# Integrated Food–Water–Nutrient Community — Build Directive

*A spec for the phased community concept: the iconic grow tower(s), the surrounding
residences, the land/periphery staple tier, and the human-waste fertilizing loop that
ties them all together. Written to be handed to a build session (Claude Code) — every
number below is a parameter, not a hardcode.*

---

## 0. How to use this directive

This describes **one closed system with four subsystems** that exchange water, nutrients,
food, and people:

1. **Residences** — the people. Both the *demand* (food, water) and the *source* (waste → fertilizer).
2. **Core Towers** — the iconic domes/silos: greens, fish, eggs, water purification.
3. **Land / periphery tier** — the staple calories the towers can't make: corn, potatoes, beans, squash, nuts.
4. **Nutrient loop** — the poo/pee + greywater cascade that fertilizes 2 and 3.

Two design rules run through all of it:

- **The honesty constraint.** Towers deliver diet *quality* (greens, fresh protein) and ~11–25%
  of *calories*. Energy calories come from land. Say this up front; it is true of every closed
  food system that isn't importing grain.
- **Two independent levers.** *Height* (stacked grow layers) drives **people fed per acre**.
  *Cluster size* (shared infrastructure) drives **cost per household**. Do not conflate them —
  they are separate sliders in the calculator.

---

## 1. Spatial model: the radial contour gradient

The site is steep Appalachian terrain (Pikeville, KY; USDA zone 6a–6b; ~150–170 frost-free
days; ~1,150 mm rain/yr). A concentric/radial plan is not just aesthetic — **the rings double
as contour lines.** Planting rows run *along* the rings (catch water, stop erosion); radial paths
are access spines and drainage ways. Build rings as shallow swales/terraces = keyline water
harvesting in the street plan.

Zonation outward, ordered by both *intensity* and *cleanliness/elevation*:

| Zone | Ring | Contents | Water grade |
|------|------|----------|-------------|
| 0 | Center | Core Tower(s): grow cylinders, trout RAS, potable purification | **Clean** (potable + RAS) |
| 1 | Inner | Intensive daily-harvest greens, high tunnels, community kitchen/processing | Clean + greywater irrigation |
| 2 | Middle | Residences (uphill of treatment), rooftop rainwater catchment | Potable in / waste out |
| 3 | Outer | Annual staples on contour terraces: Three Sisters, potatoes, squash, sweet potato | Greywater + nutrient irrigation |
| 4 | Periphery | Perennial agroforestry (chestnut, hazelnut, fruit, sunchoke), grain plots, storage core, **the nutrient cascade** (BSF, vermifilter, wetland, duckweed lagoon) | **Dirty/nutrient** (downhill, gravity-fed) |

**Elevation logic:** residences uphill → waste flows *down* by gravity through the treatment
cascade → polished/recovered outputs return to land beds (low) and pump back up only where
needed. Potable/RAS water is physically and procedurally separated from the nutrient cascade;
the two only reconnect after full treatment + polishing, and even then polished water serves
irrigation/flush, **not** drinking.

---

## 2. Residences — demand *and* nutrient source

Residences are a first-class subsystem, not background. Each household is modeled as:

- **People:** 2.5 per household (parameter).
- **Demand:** food (see §6–§8) + water ~120 L/person/day domestic (parameter; drops with
  vacuum toilets + greywater reuse).
- **Catchment:** rooftop rainwater. ~100 m² roof/household × 1,150 mm ≈ 115 m³/household/yr.
  For 40 households (one tower's population) that's ~4,600 m³/yr captured vs ~4,380 m³/yr
  domestic demand — **roughly self-supplying on rain alone**, with the tower as purification +
  buffer, not sole source.
- **Source (the integration point):** each person yields the nutrient stream in §4–§5.

**Plumbing spec per residence:** three separated outflows —
1. **Urine** (urine-diverting toilet) → struvite/N recovery.
2. **Feces + flush** (vacuum or low-flush, 0.5–1 L vs 6 L conventional) → BSF + vermifilter.
3. **Greywater** (shower, sinks, laundry) → light treatment → irrigation / duckweed pond / toilet flush.

Tie-in to existing build language: the residence prototype already in play (≈3-story, 25×25
footprint, reclaimed coal-belt cladding, standing-seam roof) is the unit; the only additions
this directive requires are the source-separating fixtures and the rooftop catchment to cistern.

---

## 3. Water: three grades, one topology

| Grade | Source | Treatment | Destination |
|-------|--------|-----------|-------------|
| **Potable** | Rainwater (roofs + dome) | Cisterns → filtration + biological + UV (tower mechanical level) | Drinking, cooking, RAS top-up |
| **Greywater** | Shower/sink/laundry | Settling + biofilter | Irrigation, toilet flush, duckweed lagoon |
| **Nutrient/black** | Toilets (separated) | BSF → vermifilter → constructed wetland → duckweed lagoon → polishing pond | Fertilizer + feed; polished water → irrigation, recirculate |

**Cleanliness firewall:** trout are cold-water and fussy; ducks and waste foul water fast.
Keep the clean RAS and potable loops *upstream and separated* from the duck/duckweed/wetland
"dirty-but-productive" stages. They never share an open surface.

---

## 4. The nutrient loop — human waste as fertilizer (fully integrated)

This is the engine that makes the community closed. Per **100 people** (one tower's population),
annual recoverable nutrients (parameters, round figures):

- **Urine:** ~50,000 L/yr, carrying ~**400–450 kg nitrogen** and ~**35–40 kg phosphorus**,
  and it comes out essentially sterile. Source-separated → **struvite precipitation** (slow-release
  P fertilizer, chemically identical to mined phosphate rock) + a dilute **nitrogen solution** for
  tower fertigation and land beds.
- **Feces + food scraps:** ~5,000 kg/yr wet feces + kitchen/crop waste → **Black Soldier Fly
  (BSF) bioreactor.** Larvae reduce feedstock mass 50–80% in days, self-harvest by crawling out,
  and are ~40% protein / 30% fat → **feed for fish, poultry, and ducks** (offsets ~20–40% of
  bought feed). Yield: several hundred kg larvae protein/yr per 100 people.
- **Frass + residual solids:** → **vermicomposting** (earthworms) → castings = premium soil
  amendment → land staple beds (§8).
- **Liquid fraction:** → **constructed wetland** (cattails, bulrush, root-zone microbes) →
  **duckweed lagoon** → polishing pond → clean enough to irrigate/recirculate.

**Closure note for the pitch:** that ~400–450 kg N/yr from 100 people's urine can supply a
*meaningful fraction* of the staple beds' nitrogen demand; legumes (beans, cowpeas) fix the rest.
The community fertilizes its calories with its own waste stream — no synthetic inputs required.

The cascade is gravity-driven and terraced (each stage drops to the next), so it wants the
downhill periphery (Zone 4). The whole thing is a designed nutrient stream, not "sewage."

---

## 5. The Core Tower module

Base unit = **1 Core Tower ≈ 100 people ≈ 40 households.** Per-100-people figures:

| Component | Value | Parameter source |
|-----------|-------|-------------------|
| Grow canopy | 250 m² (~75 rotating cylinders, 4 tiers daylit) | 150 kg produce/person ÷ 60 kg/m²/yr |
| Produce yield | ~15,000 kg/yr | — |
| Trout basin | ~95–100 m³ | 50% edible, ~1.2 cycles/yr, 35 kg/m³ |
| Trout | ~8,000 standing / ~10,000 harvested/yr / ~2,000 kg edible/yr | 20 kg edible/person |
| Laying hens | ~100 → ~27,000 eggs/yr | 1 hen/person, 270 eggs/hen/yr |
| Building footprint | ~200 m² (~2,150 ft²) | 2 m²/person |
| Capex | ~$1.5–2.5M (midpoint $2.0M) | bespoke dome; tune in calculator |

### The two-lever model (must be explicit)

**Lever A — Height → density.**
`people_per_acre = (canopy yield/m²) × (effective grow layers) ÷ (food per person)`
- **Daylit dome:** sun-driven, low energy, but self-shading caps at **~4–6 effective layers** →
  ~300–500 people/acre (produce+protein footprint).
- **LED silo:** cheap corrugated/slip-form envelope, rotary/conveyor brings cylinders to one
  base harvest station (no per-floor access waste) → **~12–20 layers** → ~1,000–2,000 people/acre,
  but ~150–350 kWh/kg LED energy (leans on the solar field).
- **Strategy:** the **dome is the icon** (fundraising image, daylit, low-opex). The **silos are the
  workhorses** (calorie/protein volume, capex-light/opex-heavy). Build both.

**Lever B — Cluster size → cost.** Shared central plant (water/waste/cold storage), standardized
fabrication (30 identical silos ≪ 1 bespoke dome per unit), one big solar array, shared
automation/labor → ~20–40% off cost/household from pilot to district. **Does not change density.**

---

## 6. Species & crop modules (master table)

Each row = a module the calculator can toggle. `water_zone` enforces the cleanliness firewall.

| Module | Output | Yield basis | kcal/kg | Edible frac | Water zone | Role |
|--------|--------|-------------|---------|-------------|------------|------|
| Leafy/fruiting produce | vegetables | 60 kg/m²/yr canopy | ~300 | — | clean (fertigation) | nutrition |
| Trout | fish | 35 kg/m³, 1.2 cyc/yr | ~1,400 | 0.50 | clean RAS | protein |
| Chicken (layers) | eggs | 270 eggs/hen/yr | ~1,440/kg (72/egg) | — | land/structure | protein |
| Duck (layers) | eggs | ~280 eggs/duck/yr | ~1,850/kg (130/egg) | **dirty pond** | calories+protein |
| Broiler / poultry meat | meat | 15 kg/person/yr (opt.) | ~1,800 | — | land/structure | calories |
| Rabbit | meat | eats trimmings | ~1,700 | — | land/structure | protein |
| **Duckweed** | feed (some food) | 10–20 t DM/ha/yr, 30–45% protein | ~2,200 (dry) | — | **dirty pond** | feed-loop closer |
| Mushrooms | food | on spent substrate/ag waste | ~330 (fresh) | — | dark silo space | protein/micronutrient |
| Crayfish / mussels | food | in/around water | low | — | clean (mussels) / dirty (crayfish) | protein, mussels filter-clean |
| BSF larvae | feed | from waste stream | — | — | nutrient cascade | feed (40% protein) |

**Duckweed is the quiet unlock:** it grows *on the nutrient water for free*, 30–45% protein,
doubles mass in 1.5–4 days, proven feed for fish/poultry/ducks (13–40% of diet). It mostly fixes
the *business case* — imported feed is what kills aquaponics economics; duckweed closes that loop.

---

## 7. Land / periphery staple tier — the calories

Zone 6a/6b: **grow in season, store for winter.** "Year-round" = succession + storage + a few
unheated high tunnels. Crops by the role the towers *can't* fill:

**Carbohydrate/calorie staples:** flint/dent corn (meal, the backbone); Irish potatoes (spring +
fall crops; use grow-bags/hilled contour rows — *towers are finicky and often don't fill*); sweet
potatoes (best calories/area, stores months, black plastic to warm soil); winter squash / pumpkin /
cushaw / **candy roaster** (cellar-keeps); grain sorghum (drought/poor-soil tolerant, grain + syrup);
buckwheat (fast 70–90 day pseudocereal, loves acidic soil, double-crops); sunchokes (perennial,
store in ground, inulin caveat).

**Storage protein:** dry beans (keystone — corn+beans ≈ complete protein; Appalachian heirloom
diversity: greasy, October, cornfield, Trail of Tears); cowpeas/field peas (heat-tolerant, N-fixing);
soybeans; **amaranth** (grain + greens, thrives in humid heat where quinoa fails).

**Fat (hardest):** sunflowers (seed + pressed oil); tree nuts — black walnut, native hickory,
hazelnut; **chestnut** (a perennial *carbohydrate* tree — Chinese/hybrids plant reliably now;
American restoration still slow, don't build the plan on it). Most fat realistically comes from the
animal side (eggs, fish, poultry).

**Winter fruit:** heritage **storage apples** (Albemarle Pippin, Arkansas Black, Winesap) — root-cellar
from October into spring; the fresh-fruit/vitamin-C answer for cold months.

### The "innovative" Three Sisters = a storage guild
Four sisters, every variety chosen for *keeping quality*: **flint/dent corn** (carbs + living trellis)
+ **pole dry beans** (storage protein, fixes N) + **winter squash** (storage carbs, living mulch) +
**4th sister sunflower or amaranth** (sturdier edge support + the seed = fat/protein the trio lacks).
Plant on contour along the rings; trellis extra pole beans up path edges; where slope steepens toward
the periphery, **alley-crop the annual guild between perennial chestnut/hazelnut rows** — annual +
perennial calories on the same ground.

### Year-round cycling
- **Cool (Mar–May, Sep–Nov):** potatoes (both), brassicas (cabbage→kraut, overwintering kale/collards),
  turnips/rutabaga, peas, fall garlic/onions, winter rye (soft red wheat is marginal — fungal head
  blight in humid summers; rye/buckwheat/corn are the reliable "bread" grains).
- **Warm (May–Sep):** Three Sisters guild, sweet potato, sorghum, amaranth, sunflower, cowpeas.
- **Bridge:** double-crop (rye → buckwheat → fall brassicas); unheated high tunnels for fresh winter
  greens/roots.
- **Storage = the winter calorie engine:** root cellar (potatoes, sweet potatoes, squash, cabbage,
  sunchokes), dry bins (corn, beans, sorghum, sunflower, amaranth), ferment (kraut, leather britches),
  sorghum syrup, cellar apples.

---

## 8. Calorie & nutrition accounting (the honest ladder)

Need: **2,250 kcal/person/day.** Build the bar up in stages so the story is legible:

| Stage | Adds | kcal/day | Running share |
|-------|------|----------|---------------|
| Base produce | 150 kg greens | 123 | ~5% |
| + Trout | 20 kg | +77 | ~9% |
| + Chicken eggs | 270 | +53 | ~11% |
| + Double trout | +20 kg | +77 | ~15% |
| + Ducks (eggs) | ~280 | +100 | ~19% |
| + Broiler/poultry meat | 15 kg | +74 | ~23% |
| + Rabbit | 10 kg | +47 | **~25%** |
| **+ Land staples** | corn/potato/squash/beans/nuts | +~1,690 | **100%** |

Tower ceiling without staples ≈ **25%, maybe 30%** if you go all-in on animal fat. The remaining
~75% is land. (Growing staples vertically is *possible* but the worst use of expensive LED space —
spend sun and dirt on calories, reserve the tower for high-value nutrition + water-coupled protein.)

Persistent gaps in any temperate closed system: cooking oil at volume (sunflower pressing narrows it)
and high-protein bread wheat (cornbread, rye, buckwheat are the regionally honest answer).

---

## 9. Phasing & cluster triggers

| Phase | Towers | People | Households | Adds | Capex (planning) |
|-------|--------|--------|------------|------|------------------|
| **Pilot** | 1 (partial loop) | ~100 | ~40 | One dome, prove the loop | $1.5–2.5M |
| **Small** | 1 (full loop) | ~100 | ~40 | Complete waste/poultry/water loop | $2–3M |
| **Medium** | 5–8 + shared plant | ~500–800 | ~200–320 | Central water/waste/cold plant, bigger solar, shared LED silos | $8–15M |
| **Large / District** | 30–50 (5–8 mediums) | ~3,000–5,000 | ~1,200–2,000 | District staple greenhouses, reservoir, energy plant | $50–120M |

**Trigger rule:** each time you reach ~5–8 of a tier, build the next-tier shared chamber
(bigger grow + processing + water). That shared infrastructure is what unlocks the next ring of
residences. One shared central plant per ~8 towers; district infrastructure at 25+ towers.

---

## 10. Parameter block (for the calculator / site)

```
# Demand
people_per_household        = 2.5
kcal_need_per_day           = 2250
produce_kg_per_person_yr    = 150
trout_edible_kg_per_person  = 20      # slider 10–35
eggs_per_person_yr          = 270
water_L_per_person_day      = 120

# Produce
produce_yield_kg_m2_yr      = 60
produce_kcal_per_kg         = 300

# Trout (cool-water, fits KY)
trout_density_kg_m3         = 35
trout_cycles_per_yr         = 1.2
trout_edible_fraction       = 0.50
trout_kcal_per_kg           = 1400
trout_avg_standing_wt_kg    = 0.45

# Eggs / poultry
egg_kcal                    = 72
duck_egg_kcal               = 130
broiler_kcal_per_kg         = 1800
rabbit_kcal_per_kg          = 1700

# Lever A: height -> density
layers_daylit_max           = 6       # daylit dome cap
layers_led_silo             = 12-20   # LED silo range
people_per_acre_daylit      = 300-500
people_per_acre_led_silo    = 1000-2000
led_energy_kwh_per_kg       = 150-350

# Lever B: cluster -> cost
cost_per_tower_M            = 2.0     # slider 1.2–3.0
cluster_cost_reduction      = 0.20-0.40
towers_per_central_plant    = 8
district_threshold_towers   = 25

# Nutrient loop (per 100 people / yr)
urine_L_yr                  = 50000
urine_N_kg_yr               = 425
urine_P_kg_yr               = 38
feces_wet_kg_yr             = 5000
bsf_mass_reduction          = 0.50-0.80
bsf_larvae_protein_pct      = 0.40
feed_offset_from_bsf        = 0.20-0.40

# Site / climate (Pikeville KY)
hardiness_zone              = 6a-6b
frost_free_days             = 150-170
annual_rain_mm              = 1150
roof_catchment_m2_per_hh    = 100
```

---

## 11. Open questions to resolve

- **Dome vs silo split:** how many daylit domes (icon) vs LED silos (volume) per cluster? Drives
  both the energy budget and the fundraising imagery.
- **Trout summer cooling:** confirm passive cooling (deep basin + ETFE shading + night flush) holds
  <18 °C through a Pikeville July, or budget active chilling.
- **Greywater vs nutrient routing:** how much greywater goes to duckweed vs direct irrigation vs flush.
- **Staple land per person:** ~300–500 m²/person in full mode — confirm against available terraceable
  acreage on the actual site (GIS slope analysis).
- **Energy:** size the solar array against the LED-silo layer count (this is the swing variable for
  both capex and opex).
- **Regulatory:** human-waste reuse for food crops has real permitting friction; the
  urine→struvite and BSF→non-food-feed paths are the cleaner regulatory routes — map which outputs
  touch food.