Coverage and Scope

This project is global in remit — the data model, the map and the editorial standard are built to extend anywhere there are diveable wrecks. The database itself is not global yet, and this page exists so that distinction is never left ambiguous.

A wreck is marked diveable when published sources show it being dived, and when its depth and condition put it within reach of some recognised qualification level — from Open Water to advanced trimix. Diveable is a statement about physical and legal access, not a recommendation, and it does not mean a given wreck is appropriate for a given diver.

A wreck can be diveable and still be closed to you. Protected sites may require a licence, war graves prohibit artefact removal, and at least one wreck in the database may not be dived at all without permission from the authority that controls it. Those constraints sit in their own fields on each record.

Records marked not diveable are retained where they matter historically or where their status could change. The database holds 113 records in total, of which 111 are marked diveable.

Every factual field is meant to point at an evidence record, and every evidence record has fields for a source name, a source URL, a source date, an excerpt and a confidence rating. Those fields are not all populated to the same depth, and the gaps matter more than a summary figure would suggest, so they are reported here directly rather than left for a reader to discover.

• 472 evidence records support 113 wreck records. Every one carries a source name, a source URL, an excerpt and a confidence rating.
• 104 of 472 carry a source date. For the remaining 368 the age of the underlying claim cannot be assessed from the record — the source is named, linked and quoted, but undated. Dates are never inferred or back-filled to close this gap.
• 111 of 113 wreck records carry an explicit coordinate visibility state. Where the state is not set, the position is withheld rather than shown.
• 1 recorded source conflicts. Each is displayed on the wreck page it affects rather than resolved silently.

Evidence depth also varies by cohort, and not in the database’s favour: the first batch of records carries roughly twice the evidence per wreck of the first expansion batch. Closing that gap is a prerequisite for further expansion, not a follow-up task.

• Wrecks with no usable published evidence. Absence here means nothing was recorded, not that nothing is there.
• Exact positions for protected sites and war graves — see the coordinate visibility states in the methodology.
• Conditions forecasting. Seasonality reflects recorded patterns, not a prediction for a given date.
• Operator listings, bookings and prices. This is a reference database, not a booking service.

New regions are added in batches: research against published sources first, then structured extraction with evidence attached per field, then review. A region goes live when its records carry enough sourced detail to answer the questions the rest of the database answers — not when the wreck names are known.

Gaps in an existing region are expected and visible. If a wreck you know is missing from a covered region, that is a gap worth reporting rather than evidence the region is finished.