Pollution, Waste & Organic Systems Design¶

Document Type: Design Document Version: 0.3 (Draft) Status: In Progress Last Updated: 2026-09-11 Vault destination: 01_Design Documents Phase: Phase 1 — Launch release Related: [25_Power System Design](<./25_Power System Design.md>) · [04_Factory & Resource System Design](<./04_Factory & Resource System Design.md>) · [06_Vertical Layers & Core Power System](<./06_Vertical Layers & Core Power System.md>) · [24_Module System Design](<./24_Module System Design.md>) · [06_Outpost/02_Power Systems](<../../06_Outpost/02_Power Systems.md>)

1. Summary¶

Three interlocking systems: pollution as the consequence of industry, waste as its physical residue, and organic farming as the renewable counterweight.

The design goal is that the clean path compounds in the player's favour and the dirty path undermines itself, with no bolted-on punishment system required.

2. The Core Insight¶

Nobody on the planet breathes. Genesis is a robot, there is no organic player character, and there are no biters. Pollution therefore needs a real victim or it is just a number going up.

The answer already exists in the systems built: air pollution blocks sunlight, and sunlight powers both Solar Panels and farms.

Burn fossil fuel  →  atmosphere thickens  →  less light reaches surface
                                              ↓
                        Solar output falls  +  Farm yields fall
                                              ↓
                        Need more power  →  burn more fossil fuel

That is a genuine feedback trap, it is physically true, and it costs nothing new to build because Solar is already variable and farms are new anyway.

The clean path self-reinforces. The dirty path self-undermines. No punishment mechanic needed.

3. Three Pollutants, Three Media¶

Pollutant	Where it lives	Emitted by	Effects
Air / Smog	Atmosphere, zonal	Biomass Burner, Coal Generator, Fuel Generator, Smelter, Chemical Plant	Solar output down, farm yields down, building decay up
Ground & Water Table	Soil, zonal, leaches downward	Waste water discharge, leaking or dumped waste, buried nuclear waste	Farm yields down, extraction yields down, water quality for nuclear cooling degraded, decay up
Solid Waste	Items on belts	Smelting slag, chemical byproduct, nuclear waste, e-waste	Inert while contained. Becomes ground contamination when dumped. E-waste attracts hostiles (§6).

Correction from an earlier draft: water contamination was previously modelled as living inside the fluid network. That is wrong. Real water pollution is in the ground and the water table, put there by discharge and leaking waste. Contaminated water inside the player's own pipes is a separate problem and an over-complication.

Air pollution is zonal, concentrating around emitters and dispersing with distance, so a dirty smelting district does not ruin a solar farm on the far side of the planet. This keeps factory layout a meaningful decision. How that dispersal actually works is §3b.

Open items for this document are tracked in docs/open-items.md.

3a. Pollution and Vertical Levels¶

The two media behave differently, and both behaviours are physically true.

Medium	Vertical behaviour
Air / smog	Does not travel between levels. Each level holds its own atmospheric state.
Ground contamination	Leaches downward. Never upward. Gravity.

Air, per level¶

Level	Dispersal	What air pollution hurts	Net character
Sky	Above the smog layer	Little to nothing. Solar and wind largely unaffected.	Refuge. Clean generation.
Surface	Normal atmospheric dispersal	Solar output, farm yields, building decay	Standard. The model in §2.
Underground	None. Sealed, no atmosphere to disperse into.	No solar, no farms, so only decay — but concentration climbs far higher, so decay hits much harder.	Dirty industry zone, at a corrosion cost.

3a.1 Sky levels sit above the smog¶

Real inversion layers hold pollution low. Sky-level Solar Collectors and Wind Turbines are therefore largely unaffected by surface smog.

Consequence: dirtying the surface increases the value of building upward. That gives Sky Levels a concrete purpose beyond the efficiency table in 06_Vertical Layers, and it emerges from the pollution system rather than being granted.

3a.2 Underground has nothing to ruin but nowhere to vent¶

There is no sunlight underground and no farms, so pollution cannot reduce output there. But a sealed level has no dispersal, so concentration climbs faster and to a much higher ceiling, and building decay is severe.

This creates a real strategic geography:

Bury the dirty industry where there is no sunlight to block
Put renewables above the smog
Pay for both in accelerated corrosion below

Counter-pressure: underground is already valuable real estate. Core Drills, Deep Scanners and the best Geothermal Taps all live there (06_Vertical Layers). Filling those levels with smelters costs the player that access.

3a.2a Ground contamination leaches downward¶

Contaminants percolate through soil into the levels below. Downward only, with each level receiving a diminished share, and slowly, so the effect arrives well after the cause.

This produces a delayed consequence that is physically intuitive. A player who dumps freely on the surface in the first hours excavates Underground L1 twenty hours later and finds it already contaminated. They were never warned, but they are not cheated either, because it is obvious in hindsight. Effects of this kind are difficult to design deliberately; here it falls out of the physics.

It also disciplines the underground-dumping strategy. Whatever is dumped on an underground level poisons every level beneath it, and the deepest level becomes the sink where everything accumulates. That is precisely where Core Drills and the best Geothermal Taps live (06_Vertical Layers).

What ground contamination does: - Reduces farm yields (irrigation) - Reduces extraction yields, including Core Drills - Degrades water quality for nuclear cooling - Contributes to building decay (corrosive soil)

Open items for this document are tracked in docs/open-items.md. Contamination is restorable. See §3a.4.

3a.4 Restoration — Soil Remediation¶

Ground contamination can be cleaned up. A poisoned planet is recoverable, not permanently ruined.

Soil Remediation building: consumes power and an organic input from farms — specifically raw_biomass, from either the Biomass Harvester or the Biomass Farm ([27_Organic Farming & Biomass Design](<./27_Organic Farming & Biomass Design.md>)) — slowly reducing ground contamination in a zone.

The guard rail: poisoning is fast, healing is slow¶

⚠️ Remediation must be far slower than contamination. If it cleared contamination at anything near the rate the player creates it, dumping would become consequence-free with a delay, and the entire waste decision (§5) would collapse into a chore.

Slow and expensive keeps the original choice meaningful. It is also true to life.

Bio-remediation closes a loop¶

Bio-remediation is real: plants and microbes that absorb contaminants. Making Soil Remediation consume an organic input gives farming a second customer beyond fuel and polymers, and it means restoring a planet requires having built the renewable side rather than simply throwing power at the problem.

Restoration is per level¶

Because contamination leaches downward (§3a.2a), cleaning the surface does not clean the levels beneath. Remediation must be run on each level, and the deepest level is the hardest, since it is the sink where everything accumulated.

The full cleanup suite¶

Building	Medium	Notes
Air Scrubber	Air	Neutralises local atmospheric pollution. Essential underground. §3a.3
Water Treatment	Water, before discharge	Prevents contamination rather than reversing it. §5
Soil Remediation	Ground, already contaminated	The only reversal tool. Slow, power-hungry, farm-fed.

Open items for this document are tracked in docs/open-items.md.

3a.3 Air Scrubber — the ventilation answer¶

Without a way to clear it, underground pollution only ever rises, and a level eventually becomes unusable. That is a dead end rather than a decision.

Air Scrubber: a building that consumes power to neutralise local pollution.

Works on any level; essential underground
Mine ventilation is a genuine engineering discipline, consistent with the teaching goal
Ties pollution to the power system the same way recycling does (§5): the clean answer costs electricity, so a player choosing it must first solve generation
Does not vent pollution to another level. Pollution never travels; the scrubber neutralises it.

Open items for this document are tracked in docs/open-items.md.

3b. The Air Dispersal Model¶

Decided 2026-09-11. Closes OI-089. Air pollution is a coarse grid, not a radius.

Each level carries its own grid, per §3a, and each cell holds an amount. Three things act on every cell, every tick:

	What it does
Sources	Emitters add to the cells around them, weighted by distance. Falloff is still here, it just shapes the contribution rather than being the whole model
Diffusion	Each cell bleeds a fraction to its neighbours, so pollution spreads over time rather than appearing at range instantly
Sinks	Trees, Soil Remediation (§3a.4) and Air Scrubbers (§3a.3) subtract, plus a small natural decay in every cell

Cells are coarse — a whole factory district sits in a handful of them. Fine cells would make pollution a per-machine problem, which invites min-maxing the placement of individual smelters. Coarse cells make it a district problem, which is the scale the rest of the game plays at.

3b.1 Why not radius falloff¶

The alternative in OI-089 was a pure radius model: intensity computed on demand from nearby emitters. It is cheaper and it fails on four counts, all of which come from the same root — it stores nothing:

Pollution could never arrive. A native zone would be inside the radius or outside it, which is a switch rather than a pressure, with no ramp and no warning.
Pollution could never accumulate. A smelter running for an hour would be indistinguishable from one switched on a second ago.
Restoration would have nothing to act on. Soil Remediation and trees would be subtracting from a number recomputed from scratch every frame, so cleanup would be a fudge rather than a mechanic.
Switching the factory off would clean the world instantly, which quietly deletes the entire consequence this document exists to create.

[37_Planet Threats & Defence Design](<./37_Planet Threats & Defence Design.md>) §5.1 makes native aggression trigger on the player's air arriving where the natives live. A model that stores nothing cannot express arrival, so the grid is not a fidelity preference. It is what the threat design already requires.

3b.2 The natural decay is the free budget¶

Every cell loses a little on its own, representing the planet absorbing what it can. That produces an equilibrium: below some emission rate the world simply handles it and nothing ever stirs; above it, pollution climbs until absorption catches up, and where it settles is the number that matters.

Two things follow, and both are worth more than the realism:

The first hour is protected. An early factory is not punished for existing, which Phase 1 needs ([00_Canonical Design Facts](<../../04_Project Management/00_Canonical Design Facts.md>) §23b).
The player gets a legible target. Get back under the line is an instruction. More is worse is a mood. It also makes Soil Remediation's value visible the moment one is built, instead of being an act of faith.

3b.3 Wind¶

Wind ships in the Launch release, for everyone. Diffusion is biased in a direction, so pollution forms a downwind tail rather than a blob.

A farm downwind of a smelter is poisoned. An identical farm the same distance upwind is clean.

Wind adds no new rule. It only changes which cells receive the air, and everything downstream — suppressed farm output, woken natives — follows from the cell value exactly as before.

Why it is not a Simulation Level feature, despite looking like a perfect one: it changes where you build, and [00_Canonical Design Facts](<../../04_Project Management/00_Canonical Design Facts.md>) §24 forbids a fidelity setting that changes what is optimal. A session with wind on and a session with wind off would be different games. So it is in for everyone, or not at all.

What it buys, beyond the farm: threat direction becomes predictable. A zone downwind wakes long before an equidistant zone upwind, so defence becomes a decision about where instead of a tax on the whole perimeter. Ringing the entire base is the boring answer, and wind removes it.

⚠️ What it costs: wind is invisible. A player who cannot see it places a farm, watches it wither, and has no way to deduce why. Shipping wind at launch therefore obliges a persistent wind indicator and a tutorial beat that names it before the first farm goes down. That is real work and it is not optional; it is the price of the decision. Tracked as OI-381.

3b.4 Per level¶

The grid inherits §3a's behaviour rather than restating it:

Level	Diffusion	Wind	Natural decay
Sky	Above the smog layer; surface cells do not reach it	—	—
Surface	Normal	Yes	Normal
Underground	None. Sealed	None	None. Nothing vents

Underground is the interesting row. With no diffusion and no decay, a cell only ever climbs, which is exactly §3a.2's nowhere to vent expressed as one coefficient rather than as a special case in the code.

3b.5 What the player sees¶

Two things with two different jobs, and they must not be merged.

The world haze is always on and deliberately weak. You get a sense that the air over the smelter row is dirty without having to ask.

⚠️ The haze must stay well below the point where it obscures the factory. Players will not accept their base becoming harder to look at as it grows. A haze that reads denser equals worse gets switched off permanently on the first bad day and is never switched back on, taking the feedback with it.

The overlay is a toggle, and it carries the real intensity. Its job is to answer one question:

Is my pollution reaching anything that matters, and where do I put the cure?

Which is why it shows three things at once, not just the cloud:

The pollution, as cell intensity, so shape and gradient are readable
The sinks, so you can see whether the cure is where the problem is
The native zones, since they are the reason the overlay was opened

Show only the cloud and it is a status light. Show all three and it is a planning tool, which is what earns the keystroke.

On the HUD, two numbers rather than a world total. The local cell value with a trend, because is this getting worse is the actual question; and an emission against absorption rate pair, because that is the equilibrium in §3b.2 made visible. A world total still exists for scoring and for the supply chain contribution multiplier, derived by summing.

3b.6 What this settles elsewhere¶

Simulation Level gets a legal dial here. Cell size and tick frequency are pure fidelity: a coarser, slower grid is cheaper and blurrier, and the same layout is still the right layout. That passes [00_Canonical Design Facts](<../../04_Project Management/00_Canonical Design Facts.md>) §24 exactly where wind failed it.

And OI-365 gets a real answer. A planet nobody is standing on runs its grid at a bigger cell and a lower tick rate. It still accumulates, still spreads, still wakes natives, so a stalled chain the player finds on return has a cause they could have predicted. That is the honest over cheap constraint [39_Space Layer Design](<./39_Space Layer Design.md>) §7.2 insists on, satisfied by turning two dials rather than by writing a second, fake simulation.

Attribution is not needed and must not be added. A cell holds an amount, not a list of who put it there. Per [00_Canonical Design Facts](<../../04_Project Management/00_Canonical Design Facts.md>) §25 the world is shared and nothing in it has a per-player owner, so whose pollution was that has no answer and needs none — in co-op the farm is the world's farm, and the message is output is suppressed here, identical to what a solo player sees.

4. Building Decay¶

Decay is included, but under one hard rule that determines whether players tolerate it.

4.1 The rule¶

Decay is caused by the player's own pollution. It is never a universal tax.

A clean factory decays at essentially zero. A dirty one corrodes. This makes decay a feedback loop the player controls, not a chore imposed on them. Many factory players actively dislike upkeep mechanics, and Factorio deliberately has none; this framing is what separates the two.

4.2 Never destructive¶

Decay never destroys a building and never loses player progress
It reduces efficiency until repaired
Genesis repairs automatically via the existing task queue, so the game stays AFK-safe
This matches the stated Outpost Mode principle of no death and no failure, keeping the two modes philosophically consistent

4.3 Counters¶

Counter	Effect
Reduce local air pollution	Root cause. Always the best answer.
Maintenance module	Roof module slot. Reduces or eliminates decay for that building.
Genesis repair task	Automatic, queued, no player input required

4.4 Decay affects productivity only¶

Decay is a gradient, not a fault. It reduces a building's productivity and does nothing else.

It therefore does not appear on the base ring. Ring status is reserved for binary fault states (stalled, under-voltage, no power). A machine running at 80% is a number, not an alarm.

Decay feeds the existing Analytics tab, which already shows per-building efficiency percentage, colour-coded and sortable worst-first (17_HUD Panel Flap Tab 3). No new UI is required.

In-world, decay dims the cyan working lights on the machine body, scaling emission energy from 2.5 down to roughly 0.9. A worn machine looks tired; a maintained one glows properly. This is the correct channel for it: decay is continuous and brightness is continuous, while the ring handles binary faults. A neglected district reads as visibly duller across the factory. See 07_Art Direction Style Bible §1.7a.

Brightness is an ambient hint, not a readout — the precise figure stays in Analytics.

This resolves what looked like a three-way contest for amber. Only two fault states remain, and they separate cleanly by system. See 07_Art Direction Style Bible §1.7b.

5. Waste Handling — Three Choices¶

All three are valid. Each has a distinct cost, so this is a real decision rather than a right answer.

This applies identically to solid waste and to contaminated water. One structure, two media, so the player learns it once.

Option	Solid waste	Contaminated water	Up-front cost	Ongoing cost
Store	Silo	Tank	Cheap	Space, and logistics forever. Fills and backs up the line.
Treat / Recycle	Recycler	Water Treatment plant	New buildings, research	Power draw, and it is significant
Dump / Discharge	Dump	Discharge	Nearly free	Ground contamination (§3a.2a), and e-waste attracts hostiles

The Water Treatment plant is the direct parallel to the Air Scrubber (§3a.3): the clean answer always costs electricity, so a player choosing it must first solve generation.

5.0a Incinerator — controlled disposal¶

Open dumping and incineration are both disposal, but they create different pollutants:

Method	Cost
Open dumping	Ground contamination. Leaches downward (§3a.2a). Hits extraction and farm yields.
Incinerator	Power draw plus air pollution. Hits solar and farm yields. Does not leach.

The player chooses which pollutant they would rather create. Neither is free.

The Incinerator also answers the inevitable "what do I do with 40,000 spare stone", giving surplus a deliberate exit that still costs something.

It recovers heat as power, but recovery is always a fraction of the item's embodied energy, so burning is never better than not producing. In normal use it remains net-negative or break-even. See 20_Belt Logistics - The Router §3b.2.

DEFERRED: per-building waste heat recovery (smelters, refineries, and similar) is a separate system for a later session.

5.1 Nuclear waste is the exception¶

Nuclear waste does not attract scavengers the way e-waste does. Its distinct character is that dumping or leaking it produces severe, long-lasting ground contamination.

It is therefore the one waste stream a player genuinely cannot be lazy about, which is the right feel for it.

Storage gives the Silo a job nobody wants, which is a good use for it. Recycling ties waste directly to the power system, so a player choosing clean disposal must first solve generation. Dumping is the tempting early shortcut that quietly compounds.

Open items for this document are tracked in docs/open-items.md.

6. Hostiles — Scavengers, Not Biters¶

Extended 2026-09-05 by [37_Planet Threats & Defence Design](<./37_Planet Threats & Defence Design.md>). This section's argument was accepted and widened to the whole game: everything that attacks is a machine, planet natives included. Two things follow. Pollution is now a trigger too, drawing the Natives, which is what makes this document load-bearing in Phase 1 rather than a number that goes up. And the family described below needs a new name, because Scavengers collides with The Salvagers faction.

Hostiles in this world are drawn to electronic waste and power signature, not to pollution generally.

Rogue machines and scrap drones assembling themselves from discarded electronics fit a robot world in a way organic monsters never would, and they connect naturally to the drone hostiles already designed for Outpost Mode.

Attraction factors: - E-waste dumped or stockpiled in the open - High local power signature (large generators, dense grids)

⚠️ Confirmed 2026-09-06: this means generation, not consumption. Power draw does not attract hostiles. Only generators and grid density do.

That is deliberate and load-bearing. It means a site powered entirely by imported Charge Cells is quiet, which [39_Space Layer Design](<./39_Space Layer Design.md>) §4a.4 turns into a real player choice: generate on a rock you do not care about and ship the charge to the site you do. The threat is relocated and paid for, not removed.

This gives dumping a consequence with teeth rather than a rising number, and it links pollution to the existing war and defence layer.

Open items for this document are tracked in docs/open-items.md.

7. Organic Farming — Feedstock, Not Food¶

Nobody eats. This is industrial agriculture for feedstock, which is a genuinely less common framing than survival farming and fits the world.

Authority: [27_Organic Farming & Biomass Design](<./27_Organic Farming & Biomass Design.md>). This section summarises it. Where the two disagree, that document wins.

7.1 Outputs¶

Two families, not three crops. Both are built as a Control Hut plus Bays, where the head is an ordinary building and the bays are extensions with no ports, no module slots and no base ring.

Family	Head + extension	Output	Feeds into
Biomass Farm	Control Hut + Greenhouse Bay	`raw_wood`	Biomass Burner (`25_Power System Design` §5.1) · Polymer Sheets at the Chemical Plant · Soil Remediation (§3a.4)
Algae Farm	Pond Head + Pond Bay	Algae	Biofuel → Fuel Generator

Fast-growth timber and bio-polymer crops are removed. Both duplicated raw_biomass, which has existed since Phase 1 and is already load-bearing through Polymer Sheets. "Wood" was a naming error: the building is the Biomass Burner. Three invented resources were deleted rather than defined. See [27_Organic Farming & Biomass Design](<./27_Organic Farming & Biomass Design.md>) §2.

Wild biomass still exists alongside cultivated. The Biomass Harvester is slow and infinite, needs no inputs and no research; the Farm is volume. The Harvester is outscaled, never outgrown.

7.2 Inputs¶

Water, land, light, fertiliser and power. Light is reduced by air pollution (§2), and greenhouses may run artificial lighting to escape that at the cost of power — an escape only as clean as the grid behind it.

Full specification, including the open question of a captured air input, is in [27_Organic Farming & Biomass Design](<./27_Organic Farming & Biomass Design.md>) §5.

7.3 The strategic axis¶

	Renewable (farm)	Fossil (mine)
Speed	Slow	Fast
Supply	Infinite	Finite
Land use	High	Low
Pollution	Low	High

Farm biomass forever, or strip-mine coal now. Both are legitimate; they simply have different shapes.

It also makes terraforming a polluted planet a meaningful late-game goal rather than a cosmetic one, feeding Soil Remediation (§3a.4) with the organic input it needs.

7.4 Restoration as a third endgame path¶

The endgame is currently open after Milestone 3, with two directions: war effort or supply chain provider.

Restoration is a natural third path, and a peaceful one. Take a planet that has been wrecked — by the player, or by someone else in the shared persistent galaxy — and bring it back.

Uses systems that now exist rather than requiring new ones
Gives players who do not want the war layer somewhere meaningful to go
Fits the shared galaxy: a wrecked planet is a problem another player can be paid to solve

Open items for this document are tracked in docs/open-items.md, area pollution.

9. Revision History¶

Version	Date	Changes
0.3	2026-09-11	§3b added: the air dispersal model, closing OI-089. Air pollution is a coarse grid per level, with sources contributing by weighted falloff, diffusion spreading over time, and sinks plus a small natural decay removing. Radius falloff rejected because it stores nothing, so pollution could never arrive, never accumulate, and never be cleaned up, and `37_Planet Threats & Defence` §5.1 needs air to arrive where the natives live. §3b.2: the natural decay is a free budget that protects the first hour and gives the player a legible target. §3b.3: wind ships in the Launch release for everyone, biasing diffusion into a downwind tail, which poisons a farm downwind and leaves an identical upwind farm clean, and makes threat direction predictable so defence is a decision about where rather than a perimeter tax. Wind is not a Simulation Level feature per Canonical §24, because it changes where you build; the price is a mandatory wind indicator and tutorial beat (OI-381). §3b.4 folds §3a's per-level behaviour into diffusion and decay coefficients, so underground nowhere to vent stops being a special case. §3b.5 separates the weak always-on world haze from the toggleable overlay, which shows pollution, sinks and native zones together so it is a planning tool rather than a status light, plus two HUD numbers instead of a world total. §3b.6: Simulation Level gets a legal dial (cell size, tick rate), OI-365 gets its answer, and attribution is ruled out per Canonical §25.
0.2	2026-08-27	§7 rewritten for two farm families instead of three crops. Fast-growth timber and bio-polymer crops removed as duplicates of `raw_biomass`; "Wood" identified as a naming error. §7.2 now points at `27_Organic Farming & Biomass Design` rather than restating inputs. §3a.4 names Soil Remediation's organic input as `raw_biomass`. §7.3 and §7.4 unchanged apart from one reference to the deleted Wood resource.
0.1	2026-08-25	Initial draft. Three-pollutant model, decay-from-own-pollution rule, three-way waste choice, scavenger hostiles, feedstock farming.