Island Lab · A partnership proposal

Co-run the first living twin of a Mediterranean coast.

The platform is built. Your science is what makes it true. Let us prove the loop together, at Pollica.

Prepared for ⟨partner⟩ · Pollica, Cilento coast · use → ← or click to move

Why

A coast goes blind between surveys, and the meadow does not wait.

Why · the stakes

A third of the meadows we have records for, gone in fifty years.

14–20 L

oxygen per m² on a good day, from a healthy meadow

up to ~700 t

carbon per hectare in the deepest matte (Corsica)

~1 cm

growth per year. Lost, it does not come back.

Posidonia oceanica is the nursery the fishery grows in, the reason the beaches hold their sand, and one of the strongest carbon sinks in the basin. Because it cannot be fully restored on a human timescale, protection is the only cheap option, and protection means seeing stress before it becomes loss.

Why · the clock

In 2003, a marine heatwave struck up to 80% of coral and sponge colonies in the Gulf of Naples.

Next door to Pollica. Those events are getting more frequent and more intense, and the same summer heat that hit the reef fauna also drives Posidonia shoot loss. Today a town learns its meadow is in trouble from an occasional dive survey, after the damage is done. What linked that heat to the die-offs was the T-MEDNet network of in-situ loggers: continuous local monitoring. It is just rarely affordable.

Why · the gap

The big models are excellent, and too coarse to see the meadow.

Copernicus and the EU Digital Twin Ocean forecast a heatwave across the Tyrrhenian Sea, and do it well. They cannot tell you whether it reached your meadow at depth, and they do not see the light and turbidity that decide whether it lives. A nearshore bay can sit several degrees off the regional average. The decision gets made on that gap, and most of the time no one ever sees it.

How

A platform that turns a place into a living, honest, local twin.

How · the loop

We make monitoring purposeful, not just pretty.

Scan what is already known. Turn it into a map of what is where and a profile of what threatens it. Build the picture from public data, every layer labeled for how real it is. Let the model say where it is blind, and put one sensor exactly there. Fuse it, twin it, decide on it, then re-score. Each measurement sharpens the model and points at the next one. The loop is what makes a sensor fleet earn its keep.

How · where we fit

We do not compete with the basin models. We complete them.

Take the global model, bend it toward what the water actually measured: data assimilation and downscaling, not a worse copy of a forecast that already exists. We ingest Copernicus, Sentinel, and EDITO as forcing and add the local half they cannot reach. The difference between the regional prediction and the local reality, the delta, is the product. Most external platforms are a feed we pull in, not a rival.

How · adaptive sampling

The twin knows where it is blind.

The model produces state, an 8-day forecast, and a per-cell map of its own uncertainty. The zones where it is least sure and the meadow matters most are exactly where one real sensor buys the most truth. The free public-data twin is the broad sweep; your sensors are the precise instrument; the twin decides where they look. Tip and cue, run at the scale of a single bay.

How · honesty

Every value carries where it came from.

Provenance is a required field, not a footnote. A project starts on open and synthetic stand-ins and is only "finished" when nothing reads synthetic: every layer has resolved to measured or calibrated truth. We would rather show an honest placeholder than a confident guess, the trait a research partner needs most.

Local our sensor Open public data Derived our model Synthetic placeholder

How · the open seam

We produce the substrate. Your model can do the predicting.

The platform fuses the evidence, builds the local climatology, computes the cross-source relationships, and exports it all as one portable bundle with units and provenance on every value. Prediction is delegated to agents and partner pipelines that read that bundle. You bring your method; we make it cheap to feed and easy to defend. No proprietary formats, no lock-in.

What

Built, deployed, and walkable today.

What · one console, many tools

A constellation of apps, each turning shared inputs into client-ready outputs.

Helmthe consoleLIVE

Tesseraecosystem mapLIVE

Reckoningscrape + fuseLIVE

Tritonsensor fleetLIVE

Kairosdecisions, loggedLIVE

Augurrisk + blind spotsSCOPED

Meridianpick a placeNEXT

Argusscan the contextNEXT

Nereustwin integratorNEXT

Proteuslocal twinNEXT

What · proof, not a promise

The thing exists. Here is the evidence.

HDOP 0.6

clean GPS fix from our buoy hardware in the field

deployed

engine returning real climatology + verdict rows from Supabase

1 step

any coast geocodes into a fresh, isolated project, live

Five apps run today. The field stack produces real fixes; the engine exports the agent-readable bundle now. The honesty is structural: "what is real today" is a property of the data, not a slide.

What · the working example

Pollica, today.

The water twin, the meadow map, the action layer, the live fleet tracker, and the Posidonia model underneath, all running on public data and honest about every number. Not a report. The start of a project a community owns.

See it live: the project dashboard → · the twin walkthrough →

Together

The platform is half the answer. Your science is the other half.

Together · what each side brings

A clean division of labor.

Island Lab brings

The platform and the living twin, free on public data from day one
The sensor fleet and the loop that places each device where it matters
The decision layer: verdict, tiered action, a logged, defensible record
The portable bundle your pipeline reads, no lock-in

You bring

Local ground truth: site access, dive surveys, the meadow's real behavior
Scientific validation of the thresholds against the Cilento meadow
Your data and your model, integrated as a first-class feed
Standing and reach with the community and the institutions that act

Together · the pilot, in three tiers

Start free on public data. Sharpen the meadow sensor by sensor.

TIER 0 · NOW

Public baseline

A real twin of Pollica on public data, at no cost, honest about its limits. Running today.

TIER 1 · THE PILOT

Ground-truthed

Your feeds wired in, your survey validating the thresholds, one buoy at the meadow. Synthetic turns measured.

TIER 2 · OPERATIONAL

A living twin

A fleet routed to where information is worth most. Forecasts the community can run on, and a result worth publishing.

Together · the ask

Co-run Pollica as the pilot. Two concrete moves.

MOVE 01

Make the twin true

Wire your data and the European open feeds into the twin, and validate the meadow thresholds against your survey. The synthetic layers go real.

MOVE 02

Put one buoy in the water

A single Island Lab sensor at the meadow, placed where the twin is most blind. The loop starts running on truth instead of stand-ins.

The outcome: Pollica becomes the first coast in the Mediterranean with a living twin of its own, and we have a joint, defensible, publishable result to show Rome, Brussels, and the next coast.

The short version

The basin model tells you what is probably happening across a region. Together, we can tell Pollica what is actually happening at its meadow, and how much to trust it.

Island Lab · A partnership proposal · Pollica, Cilento