
  • Biology eats the rot. The digester chain (MELiSSA architecture, nested section) turns faeces, inedible biomass and plant waste into fertiliser and fuel gas; pasteurised digestate goes back to the gardens. Outdoors, the black-fungus farm (Radiation Works) grows shield tiles and mats from mycelium composite — waste carbon becoming structure.
  • Nothing wasted is a rule with a ledger, not a slogan: every stream in the inventory carries a mine-later value, including the honest NONEs — hydrazine is destroyed, ethylene is scrubbed, raw faeces is never stockpiled. Treasure is what survives sorting; the rest is neutralised, not kept.

Founding II · Water — one train, from dark ice to closed loop

Flown hardware · lunar feedstock sampled once

Polar ice is a chemical feedstock, not a drink. It is mined in permanent darkness, melted and purified at the edge, and only polished water ever crosses inward. Told once, in order:

Step 1 · MineDark floor iceLCROSS (Cabeus): 5.6 ± 2.9 wt% water in icy regolith
Step 2 · MeltIce plant, at the edgeVolatiles captured as they come off — NH₃ kept, never vented
Step 3 · PurifyDistil + stripMercury is the problem passenger — removed, sealed, toxic bay
Step 4 · PolishWPA modelFiltration + catalytic oxidation + ion exchange (ISS pattern)
Step 5 · LoopUPA + brine processor85–90% → ~98% demonstrated 2023
  • The ice is dirty — measured. LCROSS abundances relative to water (=100): H₂S 16.75, NH₃ 6.03, SO₂ 3.19, ethylene 3.1, CO₂ 2.2, methanol 1.6, CH₄ 0.7. One impact, one crater — this base's own crater must be sampled before the numbers are relied on. Reference art: the ice vault is drawn in the cross-section below; the ice plant sits outside the seal in the plan view.
  • Mercury, stated exactly: detected in polar ice (LAMP); its tabulated abundance is uncertain and instrument-dependent. It is a purification problem with a known fix — sulphur-impregnated carbon removes it — held to NASA SMAC 0.01 mg/m³ in air and the EPA 2 ppb drinking-water limit. “Polar water is poisoned” and “mercury is negligible” are both unsupported; the page holds the middle line.
  • The loop's numbers, once: the ISS Urine Processor Assembly alone recovers 85–90% of its feed as water; adding the Brine Processor Assembly demonstrated ~98% total water recovery (June 2023) — a water-only milestone, not total closure. Humidity condensate and greywater rejoin the same polish step; greywater needs biological pretreatment first, because surfactant foaming breaks physicochemical processors.
  • The pretreat law: the flown pretreat (CrO₃ + acid) stops urea → NH₃ but leaves the brine Cr(VI)-bearing, pH ~2.0, tox-2. This base therefore specifies a non-chromium pretreat — or the MELiSSA biological route — so the brine salts stay mineable (Founding I) instead of becoming hazardous waste. The brine itself is a sealed process stream at the edge, handled by the plant, attached to no person's duties.

Sources: LCROSS — pubmed.ncbi.nlm.nih.gov/20966242/ · Brine / ARFTA & 98% demonstration — ntrs.nasa.gov/api/citations/20160009716/downloads/20160009716.pdf · ECLSS review, ICES-2025 — ntrs.nasa.gov/api/citations/20250003953/downloads/ICES-2025-125%20Final_Rev%20C.pdf · Mercury, OSHA — osha.gov/chemicaldata/505 · Full thresholds: the inventory table

Founding III · How the cover is built — skeleton first, one melt

Method: proposed · Materials: demonstrated parts

Loose bags cannot hold a lens shape while they wait to be fused — they slump. So nothing loose ever has to hold the shape: The skeleton holds it from minute one.

Step 1 · Cast

Poles from rock

In the melt shops: lunar rock is melted and drawn into basalt-type fibre, laid in a sulphur-glass matrix, and cast as long poles. The oxygen plant's byproduct iron (MRE) is cast for the anchor ring and fittings.

Step 2 · Raise

The eagle skeleton

Long poles go up radial by radial — one run per radial, tipi-style, leaning into the lens shape over the pressure skin, converging at the eye ring. The roof already exists as a frame before the first bag moves.

Step 3 · Stack

Every bag on the frame

Bags of regolith–sulphur mix are stacked directly on the armature, across the whole cover. The poles carry every load; nothing waits on anything curing, because nothing has been asked to cure yet.

Step 4 · Melt

One pass, 130–140 °C

A heater sweep bakes the whole cover once. Only the sulphur melts — it liquefies at 119 °C — flows through the mix, and sets on cooling as sulphur concrete: one solid shell, fused in place.

Step 5 · Keep

Ribs in the casting

The poles stay inside the sheet as ribs — rebar in the casting — stopping at the eye ring, where the glass eye sits in their converging collar. A crack is repaired by re-melting its sulphur seam shut.

  • Where the materials come from — not Earth, not trash. The poles are regolith fibre: same melt shops, same industry as the sheet itself. MRE iron is a byproduct of making oxygen. It is iron, not steel — steel wants carbon, and carbon is among the scarcest elements on the Moon; biosphere carbon is not spent on poles. Imported trash steel is seed stock at most, never a supply line.
  • The long game is wood. Paulownia — fastest-growing hardwood, pole-straight, light, and it coppices: cut the pole and the stump grows the next — is planted in Lewa from year one. The first cover cannot wait years for trees; every cover after it can be grown. Fibre poles build the first base; the grove builds its successors and its repairs.
  • Why one melt works: the bake never has to melt rock — 130–140 °C is the sulphur-concrete working range already used in the inventory, far below sintering temperatures, and the set is a freeze, not a cure: the shell is structural the moment it cools. Reference art: in the cross-section below, the cap's smooth band is this finished sheet — read the faint seams as the ghost of the bag courses and the ribs as the skeleton inside it; the film shows the poles going up before a single bag is stacked.

Sources: Sulphur-concrete working range & remelt — cdnsciencepub.com/doi/10.1139/cjce-2021-0098 · Vaniman et al., Uses of Lunar Sulfur — ntrs.nasa.gov/api/citations/19930004793/downloads/19930004793.pdf · MRE metals as O₂ byproduct — ntrs.nasa.gov/citations/20170001458 · Pressure & ballast numbers: the CAD sheet

All the verdicts, at a glance

QuestionVerdictThe one-line reason
1 · A lid over the whole craterProposed · unprovenPublished since Taylor (1992); the only crater-lid structure ever modelled is 17 m wide. Ours is ~1 km. Pressure uplift — not materials — is the dominant problem. Pressure skin + fused regolith sheet on a fibre-pole skeleton, no glazing except one small glass eye at the apex. Build sequence: Founding ↑
2 · Free atmosphere / ozone, no lidContradictedA thin released atmosphere is stripped in ~100 days; a thick one needs ~10¹⁴ tonnes of gas. A standalone ozone layer cannot exist without that thick atmosphere first.
3 · Ozone breathing into treesContradicted — abandonOzone is a high-altitude UV shield, not a plant food. Ground-level ozone damages leaves above 40 ppb and is unhealthy for people above 70 ppb.
4 · Trees under the lidPlants: supported · Trees: unprovenSmall plants photosynthesise at 10–25 kPa if CO₂ partial pressure is held up. No tree has ever been grown at low pressure or in lunar soil.
5 · Hot peak to ice basementReal in vacuum · Not passive under a lidPolar rims really are lit 80–96% of the time beside floors at 35–100 K — because there is no air. Add warm air and the cold floor survives only as a separate, refrigerated compartment.
6 · The living sheath, assembledParts real · One part contradictedNothing contradicts physics — but leaves fed gas through a contact membrane contradicts plant physiology. With a gas gap inside, the sheath is a conformal greenhouse: legitimate engineering development, not off-the-shelf. Full breakdown ↓
7 · The sorting system — every chemical & waste streamDoctrine: real practiceThe rule “nothing hazardous inside, everything flows out through the loko iʻa, stockpiled at the edge to be mined later” is the common pattern of BSL labs, cleanrooms, hospitals and ISS ECLSS zoning. Every stream is labelled, thresholded and assigned below. The inventory ↓
8 · Ahupuaʻa inside ahupuaʻa — indexed rooms & gardensDesign system · parts demonstratedEvery room and garden carries its own index and its own quarantine boundary; the channel network routes addressed resources — potassium, nitrogen, water — the way a signal network routes packets. Built from demonstrated parts: flown root pouches, lab microvascular channels, MELiSSA's compartment chain. How it works ↓
9 · Mature trees moved outsideProposed · depends on the sheathTrees are raised inside the lid, in Lewa's gardens; when mature, each is fitted with its sheath inside, carried out through the dust lock, and stands in vacuum on its tether. The move is the sheath programme's graduation test — gold, once it works. The move ↓
10 · The cover's Lewa — a radiation garden in the glass eyeProposed · trial plotOne small glass section at the cap's apex, over a dark trial plot: the radiotrophic fungus Cladosporium sphaerospermum, which grows toward radiation and shields as it grows (ISS, 2018–19). Full “radiosynthesis” is not proven — the ISS paper says so itself. The cover's Lewa ↓
Sheath component, as designedStatusBasis
Tree survives / grows in vacuum unprotectedContradictedDesiccation; only 30 min at 1.5 kPa shown survivable
Living skin survives open space like a lichenMisapplied evidenceLichens / seeds / spores survived dormant; growing tissue did not
Flexible sheath holding 10–35 kPaOff-the-shelf practiceBEAM 101 kPa flown; EMU suit 29.6 kPa flown
Plant growth at 10–35 kPa insideExists and testedGrowth / photosynthesis shown at 10–25 kPa if ppCO₂ / ppO₂ held
Root feeding by a contact pouch / mantleOff-the-shelf · flownVeggie pillows, APH substrate; fungal mantle template (roots only)
Water / nutrient circulation through channels in a materialEngineering developmentMicrovascular circulation demonstrated at coupon scale only
Self-repair of punctures in the sheath materialExists and tested (lab)Up to >35 mm damage filled; 7–30 heal cycles
Gas supply to leaves through a contact membraneContradicted as statedLeaves exchange gas via stomata to air; cuticle is a barrier
Gas supply across a membrane wall of a gas-gap envelopeEngineering developmentNumerically borderline-feasible: needs ~1,000+ GPU membranes, ≈leaf-area, enriched CO₂ feed
Fibre-optic light as supplementExists and testedCommercial daylighting; indoor-ag supplement trials
Fibre-optic light as a canopy's sole light sourceProposed · currently short69 PPFD measured at fibre end vs 200–1,000 needed — 1–2 orders short at canopy scale
A sheath that is itself living and growingProposed onlyEngineered living materials limited by viability, mass transport, long-term stability

The whole base — one animated cross-section

Lewa sits in the middle and it is not empty: the biology libraries, the houses and the forest fill the bowl — and under the fused cap, the middle is now the most shielded place in the base, under the deepest cover, which is exactly where the ahupuaʻa logic already puts the children and the living collections. The cap above is opaque: a pressure skin under one solid fused regolith sheet, its fibre-pole skeleton inside it as ribs and the bag courses surviving only as faint fused seams (build sequence: Founding III). No sunbeam crosses it — except at one small glass eye at the very apex, the cover's Lewa, where a dark trial plot sits under raw sky in a single narrow column, drawn beneath the eye. Light is born at the collector lens on the rim, sorted there — UV and excess heat stripped at collection, which is why the light inside keeps its amber tone — and carried down the fibre trunks to emitters that light the bowl from within. The eagle suit rides the still air under the cap. Toggle the layers to see what each system is really doing — the ice basement is drawn as its own insulated, refrigerated vault, as its own sealed, separately refrigerated chamber.

Postcard view, not to scale · 1 km crater · the golden band over the arc is our sulphur-glazed sheet — molten sulphur glass through fused regolith, glossy as crystal, over its fibre-pole skeleton — broken only by the glass eye at the apex; the lens stands apart on the rim

CAD — Crater Lid · Sheet 01 · The numbers that decide it

Crater diameter (design)~1,000 m
Lid ever modelled in print17 m Ø · 6 m deep (2025)
Rim seal length alone~3.1 km
Bowl volume (paraboloid)~6–8 × 10⁷ m³
Air to fill it at 1 bar~70,000–95,000 tonnes
Air at 35 kPa~30,000 tonnes
Ballast for 1,000 hPa (published)28–42 m of regolith
Ballast for 600 hPa (published)17–25 m of regolith
Crew pressure floor (NASA-STD-3001)34.5 kPa · ≥30% inert diluent
Winning lid shape at 17 mConcave — pressure → compression
Cap coverFused regolith sheet over the pressure skin — one solid shell, opaque, except one glass eye
Skeleton (built first)Long regolith-fibre poles, one run per radial, tipi-style over the skin · MRE iron anchor ring · poles remain in the sheet as ribs
Sheet constructionAll bags stacked on the skeleton first · then ONE melt pass at 130–140 °C — only the sulphur melts (119 °C) and sets as sulphur concrete · sintered regolith, sulphur glass in matrix + seams
Sheet shape & repairShell action + embedded ribs + anchor ring hold the lens shape · cracks repaired by re-fusing — sulphur seams re-melt
The glass eye — the cover's LewaSmall glass oculus at the apex · raw sky to one trial plot only
Collector lens — positionSeparate structure on the rim, ring frame on posts
Collector glassSulphur glass · n = 2.0
Critical angle30° — past it, light pipes along the shell to the rim collectors
Low sun (~1.5°, rim tilt 22.6°)Lens bends grazing light 32.5° into the fibre intakes
Sorting at collectionUV + excess heat stripped at the lens — the interior's amber is sorted light

Try the pressure — watch the lid problem shrink

Uplift, ballast, air mass and leak pressure all scale with total pressure. Drag, or tap a studied setting. The literature's practical sweet spot is roughly 35–60 kPa, oxygen-enriched but never pure O₂. The fused regolith sheet is not decoration: its mass sits on the skin and counts toward the ballast-equivalent below — shielding and hold-down in the same solid shell.

101 kPa
Total pressure inside the lid
10.3 tf/m²
Upward force on every square metre of lid
≈49 M tonnes
Ballast-equivalent the anchors / ring must hold (1 km crater, lunar weight)
≈82,000 t
Air mass in the bowl (order of magnitude)
◀ 17 m — largest lid ever analysed1,000 m — this design ▶

A shape that works at 17 m does not extrapolate to 1,000 m — the span grows ~59× and stress grows with radius. No published analysis designs a kilometre-scale crater lid. That gap is stated plainly, not papered over.

The cover's Lewa — a radiation garden in the glass eye

Reference art: the eye and the dark plot beneath it are drawn at the apex of the cross-section above — one narrow raw-sky column in an otherwise opaque cap. The cover follows the ahupuaʻa principle too. The fused sheet is the cover's shielding everywhere — except at its very centre-top. There, at the apex, one small glass section is kept: the cover's Lewa, an experimental plot under raw sky, stacked directly over the Lewa below. Everything under the rest of the cap lives on fibre light alone; in this one narrow column the sky arrives unsorted — deliberately, because the doctrine is experiments of all kinds, and some life may want exactly what everything else must be shielded from. If a creature wants to grow in radiation, it should be allowed to try.

The candidate — a black fungus that grows toward radiation

Thrives + shields: measured · “Radiosynthesis”: unproven

Cladosporium sphaerospermum — a melanin-packed black fungus found growing on the walls inside Chernobyl's destroyed reactor, its hyphae reaching toward the radiation source itself (radiotropism). It is the strongest known candidate for a crop that wants the radiation the fused sheet keeps out of everywhere else.

  • Dadachova et al. (2007): melanized fungi exposed to ionising radiation at ~500× background grew significantly faster — more biomass, ~3-fold greater incorporation of acetate — and the radiation measurably changed melanin's electronic properties, its electron-transfer capacity rising ~4-fold. Non-melanized mutants showed no such gain: the melanin is doing the work.
  • Flown — ISS, Dec 2018 – Jan 2019: C. sphaerospermum grew 1.21× faster in orbit than in ground controls, and once its mat reached 1.7 mm thick, the radiation measured beneath it fell by ~2.2% — a shield that grows itself (Averesch, Shunk & Kern 2022).
  • Why it belongs in the eye: the trial needs the one thing the fused sheet denies the whole bowl — raw, unsorted sky. The eye is small, at the apex, over a single dark plot: if the trial fails, one plot fails, sealed off from the Lewa below. Nothing else in the base is exposed.
The farming doctrine: prove it in the eye first. Once a strain is proven there, it gets farmed outdoors, under raw sky — beds beyond the seal, where radiation is free and unlimited — harvested as living shield mats and as melanin: a shield crop that grows back, trialled first in the one glass section the cover keeps.

Sources: Dadachova et al. (2007), PLoS ONE 2(5): e457, DOI 10.1371/journal.pone.0000457 — pmc.ncbi.nlm.nih.gov/articles/PMC1866175/ · Averesch, Shunk & Kern (2022), Frontiers in Microbiology 13:877625 — frontiersin.org/journals/microbiology/articles/10.3389/fmicb.2022.877625/full

The honesty line — kept exact: full “radiosynthesis” — living on radiation alone, the way a plant lives on light — is NOT proven. The published ISS paper says so itself: “Actual radiosynthesis, however, remains to be shown.” The cultures had nutrients; radiation may make the fungus use them faster rather than feed it directly. What is proven is enough to trial: it thrives where other life dies, it grows toward the radiation, and it shields while it grows.

The Radiation Works

Radiation, put to work

What the fused sheet keeps out of the bowl is not wasted. The same radiation has three jobs in this base — building, measuring, and marking — plus one extrapolation, flagged as such.

Outdoors · the farm

The mycelium yard

The black fungus farmed outdoors is also a builder. NASA's real Mycotecture programme (Rothschild, NASA Ames, NIAC Phases I–III) grows structures from fungal mycelium composites — habitats to furniture — and its melanin-rich “Black series” targets radiation-shielding furniture. Our farm grows shield tiles, mats and forms.

Honesty — lab / design stage (TRL 2–3), not flight-proven.
At the edge · by the isotope store

The scintillator garden

Crystals grown in place from bus-fed brine along channels — fungi precipitate crystals naturally; calcium oxalate is routine fungal chemistry. Their glow is dosimetry: brightness is the radiation reading. Light as information.

Honesty — ambient radiation carries microwatts: a scintillator in open air glows far too faintly to light a room. It is an instrument, not a lamp.
Throughout · egress & vaults

The backup light layer

Tritium-class radioluminescent markers (GTLS): self-powered, no charging, 12.3-year half-life — egress paths, airlock outlines, vault markers that cannot go dark when power fails.

A marker, not a lamp.
Extrapolation — grown crystal waveguides can pipe fibre light to depth — the ice-vault levels — where laying cable is costly. Flagged as extrapolation from solution crystal growth.

The people, in their places

First, the postcard: our crew in their suits, going about the day. Worlds within worlds only work if everyone lives in the right one. The cast is placed by the same law that places the chemicals: the most sensitive subjects — the children, the teachers, the living collections — sit deepest inside the seal, and the people who work the dirty edge cross back inward only through the locks. Each card names the zone, because the zone is the safety system.

SHORTS · EDGE CREW
LUMEN · LIGHT CREW
KIRI · GARDENS
ELLIE · THE PONDS
A SPACE KID · LEWA SCHOOL
A SPACE KID · GARDEN CREW
Centre · Lewa · inside every seal

The Space Kids

They live and learn at the crater's heart, in the cleanest air the base can make — the zone with the highest pressure, the strictest contamination budget, and nothing stored in it but living things and books. Their school gardens are real ahupuaʻa units (below), each indexed, each growing part of the base's food.

Centre · Lewa · libraries of biology

The Teachers

Keepers of Lewa: the seed vault, the living collections and the growth records. They run the indexing — every garden, bed and culture carries a code they maintain — and they decide what the evidence says is ready to move outward, ring by ring.

Middle ring · gardens & sheath fitting

Kiri

Gardener of the terraces. Kiri raises the trees under the fused cap — in amended soil, at garden pressure, under fibre light — fits each mature tree's sheath while it is still inside, and walks it out through the dust lock to stand in vacuum. The tree-moving programme is Kiri's route, walked tree by tree.

Middle ring · the loko iʻa ponds

Ellie

Ellie keeps the loko iʻa — the fishponds that form the base's living seal. She tends the fish, reads the water, and certifies the air and water coming back inward: clear, oxygenated, safe to breathe and drink. Her rule is the pass-through rule — nothing crosses inward except through a verified step: settled, filtered, polished, or cleaned. If it cannot pass, it does not cross.

Under the lid · light & fibre

Lumen

Lumen tends the light: the collector lens and rim collector heads on the rim, the fibre trunks down the wall, and the fibre-optic runs and emitters that light the gardens, Lewa and the sheathed trees — plus the wing-suit inspections of the anchor ring. No light enters any other way: where light has to be carried like a resource through a channel, it is Lumen's channel.

Edge · sulphur plant & tank farm

Shorts

Shorts works outside the seal — the sulphur plant, the oxygen plant, the edge tank farm and the five stockpile bays — in a suit, on the dirty side of every lock. What Shorts stockpiles today, segregated and labelled, is the mine the base digs later. Shorts never carries dust, brine or gas back through the lock uncleaned.

Under the lid · the trapped air

The Wing Suit

The eagle suit glides the still air under the lid — commuting, canopy survey, lid inspection (see Life under the lid). It belongs to the pressurised bowl only: outside the lid there is no air to glide on, and the suit's job passes to pressure suits and to Shorts' edge crews.

Centre · Lewa · the living memory

The gardens themselves

The intentional biodiversity of Lewa is a resident, not a decoration: seed vault, living collections, school gardens and forest terraces, each one an indexed ahupuaʻa unit with its own inputs, outputs and quarantine boundary — detailed in Ahupuaʻa inside ahupuaʻa.

Outside · the loko iʻa & the bays

What nobody does

Nobody stores anything hazardous in the middle. Nobody carries waste through a clean corridor. Nobody opens a brine tank, a sulphur line or a battery quarantine box inside the seal. The cast list and the chemical list obey one law — that is the whole system.

Schematic plan, not to scale · the dashed ring is one boundary doing two jobs: it holds the air in, and it keeps pollution out of the middle

The living sheath — the mechanism, tested item by item

The mechanism: a tree stands outside the pressurised base, in lunar vacuum, wrapped in a living fibre-optic nutrient-sorting channel system — a smart material that hugs the tree, protects it from vacuum and radiation, does not choke it, and feeds it O₂, CO₂, water and nutrients through the material itself, by membrane / osmosis-like transport from proximity — the way a fungal mantle feeds a root — with fibre-optic channels delivering light. Research read 14 Nov 2026. Verdict key: exists and tested / proposed or partial / contradicted by evidence.

Schematic, not to scale · the tree is never independent of the base — gas, water and power run through the tether

The honest version — re-describe it slightly and it becomes a legitimate engineering programme

Nothing in the assembled mechanism contradicts physics. As literally stated — a living sheath hugging the tree and feeding its leaves gases through the material by osmosis-like proximity, on the lichen / mycorrhiza precedent — it fails on leaf physiology and misreads the space-survival evidence. The redesign that the evidence supports:

  1. A low-pressure (10–35 kPa) conformal inflatable greenhouse envelope with a thin internal gas gap — not a contact membrane on leaves.
  2. A root pouch fed by pumped, oxygenated nutrient channels — flown practice (Veggie pillows, Advanced Plant Habitat substrate).
  3. A synthetic microvascular service layer for water, thermal and self-healing functions — lab-demonstrated, not yet at tree scale.
  4. Best-in-class gas-permeable membrane sections in the envelope wall, or conventional ECLSS-style gas circulation from the habitat.
  5. Fibre-optic light only as a supplement to light admitted through translucent envelope sections — a tree outdoors already stands in full, unfiltered sunlight.

Trees being moved outside once they are mature — gold

This is where the sheath, the gardens and Lewa meet. A tree is never planted in vacuum and hoped for: it is raised inside the seal, in the indexed garden system, and only graduates outward when it — and its sheath — are ready. The move is the programme's gold standard: the moment the base stops being a greenhouse with trees in it and becomes a forest that can stand on the Moon itself.

Step 1 · Lewa

Seed & record

Seed from Lewa's vault, germinated in the libraries' nursery units; the tree gets its index code at germination and keeps it for life.

Step 2 · Garden

Grow up inside

Years on a forest terrace under the fused cap, lit by the fibre emitters, fed by its garden's own channel index — water, potassium, nitrogen dosed by signal, logged by the Teachers.

Step 3 · Fitting

Sheath fitted indoors

Kiri's crew fits the conformal envelope, root pouch and tether connections inside the fitting bay — pressure-tested in air before it ever meets vacuum.

Step 4 · The lock

Out through the dust lock

The sheathed tree crosses the loko iʻa line through the large lock — the one outward journey a living thing makes, sealed inside its own little atmosphere.

Step 5 · Outside

Standing in vacuum

Planted at its indexed outside plot, tethered to the base for gas, water and power, in full unfiltered sun. Gold — when the first one survives a full lunar year, the design is proven at tree scale for the first time anywhere.

Honest flag, kept from the verdicts above: no growing tree has ever survived hard vacuum, with or without a sheath — the 30-minute, 1.5 kPa seedling test (S1) is the closest evidence. The move is a programme to run, not a result to claim.

S1 · A growing plant in vacuum, unprotected

Contradicted — for a growing plant

The killer is boiling / desiccation, not “explosion” — and the sheath must therefore be a genuine pressure envelope with zero tolerance for sustained leaks.

  • Why it dies: liquid water cannot remain liquid below its vapour pressure (~2.3 kPa at 20 °C); lunar surface pressure is a hard vacuum of ~10⁻¹² torr, effectively zero, with severe temperature fluctuation, lethal radiation, abrasive dust and micrometeoroids. Exposed leaf water evaporates / boils off, stomata cannot regulate the loss, turgor collapses, tissue desiccates. With no ambient CO₂ photosynthesis stops; with no O₂, root and stem respiration stops. Hypobaria research on Arabidopsis shows low pressure is a distinct stress response, not just hypoxia, and that “water movement is a paramount issue in plants” under hypobaria.
  • The one controlled experiment (Wheeler et al. 2011): radish, lettuce and wheat grown 20 days at normal pressure were plunged to 1.5% of sea-level pressure (~1.5 kPa) for 30 minutes. All recovered on re-pressurisation and grew to harvest a week later with no significant weight difference; they wilted temporarily from leaf water loss, and the authors identified dehydration, not oxygen lack, as the likely eventual killer. Note carefully: 1.5 kPa is near-vacuum, not lunar hard vacuum — still ~10 orders of magnitude above lunar pressure — and 30 minutes is not growth.
  • The survivable floor for growth is ~10 kPa, not zero: Arabidopsis at 101, 75, 50, 25 and 10 kPa showed net photosynthesis insensitive to total pressure provided CO₂ partial pressure was kept non-limiting (0.07–0.10 kPa); the authors concluded hypobaric modules “might be operated around 10 kPa.” At 10 kPa, lettuce water loss rose 6.1-fold versus ambient (Richards et al., Astrobiology 6:851–866, 2006). Lettuce also grows comparably at 25 kPa.
  • Honest timescale: minutes at ~1.5 kPa are survivable; sustained hard vacuum desiccates and kills active tissue, the rate set by water loss — faster for thin leaves, slower for bark-covered wood, whose living cambium would still desiccate through lenticels and any wound. Claims of death “within seconds” for a whole plant, found on low-quality web pages, are not supported by the experimental record and are not used here. No experiment was found exposing a growing tree — or any actively growing vascular plant — to hard vacuum for hours / days and recovering it.
What this means for the sheath: it does not need sea-level pressure — 10–35 kPa internal pressure is a biologically sensible target — but it must be a real pressure envelope.

Sources: Wheeler et al. 2011, Advances in Space Research, DOI 10.1016/j.asr.2010.12.017, via New Scientist — newscientist.com/article/1956405-vacuum-of-space-no-match-for-the-mighty-radish/ · Richards et al. 2006 via Science.gov — science.gov/topicpages/p/pressure+plant+growth · Paul & Ferl 2006 — researchgate.net/publication/228351730 · NASA Lunar Sourcebook ch. 5 — eng.auburn.edu/~dbeale/ESMDCourse/Chapter5.htm

S2 · Symbiotic skins in open space — the lichen precedent

Exists and tested — dormant only

The most commonly over-read result in the whole proposal. Lichens, seeds and spores survived open space dormant and desiccated, reviving only after return to Earth. No active, growing, photosynthesising plant tissue has survived open vacuum.

  • Foton-M2 / BIOPAN (2005): Rhizocarpon geographicum and Xanthoria elegans exposed outside the Foton capsule for 14.6 days to vacuum, temperature fluctuation, full solar UV and cosmic radiation — full rate of survival and unchanged photosynthetic ability, measured after return to Earth. (ESA)
  • Foton-M3 / BIOPAN-6, Lithopanspermia (2007): the same lichens plus Aspicilia fruticulosa, on natural rock, exposed 10 days: all three lichens quite resistant; ~80–90% of ascospores germinated and grew post-flight. Resistance is attributed to the lichens' symbiotic nature and protection by the upper pigmented layer, the cortex. The rock / halite bacteria in the same experiment were severely damaged, and a Rhizocarpon sample that went through atmospheric re-entry had no survivors (de la Torre et al., Icarus 2010).
  • EXPOSE-E / LIFE, ISS (2008–2009): Xanthoria elegans exposed 1.5 years outside the ISS: LIVE/DEAD staining gave average viability 71% for the photobiont (alga) and 84% for the mycobiont (fungus). The paper discusses this in the context of anhydrobiosis and UV-screening pigments — the lichen was in an inactive, desiccated state of latent life for the exposure (Brandt et al., Int J Astrobiology 14:411–425, 2015).
  • Plant seeds — SEEDS, EXPOSE-E: of 2,100 wild-type Arabidopsis thaliana and Nicotiana tabacum seeds exposed 18 months outside the ISS, 23% produced viable plants after return; highest survival was tobacco at 44%; survival was lower in mutants lacking flavonoid UV screens, and seeds shielded from solar light showed full survival — suggesting opaque embedding is the protective factor (Tepfer, Zalar & Leach, Astrobiology 12:517–528, 2012).
  • Moss spores (2025, newest data point): Physcomitrium patens spores encased in their sporangium spent 283 days (9 months) outside the ISS; over 80% survived and germinated on return. Unprotected vegetative tissue (protonemata) was killed outright by 10 kJ/m² UVC and by 4 days at −80 °C in simulation, while encased spores tolerated ~1,000× more UVC (Maeng et al., iScience 2025, DOI 10.1016/j.isci.2025.113827).
What protected them, in every case: dormancy + desiccation (anhydrobiosis), a pigmented or opaque outer layer, and — where survival was highest — shielding from UV. A growing tree is hydrated, metabolically active tissue — physiologically the protonemata case, not the spore case. The lichen result supports a protective pigmented skin as a design motif, but provides no evidence a living symbiotic skin can keep a tree alive and growing in vacuum — and a lichen-style living layer on the sheath would itself have to be kept dormant / desiccated outside, or fed / pressurised inside.

Sources: ESA — esa.int/Science_Exploration/Human_and_Robotic_Exploration/Lichen_survives_in_space · de la Torre et al. 2010 — oa.upm.es/7093/ · Brandt et al. 2015 — resolve-he.cambridge.org (PDF) · Tepfer et al. 2012 — web archive of ESA attachment · Maeng et al. 2025 — pmc.ncbi.nlm.nih.gov/articles/PMC12767186/

S3 · A conformal pressure skin at 10–35 kPa

Exists and tested

Within demonstrated materials practice — a flexible membrane holding far more than the proposed pressure has flown for a decade. The specific geometry (branching, growing, conformal) is proposed, not demonstrated.

  • BEAM (flown): expanded and pressurised on the ISS on 28 May 2016, pressure equalised with the station (~101.3 kPa / 14.7 psi), giving 565 cu ft (16 m³) from a packed module ~7 ft long; multiple layers of flexible fabric with an internal restraint and bladder system, plus closed-cell foam layers functioning as a Whipple-type micrometeoroid shield. Later certified to remain attached until 2028. A flexible membrane holding 101 kPa for a decade is ~3–10× the proposed sheath pressure. (NASA)
  • Spacesuits (flown, conformal): the NASA EMU is a human-shaped flexible pressure vessel operated for decades at 29.6 kPa (4.3 psi) pure O₂; the Russian Orlan operates at ~40 kPa — almost exactly the proposed 10–35 kPa band, worn skin-tight in vacuum.
  • Mechanical counterpressure (prototype only — important caveat): MIT's BioSuit aims to replace gas pressure with a skin-tight elastic garment applying ~30 kPa directly; active-coil cuffs produced the required pressure in tests and prototypes reached ~20–30 kPa, but no full counterpressure suit has flown. And the catch that transfers directly to the tree: counterpressure squeezes — it supplies no gas. A tree still needs a gas-filled envelope for stomatal exchange (S4), so the workable sheath is a gas-pressurised envelope that happens to lie close to the plant, not a pure squeeze garment.
  • Scaling works in the design's favour: membrane stress scales with pressure × radius, so a small-radius sheath around a trunk / branch at 10–35 kPa is a far easier pressure vessel than BEAM. The unsolved part is not strength — it is geometry and growth (hardest problems, below).

Sources: NASA, BEAM fully expanded and pressurized, 28 May 2016 — nasa.gov/blogs/spacestation/2016/05/28/beam-fully-expanded-and-pressurized/ · BEAM construction — alchetron.com/Bigelow-Expandable-Activity-Module · NASA NTRS BEAM demo — ntrs.nasa.gov · Spacesuit pressures review — mdpi.com/2226-4310/13/11/1047 · MIT BioSuit — news.mit.edu/2014/second-skin-spacesuits-0918

S4 · Feeding by proximity — roots yes, leaves no (as stated)

Split verdict by organ

Roots fed by proximity: exists and tested, including in spaceflight. Water / nutrient transport through membranes: exists industrially, at fluxes adequate for roots. Leaves fed CO₂ / vented O₂ “through the material” by contact: contradicted as stated.

  • Roots — Veggie (ISS, flown since 2014): each plant grows in a “pillow” of clay-based media and fertiliser that distributes water, nutrients and air in a healthy balance around the roots by wicking / capillarity. Advanced Plant Habitat (ISS, flown since 2018): “a porous clay substrate with controlled release fertilizer to deliver water, nutrients and oxygen to the plant roots,” fully enclosed and automated. (NASA) This is feeding by proximity — for roots, in a pouch, exactly as imagined. Caveats: never scaled beyond small herbs / vegetables (~0.16–0.2 m² systems), and both systems enclose the shoot in cabin air — neither feeds a leaf through a membrane.
  • Root-zone oxygen is a hard, quantified constraint: hydroponic practice targets dissolved O₂ above ~6 mg/L (5–8 mg/L “good”; below ~3 mg/L roots are oxygen-deprived and uptake fails), delivered by continuous aeration, because roots respire and suffocate in stagnant solution. Flag: these thresholds come from horticultural practitioner sources, not peer-reviewed papers — consistent across sources, flagged accordingly. A sheath feeding roots osmotically must deliver oxygenated solution at a rate matching root respiration — doable with pumped, aerated channels, not with passive osmosis alone.
  • Leaves — the physiology problem: a leaf's outer cuticle is a waterproof barrier that “reduces water loss but also prevents movement of carbon dioxide and oxygen”; essentially all leaf gas exchange occurs through stomata into the leaf's internal air spaces, by diffusion from / to surrounding air. There is no root-like absorptive membrane on a leaf for a sheath to feed through. A gas-tight sheath pressed onto a leaf blocks its stomata. The design therefore needs a thin gas layer between sheath and leaf — the sheath is a conformal greenhouse, and “membrane feeding” can only describe transport across the greenhouse wall, not across the leaf surface.
  • The flux arithmetic (researcher's calculation from published membrane data; order-of-magnitude only): demand — a eucalypt canopy photosynthesises at 28 µmol CO₂ m⁻² s⁻¹ (ground area) at ambient CO₂, saturating at about half full sunlight (CSIRO); individual tree leaves reach ~30 µmol m⁻² s⁻¹. O₂ release is equimolar. Water loss is far larger: ~400 water molecules lost per CO₂ fixed; a large tree transpires hundreds of litres per day (cottonwoods ~120 L/tree/day; a large Amazon tree up to 1,180 L/day).
  • Supply through a membrane wall: the best reported composite hollow-fibre membranes reach CO₂ permeance 3,700 GPU and O₂ permeance 860 GPU (PDMS/PAN; CO₂/N₂ selectivity 10, O₂/N₂ 2.2). Converted (1 GPU ≈ 0.335 µmol m⁻² s⁻¹ per kPa of partial-pressure difference), a 1,000-GPU-class membrane driven by a CO₂ partial-pressure difference of 0.1 kPa passes ≈ 33 µmol m⁻² s⁻¹ — the right order of magnitude for photosynthesis if membrane area ≈ leaf area, the habitat side is held at elevated CO₂ (~0.1 kPa ppCO₂ — the same condition Richards et al. found necessary at low pressure), and humidity / fouling do not degrade the membrane. Commodity films are 1–2 orders of magnitude worse (e.g. a patented co-extruded film at ≥20 GPU CO₂ flux, CO₂/O₂ selectivity ≥4; cellulose-acetate membranes at CO₂ permeability 4.59 Barrer, selectivity 5.68) — those would starve the leaf.
  • The water-vapour problem is bigger than the CO₂ problem: in a sealed sheath, transpired vapour (potentially 100+ L/day for a mature tree) saturates the internal air, after which transpiration — which drives nutrient ascent in the xylem — stalls, unless the sheath continuously condenses and returns that water and rejects its latent heat. No flown plant system does this at tree scale; Veggie / APH operate inside a station whose ECLSS does the condensing for them.
Conclusion: membrane-wall gas supply is not physics-barred, but it requires best-in-class membranes, large area, active CO₂ enrichment on the feed side, and active O₂ / water-vapour removal — an engineered life-support loop, not passive “osmosis-like” proximity.

Sources: NASA, Growing Plants in Space — nasa.gov/exploration-research-and-technology/growing-plants-in-space/ · PAN–PDMS membranes — sciencedirect.com/science/article/abs/pii/S0376738813000793 · Co-extruded film patent — patents.justia.com/patent/8911540 · Cellulose-acetate membrane — doi.org/10.3390/polym18060740 · Eucalypt photosynthesis — publish.csiro.au/FP/ExportCitation/PP9870619 · Transpiration — ebsco.com/research-starters/environmental-sciences/evapotranspiration-et · Root-zone O₂ (practitioner) — hydrohomegardens.com/learn/root-zone-oxygen-impact-on-yield/ · Leaf gas exchange — brainscape.com/flashcards/gas-exchange-and-transport-scr-16940904/packs/21587343

S5 · The biological template — the fungal mantle

Root trade real · Tree-to-tree contested

Exists and tested for the root–fungus trade itself. The popular tree-to-tree extension is contested — two 2023 reviews find its key claims insufficiently supported, with a published rebuttal; the debate is live, not settled in either direction.

  • What is established: an ectomycorrhiza has three parts — a fungal mantle (sheath) enclosing the root tip (mantle hyphae up to ~100 µm thick, which also protects the root from water loss and pathogens), the Hartig net of hyphae growing between root cortical cells, and extraradical mycelium exploring the soil. “The Hartig net is the place of massive bidirectional exchanges”: the fungus transfers nitrogen, phosphorus, potassium and water to the host, and the host allocates a significant share of its photosynthate to the fungus. Crucially for the analogy: this exchange is root-only, in a liquid film, across living cell membranes with active transporters — the fungus does not feed leaves, seal anything against a vacuum, or move gases, and it charges the tree a large carbon fee for the service.
  • Contested — Karst, Jones & Hoeksema (2023), Nature Ecology & Evolution 7:501–511: the claims that common mycorrhizal networks (CMNs) are widespread in forests and that resources transferred through them increase seedling performance are “insufficiently supported” (field results vary too widely, have alternative explanations, or are too limited); the claim that mature trees preferentially send resources / defence signals to offspring through CMNs “has no peer-reviewed, published evidence”; and unsupported claims in the citing literature have doubled in 25 years with a positive citation bias.
  • Contested — Henriksson et al. (2023), New Phytologist 239:19–28: “evidence of a significant net C transfer via common mycorrhizal networks that benefits the recipients is still lacking”; a fungal C pipeline between trees is hard to reconcile with any adaptive advantage for the fungus; isotope studies risk inferring much from “miniscule shifts in isotopic abundance.”
  • Pushback exists: Klein et al. (2023) argue belowground C transfer across CMNs was demonstrated in the lab (1969) and in the field (Simard et al. 1997, birch↔Douglas-fir isotope transfer) and that the evidence is “solid and accumulating,” while conceding the importance of CMNs remains unclear.
Design consequence: using the mantle as a template for a root pouch is legitimate and matches flown hardware (S4). Using “the fungus feeds the tree through a sheath” as a template for feeding the whole tree, including its canopy, through a sheath extends the analogy beyond anything the biology does — even before the contested tree-to-tree claims are reached.

Sources: Ectomycorrhizal review — thefreelibrary.com (multifunctional role of ectomycorrhizal associations) · Karst et al. 2023 — pubmed.ncbi.nlm.nih.gov/36782032/ · Henriksson et al. 2023 — pubmed.ncbi.nlm.nih.gov/37149889 · Klein et al. 2023 — pmc.ncbi.nlm.nih.gov/articles/PMC10751480/

S6 · Fibre-optic light delivery

Real for buildings · Partial for plants

Exists and tested as building daylighting; proposed or partial as a plant-growth light source — demonstrated only as a supplement, with measured delivered PPFD typically well below growth requirements.

  • Commercial systems (delivered light): Parans SP4 (manufacturer specification) — 700 lm per fibre-optic cable (current SP4 sheet; 900–1,300 lm/cable on another SP4 sheet), system totals 5,600–28,000 lm, all at 100,000 lux solar input, cables up to 100 m, with no IR and UV transmitted. Independent evaluation of an SP4-24.2 by UC Davis CLTC measured 900–1,200 lm per bundle consistently, maximum 1,603 lm. A Himawari system is quoted at ~4,000 lm total under 100,000 lux (comparative study quoting developer data — no independent test located). A Parans SP3 study-hall installation delivered ~4,500 lm at 130,000 lux input over 10 m of fibre, giving nadir illuminance >500 lux bare and >4,700 lux with a spot diffuser.
  • What plants need: roughly 100–400 µmol m⁻² s⁻¹ PPFD for seedlings / young plants and 400–1,000 for vigorous / full-sun growth; full sunlight is ~2,000 µmol m⁻² s⁻¹. In illuminance terms, growth light is tens of thousands of lux spread over the canopy — a 700-lm cable spread over even 0.25 m² gives only ~2,800 lux, low-light-houseplant territory; concentrated into a spot it can be intense, but a tree canopy is square metres, not spots.
  • Direct fibre-to-plant measurement: in a PMMA-fibre indoor-plant study, direct sunlight of 2,121 PPFD fell to 69.3 PPFD at the fibre output (no filter) and 15.1 PPFD with a UV filter — 30–140× losses, below the growth range for all but shade plants.
  • Best case for the defence: an optimised Fresnel-lens / plastic-fibre system for indoor agriculture achieved ~16% optical efficiency in practice (19% simulated) and, installed in an indoor farm, cut LED energy use 54.3% on a sunny day and let plants grow better than LED-only — as a supplement to electric light, not a replacement.
  • Physics note: fibre delivery cannot beat conservation of étendue — to light a canopy of area A at near-sunlight PPFD through a system of ~16% efficiency, the collector must be several times A. And on the Moon: a tree standing outdoors is already in full, unfiltered sunlight (~2,000+ PPFD). Fibre delivery only earns its losses if the pressure sheath must be opaque (for radiation / micrometeoroid protection) — a transparent or translucent sheath section over the canopy would deliver more light for free, at the cost of UV management. Fibre also delivers nothing during the ~14-Earth-day lunar night (except at specially lit polar sites), so stored energy + LEDs are required regardless.

Sources: Parans SP4 spec — parans.com/wp-content/uploads/2021/12/Parans_Product_information_2.3.pdf · UC Davis CLTC evaluation — cltc.ucdavis.edu · PMMA fibre plant study — hindawi.com/journals/jre/2018/9429867/fig7/ · Fibre daylighting for indoor agriculture — researchgate.net/publication/364476677 · PPFD requirements — gorillagrowtent.com/blogs/news/ppfd-for-seedlings

S7 · Living / vascular smart materials — the channel system itself

Exists at lab scale · Living version proposed only

Exists and tested at laboratory coupon scale. Nothing demonstrated is a living, growing, pressure-holding skin — that combination is proposed only.

  • 2001: first autonomic self-healing polymer — microencapsulated healing agent + catalyst in an epoxy, healing a crack once, locally (Nature 409:794–797).
  • 2007 (Toohey et al., Nature Materials 6:581–585): 3D microvascular networks of 200 µm channels (direct-write assembly of fugitive ink) filled with liquid healing agent under a brittle epoxy coating: up to 70% healing efficiency, repeated up to 7 cycles at the same location. Interpenetrating dual networks (resin + hardener separate) later reached 30 cycles.
  • Pressurised / pumped vasculature: actively pumped two-part healing chemistry achieved nearly full fracture-toughness recovery through 15 damage / heal cycles with the network occupying just 0.1 vol% of the specimen (Hamilton, Sottos & White, J. R. Soc. Interface 2011).
  • 2014 (White et al., Science 344:620–623): a vascular system delivering two gel-forming fluids filled impact damage exceeding 35 mm diameter within 20 minutes and restored mechanical function within 3 hours, recovering 62% of absorbed energy — “regeneration,” not just crack-sealing.
  • Circulation as a service, not just healing: vascular fibre-reinforced composites have circulated different fluids to demonstrate temperature regulation, chemistry, conductivity and electromagnetism in one material system (Illinois News Bureau). This is the closest demonstrated analogue to a “channel system circulating fluids through a sheath” — in rigid epoxy / fibre coupons, at bench scale, with external pumps.
  • Engineered living materials (ELMs) proper — living cells embedded in a matrix that sense, respond, adapt and in some cases self-repair — exist as demonstrations, but a 2026 review lists the field's limits plainly: “trade-off between mechanical robustness and cellular viability, restricted mass transport, unstable long-term function, bio-abiotic interfacial complexity, biosafety risks, and scalable manufacturing.”
Gap to the proposal: demonstrated channel networks move millilitres of healing resin in centimetre-scale coupons. The tree sheath must move litres-to-hundreds-of-litres per day of water, plus gases, over metres, for years, while flexing, in thermal extremes, and holding pressure. The channel-network concept transfers; every scale, lifetime and environment parameter is un-demonstrated. None of the demonstrated “vascular” materials is alive — “living” currently buys nothing demonstrated beyond what pumped synthetic vasculature already does, and adds the ELM field's viability / mass-transport problems.

Sources: Toohey et al. 2007 via review — hindawi.com/journals/amse/2012/854203/fig4/ · Hamilton, Sottos & White 2011 — whitegroup.beckman.illinois.edu (PDF) · White et al. 2014 via Phys.org — phys.org/news/2014-05-regenerating-plastic-video.html · Illinois News Bureau — news.illinois.edu/vascular-composites-enable-dynamic-structural-materials/ · Illinois Chemistry — chemistry.illinois.edu/news/2007-12-31 · ELM review 2026 — pubmed.ncbi.nlm.nih.gov/42716398/

The hardest unsolved problems — flagged honestly

Problem 01

Growth and girdling

A tree thickens (secondary growth), extends branches, sheds and regrows leaves. Any sheath that “hugs” will girdle unless it actively expands at the right rate everywhere — no demonstrated material does this, and a loose sheath loses the “conformal” premise.

Problem 02

Sealing at branch junctions

Every crotch is a complex, moving, bark-shedding seal surface; seals are the classic failure point of far simpler pressure garments (spacesuit joints are engineered bearings, not bark seals).

Problem 03

Transpiration recycling and heat

Condensing and returning up to hundreds of litres/day of water vapour inside the envelope, and rejecting the latent heat, in an environment swinging roughly +127 °C to −173 °C, with no convective cooling outside.

Problem 04

Lunar night

~14 Earth days of darkness at most sites: fibre optics deliver nothing; the tree needs stored-energy lighting / heating or a polar peaks-of-eternal-light site.

Problem 05

Radiation

A thin membrane stops UV (good — and Parans-type fibres also filter UV/IR) but provides negligible galactic-cosmic-ray / solar-particle-event shielding compared with BEAM's multi-layer foam stack or regolith; chronic dose to cambium over a tree's lifespan is unstudied.

Problem 06

Leak economics

Whole-canopy surface area + micrometeoroid punctures (self-healing helps, S7) means a continuous make-up gas / water budget from the habitat; the tree is never truly independent of the habitat it stands “outside” of.

Problem 07

Ethylene and internal air quality

Enclosed plants accumulate ethylene and other volatiles that distort growth (low pressure mitigates this ~3-fold in lettuce studies) — the envelope needs scrubbing / venting, another habitat tether.

Sheath — could not verify, marked plainly

  • No published experiment exposing a growing tree (or any growing vascular plant) to hard vacuum was found; the hard-vacuum timescale in S1 is inference from the 1.5 kPa experiment plus vapour-pressure physics, and is labelled as such.
  • No published system demonstrating a conformal pressurised sheath on any plant was found — S3's verdict rests on BEAM / spacesuit analogy, not on a plant demonstration.
  • Dissolved-oxygen thresholds in S4 come from horticultural practitioner sources, not peer-reviewed papers; consistent across sources but flagged accordingly.
  • The GPU→flux conversion in S4 is the researcher's arithmetic from published permeances, not a figure quoted from a source; order-of-magnitude only, assuming ideal, unfouled membranes at steady partial-pressure differences.
  • Himawari performance figures come from a comparative journal paper quoting developer data, not from an independent test located in this run.

The ozone question

ABANDONED

Ozone does not breathe into trees. It never did — on Earth or anywhere.

Ozone (O₃) is not a food, a nutrient, or a breathing gas. It is a trace gas that lives 10–50 km overhead, where it works as a sunscreen. The moment ozone touches a leaf or a lung, it is a damaging oxidant. A “coating” of it around trees and people would injure the biology long before it shielded anything.

  • What ozone actually is: 90% of Earth's ozone sits in the stratosphere (10–50 km up, peak ~32 km at up to 15 ppm). All of it together is 0.00006% of the atmosphere. It absorbs all UV-C and most UV-B — from a distance. (NASA Ozone Watch)
  • What plants actually breathe: CO₂ in through the stomata for photosynthesis; O₂ for respiration. There is no ozone-uptake pathway in any plant science found. None.
  • What ozone does to a leaf: it slips in through those same stomata, breaks into reactive oxygen species, and burns the cell from inside — less photosynthesis, damaged Rubisco, stippled and dying leaves, smaller harvests. Plant damage accrues above 40 ppb (the AOT40 threshold). (EPA)
  • What ozone does to people: natural background is ~10 ppb. The US health standard is 70 ppb over 8 hours — above that: irritated throats and lungs, worse asthma and emphysema. (NASA / EPA)
  • And with no lid at all? A standalone ozone layer is impossible: ozone only forms inside a thick O₂ atmosphere (the Chapman cycle needs pressure and a third body to collide with). The Moon can't hold the thick atmosphere — see Verdict 2 — so it can't hold the ozone either.
10 ppb · natural background
40 ppb · plants start taking damage
70 ppb · unhealthy for people (8-hr standard)
Scale: 0–140 ppb. A shielding “coating” would need concentrations orders of magnitude higher than everything on this bar — deep inside the poison zone. That is why there is no version of this idea to rescue.

What actually breathes into the trees — the real mechanism

CO₂ →

People & soil breathe out

Crew, animals and soil microbes exhale carbon dioxide into the crater's air. In a sealed bowl, that CO₂ doesn't escape — it becomes the garden's feedstock. This, not ozone, is the “breathing into trees” the design was reaching for.

photosynthesis⇄respiration
← O₂

Trees breathe it in — and answer

Leaves take in CO₂ and return oxygen. The one number that matters is CO₂ partial pressure: hold it near 0.07–0.10 kPa and photosynthesis runs normally even at low total pressure. Scrub the air “clean” of CO₂ for crew comfort and you starve the forest — the loop has to be engineered, together.

And ozone's real job — blocking UV — is done by a solid: the fused regolith sheet over the cap's pressure skin, metres of fused rock between the sky and everything living, with the collector sorting UV out of the delivered light before it ever enters. A well-run greenhouse would actually scrub stray ozone (from electrical gear and UV lamps) out of its air.

The ahupuaʻa, standing on end

An ahupuaʻa runs from the uplands to the sea, one watershed, every zone feeding the next. The crater does the same thing vertically — rim to floor — which is why the design maps so cleanly. Each zone below carries its honest physics flag.

Uka · The uplands

The rim & the peak

Solar towers, the CAD compound, the cap's anchor ring — and the collector lens, standing apart on the rim in its ring frame, drinking the grazing light and pouring it down its fibre trunk. The rim's gift is not heat — polar sun arrives at a grazing ~1.5° — it is continuity: near-constant light and power.

  • Most-lit measured point: sun 86% of the year (Bussey et al. 2010)
  • At 10 m up a tower: 95.65%; longest darkness only 3–5 days (Gläser et al. 2014)
  • Rim temperatures ~200–300 K — stable, not “hot”
⚠ NASA (2022), verbatim: “there are no ‘peaks of eternal light’ on the Moon.” Every site is shadowed at some point in the cycle. Plan for the dark days; don't name them away.
Waena · The middle lands

Lewa, forest & houses

The temperate living bowl: Lewa and the libraries of biology at the centre, forest terraces and gardens on the slopes, small houses under the fused cap. The zoning now completes itself: the centre sits under the deepest regolith cover, so Lewa — children, teachers, living collections — is the most shielded place in the base, not the least. Walk outside your house; you're still inside the world.

  • Climate here is made, not given: whatever the lid's energy balance and thermal system hold it at
  • Temperate trees want roughly 5–35 °C, frost-free, with seasonal cues
  • Air sweet spot ~35–60 kPa, ≥30% inert diluent (fire rule)
✓ This is the zone the science supports building — as an actively managed greenhouse, not a passive one.
Kai · The lowlands & the sea

The ice basement — and beyond

In an open polar crater the floor is a natural deep-freeze holding real ice. Under a lid it becomes an engineered cold vault. And the ahupuaʻa completes itself outside: a chute down the crater's outer wall to the loko iʻa, the fishpond — the sea at the end of the watershed.

  • Measured PSR floors: 35 K detected; 26 K in Hermite — colder than Pluto's surface
  • Shackleton's floor: ~90 K; ice is stable below ~110 K
  • LCROSS found 5.6 ± 2.9% water ice by mass in Cabeus
⚠ Warm air destroys a passive cold floor — convection sees to that. The vault must be insulated, refrigerated and airlocked: a cryogenic work zone entered in protective gear, not a chilly stroll.
The isolation is the point. Rim, middle and basement are sealed off from the vacuum together — one watershed, one atmosphere, one fused cap between the garden and a Moon that would strip an unprotected atmosphere away in about a hundred days.
The loko iʻa is the seal, not just the pond. The ring at the crater's edge does both jobs at once: it is the air seal (the lid's rim line, holding the pressure cascade) and the pollution seal (the one-way boundary every waste stream crosses outward and nothing dirty crosses back). Inside it, the ahupuaʻa repeats — zone inside zone, room inside room, each with its own index and its own smaller seal (the nested system). Outside it: the edge plants, the tank farm, the five sorted bays, the sulphur works — and, at the end of the chute, the fishpond the ring is named for. Full sorting table: tanks, chemicals & waste.

Every tank, chemical and waste stream — labelled and sorted

The sorting table for the whole base, built from real programmes' practice (research read 14 Nov 2026). The law: the loko iʻa is the whole air seal and the whole pollution seal — nothing bad ever goes into the middle, all storage of anything hazardous sits at the edge, and every stockpile is segregated by stream so it can be mined later instead of becoming a mixed dump. Clean consumables cross inward only in dilute, closed, engineered form: potable water, cabin O₂ at ~21%, dilute nutrient solution in a sealed loop.

Highest pressureCentre — LewaHomes, clinic, kids, teachers, libraries
↓ air leaks outwardGreenhouse ringGardens, forest, houses
↓Utility nodeToilets, UPA/WPA, digesters — one node, own ventilation
↓Airlock / dust-lock ringThe pass-through boundary
Lowest pressureEdge plantsSulphur, MRE, tanks, bays — air flows plant-ward, never home-ward

The pressure cascade is the pneumatic version of the loko iʻa: air always leaks outward, carrying contaminants away from the centre. It copies BSL laboratory zoning (clean → semi-contaminated → polluted, negative pressure in the dirty zone), planetary-protection cleanrooms (clean is a measured budget, not an intention), hospital clean/dirty corridors (two circulation systems that never cross), and ISS ECLSS (the dirty machinery concentrated in one node — Node 3 — away from sleeping quarters).

StreamComes fromHazard — with the threshold that decides itThe ruleMine-later value
Sulphur system — the sulphur mines & concrete plant
Elemental sulphur, moltenClaus conversion of extracted SO₂ + H₂S; melting for concreteWorking range for concrete only 130–140 °C; stiffens above ~148–160 °C; ignites >~200–207 °C. Never add to a vessel with trace water (steam explosion) or hydrocarbons (generates H₂S)Process at edgeHIGH — the product itself; solidify & stockpile as construction feedstock
Sulphur concrete (finished)12–22 wt% S + aggregate (terrestrial); lunar proposals up to 35% S / 65% regolith; ~70–80% strength in 24 h; remeltable ~120 °CCannot be used where temperature exceeds ~120 °C — equatorial daytime (+123 °C) is marginal. Vacuum sublimation (Grugel, NASA TM-2008-215250): ~<2 h to sublime 1 cm at 120 °C, years at ~20 °C; thermal cycling to −191 °C cut strength to ~20% in one test seriesUse buried / shaded onlyHIGH — fully recyclable by remelting; demolition blocks are feedstock
SO₂ — sulphur dioxideTroilite oxidation; sulphur burning; Claus feedSevere respiratory irritant. OSHA PEL 5 ppm TWA; NIOSH REL 2 ppm; ACGIH STEL 0.25 ppm; IDLH 100 ppm; NASA 7-day SMAC 1.0 ppm. Smell is not a safe monitorNever store insideMEDIUM — capture & feed to Claus → S + water; never stockpile as gas
H₂S — hydrogen sulphideTroilite reaction; molten-sulphur headspace; anaerobic wasteRapid systemic poison. Smell paralysed ~200 ppm; IDLH 100 ppm; OSHA ceiling 20 ppm; flammable LEL 4.0–4.3%. Heavier than air — pools low; enclosed headspace above molten S is assumed toxic and explosive until measuredNever store insideMEDIUM — Claus feedstock
Sulphur dust (fines)Grinding, solidification, block cuttingDust explosion: LEL ~35 g/m³, autoignition ~232 °C; ignites from friction / static / spark. Combustible dust, not inert sand (NFPA 655)Process at edgeHIGH — collector fines remelt directly
Slag / tailings, S extractionRegolith heated 750–1,200 °C. Lunar sulphur is only ~0.05–0.13 wt% (avg ~0.10%), mainly troilite FeS — a by-product industry, not an ore bed. Extraction releases 12–30% of S at 750 °C, 85–95% at 1,100 °CHot, dusty, residual adsorbed SO₂/H₂S when fresh; co-volatiles (Zn, Na, K, Cl, F) on fresh material. Cool & degas in sealed edge bays firstStore at edge, by feed typeMEDIUM–HIGH — magnetite for iron, bulk residue for shielding / aggregate
Oxygen production — ilmenite reduction & molten regolith electrolysis
H₂-reduction loop / MRE meltFeTiO₃ + H₂ → Fe + TiO₂ + H₂O at 900–1,050 °C; MRE silicate melt at ~1,575–1,600 °C. ~40 t high-Ti feed per 1 t O₂ (H₂ route)Extreme thermal / molten-oxide splash; hot hydrogen loop; very large tailings stream. A polar base needs beneficiation or imported mare feed (ilmenite <1% highlands vs >10% high-Ti mare)Process at edge — roboticThe melt is the ore body — process, don't store molten
Iron, silicon, aluminium, titaniumMRE cathode sequence Fe → Si → Al; TiO₂ from the H₂ routeMolten handling; Si/Al fines combustible; Ti fines pyrophoric as powderCast to ingot at edgeHIGH — first structural metal; Si for solar cells, Al for wire, TiO₂ pigment / photocatalyst. Ingots are clean and may travel inward
MRE / carbothermal slag & glassResidual Ca-Mg-Al silicate melt, solidifiedHot when fresh; thereafter inert glass, dusty if crushedStore at edgeHIGH — glass fibre, aggregate, shielding blocks. A primary mine-later pile
Hydrogen GH₂ / LH₂Electrolysis intermediate; recycled reductantFlammable 4.0–75 vol% in air; detonable 18.3–59%; LH₂ boils 20.3 K — frostbite, ice-blocked vents, near-invisible flame, embrittlement (NASA Glenn Hydrogen Safety Manual)Never insideConsumable — recycle; working inventory only, at the edge, vented
Oxygen GOX / LOXMRE, H₂-route electrolysis, OGA-type electrolysisNot flammable — it makes everything else flammable. Oxygen-enriched atmosphere = >23.5 vol%: materials non-flammable in air burn at 23%; LOX at −183 °C soaks into porous material → impact-sensitive. Every O₂-service material needs a NASA-style compatibility assessment; O₂-clean, no hydrocarbonsBulk never insideConsumable. Only breathing gas at normal cabin composition (~21%, <23.5%) belongs inside
Water & ice — the polar ice is a chemical feedstock, not water
Raw polar ice / icy regolithMining the permanently shadowed floor. LCROSS (Cabeus): water 5.6 ± 2.9 wt%; abundances relative to water (=100): H₂S 16.75, NH₃ 6.03, SO₂ 3.19, ethylene 3.1, CO₂ 2.2, methanol 1.6, CH₄ 0.7The whole volatile suite below rides along. One impact, one crater — not a polar-wide assay; the base's own crater must be sampled before these numbers are relied onMelt & distil at edgeHIGH — water plus every separated volatile below. Only polished potable water crosses inward (the ISS WPA model)
Ammonia NH₃ (in ice)~0.31–0.34 wt% of ejectaNASA SMAC 7 mg/m³ (7–180 d); eye / respiratory irritantCapture at ice plantHIGH — the base's scarcest nutrient: nitrogen for the greenhouse. Capture, never vent
Mercury Hg (in ice)LAMP detection; one compilation tabulates 0.24 wt% — treat as uncertain / instrument-dependent: instruments normalised differently, and “mercury ≈ water” mixes two different tablesNeurotoxic vapour. NASA SMAC 0.01 mg/m³ — 10× below the OSHA ceiling (0.1 mg/m³); EPA drinking-water limit 2 ppb. Never heated in an open systemNever insideLOW / NEGATIVE — a purification liability, not an ore. Sealed toxic-bay stockpile, labelled hazardous. Removable (sulphur-impregnated carbon) — “polar water is poisoned” is as unsupported as “mercury is negligible”
CO · CH₄ · methanol · ethylene (in ice)LCROSS volatiles, each ≤3.1% of water abundance (LAMP ranks CO most abundant by mass — different normalisation)CO toxic (SMAC ~10 mg/m³ long-duration); CH₄ flammable 5–15%; methanol toxic (SMAC 9 mg/m³)Separate at ice plantMEDIUM — CO reductant, CH₄ propellant feedstock, methanol chemical feedstock. Small quantities — don't design the base around them
GreywaterShower, handwash, laundry — a lunar base adds what ISS avoidsSurfactants → foaming that breaks physicochemical processors; microbial growth. Needs biological pretreatment (membrane-aerated bioreactors + RO treated >20,750 L in NASA exploration tests)Process at utility nodeLOW as a stockpile — value is water + N/P recovered immediately to hydroponics; stored greywater putrefies
Humidity condensateCabin dehumidification — breath, sweat, greenhouse transpirationLow chemical load, real microbial / biofilm load (ISS biofilms clog the water system)Polish & recycleWater only — recycle immediately through WPA-type filtration + catalytic oxidation + ion exchange
Greenhouse & biology — Lewa's working chemistry
Nitrate salts — KNO₃, Ca(NO₃)₂, ammonium nitrateFertiliser stock; MELiSSA nitrification product is NO₃⁻ in solutionDilute in solution: low hazard. Dry bulk: oxidiser — AN fertilisers are IMDG Class 5.1 (UN 2067); heated / contaminated bulk AN is the classic explosion precursor. Keep Ca stock separate from phosphates / sulphates (precipitation clogs lines)Dry store at edgeHIGH — nitrogen is imported / ice-limited. Only pre-mixed dilute solution is piped inward
Phosphate salts & phosphate sludgeFertiliser; phosphoric-acid urine pretreat; water-plant precipitatesConcentrated acid corrosive; phosphate leaks bloom any open waterConcentrate at edgeHIGH — phosphorus has no lunar source yet demonstrated; every gram is imported or recycled. The sludge pile is arguably the most valuable mine-later pile per kilogram
CO₂ — enrichment & scrubbing productCrew metabolism (~1.0 kg / crew-day); Sabatier / Bosch; greenhouse dosingISS SMAC 5.3 mmHg (~7,000 ppm) for 180 d, with effects modelled below it. Plants want ~1,000–1,500 ppm, crew want <~1,000–2,500 ppm — greenhouse and crew air are managed separately, never one mixed atmosphereCapture & dose by zoneConsumable — carbon is food. Bosch solid carbon: MEDIUM (filler / structural)
Sabatier methane CH₄CO₂ + 4H₂ → CH₄ + 2H₂O (~515 °C, ISS unit)Flammable 5.0–15.0 vol%. ISS vents CH₄ overboard specifically because storing flammable gas onboard is considered hazardous — that venting is what keeps ISS open-loopNever tank insideHIGH at the edge — propellant (with LOX), fuel, H₂ via pyrolysis. Pipe it straight out
Digester biogas + digestateMELiSSA CI-type liquefaction of faeces, inedible biomass, urineBiogas typically 50–70% CH₄, 30–50% CO₂, trace H₂S / NH₃; volatile fatty acids corrosive; the digester is a live pathogen cultureSealed, utility edgeHIGH — digestate (pasteurised) is the fertiliser loop; biogas is fuel. A process, not a stockpile
Ethylene C₂H₄ + plant volatilesThe plants themselves; ripening fruit; plastics off-gassingThe tightest threshold in this inventory: plant hormone active at ppb. ~60–100 ppb flecked wheat leaves (NASA KSC); control to ≤50 ppb; 250 ppb → 60% wheat yield loss; 500 ppb → no seed. The crop fails long before the crew noticesDestroy in scrubberNONE — KMnO₄ / photocatalytic scrubbers in the greenhouse loop; spent media is Mn-bearing edge waste, LOW
Culture media, agar, microbial culturesBiology labs; MELiSSA CII (Rhodospirillum), CIII (nitrifiers), CIVa (Limnospira)Containment hazard, not toxicity: cultured organisms must not enter the greenhouse or crew flora uncontrolled; nutrient-rich media amplifies any contaminationLab only — autoclave outSpent media after autoclaving: LOW–MEDIUM (N/P to digester)
Algae photobioreactor wasteCIVa-type O₂ / food production (Limnospira / Chlorella-type)Culture crash → NH₃ / VFA spike; decaying algae make H₂S in anaerobic pocketsSealed reactorsMEDIUM–HIGH — harvested biomass is food / feed / fertiliser; never open troughs in living air
Human waste — the state of the art, stated plainly
Pretreated urine~1.5 L + 0.5 L flush / crew-day; pretreat CrO₃ + H₂SO₄, later CrO₃ + H₃PO₄, to stop urea → NH₃ and microbial growthPretreat makes it oxidising, acidic and Cr(VI)-bearing — Cr(VI) is carcinogenic. Untreated urine grows microbes “relatively quickly” (NASA)Sealed — process at onceWater + N / P / K — process immediately, never open storage
Urine brine — the key stockpileUrine-processor residue; ~22 L fills an ARFTA tank roughly every 3–4 weeks (6 crew). Brine processing demonstrated 98% total water recovery (June 2023) — a demonstration milestone, water-only, not total closurepH ~2.0, high oxidation potential, Cr(VI) content → ISS toxicity level 2 (tox-2). Brine processors are designed for “safe containment of this toxic substance” with minimal crew manipulationNever inside — sealed ARFTA at edgeMEDIUM, with a poison asterisk: dried salts hold N, P, K, NaCl — but Cr(VI) pretreat contaminates the whole stream. Design lesson: choose a non-chromium pretreat (or the MELiSSA biological route) if brine salts are meant to be mined later; otherwise the pile is hazardous waste, not ore
Faecal waste (raw)~0.03 kg solids + 0.09 kg water / crew-dayPathogen load, odour / VOCs, gas in storage — the highest biological-risk routine stream. ISS does not recycle faeces: bagged, sealed, burned up on re-entryNever stockpile rawNONE raw. Only the sterilised product may be stockpiled
Torrefied faecal / solid-waste charTorrefaction (mild pyrolysis) of faeces + wipes, paper, food waste; ultrasonic drying recovers 80–100% of faecal waterDry, free-flowing, hydrophobic, does not support bacterial growth in tests; dust controls applyStore at edge, sealedMEDIUM — soil amendment / carbon stock / shielding filler
Regolith & dust — the boundary is the dust lock
Lunar dust (<10 µm)All surface work, EVA ingress, mining, construction. Apollo: abraded seals, floated in the LM, “hay-fever” symptomsNASA-STD-3001 limit: 0.3 mg/m³ in habitable air (up to 6 months intermittent). Freshly ground dust generates 2–3× the hydroxyl radicals of simulant, 10× quartz; LADTAG rat study: moderately toxic, NOAEL 6.8 mg/m³. Not a silicosis dust — ~50% amorphous silica; 2025 primary-cell work found no IgE allergy response. And no perchlorate hazard is established for the Moon — that is a Mars finding; don't import itNever inside — suits stay outNONE as a separate product — it is regolith feedstock. Keep it out of lungs, not in a stockpile
Bulk regolith tailingsBeneficiation reject, excavation spoil, sintered-block offcutsDust hazard only; otherwise inert silicateStore at edge, graded by sourceHIGH (bulk) — radiation shielding by the metre, roadbed, landing pads, aggregate, MRE feed. The largest mine-later pile by mass
Propellants & batteries — if present, they live at the landing zone
Hydrazine / MMH / NTOVisiting vehicles only — a base that makes its own H₂/O₂ needs none insideToxic, corrosive, probable carcinogen, skin-absorbed. OSHA 0.1 ppm TWA (skin); NIOSH ceiling 0.03 ppm / 2 h; odour threshold ~3.7 ppm — over the limit before smell can warn you. Leakage classed catastrophic under range safety; fuel and oxidiser in separated bunkersNever inside — SCAPE suits onlyNONE — destroyed, never mined. Contaminated PPE is hazardous edge waste
Green propellants — AF-M315E / LMP-103SLander / small-vehicle alternative (HAN- / ADN-based)Lower-toxicity, not non-toxic: energetic ionic liquids, potentially explosive crystalline (kept in aqueous blend), catalyst preheat >350 °C, decomposition ~1,600–1,827 °C vs ~900–927 °C for hydrazine. The gain: ambient handling without SCAPELanding-zone storageNONE
Lithium-ion batteries (in service)Rovers, tools, suits, backup power (NMC / LFP / LCO)Thermal runaway: vent gas H₂ 36–54%, CO ~24%, CO₂ ~22–27%, CH₄ ~6%, plus HF / POF₃; onset ~190 °C in heated-cell tests; single small cells released ~3 L (LFP) to ~41 L of gas. Chemistry changes the hazard — LFP propagated without flame in some tests, NMC through every cell with fire — it does not remove itCertified packs, vented lockersIn service — n/a
End-of-life / damaged Li-ionReplacements, rover & tool retirementDamaged cells are more likely to run away; Co / Ni content is toxic-metal waste. Never in general trash, never compacted — compaction can trigger runawayFire-proof edge containerMEDIUM — Co, Ni, Li, Cu, Al are genuinely scarce on the Moon. A legitimate mine-later pile if stored runaway-safe, segregated by chemistry (NMC ≠ LFP value)

Sources: Vaniman et al., Uses of Lunar Sulfur, NASA NTRS — ntrs.nasa.gov/api/citations/19930004793/downloads/19930004793.pdf · Sulphur concrete review — cdnsciencepub.com/doi/10.1139/cjce-2021-0098 · Grugel, NASA TM-2008-215250 — ntrs.nasa.gov/api/citations/20080022947/downloads/20080022947.pdf · Molten-sulphur SDS via EPA — response.epa.gov/site/download.ashx?counter=420419 · H₂-reduction / MRE — ntrs.nasa.gov/citations/20170001458 · NASA Glenn Hydrogen Safety Manual ch. 6 — www1.grc.nasa.gov/wp-content/uploads/chapter_06.pdf · MELiSSA — frontiersin.org/journals/microbiology/articles/10.3389/fmicb.2021.700810/full · APMBR, ICES-2024-347 — ntrs.nasa.gov/api/citations/20240007366/downloads/ICES-2024-347.pdf · Ethylene, NASA KSC — ntrs.nasa.gov/api/citations/19960000867/downloads/19960000867.pdf · CO₂ SMAC — ntrs.nasa.gov/citations/20160014530 · Brine / ARFTA — ntrs.nasa.gov/api/citations/20160009716/downloads/20160009716.pdf · ECLSS review, ICES-2025 — ntrs.nasa.gov/api/citations/20250003953/downloads/ICES-2025-125%20Final_Rev%20C.pdf · Dust — ntrs.nasa.gov/citations/20090001228 · Hydrazine, ATSDR — wwwn.cdc.gov/TSP/PHS/PHS.aspx?phsid=499&toxid=89 · Batteries, ICES — ntrs.nasa.gov/api/citations/20250005133/downloads/ICES_Fina_426.pdf · Cleanroom doctrine — planetaryprotection.jpl.nasa.gov/mission-implementation · Mercury, OSHA — osha.gov/chemicaldata/505

The edge is sorted too — five bays, never one dump

A single mixed “mine later” pile destroys both safety and value. Incompatibles are separated by distance and berms, exactly as industrial chemical zoning separates oxidisers from fuels on Earth.

(a) Oxidiser bay

Dry nitrate salts, LOX-side hardware. Away from fuels, organics and sulphur.

(b) Fuel bay

Sulphur stock, CH₄ / H₂-side hardware, battery quarantine in its own fire-proof container.

(c) Toxic bay

Sealed ARFTA brine tanks, mercury waste & Hg filters, hydrazine-contaminated material — each labelled by what it holds.

(d) Inert bay

Slag, glass, tailings, ingots, sulphur-concrete blocks — the bulk mine-later piles.

(e) Biological bay

Torrefied char, pasteurised digestate, autoclaved spent media — sterilised product only, never raw waste.

Never stored inside a living zone — under any circumstances

  1. Hydrazine / MMH / UDMH and NTO — carcinogenic at 0.1 ppm, skin-absorbed, SCAPE-only. No quantity is a safe quantity in recycled air.
  2. Bulk LOX or high-pressure GOX — anything that could push O₂ above 23.5% turns furnishings into fuel.
  3. Bulk hydrogen — 4–75% flammable range, invisible flame, leaks accumulate in enclosed volumes.
  4. Molten sulphur, bulk solid sulphur, sulphur dust — H₂S headspace to toxic and explosive levels; dust-explosion risk; SO₂ on any fire.
  5. SO₂ / H₂S in any stored quantity — IDLH 100 ppm for both; H₂S paralyses the sense of smell that would warn you.
  6. Urine brine and pretreat chemicals (CrO₃, concentrated acids) — pH ~2, Cr(VI) carcinogen, tox-2. Raw urine in open or long storage likewise (microbial growth, NH₃).
  7. Raw faecal waste — pathogen load and gas production; only sealed bags in transit outward, only sterilised product stockpiled.
  8. Mercury and mercury-bearing waste / filters — SMAC 0.01 mg/m³; a spill in a closed habitat is a long-term contamination event, not a clean-up job.
  9. Dry bulk ammonium nitrate & oxidiser fertiliser salts — explosion precursor when heated or contaminated; only dilute solution belongs inside.
  10. Damaged, swollen or end-of-life lithium-ion batteries; bulk battery charging — runaway releases H₂ / CO / HF into closed air and propagates cell to cell.
  11. Methane in bulk — 5–15% flammable; ISS vents it rather than store it. Pipe it out.
  12. Uncleaned regolith, dusty suits, EVA equipment — the inside dust limit is 0.3 mg/m³. Cleaned, cast or sintered products are building materials; raw regolith is not.
  13. Live digester / MELiSSA CI cultures and un-autoclaved lab waste — an active pathogen-and-VFA reactor; cultures cross boundaries only after autoclaving.

Inventory — claims checked, and what could not be verified

  • “The Moon has rich sulphur deposits.” Not supported: ~0.1 wt% in mare soils, ~0.05% in many highland soils. Sulphur concrete works despite this, because concrete needs only 12–35% S and regolith throughput is already large for O₂.
  • “ISS recycles everything / 98% closed.” False: 98% was a demonstrated water-recovery milestone (2023, with the Brine Processor); faeces and trash are disposed, Sabatier CH₄ was vented, food and O₂ are resupplied. MELiSSA full closure is a ground pilot plant (UAB Barcelona), not flight-proven.
  • “Green propellants are non-toxic.” They are lower-toxicity, low-vapour alternatives that remove SCAPE handling — and still energetic, and ~700 °C hotter in decomposition than hydrazine.
  • “Ammonium nitrate = bomb in the greenhouse.” A conflation: dilute hydroponic solution is routine and low-hazard; the hazard belongs to dry bulk storage, heating and contamination — which is exactly why the dry store is at the edge.
  • Open: no published assay exists for polar ice at any specific crater other than the LCROSS Cabeus point — Hg and CO at this base's own crater must be sampled first. No lunar-specific storage standard exists for torrefied char, MRE slag or brine salts as long-term outdoor stockpiles; terrestrial hazardous-waste segregation is the model used here.

Ahupuaʻa inside ahupuaʻa — every room, every garden, indexed

Not just dome inside dome. The watershed logic repeats at every scale: the crater is one ahupuaʻa; inside it, every zone is one; inside that, every room and every garden is its own ahupuaʻa — its own boundary, its own inputs and outputs, its own quarantine rule, and its own index. Because every unit is indexed, the channel network can address it: if a signal comes through asking for a resource, the channel sends that resource to that address. Minerals are routed the same way as electrical signals — a request is a packet; the dose is the payload; the index is the address.

Worked example — a signal asks for potassium

1 · Sense

The bed reports

Bed G03-B12's sensors read potassium falling in its loop. The bed — its own ahupuaʻa — raises a request in its own name.

2 · Address

The request is a packet

“K⁺ · measured dose · to G03-B12 · via clean loop.” The index is the address — the same code the Teachers log in Lewa's records.

3 · Pass-through

Dry salt never travels

The utility node draws KNO₃ from the edge dry store and mixes a dilute solution. The oxidiser stays at the edge; only the dilute, verified solution crosses inward — the pass-through rule, applied to a mineral.

4 · Deliver

The channel routes it

The channel network delivers the dose to B12's root pouch — feeding by proximity, the flown Veggie / Advanced Plant Habitat pattern (S4) — not to the garden generally. Addressed, not broadcast.

5 · Return

The loop closes outward

Spent solution drains one way — bed → garden drain → digester / water plant — where potassium is recovered and logged. Nothing pools, nothing crosses a clean corridor, nothing is dumped mixed.

The same addressing carries nitrogen from captured ice ammonia and urine, phosphate from the sludge-recovery line, CO₂ from the crew loop to a garden that asks for enrichment, and fibre-carried light to a shaded bed. One network, two kinds of traffic — signals about matter, and matter itself — which is why the index has to be exact: a dose sent to the wrong address is not a typo, it is a contamination event in a unit that keeps its own quarantine.

The living-materials studies, built in — at the scale each one has earned

Root pouches in every bed — flown practice

Exists and tested

Veggie pillows and the Advanced Plant Habitat's porous clay substrate already feed roots water, nutrients and oxygen by proximity on the ISS (S4). Every Level-3 unit in the nested system is this device, indexed: a pouch, a pumped oxygenated feed, a drain.

Microvascular channel mats — the network's material

Lab scale · engineering development

The channel walls themselves copy the demonstrated vascular materials (S7): 200 µm channel networks (Toohey et al. 2007) that circulate fluids for temperature regulation, chemistry and self-repair, and pumped systems that filled >35 mm damage in 20 minutes (White et al. 2014). In the base they line greenhouse panels, garden beds and the tree sheaths — circulating water and nutrients, carrying heat away, and sealing their own punctures. Demonstrated at coupon scale; the base is the scale-up programme, and the page says so.

MELiSSA — the utility node is a chain of living units

Ground pilot · not flown

ESA's MELiSSA is the nested idea already drawn by someone else: CI liquefaction (faeces, urine, plant waste → VFAs, CO₂, NH₄⁺) → CII Rhodospirillum rubrum removes VFAs → CIII nitrifiers turn NH₄⁺ into NO₃⁻ → CIVa Limnospira algae → CIVb higher plants → CV crew. Each compartment is a sealed, indexed unit passing only its product to the next — an ahupuaʻa chain. It is a ground pilot plant, not a flown closed loop; the base adopts its architecture, not a claim of proven closure.

Source: MELiSSA review — frontiersin.org/journals/microbiology/articles/10.3389/fmicb.2021.700810/full · NASA KSC anaerobic–phototrophic membrane bioreactor (APMBR), ICES-2024-347 — ntrs.nasa.gov/api/citations/20240007366/downloads/ICES-2024-347.pdf

The fungal mantle — template for pouches, not for canopies

Root trade real · tree-to-tree contested

The ectomycorrhizal mantle (S5) is the design template for Level-3 root units: a sheath around the root that trades nitrogen, phosphorus, potassium and water across living membranes. It is not a template for feeding leaves, and the popular tree-to-tree network claims remain contested (Karst et al. 2023) — the base builds on the root trade only.

Engineered living materials — used where they are honest

Proposed only

Truly living channel linings — cells embedded in the material that sense and respond — are demonstrations, limited by cellular viability, restricted mass transport and long-term stability (2026 review, S7). The base specifies synthetic pumped vasculature where a function is load-bearing (pressure, dosing, self-heal) and treats living layers as monitored experiments inside Lewa's indexed units — where a failure is quarantined by design instead of spreading.

Life under the lid

Flight · inside the air

The eagle suit

The suit belongs to the pressurised bowl: with real air trapped under the lid and lunar gravity at 1/6 g, a winged suit glides the long, slow traverse from rim terrace to Lewa — commuting, surveying the forest canopy, reaching the anchor ring for lid inspection.

  • It only works because of the cap: no trapped air, no glide. Outside, it's a pressure suit's job.
  • Under the lid it needs no pressure shell — the crater itself is the suit's outer layer. That's the double protection working for people, too.
  • Watch it in the cross-section above — the gold figure circling under the cap.
Lewa · the centre is full

The libraries of biology

Lewa, at the crater's heart, holds the base's living memory: seed vaults, living collections, and the records of everything that grows in the bowl — the reference library the gardeners work from, lit green in the cross-section.

  • Seed vault — the reserve if a terrace crop fails.
  • Living collections — the species library, growing, not shelved.
  • Growth records — what actually thrived at 35–60 kPa, in amended soil, under polar light: the base's own tested science, replacing guesses year by year.

Why it matters scientifically: the published record today holds small plants in chambers — nothing at ecosystem scale. A library that logs real results is how “proposed but unproven” becomes “tested.”

The verdicts in full

And here's the science, for the professionals — each question judged separately, evidence first, wherever an idea fails. Visitors: you've already had the tour. This is the paperwork that lets us give it.

1 · Paraterraforming — the worldhouse lid

Proposed · unproven

A real, named literature concept — not established engineering. Established physics makes a 1 km, 1 bar lid an anchoring problem first and a materials problem second.

  • Who proposed it: Richard L. S. Taylor (1992), “Paraterraforming – The Worldhouse Concept,” JBIS 45:341–352 — a roof a kilometre or more up, pressurised, “anchored with tension towers and cables.” Recounted in Fogg, Terraforming (1995). Taylor's roof was drawn transparent; this design's cap is not — the optics moved to the collector lens on the rim (see the CAD sheet), and the cap carries shielding instead of windows.
  • Crater-specific, modern: Mrozek et al. (2025), Scientific Reports 15:24744 — a regolith-geopolymer lid “to maintain atmospheric pressure conducive to human life.” Scale caveat, stated in the paper's own terms: the crater analysed is 17 m across, 6 m deep. Concave lids won: internal pressure puts them in compression, where geopolymer is strong.
  • Uplift (established physics): pressure pushes up on every square metre with vacuum outside — 101 kPa ≈ 10.3 tonnes-force/m². Published ballast arithmetic (Mrozek et al.): balancing ~100 kPa with dead regolith alone takes 28–42 m of it; at 600 hPa, 17–25 m. Roberts' NASA inflatable-habitat work agrees: 101.4 kPa “is capable of supporting 40 m of soil on the Moon.”
  • At our scale (calculation, labelled as such): a 1 km crater at 101 kPa feels ~7.9×10¹⁰ N of uplift — ~49 million tonnes of lunar-weight ballast-equivalent. At 50 kPa, half. At 30 kPa, ~15 million tonnes. Anchors, a compression ring, or a heavy lid must supply it, continuously, for decades.
  • Leaks & impacts: tightness is “a priority issue” that “determines the safety level of the entire engineering structure” (Mrozek et al.). The Moon takes ~1.4 t/day of micrometeorite flux at 5–65 km/s. This is the conflict: a transparent lid can't also carry metres of shielding — transparency, radiation protection and puncture resistance pull in opposite directions — so the cap gives up transparency — bar one small glass eye at the apex, kept deliberately as a trial plot, the cover's Lewa (below). The fused regolith sheet — raised on its fibre-pole skeleton and fused in one melt pass (Founding III) — does the shielding and the ballast; the separate collector lens does the optics. The big volume is a genuine safety plus: a puncture is a chronic leak to make up, not instant decompression.
  • Lower pressure helps — linearly: NASA-STD-3001 sets indefinite crew exposure at 34.5–103 kPa with ≥30% inert diluent gas (fire and lung rules); body fluids boil at 6.27 kPa. Practical sweet spot in the literature: ~35–60 kPa, O₂-enriched, never pure O₂.
Keep, with eyes open — and fused: the cap is the design's strongest idea — one roofed crater needs ~10⁵ tonnes of air where a whole-Moon atmosphere needs ~10¹⁴. The design: pressure skin + fused regolith sheet, fibre light only, the lens on the rim as collector. But nobody has engineered it past 17 metres.

Sources: phys.org summary of Taylor/Fogg — phys.org/news/2016-08-colonize-mercury.html · Mrozek et al. 2025 (full text) — pmc.ncbi.nlm.nih.gov/articles/PMC12241417/ · NASA-STD-3001 brief — nasa.gov/wp-content/uploads/2023/12/ochmo-tb-003-habitable-atmosphere.pdf

2 · A free atmosphere — or ozone — with no lid

Contradicted

Contradicted as a permanent or standalone solution. The lid isn't decoration; it's the only thing standing between the garden's air and the solar wind.

  • Today's Moon: an exosphere under 10 tonnes total, ~3×10⁻¹⁵ atm, needing “constant replenishment.” Escape velocity 2.38 km/s.
  • The thin-atmosphere killer (established): sunlight ionises released gas and the solar wind sweeps the ions away in ~100 days (Landis 1990, Acta Astronautica 21:183–187; independently stated in NASA NTRS lunar-environment work: photoionisation “within a period of weeks to months”). Light gases go faster still — H₂/He in ~15 minutes on the sunlit side.
  • A thick one? Possible in principle, planetary in scale: Landis's breathable Moon needs ~200 trillion tonnes of O₂ — a ~50 km cube of rock reduced — generated at something like 10,000 tonnes/day before shielding even begins. Precedent: the Moon's own volcanic atmosphere of 3–4 billion years ago lasted ~70 million years and was still stripped away (NASA, Needham & Kring 2017).
  • Lifetimes disagree — reported honestly: Landis estimates ~3,000 years for a thick artificial atmosphere; Fogg-derived figures for thermal escape alone run 10⁷–10⁸ years. The literature does not reconcile these. Settled points only: thin ≈ 100 days; thick = millennia at least, never forever.
  • Ozone without the lid: impossible on its own — it needs the thick O₂ atmosphere first (Verdict 3), inherits all of the above, and would be a trace by-product (0.00006% on Earth), not a mechanism that holds anything up.
Closest real mechanism: the fused cap itself does ozone's protective job locally — metres of fused regolith over the pressure skin, and UV sorted out at the collector before light is piped in. Outside it, suit and habitat shielding.

Sources: Landis, “Air Pollution on the Moon” (full text) — geoffreylandis.com/moonair.html · NASA NTRS — ntrs.nasa.gov/citations/19930008266 · Vondrak (1974) Nature 248:657–659 is the early published proposal in this line — a proposal, not a demonstration.

3 · Ozone and trees

Contradicted — abandon completely

No such mechanism exists in any real science found. Full evidence in the ABANDONED panel above.

  • Plants take in CO₂ and O₂. Ozone is a substrate for neither photosynthesis nor respiration.
  • Ground-level ozone is commonly called the most damaging air pollutant for vegetation (USDA/EPA summaries); a California modelling study predicted >5% biomass loss for 17 of 20 herbaceous species in at least one county, >10–20% for the most sensitive.
  • Even Earth's stratospheric peak is only ~8–15 ppm — protective precisely because nothing living is in contact with it, 30 km overhead.
What does its intended job instead: UV protection → the fused regolith sheet, plus sorting at the collector. Feeding trees → the enclosed air itself: CO₂ from people, O₂ from leaves, water, nutrients, and fibre-delivered light.

Sources: NASA Ozone Watch (full text) — ozonewatch.gsfc.nasa.gov/facts/ozone_SH.html · NASA Aura — science.nasa.gov/mission/aura/ozone/ · EPA, Our Nation's Air 2016 — gispub.epa.gov/air/trendsreport/2016/ · Ozone biomass study — pmc.ncbi.nlm.nih.gov/articles/PMC10543089/

4 · Trees under the lid

Plants supported · Trees unproven

Pressure is not the blocker. CO₂ partial pressure, water, light across the lunar cycle, and soil chemistry are the real constraints.

  • Low pressure works for small plants (tested): NASA/Kennedy review — some plants grow below 20 kPa, perhaps at 5–10 kPa. Richards et al. (2006, Astrobiology 6:851–866) grew Arabidopsis at 101 down to 10 kPa: with CO₂ partial pressure held non-limiting (0.07–0.10 kPa), photosynthesis was insensitive to pressure down to 10 kPa.
  • The catch: below ~50 kPa, Arabidopsis switches on drought- and pathogen-response genes — low pressure reads to a plant as water stress. Humidity and watering must be actively managed. (Paul/Ferl, NASA Task Book)
  • CO₂ fraction must rise as pressure falls: at 30–50 kPa total, holding ppCO₂ near 0.05–0.10 kPa means ~0.15–0.3% CO₂ — several times Earth's 0.042%, still far below human toxicity limits. Compatible, but engineered.
  • Real lunar soil (tested, with a warning): Paul, Elardo & Ferl (2022, Communications Biology 5:382) — Arabidopsis germinated in actual Apollo 11/12/17 regolith, but grew slowly with “severe stress morphologies” and salt/metal/oxidative-stress gene signatures. Raw regolith has no carbon, no nitrogen, no microbiome. Trees in unamended crater dirt are contradicted by this evidence; amended soil or hydroponics is the route.
  • Light: non-polar sites swing between 14 days of light and 14 of dark. Polar sites get grazing light — and none of it enters directly: the collector lens (sulphur glass, n = 2.0, critical angle 30°) gathers the low sun at the rim and the fibres deliver it, sorted, to the emitters. Collector area, fibre losses and emitter placement — not cap transparency — are the light budget now. Measured photosynthesis optima run ~11–31 °C across studied species (USDA ARS).
  • Trees specifically: no experiment found — anywhere — growing a woody tree at low pressure or in lunar regolith. Direction is reasonable; proof does not exist. Tree risks are their own: xylem water transport under changed transpiration, multi-year nutrient cycling, fungi and pollinator partners, wind loading in 1/6 g.
Build order the evidence supports: herbs and crops first at 10–25 kPa in chambers → amended-soil garden under the fused cap at 35–60 kPa, lit by fibre → trees as the long experiment, logged in Lewa's library.

Sources: NASA hypobaria review — ntrs.nasa.gov/citations/20040088353 · Paul/Ferl Task Book — taskbook.nasaprs.com · Apollo-regolith study DOI 10.1038/s42003-022-03334-8 · USDA ARS — ars.usda.gov/research/publications/publication/?seqNo115=304049

5 · The peak and the basement

Real in vacuum · Not passive under a lid

Supported by measurement — for an open, airless polar crater. The design's one true conflict lives here: the temperature contrast exists because there is no air. You cannot keep the contrast and add the air.

  • The type example is real: Shackleton — 21 km wide, 4.2 km deep, pole on its rim — holds lit rim points and a ~90 K floor in one crater: a rim-to-floor contrast of roughly 150–210 K. Measured, not hypothetical. (Our 1 km crater, ~150–200 m deep, has far less room for three climates.)
  • Ice evidence: LCROSS (Cabeus, 2009) excavated 155 ± 12 kg of water in view, 5.6 ± 2.9% ice by mass; LEND found Cabeus hydrogen at 0.5–4.0% water-ice equivalent; Moon Mineralogy Mapper saw exposed surface ice in only ~3.5% of cold traps — real, percent-level, patchy, mixed into regolith like icy concrete. Not a clean glacier, and not in every shadowed crater.
  • Under the lid: warm air convects and conducts heat down to any cold floor in contact with it. Stratification helps a little; people, lights, ventilation and a sun-warmed cap mix it away. Below-110 K ice storage under ~290 K air requires a thermally separate, insulated, actively refrigerated compartment — exactly how the vault is drawn in the cross-section.
  • Zonation survives as architecture, not as a gift: separate decks, insulation, refrigeration, airlocks between zones — plausible engineering. Passive zonation inside pressurised air is contradicted by basic heat transfer.
The honest version of the design: brightest peak (power) → temperate middle (life) → engineered ice vault (storage) — three zones, deliberately built and actively held. Same picture. True physics.

Sources: Paige et al. 2010 (Diviner) — ntrs.nasa.gov/citations/20100005359 · LCROSS — pubmed.ncbi.nlm.nih.gov/20966242/ · LEND — pubmed.ncbi.nlm.nih.gov/20966247 · Bussey et al. 2010 — spudislunarresources.nss.org/Bibliography/p/98.pdf · Gläser et al. 2014 — ui.adsabs.harvard.edu/abs/2014Icar..243...78G · NASA OTPS 2022 — nasa.gov/wp-content/uploads/2023/05/nasa-otps-lunar-landing-and-operations-policy-analysis-final-report-2.pdf

Open questions — marked plainly, not filled in

  • No published structural analysis of a crater lid anywhere near 1 km span exists. Largest found: 17 m (Mrozek et al. 2025). Taylor/Fogg treatments are conceptual, at planetary scale.
  • No published model of a lunar ozone layer under an artificial thick atmosphere was found; the ozone conclusion is an inference from Chapman chemistry plus published escape rates — labelled as inference.
  • Thick-atmosphere lifetimes in print disagree by orders of magnitude (~3,000 years vs. 10⁷–10⁸ years for thermal escape alone). Neither figure should be quoted alone as settled.
  • No tree — no woody plant of any kind — has been grown at low pressure or in lunar regolith. All plant evidence is herbaceous model and crop species.
  • Exact Shackleton rim maximum temperature varies by source (~220–300 K); the range is used, not a single value.