What “No Alerts” Actually Means — Why an Empty Map Is Not a Quiet Day

Direct answer: an empty screen means nothing reached the feeds we monitor. It does not mean nothing happened. Global disaster data is a record of what agencies detected, judged significant, and published — and each of those three steps drops things. The gaps are not random either: they are largest in remote areas, in places with thin monitoring networks, at night, and for hazards that arrive slowly instead of suddenly. Read a quiet map as an absence of information, never as an all-clear.

A feed can only carry what somebody published

Everything on this site arrives from an agency that decided to publish a record: a seismic network, a satellite programme, a national meteorological service, a humanitarian reporting system. Nothing appears here because it happened. It appears because it was measured, written down and released.

That chain is reliable for some hazards and thin for others. A significant earthquake is detected by instruments on the other side of the world within minutes, whether or not anyone nearby reports it. A landslide that destroys a village on a mountain road is detected by nobody at all until a person tells somebody, and that report may take days to reach a system we read — if it ever does.

Detection is uneven, and the unevenness is physical

Seismic networks are dense in some countries and sparse in others, so the smallest earthquake that gets reliably recorded differs from place to place. The same modest tremor can be a published record in one region and invisible in another, purely because of how many instruments happen to be nearby.

Satellites have their own blind spots. Thermal sensors that spot wildfires pass overhead on a schedule and cannot see through thick cloud, so a fire burning under a storm system is a fire nobody is watching. Flood mapping has the same problem: the water is often clearest to a satellite well after the night the river came up.

Some disasters never have a moment to report

Feeds are built around events — a thing that happened, at a time, in a place. Several of the deadliest hazards do not have that shape.

• Drought has no first day. It is a slow divergence from normal that becomes a disaster months later, which is why it is chronically under-represented in event-shaped databases. We have written separately on why the rain returning does not end one.
• Heat kills largely through excess mortality that only becomes visible in statistics weeks or months afterwards, rather than through a single reportable incident.
• Air quality is a condition rather than an event, and a city can sit at genuinely dangerous levels for weeks with no moment any feed would mark. See what the AQI actually measures.

So a screen showing nothing during a severe heatwave or a long drought is not malfunctioning. It is showing you the limits of the format.

Reporting capacity is not evenly distributed

The systems that produce humanitarian records depend on people, organisations and communications infrastructure functioning well enough to file a report. In a large disaster those are precisely the things that stop working.

The result is a bias that runs the wrong way. The events most likely to be under-reported in the first hours are severe ones in hard-to-reach places — the opposite of what a reader instinctively assumes when a map looks sparse. Reports also arrive in local languages and on local schedules, and translation and collation take time. Nights and weekends are slower than weekday mornings almost everywhere.

None of this is a criticism of the agencies. It is what reporting looks like when it is done by people under pressure rather than by instruments.

Something has to count as an event before it can appear

Every feed applies a threshold, because without one it would publish continuously and mean nothing. Those thresholds are reasonable in isolation and they compound: a flood that displaces a few dozen families can be a serious local emergency and still sit below the bar of every global system watching.

The reverse also happens, and it is why this site filters. Some feeds publish administrative units rather than events — one storm split across dozens of county zones, or satellite fire detections as grid squares. A screen busy with those does not show a busier world; it shows a noisier feed. What we exclude and why is set out in our methodology.

What to do with an empty screen

1. Check the national agency. They will know before any global aggregator does, and for a fast-moving local hazard the gap can be hours. Every one of our country pages links that country's own meteorological and disaster authorities.
2. Trust local information over a quiet map. If people around you are reporting water rising, or you can smell smoke, that is better data than anything on this screen.
3. For fast-onset hazards, act on the physical signs. Strong or long shaking near a coast is itself a tsunami warning, and waiting for a screen to agree is time spent badly — see how tsunami warning systems work.
4. Do not read absence as reassurance about a place you cannot see. This is the hardest one, because a quiet map is exactly what someone worried about family abroad is hoping to find.

What this site does about it

• We never publish an all-clear. Nowhere on this site, in the Android app, or in the embeddable widget will you see a claim that a place is safe. The strongest thing we say is that nothing has been recorded — a statement about our records, not about the world.
• We never present a stale fetch as current. A cached “nothing happening” is the most dangerous thing this product could display, so no summary is published at all when the sources could not be reached.
• We say when we last managed to check. A long gap since the last successful sync is information, and hiding it would turn a known unknown into a false reassurance.
• We name the agency on every event and link the original record. The data sources page lists every feed we read — and by implication, everything we do not.

Where figures conflict rather than go missing, a different set of reasons applies; that is covered in why two agencies report the same disaster differently, and for casualty figures specifically in how death tolls are counted.

The honest summary

A disaster map is a map of reports. Where reporting is strong it comes close to being a map of events; where reporting is weak the two diverge — and reporting is weakest under exactly the conditions that make a disaster serious. Read a quiet screen as “we have not been told anything,” which is true, rather than “nothing is happening,” which nobody is in a position to know.

This article explains the limits of aggregated disaster data in general. This site is not a warning authority and its coverage is not complete; for warnings where you live, use your national agency.

Frequently asked questions

Does an empty disaster map mean my area is safe?

No. It means nothing has reached the feeds being monitored. Absence of a record is not evidence of absence of an event, and the gap is widest exactly where reporting is hardest — remote areas, places with sparse monitoring networks, at night, and during the disruption a large disaster itself causes. Check your national agency, and trust local information over a quiet screen.

Why do some countries show far fewer disasters than others?

Mostly because of detection and reporting capacity rather than because less happens there. Seismic and weather networks are dense in some countries and thin in others, and humanitarian reporting depends on organisations and communications that are strained precisely when a disaster is severe. A sparse map often reflects who is able to publish, not what occurred.

Why does a heatwave or a drought show almost nothing on the map?

Because feeds are built around events — something that happened, at a time, in a place — and slow-onset hazards do not have that shape. A drought has no first day, and heat kills largely through excess mortality that only becomes visible in statistics weeks later. These are among the deadliest hazards there are, and they are chronically under-represented in event-shaped databases.

How quickly does a real disaster appear here?

It depends entirely on the hazard. A significant earthquake is detected by instruments worldwide and can appear within minutes. A landslide or a flood in a remote area depends on somebody reporting it, and that can take hours or days, or never happen at all. Never infer from a fast earthquake feed that everything else is equally fast.

Does this site ever say that a place is safe?

No, and that is deliberate. Nowhere on the site, in the Android app, or in the embeddable widget will you find an all-clear. The strongest statement made is that nothing has been recorded, which is a claim about our records rather than about the world. We also refuse to publish a summary built from a failed fetch, because a stale nothing-happening is the most dangerous thing this product could display.

Can I use this data to work out whether disasters are increasing?

Not safely, and it is a common mistake. Reporting coverage has improved substantially over time, so recent years contain more records partly because more is now detected and published. A rising count can reflect better observation rather than more events. Trend analysis needs datasets built for it, with reporting bias explicitly corrected.

What should I do if I think something is happening and nothing is shown?

Act on the local information. Your national meteorological or disaster agency, local radio, and what people around you are reporting are all faster and more authoritative than a global aggregator. For fast-onset hazards, act on the physical signs — strong or long shaking near a coast is itself a tsunami warning, and waiting for a screen to confirm it is time spent badly.

Sources

• CRED / EM-DAT — The International Disaster Database and its documented reporting bias
• UN OCHA ReliefWeb — How humanitarian situation reports are collected
• GDACS — Global Disaster Alert and Coordination System (EC / UN)
• USGS — Earthquake catalogs, network coverage and completeness
• NASA FIRMS — Active fire data: resolution, overpass timing and cloud limitations
• WMO — Multi-hazard early warning systems and observation gaps
• UNDRR — Disaster data, under-reporting and the Sendai monitoring framework

Found an error? See our corrections policy.