How Accurate Are Lightning Maps?

Every lightning map is a picture of something a sensor inferred, not a photograph of a bolt. How much you should trust the dot on the screen depends entirely on which kind of sensor drew it.

There are two families in wide use, and they are good at different things.

Satellite optical detection

This is what Lightning Tracker uses, and what most consumer lightning maps in the Americas are built on. A NOAA weather satellite in geostationary orbit stares at the hemisphere continuously and watches for the optical flash of lightning against the cloud tops.

Location accuracy: roughly 5 to 10 km. The instrument sees light emerging from the top of a cloud, then works back to a position on the ground. Because that light has scattered through kilometers of cloud on the way out, the apparent center of the glow is not exactly above the channel that produced it.

It sees total lightning. This is the underrated advantage. Optical detection picks up in-cloud flashes as readily as cloud-to-ground strikes, and in a typical storm most lightning never touches the ground at all. In-cloud activity often ramps up before the first ground strike, which makes total lightning a better early indicator that a cell is becoming dangerous.

It works everywhere in view, equally. There are no ground stations to be near or far from. Coverage over open ocean, mountains, and empty desert is the same as coverage over a city.

Ground-based radio detection

The other approach uses networks of ground antennas that detect the radio pulse a strike emits, and triangulates the source from the arrival times at several stations.

Location accuracy: often 100 to 200 meters for cloud-to-ground strikes inside a well-covered network. That is an order of magnitude better than satellite, and it is why insurance investigations and utility fault analysis use this data to decide whether a specific strike hit a specific structure.

The trade-offs are real. Accuracy degrades as you move away from the station network, so coverage is excellent over well-instrumented regions and poor over oceans and remote terrain. These networks are also commercial products with licensing costs that consumer apps generally cannot carry.

Which accuracy actually matters

The gap between 200 meters and 8 kilometers sounds decisive, and for some purposes it is. If you are determining whether lightning struck a particular transformer, satellite data cannot answer that and ground network data can.

For deciding whether to clear a field, it changes almost nothing.

The relevant safety threshold is measured in miles, not meters. The 30-30 rule puts the line at about six miles. Lightning can strike ten to fifteen miles from its parent storm. Against decisions made at that scale, a 5 to 10 km uncertainty on any individual strike is inside the noise. What you need to know is that there is active lightning in your vicinity and roughly where the cell sits, and satellite detection answers that well.

Put differently: the precision matters for forensics, and the coverage and latency matter for safety.

What "detection efficiency" means

A separate question from location accuracy is how many flashes a system catches at all. No system catches everything.

Satellite optical detection can miss flashes buried deep in thick cloud where not enough light escapes the top, and its performance varies somewhat with viewing angle toward the edges of the disc. Ground networks miss flashes that fall outside their station geometry.

This has one practical consequence worth internalizing: an empty map is not an all-clear. It is strong evidence that there is no significant lightning near you, and that is genuinely useful, but it is evidence rather than proof. If you can hear thunder, believe the thunder over the map.

Latency is the other half

Accuracy is only useful if it arrives in time. A perfectly placed strike reported twenty minutes late is worthless for deciding what to do next.

Lightning Tracker reads NOAA's GOES-19 satellite, which publishes a new batch of detections every 20 seconds. End to end, from the flash happening to it appearing on your map, is typically 30 to 60 seconds. An alert for a strike inside your radius lands in roughly the same window.

That combination, a position good to several kilometers delivered in under a minute, is the right shape for the actual decision. You are not surveying a strike point. You are deciding whether to get off the water.

The short version

Satellite lightning maps place a flash within about 5 to 10 km, see in-cloud and cloud-to-ground activity alike, cover everything the satellite can see, and report in well under a minute. Ground networks place cloud-to-ground strikes within a couple of hundred meters where their coverage is good, at a cost that keeps them out of consumer apps.

For the question "should I be outside right now," the satellite answer is the one you want, and it is accurate enough that the limiting factor is how fast you act on it.