Global lightning climatology from optical sensors in geostationary orbit
High-frame-rate optical sensors in geostationary orbit detect the 777.4 nm oxygen triplet emission from lightning, enabling continuous global flash mapping tied to convective intensity and climate variability.
Sensors
- Geostationary Lightning Mapper (GLM) on GOES-16/17/18: Full-disc optical sensor centred on 777.4 nm, 2 ms frame rate, spatial resolution approximately 8 km at nadir degrading to around 14 km at the disc edge. Continuous day-and-night coverage of the Americas and adjacent oceans. Detects events, groups and flashes in near-real-time with latency under 20 seconds.
- Lightning Imaging Sensor (LIS) on the International Space Station: Low-Earth-orbit instrument at roughly 400 km altitude, 4 ms frame rate, spatial resolution approximately 4 km. Covers latitudes 55°S to 55°N in swaths, giving higher spatial fidelity than GLM but non-continuous revisit. Valuable for cross-calibrating geostationary sensors and sampling regions outside GLM coverage.
- MTG Lightning Imager (LI) on Meteosat Third Generation: European geostationary lightning mapper, operational from 2024 onward, covering Europe, Africa and the Indian Ocean sector. Frame rate and spectral approach similar to GLM. Fills the major geographic gap left by the GOES constellation and will eventually provide a near-global geostationary lightning record alongside GLM.
- Optical Transient Detector (OTD) — historical: Low-Earth-orbit precursor instrument operating 1995–2000. Spatial resolution approximately 10 km, detection efficiency around 40–60%. Provided the first statistically robust global lightning climatology and remains the baseline against which later sensors are validated.
Why 777.4 nm and not a broadband camera
Lightning produces a brief, intense emission from excited atomic oxygen at 777.4 nm, lasting roughly one millisecond. A narrowband filter centred on this line, combined with a frame rate fast enough to catch the transient, lets the sensor subtract the slowly varying background of reflected sunlight from the rapidly changing lightning signal. Without that spectral selectivity, daytime scenes would bury all but the most energetic strokes in noise.
The physics sets a hard detection floor. GLM's 2 ms frame integration means the instrument accumulates background photons continuously, and a weak flash that deposits only a small radiance increment above that background will fall below the detection threshold. Published analyses of GLM data suggest detection efficiency for weak flashes in bright daytime scenes is meaningfully lower than for the same flash at night, with some estimates placing daytime efficiency 10–30% below nocturnal efficiency depending on scene brightness and flash energy. Buyers who need a complete flash count rather than a relative climatology should treat GLM flash-rate products as a lower bound, not an absolute census.
Geostationary persistence versus low-orbit resolution
GLM's defining advantage is continuity. Sitting at 35,786 km, GOES-16 watches the same disc without interruption. A convective cell that spawns a mesoscale convective system over the Gulf of Mexico at 0300 UTC is observed through its entire lifecycle, from the first isolated flashes to the mature stratiform anvil. No polar-orbiting instrument can do that: LIS on the ISS revisits a given point perhaps once or twice per day, capturing a snapshot rather than a film.
The trade-off is spatial resolution. LIS at 400 km altitude resolves roughly 4 km, GLM at geostationary altitude resolves 8 km at best and worse toward the disc limb. For most climatological applications, 8 km is adequate. For studies trying to associate individual strokes with specific convective towers or hail cores, the resolution floor matters. MTG-LI is expected to achieve roughly 10 km resolution over Europe and Africa, broadly comparable to GLM.
What flash rates actually measure about the atmosphere
Lightning flash rate is not simply a proxy for rainfall. It is more specifically a proxy for the vigour of the mixed-phase updraught, the region of a storm where ice crystals and graupel collide and charge separation occurs. Studies using LIS and OTD data have shown that flash rate scales steeply with updraught speed, often as a power law, making it a sensitive indicator of convective intensity that rainfall retrievals can miss when precipitation is entirely glaciated.
At the climatological scale, the global distribution of lightning is strongly land-biased: the Congo Basin, the Maracaibo Basin in Venezuela, and the central United States are among the most active regions on Earth. Maracaibo in particular has documented flash rates exceeding 200 flashes per km² per year in some analyses of LIS and OTD data. Seasonal and interannual variability in these hotspots tracks ENSO, monsoon strength and land-surface temperature, giving lightning climatology genuine utility for climate-trend attribution studies.
One structural limit: optical sensors in geostationary orbit detect cloud-top optical emissions, not ground strokes specifically. The GLM flash product aggregates all in-cloud and cloud-to-ground activity into a single event hierarchy. Separating cloud-to-ground strokes, which are the primary safety and infrastructure concern, requires fusion with ground-based lightning location networks such as the World Wide Lightning Location Network or regional Vaisala systems.
Operational uses: convective warnings and aviation
Meteorological services use GLM flash-rate trends in near-real-time to identify rapidly intensifying convection before radar reflectivity or satellite-derived cloud-top temperature alone would trigger a warning. A sudden jump in flash rate, sometimes called a lightning jump, has been documented as a precursor to severe weather including large hail and tornadoes, with lead times of 10–20 minutes in published case studies. NOAA's operational guidance incorporates GLM data for exactly this purpose.
Aviation route planning benefits from continuous lightning mapping over oceanic areas where weather radar has no coverage. The North Atlantic and equatorial Pacific are the clearest examples: ground networks are sparse or absent, and GLM or MTG-LI data provide the only real-time convective hazard picture available to dispatchers. The latency of GLM products, typically under 20 seconds from observation to dissemination, is fast enough to be operationally meaningful.
Building a climate record: archive depth and cross-calibration challenges
GLM on GOES-16 has been operational since early 2017, giving a continuous geostationary lightning record now approaching a decade. GOES-17 and GOES-18 extend coverage and provide redundancy. Splicing this record to the OTD and LIS heritage data requires careful cross-calibration because detection efficiency differs between instruments. NASA's Global Hydrology Resource Center maintains the combined LIS/OTD climatology dataset, which is the standard reference for pre-GLM flash-rate statistics.
The honest assessment for long-term trend detection is that a single decade of GLM data is short relative to the interannual variability driven by ENSO and the Pacific Decadal Oscillation. Detecting a statistically significant anthropogenic trend in global flash rates will require the full multi-decade record that MTG-LI, combined with GLM successors, is intended to build. For now, the most defensible use of the archive is characterising regional climatology and interannual variability rather than attributing multi-decadal trends.
Satellize ingests GLM and LIS products for clients who need lightning climatology layered into broader atmospheric or infrastructure risk assessments. The Overhead column has covered convective-intensity proxies in the context of agricultural risk, a thread that connects to the crop-estimation work done for the Kingdom of Tonga, where convective rainfall variability is a first-order driver of yield uncertainty.
What the data cannot do
Optical lightning sensors are blind to any flash that does not produce sufficient radiance at the cloud top to exceed the detection threshold. Deeply embedded strokes in optically thick anvils may be partially attenuated. The sensors also cannot distinguish polarity: positive versus negative cloud-to-ground strokes have different physical implications for fire ignition and power-line damage, but GLM provides no polarity information at all.
Geographic coverage gaps remain real. GLM covers the Americas. MTG-LI covers Europe, Africa and the Indian Ocean. The Asia-Pacific sector, including the western Pacific warm pool, one of the most convectively active regions on Earth, has no dedicated geostationary lightning mapper as of 2024. China's FY-4 series carries a Lightning Mapping Imager, but its data are not yet freely available in processed form for international research. Any global lightning climatology built today is therefore a patchwork, strongest over the Americas and increasingly reliant on low-orbit LIS swaths elsewhere.
Typical figures
| Spatial resolution (GLM at nadir) | ~8 km |
| Spatial resolution (GLM at disc edge) | ~14 km |
| Spatial resolution (LIS/ISS) | ~4 km |
| Frame integration time (GLM) | 2 ms |
| Frame integration time (LIS) | 4 ms |
| Coverage (GLM) | Full disc, Americas and adjacent oceans, continuous |
| Coverage (MTG-LI) | Europe, Africa, Indian Ocean sector, continuous from 2024 |
| Spectral band | 777.4 nm (atomic oxygen triplet emission), narrowband filter |
| Near-real-time latency (GLM) | <20 seconds from observation to product dissemination |
| Archive depth | GLM from early 2017; LIS/OTD combined climatology from 1995 |
Analytics Satellize can run
| Regional flash-rate climatology maps | Aggregation and spatial gridding of GLM event/group/flash hierarchies over monthly and annual windows, cross-referenced to LIS/OTD baseline statistics | GIS raster layers (GeoTIFF) at 0.1° grid, annual and monthly means with interannual anomaly fields |
| Convective intensity proxy time series | Flash-rate trend analysis per convective season, using published power-law relationships between flash rate and updraught intensity | Time-series report with regional breakdowns, suitable for agricultural or infrastructure risk assessment |
| Lightning-jump detection for severe convection alerts | Rate-of-change thresholding on GLM flash counts per storm cell, following published lightning-jump algorithm literature | Near-real-time alert feed (GeoJSON) with cell identifier, flash-rate trend and timestamp |
| Oceanic convective hazard layer for route planning | Spatial aggregation of GLM and LIS flash detections over oceanic domains lacking radar coverage, binned to 6-hour windows | Gridded hazard layer updated every 6 hours, delivered as GeoTIFF or WMS tile service |
| ENSO-modulated flash-rate anomaly analysis | Correlation of regional monthly flash-rate anomalies against Niño-3.4 index using GLM archive and LIS/OTD climatology baseline | Analytical report with maps and correlation statistics, updated annually |
| Multi-sensor cross-calibration assessment | Matched-overpass comparison of GLM and LIS flash counts over common scenes, following published efficiency-correction methods | Calibration coefficient table and uncertainty estimate, delivered as a technical note |
Who does the work
We can get this done for you. Satellize runs its own analyst desk and a strong science team. You do not buy a data feed and work out what it means; our people source the imagery, run the analysis described on this page, and hand you the answer with its confidence limits stated. Discuss this requirement.