Content Policy

Applies to reports sent by members of the public through the World Natural Disasters Android app and shown on this website.

Why reports have no message box

A reporter chooses from a fixed list of observations — "water in the street", "road blocked", "power is out here". There is no field in which free text can be written, and that is deliberate rather than a limitation we intend to remove.

The most harmful thing after a disaster is usually not a photograph. It is the text that travels with one: a predicted aftershock, a rumour that a dam is about to fail, a claim that the water is contaminated or the supplies are gone. Each of those moves frightened people at the moment a wrong decision costs the most. We wrote about this in why the wrong images spread after every event, and it would be indefensible to publish that and then build the exact channel it warns about.

So the vocabulary is the safeguard. A rumour cannot be expressed in it.

What is not allowed

• Anything that is not a first-hand observation from where the sender is standing. Not a forecast, not a warning, not a casualty figure — those come from agencies, and how death tolls are counted explains why a figure needs standing behind it.
• Photographs of identifiable people, of casualties, or of anyone in distress. A person having the worst day of their life has not consented to appear here.
• Graphic injury or death.
• Images that are not yours, or that were taken somewhere or sometime else. A recycled photograph is the single commonest form of disaster misinformation.
• Anything identifying another person — a face, a vehicle plate, a house number, a document.
• Advertising, spam, or anything unrelated to the event.

What happens to a report

• It appears immediately, labelled unverified. Pre-approval would make reports useless during the event, which is the only time they matter. The label, the flagging and the separation from agency data are what carry that risk instead.
• Anyone can flag it. Two separate people flagging a report hides it automatically, pending human review. One person cannot hide a report by flagging it repeatedly.
• Anyone can block a reporter. Blocking hides that person's reports from you alone. It does not remove anything for anyone else, and it does not count as a flag — otherwise any single user would hold a silent delete button.
• You can delete your own report at any time from the app.
• We can remove a report that breaks this policy, and can stop a device sending further reports. Losing the ability to report does not affect Family Circle — a person who posted a bad photograph has not forfeited their family's ability to tell each other they are safe.

What we store, and what we deliberately do not

• No name, no email, no account. A report is attached to an anonymous device identifier, so it cannot be traced back to a person by anyone reading our database — including us.
• Location is rounded to about 110 metres before it is saved. A disaster report comes from wherever the sender is standing, very often their home. Rounding says which street is affected without saying which building. The precise value is never stored, not merely never shown.
• Photographs are re-encoded, and their metadata is discarded. A phone photograph carries the exact coordinates it was taken at. We decode every image and write a new one, so no EXIF — including GPS — survives. Rounding the location while publishing the metadata would be worse than not rounding at all, because it would look careful.
• Sharing your location is optional. A report without one is an ordinary report.

Reporting something to us

The fastest route is the flag control on the report itself, in the app — that is seen immediately and can hide a report without waiting for us. For anything that needs a person, including a request to remove a photograph of yourself or your property, contact us and say which event and which report. We treat requests about identifiable people and about images of casualties as the most urgent category.

What this does not change

Everything else on this site continues to come from monitored agency feeds, with the agency named and its original record linked. We do not average conflicting figures, we do not publish photographs of events ourselves, and a user report never becomes a data point in a count, a map marker or an alert. It sits in its own section, under its own heading, with its own label.

This policy will change as the feature does. It was written when public reports were introduced in August 2026.