Lightning strike climatology for offshore wind risk assessment

Spaceborne lightning sensors on TRMM, the ISS, and GOES geostationary platforms map flash density at 0.5-degree resolution across decades, giving wind developers and insurers a quantified strike-risk baseline before a single foundation is poured.

Sensors

TRMM Lightning Imaging Sensor (LIS): Operated 1997 to 2015 on the Tropical Rainfall Measuring Mission. Detected optical pulses from lightning at roughly 0.5-degree spatial resolution over the tropics and subtropics (38°S to 38°N). Provides the foundational 18-year climatological archive used in most published flash-density maps.
ISS Lightning Imaging Sensor (ISS-LIS): Mounted on the International Space Station since 2017. Shares the same optical-pulse detection principle as TRMM LIS but benefits from the ISS inclination of 51.6°, extending coverage to higher latitudes relevant to North Sea and Baltic offshore wind zones. Revisit is non-uniform due to ISS orbital precession.
GOES-16/17 Geostationary Lightning Mapper (GLM): Continuous full-disc lightning detection over the Americas and adjacent oceans at roughly 8-km nadir resolution and sub-millisecond temporal sampling. Provides near-real-time flash, event and group data. Coverage does not extend to European or Asian offshore zones.
GPM IMERG (Integrated Multi-satellitE Retrievals for GPM): Precipitation climatology product derived from the Global Precipitation Measurement constellation. Spatial resolution of 0.1 degrees, 30-minute temporal resolution. Used as a convective-intensity proxy to contextualise lightning climatology where LIS spatial resolution is too coarse for site-level discrimination.

Why a turbine is not just another tall structure

A wind turbine presents a continuously rotating, electrically grounded conductor at heights commonly exceeding 150 metres for modern offshore machines. That geometry actively initiates upward leaders under strong electric field conditions, meaning turbines do not merely intercept lightning, they provoke it. The IEC 61400-24 standard for wind turbine lightning protection acknowledges this by requiring site-specific flash-density input (Ng, ground flash density in flashes per square kilometre per year) to calculate the expected annual strike rate per turbine.

The problem for developers is that offshore sites rarely have nearby ground-based lightning location network (LLN) sensors with sufficient detection efficiency over open water. Coastal LLN stations lose sensitivity and location accuracy beyond roughly 100 to 200 kilometres offshore, and detection efficiency can drop below 50 percent in some maritime configurations. Satellite-derived climatologies fill that gap directly.

What the Lightning Imaging Sensor actually measures, and what it does not

LIS detects optical pulses produced when a lightning channel illuminates the cloud top. It records both cloud-to-ground (CG) and cloud-to-cloud (CC) flashes without distinguishing between them. Published studies using combined LIS and ground-truth LLN data suggest that roughly 25 to 35 percent of total lightning flash counts over continental areas are CG, though the ratio varies significantly by meteorological regime and season. Over warm oceans the proportion of CG lightning tends to be lower than over land, which matters when converting total flash density to the Ng value that IEC 61400-24 requires.

Critically, LIS cannot measure strike current. Peak current determines a large fraction of the mechanical and electrical damage risk to blade spar caps, pitch-control wiring and power-conversion equipment. Characterising the current distribution requires either a local LLN with waveform-recording capability or adoption of regional current statistics from published studies. Satellite data sets the frequency; ground networks and physical models set the severity distribution. Neither is sufficient alone.

Building a site climatology from 0.5-degree pixels

The standard NASA LIS/OTD gridded climatological product delivers mean annual flash rate density at 0.5-degree resolution, approximately 55 km at the equator, coarser in latitude. For a typical offshore wind lease area of perhaps 200 to 500 square kilometres, a single LIS pixel may cover the entire site and its surrounding sea. That is not a flaw to hide; it is the honest spatial scale of the product, and it is sufficient for the climatological question being asked: what is the long-term average convective lightning environment at this location?

Refinement is possible. GPM IMERG convective precipitation fractions at 0.1-degree resolution can be used to spatially disaggregate LIS flash density within a grid cell, on the assumption that convective rain rate is a reasonable proxy for lightning-producing storm frequency. The TRMM archive spans 1998 to 2015, long enough to characterise interannual variability including ENSO-driven anomalies. Combining TRMM LIS with ISS-LIS data from 2017 onward extends the record but requires intercalibration because the ISS viewing geometry and dwell time differ from TRMM's sun-synchronous-adjacent orbit.

Convective storm tracking and seasonal structure

Raw flash density maps conceal the seasonal and diurnal structure that matters for operational risk. A site with moderate annual flash density concentrated in a short summer convective season presents a different engineering and maintenance-scheduling problem than one with low-intensity but year-round activity. GOES GLM, where its coverage is relevant, provides the sub-hourly temporal resolution needed to characterise diurnal cycles. For European and Asian offshore zones, Meteosat SEVIRI data and reanalysis convective available potential energy (CAPE) fields from ERA5 provide the temporal structure that satellite lightning sensors alone cannot resolve at adequate latency.

Convective storm tracking algorithms applied to geostationary infrared imagery can identify mesoscale convective systems and their preferred tracks, giving developers a physical explanation for spatial gradients in flash density that a gridded climatology alone does not supply. A site positioned in the preferred outflow region of a recurring convective system will see systematically higher strike rates than a site two grid cells away.

From climatology to insurance and protection design

Insurance underwriters pricing offshore wind lightning risk need two numbers: expected annual strike frequency per turbine and a plausible loss severity distribution. The satellite climatology supplies the first, via the IEC 61400-24 collection-area method applied to the site-specific Ng derived from LIS total flash density adjusted by a CG fraction appropriate to the meteorological regime. Severity distribution draws on published damage statistics and current waveform libraries, not satellite data.

For protection system design, the Ng value feeds directly into the rolling sphere method and the lightning protection level (LPL) selection in IEC 61400-24. Sites with Ng above roughly 2 to 4 flashes per square kilometre per year, common in parts of the Gulf of Mexico and South-East Asian offshore zones, will typically require LPL I protection on blade receptors and enhanced shielding on nacelle electronics. Sites in the North Sea typically see Ng values below 1, though local topographic channelling of convection near coastlines can produce local enhancements that a 0.5-degree climatology will smooth over. That smoothing is an honest limit of the satellite product.

Satellize builds site-specific lightning climatology reports drawing on the NASA LIS/OTD archive and ISS-LIS data, cross-referenced against ERA5 convective indices, as part of its renewable energy siting analytics. The methodology is the same one applied in its broader satellite-data work, including the Kingdom of Tonga crop-estimation programme, where combining open-archive sensors with physical contextualisation is standard practice.

What satellite data cannot replace

No satellite lightning sensor currently operational provides the location accuracy needed to assign a strike to a specific turbine within a wind farm. GOES GLM achieves roughly 8 km nadir accuracy; LIS is coarser still. Ground-based LLN sensors with time-of-arrival or magnetic direction-finding networks routinely achieve sub-kilometre accuracy over land, falling to perhaps 2 to 5 km offshore depending on network geometry. For post-event forensic analysis of a specific turbine failure, a ground network is necessary.

Satellite data also cannot resolve flash multiplicity (the number of return strokes per flash), which influences cumulative charge transfer and mechanical fatigue on blade receptors. The combination that actually works for a rigorous site assessment is: satellite climatology for the long-term frequency baseline, a temporary offshore LLN deployment or nearest permanent coastal network for location accuracy during the development phase, and published regional current statistics for severity modelling. Satellite data is the starting point, not the complete answer.

Typical figures

LIS/OTD climatological spatial resolution	0.5 degree (~55 km at equator)
TRMM LIS archive depth	1998 to 2015 (approximately 17 years)
ISS-LIS coverage period	2017 to present; latitude coverage to ~51.6°N/S
GOES GLM spatial resolution	~8 km at nadir, degrading toward limb
GOES GLM temporal sampling	Sub-millisecond event detection; 20-second full-disc imagery cycle
Flash type detected	Total lightning (cloud-to-ground and cloud-to-cloud); CG/CC discrimination requires ancillary ground network
GPM IMERG convective proxy resolution	0.1 degree, 30-minute intervals; archive from 2000
Minimum detectable flash radiance (LIS)	Approximately 10 µJ/sr/µm; detection efficiency ~90% for strong flashes, lower for weak or daytime events
Deliverable coordinate system	WGS84 geographic; output as GeoTIFF or NetCDF climatological grids
IEC standard interface	Ng (ground flash density, fl/km²/yr) for IEC 61400-24 collection-area calculations

Analytics Satellize can run

Site flash-density climatology map	NASA LIS/OTD gridded product extraction and ISS-LIS intercalibration; GPM IMERG spatial disaggregation within 0.5-degree cells	GeoTIFF raster and tabular Ng estimate with uncertainty range for the lease boundary, formatted for IEC 61400-24 input
Seasonal and interannual flash-density decomposition	Monthly LIS flash-rate anomaly analysis against ENSO index (MEI or ONI); ERA5 CAPE climatology overlay	PDF report with monthly flash-density profiles and ENSO sensitivity assessment for P50/P90 risk bracketing
Convective storm track density map	Mesoscale convective system tracking applied to Meteosat SEVIRI cold-cloud-top brightness temperature archive	GIS polygon layer of preferred storm tracks and initiation zones within 300 km of the site
CG flash fraction estimation	Regime-based CG/total ratio drawn from published LIS-versus-LLN comparison studies for the relevant ocean basin	Tabular CG fraction range (low/central/high estimate) with literature citations, for use in IEC 61400-24 LPL selection
Multi-site comparative lightning risk ranking	Zonal statistics of LIS flash density across candidate lease polygons; percentile ranking within regional climatological distribution	Ranked comparison table suitable for portfolio-level insurance underwriting or site-selection screening
Historical extreme-flash-rate event identification	Percentile-threshold exceedance analysis on annual LIS flash-rate time series; cross-reference with ERA5 reanalysis storm events	Event log of years with anomalously high flash density, with meteorological context, for stress-testing insurance loss models

Sources

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.