What a comune runs, and what it costs

Pollica, August 2026. Three questions in the order they were asked: who runs what, what the technology actually is, and what it costs to stand up and keep running. Built for the Future Food Institute presentation and written to survive due diligence rather than to win a room.
v0.1
The pitch is not a tool. It is a method a community learns and keeps, which means the interesting question is not what the software does but who is accountable for each part of it once we are gone. This page answers that first, describes the technology second, and prices it third. Where we do not have a number, it says PENDING rather than an estimate, because a placeholder that reaches a due-diligence reader costs more than a gap does.
01 The municipal model: who is accountable for what
Function → who is accountable Community
and fleet
Comune Island Lab FFI External
verifier
Set the survey scheduleHow often the coast is re-surveyed, and when in the season. consulted owns advises · ·
Plan the linesWhich ground gets covered, at what spacing, in what order. advises · does · ·
Run the surveyA working boat runs the line on a day it was already out. owns · trains · ·
Sign off a runThe quality layer that makes an untrained operator's data defensible. The director role. · unresolved unresolved · ·
Hold the raw dataCustody of the readings as collected, before any derivation. owns copy processes · ·
Archive the recordThe dated, published, immutable version that a decision can be defended against. · owns supplies · ·
Method, tooling, quality controlThe protocol, the software, and the trust layer that scores every reading. · · owns · ·
Teach the methodBootcamp, schools, the Paideia Campus layer that makes it survive the grant. does hosts does owns ·
Convene and brokerThe relationship with region, university and investor. · party · owns ·
Verify a credit claimIndependent confirmation that the measurement supports the claim. · · · · owns
Decide and actAnchoring zones, restoration priority, fishing practice, budget lines. consulted owns · · ·
owns accountable for the outcome does executes the work consulted / advises / hosts supports without carrying it unresolved the honest state, see below

One row is genuinely unresolved, and it is the load-bearing one

Nobody formally holds sign-off. The whole premise is that an untrained operator produces defensible data because a director checks and signs the run. That role carries authority in the data schema and does not exist in the data model today: the operator is currently a hardcoded provenance string. Until a named person with a named permission sits in that row, the quality claim rests on nothing, and it is the first thing a verifier will pull on.

The decision is whether the director is a comune employee, an Island Lab role, or a university secondment. Each answer changes who is liable when a run is wrong, which is why it is a governance question before it is a technical one.

02 The cadence: what a year actually looks like
Step 1
Plan
The comune sets the window. Lines are drawn against last season's coverage, prioritising ground nobody reached.
Step 2
Survey
Boats run assigned lines across the season, on days they were already working. Coverage accumulates rather than arriving in one campaign.
Step 3
Sign off
The director reviews each run and signs it, or sends it back. Readings are down-weighted, never silently discarded.
Step 4
Publish
The season's sessions are versioned and frozen. Once published, a record cannot be quietly altered.
Step 5
Archive
The comune takes custody as the official record, and can produce it on request without asking us.
Step 6
Compare
This season against last, with uncertainty stated. A cell that changed because the weather was worse is distinguishable from one where the meadow moved.
Step 7
Act
The comune routes what the comparison found: an anchoring proposal, a restoration priority, a practice change, a budget line.
03 The stack, layer by layer, with what is missing

Statuses below come from the platform registry rather than from ambition. Two layers are honestly absent, and both are named here rather than left for a reader to discover.

L1

Instrument

What is physically measured on the water. The fleet carries three sensors: water temperature, light, and GPS position.

temperature · light · GPS
built
L2

Habitat sensing (sonar)

Side-scan or single-beam returns, which is what turns a temperature-and-position survey into a habitat survey. Designed, costed, and not built. There is no sonar ingest anywhere in the system and no log reader for any sonar format. Every depth currently shown is either an open bathymetry grid or a synthetic distance-to-coast stand-in, and is labelled as such at the point of display.

nothing built · requires a hardware purchase plus a log reader
absent
L3

Telemetry and ingest

Device to cloud. A versioned contract governs the mapping, so a field change cannot silently break a consumer: columns first, mapping second, app version third.

Triton (fleet) · Reckoning (fusion) · contract v1.0.4
built
L4

Trust and provenance

Every reading carries where, when and how well it was taken, combined into one score. Weak readings are down-weighted rather than deleted, a lone weak reading is flagged as needing replication rather than dropped, and a bad reading collapses toward zero instead of poisoning the average. Provenance is an explicit vocabulary, so a modelled number can never be shown as a measured one.

strength model · curation gate · provenance: local, open, derived, calibrated, synthetic
built
L5

Survey execution and planning

Planning the lines, and running them. The simulator models a working day honestly: quality degrading with speed, depth and sea state, coverage accumulating, fuel and time draining. It persists nothing, so a run currently evaporates on reset. That is the active build.

Portolan (plans, mvp) · Periplus (runs, experimental)
partial
L6

Map and comparison

The per-cell observation surface, and the comparison against the known 2017 extent. This is the screen the whole argument rests on, because it is where a community sees its own coast disagree with the official map. In build now.

Tessera (ecosystem map) · comparison panel in progress
partial
L7

Model and forecast

What the meadow does next under a given plan. The Posidonia model is a single-box model today, not a per-cell simulation, so the honest interim product is a per-cell surface of observations rather than predictions.

Proteus (local twin) · Nomos (commons futures) · Meadow Futures
partial
L8

Decision and action

Routing what the monitoring found to whoever can act on it. Monitoring nobody acts on is a hobby, and this is the layer that stops it being one.

Kairos (actions) · Helm (project console)
partial
L9

Sovereign custody

Publication, immutability, and municipal archive. A published session that cannot be silently altered is what separates a record from a report. Not built. Persistence exists as an unapplied migration; nothing today makes a record immutable or transfers custody.

nothing built · migration written, not applied
absent
04 The cost, and what we do not yet know

Two halves. The comparable programmes below are sourced and public. Our own operating cost is not yet assembled, and every slot for it is marked pending rather than guessed.

Comparable programmes, sourced

ProgrammeWhat it boughtCostDerived
LIFE Posidonia Andalucia 41 mooring buoys and 27,022 ha mapped 3.74M EUR About 138 EUR per hectare mapped if the whole budget is attributed to mapping, which overstates it, since the buoys and enforcement are inside the same figure.
LIFE SeResto About 400 ha of continuous meadow restored, 75,161 rhizomes 1.56M EUR About 3,900 EUR per hectare restored. See the warning below: this does not reconcile with ARTEMIS.
ARTEMIS (Interreg Euro-MED) Posidonia restoration pilots, Menorca, Sardinia, Crete, Monfalcone 516k to 942k EUR/ha Against projected credits of 34k to 41k EUR/ha. The nearest peer to what is proposed here, and the least comfortable number on this page.
Reef Check Med 692 certified volunteer divers, 50,255 records non-profit budget The evidence that community-run monitoring at scale is cheap and durable. It is the model this work extends from divers to boats.
Torre Guaceto co-management Fisher co-management of a protected area n/a Catch per unit effort rose 2 to 5x, with a Slow Food Presidium price premium averaging about +81%. Fishers complied because it paid, which is the mechanism this model depends on.
Comune di Pollica, annual budget Scale anchor, all titles 24.31M EUR (2026) 23.57M in 2025. A LIFE-scale protection programme is a fraction of one year's municipal budget, and most of it would be EU or grant money rather than the town's own.

Two sourced restoration costs disagree by two orders of magnitude, and we have not reconciled them

LIFE SeResto implies about 3,900 EUR per hectare restored. ARTEMIS models 516,000 to 942,000 EUR per hectare. Both are real, published, EU-funded programmes.

The likely explanation is that they are not the same operation: SeResto transplanted rhizomes of fast-colonising seagrasses in a shallow lagoon, while ARTEMIS models Posidonia oceanica, which grows on the order of a centimetre a year and is far harder to establish. If that is the reason, ARTEMIS is the relevant number for Pollica and SeResto is not a comparable at all.

We have not verified that explanation, so this page states the gap rather than picking the friendlier figure. It matters commercially: at the ARTEMIS number, credits do not fund restoration and are not close, which is why the position here is protection first with restoration costed separately and funded by grant or capital rather than by credit revenue.

Our own operating cost

These four numbers are what the cost model needs and does not have. Nothing on this page fills them with an estimate.

LineBasisValue
Survey kit Capex. Tracker, sensors, mounting, spares, per boat. PENDING
Platform running cost Opex, per year, at this coast's scale. Database, hosting, device connectivity. PENDING
Survey day rate Opex. What a boat is paid per survey day, and how many days a season of coverage takes. The hinge of the whole stewardship argument. PENDING
Training and schools Opex. Bootcamp weeks, director time, school hours. PENDING
05 What the credit side could carry, at sourced prices

The revenue ceiling the operating cost has to sit under. These are ceilings, not projections: both lines are gated on verification being funded, and the band is the finding rather than the central value.

Carbon, per ha per year
9.7 EUR
band 4.5 to 11.1
0.41 tCO2e/ha/yr durable, at a sourced central of 24 EUR/tCO2e. The live materials still show 40 EUR/tCO2e and therefore 16.40, which is too high and is being corrected.
Biodiversity, per ha per year
150 EUR
band 50 to 500
The larger half and the softer half. It carries a tenfold uncertainty, and it is a market that does not exist yet in the EU on a voluntary basis.
Stacked, per ha per year
159.7 EUR
band 54.5 to 511
Carbon plus biodiversity on the same protected hectare. Carbon alone does not pay for this; stacked, it starts to.
Whole mapped extent, per year
~780k EUR
band 266k to 2.49M
4,879 ha on the 2017 ISPRA extent, if every hectare were protected and both verifications funded. A ceiling on a nine-year-old map, which is itself the reason for the survey.

Say this before you are asked

No Posidonia carbon credit has ever issued, anywhere. Every published blue-carbon price above, and every price in the wider market, is a mangrove series wearing a blue-carbon label. A counterparty running a mangrove project in Madagascar will know this better than most, so stating it first is worth more than hoping it does not come up.

06 What this page does not know

The sourced prices are not yet signed off

The credit price research describes itself as a research pass that has not been human-reviewed, and says to treat the central values as draft. Publishing 24 and 150 to an investor is that sign-off, and it has not been given.

The extent is nine years old

Every per-hectare figure multiplies through 4,879 ha from a 2017 survey. If the meadow has moved, the revenue ceiling moves with it. This is the argument for the survey, not a flaw in it, but it should be said rather than absorbed.

Nothing here has run on real survey data

The demonstrated pipeline runs on synthetic data. The first real outing has not happened, and the automatic labelling that removes the specialist from the middle is the next stage rather than a shipped capability.

The entity question is open

Island Lab is a Washington State LLC, founded by Sean Kolk and Kate Schnippering. Whether a US entity is the right vehicle to receive funds inside an Italian municipal programme is a due-diligence question with no answer yet.