US lightning statistics are not one number: a useful statistic names its sensor, unit, geography, and observation period. Optical flashes, ground-network detections, thunderstorm days, and a live state leaderboard answer different questions and should not be merged into one ranking.
Choose source before comparing numbers
| Source | What it measures | Best use | Not interchangeable with |
|---|---|---|---|
| GOES GLM | Satellite optical total-lightning flashes | Broad recent activity | Surveyed ground-strike locations |
| Ground network / NLDN climatology | Network observations under a published historical definition | Long-term thunderstorm-day patterns | Today’s satellite count |
| Daily state leaderboard | Returned GLM optical flashes for each state’s closed local calendar day | Recent closed-day state context | A unique US total or climatology |
Lightning Tracker’s daily leaderboard omits missing rows rather than filling them with zero. State bounding boxes can overlap, so its rows cannot create a unique national total. See methodology for the source, boundaries, and cadence behind map data.
What the Koehler thunderstorm-day study says
The peer-reviewed Koehler et al. study uses 1993–2018 National Lightning Detection Network (NLDN) observations. Its thunderstorm day means at least two NLDN flashes within 10 nautical miles during a day. That is a useful reproducible climatology definition, not a state ranking or a claim about current conditions.
Its Florida local grid cells reach roughly 80–105+ thunderstorm days per year. Those are grid-cell values under that study’s method; they should not be relabeled as a statewide annual average, a city value, or a live-app count.
What GLM can and cannot tell you
NOAA’s Geostationary Lightning Mapper observes total lightning optically, including in-cloud activity. Its 8–14 km spatial characteristics are instrument properties, not a guaranteed map-marker error. NOAA’s 70–90% detection characteristic is not a per-storm promise, and its 20-second product latency is not Lightning Tracker end-to-end latency.
GLM reports events, groups, and flashes; map points are optical-flash centroids, not exact ground-strike records. Read how accurate lightning maps are before using a marker for distance or safety decisions. An empty map is never an all-clear: if thunder is audible, get into a substantial building or enclosed hard-topped vehicle.
Use the right statistic for the question
- “What happened yesterday in returned states?” — use the closed-day leaderboard.
- “What has a defined long-term thunderstorm-day pattern?” — use the cited NLDN climatology and retain its period, radius, and grid unit.
- “Is a storm active nearby now?” — use current detections with official warnings, never as an all-clear.
Keep frequency, density, population exposure, fatalities, and recent activity in separate columns. A larger area can collect more detections than a smaller one without having higher density, and historical patterns cannot predict the next strike.