How accurate is lightning detection?
Updated · 6 min read
Modern ground networks locate cloud-to-ground strikes to within a few hundred meters in well-covered regions and detect most of them. Accuracy degrades with distance from the nearest stations, so a strike over open ocean may carry several kilometers of position error.
How close to the real strike is a plotted position?[1]
In regions with dense station coverage, typically within a few hundred meters to about two kilometers. The error grows where stations are sparse or badly arranged around the strike, and can reach tens of kilometers in the least covered parts of the world.
The published evaluation of the Blitzortung network over Japan compared its locations against an independent reference and found median errors in the low kilometers, with detection efficiency varying by region. That is representative of what volunteer-operated networks achieve.
Commercial networks with denser, professionally sited sensors do better, often quoting a few hundred meters. The trade-off is that their data is licensed rather than open, which is why most free public maps are built on volunteer networks.
What is detection efficiency?[1][2]
The fraction of actual flashes a network successfully locates. It is always below 100%, varies by flash type, and drops sharply outside the network's core coverage. A map showing 40 strikes may represent 60 that actually occurred.
Efficiency is much higher for cloud-to-ground flashes than for intracloud ones on low-frequency networks, because ground strikes emit stronger radio pulses. A storm producing mostly intracloud lightning can look far quieter on the map than it is in the sky.
Weak first strokes and low-current flashes are the ones most often missed. That biases the record toward stronger events, which matters when comparing strike counts between regions with different network density.
Why does coverage vary so much by region?[1]
Because ground networks need physical stations, and stations follow people, power and internet access. Europe, North America, Japan and Australia are densely covered. Central Africa, the Amazon, Siberia and the open oceans are not, despite some of them being the most lightning-active places on Earth.
This produces a systematic bias in any ground-based archive. Comparing raw detected strike counts between a well-instrumented country and a poorly instrumented one measures the networks as much as the weather.
Satellite climatologies exist precisely to correct for that. Instruments in orbit sample every longitude equally, which is why the authoritative rankings of the world's most lightning-prone places come from satellite data rather than ground networks.
How does accuracy affect the distance you see?[1][3]
A position error of one kilometer becomes a one-kilometer error in any distance calculated from it. For a strike reported 12 km away that is minor. For one reported 800 meters away it is the difference between close and overhead.
This is a good argument for treating the nearest-strike distance as an estimate with error bars rather than a measurement. The trend across several strikes is far more robust than the figure attached to any single one.
Counting thunder is subject to its own uncertainty, roughly 20% from temperature, humidity, wind and reaction time. The two methods are independent, so agreement between them is reassuring and a large disagreement usually means one strike was misplaced.
References
- Characteristics of the Blitzortung.org Lightning Location Catalog in JapanAtmosphere 14(10), 1507, 2023 · doi:10.3390/atmos14101507
- Gridded lightning climatology from TRMM-LIS and OTD: Dataset descriptionAtmospheric Research 135-136, 404-414, 2014 · doi:10.1016/j.atmosres.2012.06.028
- The variation of the specific heat ratio and the speed of sound in air with temperature, pressure, humidity, and CO2 concentrationJournal of the Acoustical Society of America 93(5), 2510-2516, 1993 · doi:10.1121/1.405827
Related guides
- How lightning detection networks find a strikeLightning maps locate strikes by timing radio pulses at multiple stations. See how time-of-arrival triangulation works and how accurate the results really are.
- Why do two lightning maps show different strikes?Two trackers, two different strike counts for the same storm. Detection method, network density and what counts as a flash all explain the gap.
- How to read a live lightning mapStrike dots, colour-coded ages and detection delay explained. What a live lightning map can tell you about a storm, and what it cannot.
- How far away is lightning? Count the seconds between flash and thunderCount the seconds between the flash and the thunder, then divide by 3 for kilometers or 5 for miles. See why the rule works and where it goes wrong.