Lightning monitoring: how to avoid unplanned downtime

In our experience with weather monitoring, lightning is that transversal threat: every business with open-air operations is affected and, of course, so is our daily life.

If we should not run outdoors on a day of electrical storms, nor cycle, swim in open water, or surf, we also should not keep open-air maintenance running, or leave workers unsheltered on a construction site or at the open-pit face of a mine.

In the field, that same prudence becomes a maintenance and operations stoppage: the fleet is recalled, the lift is suspended mid-cycle, work at height on the tower or the turbine waits, the track crew leaves the section. The protocol is right to stop. What is missing, in most operations, is knowing in advance when the storm window opens and when it closes, so the work can be rescheduled instead of the suspension being improvised.

One sector is particularly affected and needs even more auditable measures: power transmission and distribution. The grid crosses the entire territory, and every tower and every kilometer of line stays exposed all year round.

Regional data gives the scale. Climate catastrophes cost Latin America US$ 11.6 billion in 2024, and only US$ 1.5 billion of that total was covered by insurance. Lightning does not appear on its own in that count: it sits inside the region's grouped storm category.

In Brazil, the region's largest market, the INPE space research institute counts 100 to 150 million atmospheric discharges per year on average over the last decade, and projects that incidence will grow 15% to 40% for every 1 °C of increase in global average temperature. On the other side of the Atlantic, the Netherlands broke its national daily record on June 27 and 28, 2026, with more than 300,000 discharges in 24 hours.

The effect on operations shows up fast. That same storm sequence caused flight delays and cancellations at Heathrow and Gatwick and left thousands of customers without power in several European countries.

Without that warning, lightning collects on three fronts at once:

Life: whoever is exposed in the open at the moment of the discharge. The crew at height on the tower or the turbine, the operator at the pit face, the track team on the section.
Operations: everything that stops with it. Fleet recalled, lift suspended mid-cycle, line tripped by protection systems, loading yard at a standstill.
Budget: every hour of that stoppage, and every hour of late resumption, lands in the month's operating cost.

In Brazil, the lightning-specific view exists on the insurance side. In a FenSeg survey published in 2023, electrical damage accounted for 35% of all property insurance payouts, and claims of that type grew 63% between the first quarter of 2020 and the first quarter of 2023. The electrical damage category is broader than lightning, so treat the number as a ceiling, not as the closed account of lightning losses.

Lightning is the climate hazard that almost always has a formal stop protocol and almost never has a forecast window. Mining, rail, transmission, wind farms, heavy construction, and industrial plants all have a written procedure for atmospheric discharge, audited and enforced in inspections.

The cost, however, dissolves into different categories: corrective maintenance, burned equipment, downtime hours, workplace accidents, schedule delays. The real pain is spread across several cost lines and rarely shows up consolidated in a single report with the word lightning under it.

In this article

The lightning hazard: how it forms in Brazil, Chile, Mexico, and Europe.
How to turn the lightning hazard into risk calculated by asset.
Lightning and atmospheric discharge monitoring: detecting is not anticipating.
What the same electrical storm impacts in each sector, starting with the table that sums it all up.
- Mining: the stop protocol already exists; what is missing is the time to resume.
- Railways: the exposed wiring, the stopped train, and the crew off the track.
- Power transmission and distribution: the crew arrives before the outage.
- Renewables: the blade is the most expensive asset exposed to the open sky.
- Heavy civil construction: rescheduling costs less than stopping mid-cycle.
- Industry: the production line stops without lightning touching the plant.
The return on investment from anticipating the hazard.
The method: three steps to apply starting tomorrow.
Where this becomes a daily decision.

The lightning hazard: how it forms in Brazil, Chile, Mexico, and Europe

Lightning is the electrical discharge the atmosphere uses to undo a charge imbalance built up inside a storm cloud. Inside the cumulonimbus, strong updrafts carry droplets, ice crystals, and hail upward. Collisions between those particles transfer charge, the top of the cloud turns positive and the base turns negative. When the potential difference beats the insulating capacity of the air, the discharge comes, to another cloud or to the ground.

The recipe is always the same: heat at the surface, available moisture, and a mechanism that pushes that air upward. What changes from country to country is the atmospheric system doing the pushing.

Brazil: abundant heat and humidity sustain convection almost year-round, and warming amplifies the bill. INPE researcher Kleber Naccarato sums up the mechanism: the warmer the environment, the more energy is available to form storms and the greater the electrical activity. Deforestation, air pollution, and urbanization also redistribute where discharges fall.

Chile: the electrical storm usually comes from a cut-off low aloft, the cold-air core that detaches from the main circulation and forces surface air to rise. On February 1, 2026, a storm of that type over the Santiago Metropolitan Region recorded about 1,100 lightning strikes in two hours and 17 millimeters of rain in 20 minutes, at the height of summer.

Mexico: in the rainy season, instability comes from tropical waves and low-pressure troughs crossing the country. On June 25, 2026, tropical wave number 11 generated rain with hail and electrical discharges over Mexico City, with four short circuits recorded and rainfall of 30 to 49 millimeters in four regions of the city.

Europe: the continent's most electric pattern appears when a cold air mass advances over the warm, humid air built up by a heat wave. That is what happened on June 27 and 28, 2026: the Netherlands recorded more than 300,000 discharges in 24 hours, the highest daily total in the country's records, with a peak of about 30,000 discharges in 30 minutes between Dutch territory and northern Germany.

Four regions, four different atmospheric triggers, the same physics inside the cloud. That is why the lightning hazard applies to any operation exposed to the open sky, in any of these markets.

How to turn the lightning hazard into risk calculated by asset

Two operations receive the same storm cell. The one that knows which of its assets enters the critical window, and at what time, reschedules the agenda and protects the crew without losing the shift. The other pays for improvisation. In practice, the difference shows up as a lift moved from afternoon to morning, sheltered maintenance pulled into the risk window instead of wasted under clear skies, and a field crew positioned near the section before the line trips.

Lightning is the hazard. Losses and risk to life are the risks. Between the hazard and the risks are factors that multiply or reduce the impact, and they decide whether the same discharge lands in an empty field or on the turbine blade that had a crew at height at that moment. i4sea follows three layers to make that conversion.

Layer 1, the data. The base is a proprietary numerical model with 1 to 3 km resolution across all of Latin America, calibrated with more than 10 years of real climate history. On top of that base run more than one hundred AI scenarios, used to understand each storm's behavior: which kilometers will be affected and which areas are most prone to trouble. Where public forecasting operates at roughly 25 km resolution and answers storm in the region, this reading starts at the scale of the pit, the grid section, and the job site.

For lightning, a measurement layer comes in as well: satellite data and ground sensors identify every discharge that fell around the operation, within radii ranging from 10 to 100 kilometers. The distance between the discharge point and the operation can be set as one of the protocol's decision triggers. What separates two machines in the same wind farm is not the model grid, but the exposure layer described next.

Layer 2, the interaction with territory and business. The same storm does not produce the same result at two different points. The reading crosses the hazard with:

Topography and the asset's relative height: a tower, crane, blade, or isolated steel structure at the high point of the terrain takes discharges more often than its surroundings.
Historical discharge density at that point: a line section with a high incidence history has a different probability from a neighboring section with a low one.
Soil resistivity and grounding condition: high-resistivity ground turns the same discharge into greater damage to connected equipment.
What is exposed in the open at that hour: a crew at height, a fleet in operation, a suspended load, and a concrete pour in progress change the risk without changing the hazard.
What depends on the point struck: the more people, assets, and revenue pass through that specific section, the higher the risk, at the same discharge probability.

Layer 3, the risk matrix. i4sea connects that threat's index, intensity, and hazard window at the exact point with the impact it causes in that specific business. That connection turns electrical storm in the region into a risk reading calculated by asset, ready for decision.

Lightning is a probabilistic phenomenon, and no model says where the discharge will fall. What exists is the probability of a critical window per asset, with the uncertainty declared, and that is already enough to decide.

These layers sustain the four gains the operation feels: lower cost, less exposure of lives, crews allocated where the risk is, and planning closed before the event.

Lightning and atmospheric discharge monitoring: detecting is not anticipating

Whoever searches for lightning monitoring finds, in most results, detection systems: sensor networks and alarms that report where a discharge just fell and how far away the storm is right now. That layer works, has its place in the protocol, and saves lives. What it does not answer is the question that defines the day's cost: which of my work fronts enter a critical window today, and at what time?

There are two clocks in play, and they trigger different decisions. The imminence clock, of tens of minutes, belongs to real-time detection and serves to get people out of harm's way. The window clock, of hours to days, belongs to storm modeling and serves to reschedule shifts, rosters, lifts, and crew mobilization. An operation working only with the first protects people and pays for the entire schedule through improvisation.

Complete weather monitoring adds the two layers together. Detection protects whoever is in the field when the cell arrives. Anticipation decides, with hours or days to spare, what each work front does before the storm even forms. The second layer is the one usually missing, and it is where the bill for unplanned downtime is decided.

What the same electrical storm impacts in each sector

The physics inside the cloud is the same. What stops, which indicator suffers, and who signs off change completely.

Table 1 · What lightning charges each sector

Sector	What the discharge interrupts	Where the bill shows up
Mining	Fleet recalled, work at height suspended, pit face stopped	Unplanned stoppage, disrupted shift, delayed dispatch
Railways	Overhead contact line and signaling struck, train stopped, track and maintenance crews off the section	Traffic window, maintenance backlog, cascade effect across the network
Power T&D	Line tripped by protection systems, burned equipment	Outage duration, continuity indicators, emergency cost
Renewables	Blade struck, work at height canceled across the whole farm	Asset service life, unavailability hours, wasted mobilization
Civil construction	Lifting, concrete pours, and assembly interrupted mid-cycle	Rework, schedule days, risk to life
Industry	Open-air maintenance and work suspended, plant stopped by a line tripped by voltage sag	Maintenance backlog, extra contractor cost, maintenance OPEX, electrical damage claims

Mining: the stop protocol already exists; what is missing is the time to resume

Start with the result. A mine that receives, the night before, the discharge probability per pit area pulls sheltered maintenance, in the workshop and the electrical room, into the high-risk window, and saves the clear sky for production and open-air service. That way it loses one window, not two: the production window taken by lightning, and the maintenance window that had been scheduled for a storm-free period. The asset stops either way. The hour is not lost twice.

The protocol fires and is right to fire: the fleet is recalled, discharge-sensitive activities stop, work at height stops. What is usually missing are the two ends. The half hour before, when the recall order arrives at the last minute and disrupts the entire shift. And the hour after, when nobody knows whether to release the front or wait for the next cell.

By the estimates i4sea uses in its benefit studies, an unanticipated stoppage in mining costs between US$ 150,000 and US$ 500,000 per event, with an expected return of 25 to 50 times for those who anticipate.

Railways: the exposed wiring, the stopped train, and the crew off the track

On an electrified railway, the most exposed asset runs suspended above the track: the overhead contact line, the catenary that powers the train. It crosses kilometers in the open, at the high point of the alignment, and shares the same corridor with signaling and telecommunications. A discharge into that set switches off more than a piece of equipment: it brings down the logic that authorizes the train to move, and the train stops.

That is what happened on the West Coast Main Line, in the United Kingdom, on May 28, 2026. Lightning damaged signaling equipment at Weaver Junction and the operator went down to one train per hour between Crewe and Liverpool during repairs, with cascading delays and cancellations.

On a diesel railway, with no overhead line, lightning charges a different target: people. The maintenance team and the operations crew working on the track cannot stay on the section with an active electrical storm. Inspection stops, open-air maintenance stops, gantry loading stops. Every suspension without a forecast window consumes the sector's two currencies: the maintenance window, which pushes work into the backlog, and the traffic window, which does not come back.

With the window forecast the day before, the control center flips the agenda: it concentrates loading in the morning, places sheltered maintenance inside the risk window, and schedules the track crew to return to the section as soon as the cell passes. The train that would not load at 3 p.m. loads at 10 a.m.

An unanticipated line closure costs between US$ 50,000 and US$ 200,000 per day, with an expected ROI of 10 to 20 times for those who anticipate.

The linear asset is the one most penalized by generic forecasting. A long network crosses several microclimates, and the storm cell that matters covers a handful of its kilometers. A control center can do nothing with storm in the region, and a great deal with high risk on section X between 2 p.m. and 6 p.m.

Power transmission and distribution: the crew arrives before the outage

A distribution utility that receives high discharge risk per section for the afternoon window moves the crew to the base closest to the exposed corridor while it is still morning. The outage happens anyway. The fault-location time drops, because the crew is already 20 minutes away instead of two hours. The indicator that improves is outage duration.

This is the sector with the best public documentation of the hazard and the worst track record of anticipation. Between 2014 and 2023, weather events accounted for 43% of the causes of transmission line outages in Brazil, according to an EPE survey based on ONS data. On the distribution side, the concentration effect is visible in a single Brazilian concession: more than 114,000 consumer units lost power in electrical storm episodes in the first quarter of 2026 alone, 23,600 more than in the same period of 2025.

Nobody promises to prevent the discharge, and the gain is not there. It lies in three decisions that only exist with warning:

Positioning: crew moved before the cell reaches the section.
Switching: grid assessment to preserve redundancy in the exposed corridor.
Communication: warning the sag-sensitive industrial customer that the corridor is entering a high-probability discharge window, before the event, not after. The voltage sag itself is not predictable; the risk window is.

A day of unavailability in power costs US$ 40,000, with an expected ROI of 6 to 12 times for those who anticipate. The Brazilian utility behind that quarterly figure responded to the rise in discharges with automatic reclosers and a tripling of field crews, faster-reaction measures. The temporal layer, which says in which section and in which window the cell will pass, is what decides where those crews wait.

Renewables: the blade is the most expensive asset exposed to the open sky

An O&M team that receives the risk window crossed with each machine's exposure, rather than a warning by municipality, suspends work at height on the exposed machine at the top of the hill and keeps internal inspection going on the machines in the valley, in the same shift. The alternative is what happens today in most wind farms: a storm-in-the-region warning, the whole shift canceled, a maintenance window lost on machines that were never at risk.

In the wind farm, lightning produces what the other sectors do not have: capital damage, on top of time. A struck blade becomes a high-cost repair and a turbine out of operation until inspection clears it.

Brazil has a high discharge density and a relevant installed wind base, and the combination puts blade service life and O&M cost into the same decision. The same reading holds for wind farms in Mexico and across Europe: the machine at the high point of the terrain concentrates the exposure.

Less blade exposure is asset service life. Less unnecessary mobilization is pure O&M cost. Both depend on the same thing: a reading by asset, not by municipality.

Heavy civil construction: rescheduling costs less than stopping mid-cycle

A work front that knows the day before that discharge probability will rise at 3 p.m. does not start the concrete pour at 1:30 p.m. It starts at 7 a.m. or reschedules for the next day. The weather event is identical in both scenarios. In the first, the project loses a schedule day. In the second, it loses an afternoon already accounted for.

Crane lifts, concrete pours, work at height, and steel structure assembly have a point of no return. Stopping midway costs material, rework and, in concrete, quality. The lightning protocol suspends the front and is right to suspend it.

A day of delay on a large project costs between US$ 50,000 and US$ 150,000, with an expected ROI of 10 to 15 times for those who anticipate. On a job site, though, money is the second consequence. The first is the crew at height on a steel structure with a storm cell approaching, and no contract amendment compensates for that.

The project's legal team gains something the site rarely notices. Suspending a lift or a pour under the atmospheric discharge protocol consumes a schedule day, and that day only becomes a claim when there is a record showing the electrical storm warning arrived before the suspension, with time, responsible person, and action taken. Without that record, the lightning stoppage enters the report as contractor delay.

Industry: the production line stops without lightning touching the plant

A plant that receives, the day before, its region's electrical storm window postpones the start of the batch that cannot be interrupted midway, pulls sheltered electrical maintenance into the risk window, and checks grounding and protection before the cell arrives. The discharge happens anyway. What changes is at which point in the cycle it finds production.

Lightning also charges the factory what it charges the job site and the mine: open-air maintenance and work suspended when the cell approaches. Every suspension without a forecast window pushes work into the maintenance backlog, and a tight backlog turns into extra contractor cost inside maintenance OPEX.

Lightning does not need to hit the roof to stop the factory. A discharge that lands on the grid kilometers from the plant produces a voltage sag, the same phenomenon the transmission section described from the utility's side, and a sensitive process trips on a sag lasting a fraction of a second. A furnace, an extruder, a filling line, and a boiler do not restart at the push of a button: every trip charges hours of resumption, scrap from the batch in progress and, at the limit, burned equipment.

In the aggregate, industry's stopped hour is expensive. An ABB survey with Sapio Research, run in July 2023 with 3,215 plant maintenance decision-makers across several countries and sectors, measured a typical cost of US$ 125,000 per hour of unplanned downtime, and more than two-thirds of companies reported at least one stoppage per month. The number covers all causes of downtime, not just lightning, and enters here as an order of magnitude for the lost hour. On the insurance side, the FenSeg survey cited in the opening points the same way: electrical damage accounted for 35% of property insurance payouts in 2023.

With a forecast window, the decisions are the same ones this article describes for the other sectors: reschedule what has a point of no return, protect sensitive equipment, coordinate communication with the utility on the exposed corridor, and record warning, action, and outcome for the claims discussion.

The return on investment from anticipating the hazard

No system prevents lightning. What changes is the cost of dealing with it. The cases below measure that delta, and none of them measures lightning: they are here because they measure the mechanism this text defends, trading reaction for anticipation on critical assets, with an auditable number.

Vattenfall, wind power. Operating with hyperlocal forecasting instead of regional-scale forecasting, the case records an ROI of 27 times in the comparison against the global ECMWF model, with fewer unnecessary mobilizations and more maintenance windows used. What that number measures is per-asset resolution against regional-scale forecasting, the same delta that decides whether work at height stops across the whole farm or only on the exposed machine.

Puerto Mejillones, port terminal in Chile. With 426 alerts issued in the first quarter of 2026, the annualized benefit reached US$ 305,000. The same results table records preventive preparation in 28 events, about 1 serious incident avoided per year, and about 40 hours recovered per year. The phenomenon there is maritime, and the mechanism is identical: a forecast event becomes a planned event.

The logic behind both cases has independent academic validation. An NBER working paper measured that the improvement in hurricane forecast accuracy in the United States between 2007 and 2020 cut total cost by 23%, an average of about US$ 2 billion per hurricane in damage and emergency spending. The hurricane kept happening. The quality of the preparation is what changed.

The method: three steps to apply starting tomorrow

These steps work with a spreadsheet, an in-house system, or a vendor. None of them depends on hiring i4sea.

List the decisions that stop for lightning and who signs each one. List decisions, not assets: recall the fleet, suspend the lift, interrupt the pour, cancel work at height, move the maintenance crew. For each one, write the threshold that triggers it, the person who decides, and the minimum time they need to execute without improvising. That minimum time is your lead-time requirement, and it is almost never minutes.
Separate the two clocks and state which decision each one triggers. The imminence clock, of tens of minutes, triggers protection of people. The window clock, of hours to days, triggers rescheduling, rosters, mobilization, and communication with customers or the chain. Write in one line, per decision, which clock commands it. An operation using only one clock is paying for the other without knowing it.
Record forecast risk, action taken, and observed outcome, event by event. A simple record with date, time, asset, forecast risk, decision, and outcome delivers three things: calibration of your own threshold, evidence for insurance, amendments, and contract discussions, and the delta calculation. Add up, over a season, the difference between the cost of the planned stoppage and the cost of the emergency stoppage. That number is your ROI, and it is yours, not the vendor's.

The gain from the three steps is the same across the six sectors: the structurally lower cost of preparation replaces the cost of reaction. It is the Delta Framework i4sea uses in its benefit studies, and it fits in one line: benefit is the sum of the difference between unplanned cost and planned cost, multiplied by the number of events.

Where this becomes a daily decision

A lightning protocol written into a procedure only protects the operation when the team receives the warning in time, on the channel it already uses, with the recommended action attached. Weather monitoring that arrives as one more map to interpret does not solve that: the bottleneck is the distance between the data and the person deciding at 1:40 p.m., with the work front stopped and the phone in hand.

The AI Climate Agent delivers the decision to the team before the event: which asset, which window, which action. It already knows your operation's limits, protocols, and history, arrives on the channel the team already uses, WhatsApp, Teams, or email, and records in every answer the source of the data behind it.

Try the Climate Agent on your operation, with a free 14-day trial and no credit card:

If you prefer to start with a diagnosis, we map your assets' exposure to climate hazards and show you what the system would see today, with real data.

Request a free climate exposure diagnosis by asset:

The hours of warning your operation has before the next electrical storm, and what it does with them, decide the cost of the next lightning strike.

Sources

Climate catastrophes cost Latin America US$ 11.6 billion in 2024, with only US$ 1.5 billion covered by insurance. Swiss Re, sigma NatCat 2025. Scope: lightning is not isolated; the figure is a regional catastrophe aggregate.
Brazil records 100 to 150 million atmospheric discharges per year on average over the last decade; incidence grows 15% to 40% per 1 °C of global average temperature increase. Kleber Naccarato, INPE. INPE, Mar 26, 2026.
Federal District (Brazil), Q1 2026: more than 114,000 consumer units without power in electrical storm episodes, 23,600 more than in the same period of 2025; utility's reading; automatic reclosers and tripling of field crews. Neoenergia Brasília, Apr 23, 2026.
Chile, Feb 1, 2026: about 1,100 lightning strikes in two hours over the Santiago Metropolitan Region and 17 mm of rain in 20 minutes. La Tercera.
Mexico, Jun 25, 2026: tropical wave number 11 generated rain with hail and electrical discharges over Mexico City, with four short circuits and 30 to 49 mm of rainfall in four regions of the city. La Jornada, Jun 26, 2026.
Netherlands, Jun 27 and 28, 2026: more than 300,000 discharges in 24 hours, the highest daily total in the country's records. Metbeat News, citing KNMI. European effects (Heathrow and Gatwick flights, customers without power): Business Today, Jun 28, 2026.
Electrical damage accounted for 35% of property insurance payouts in Brazil, with a 63% rise in claims between Q1 2020 and Q1 2023 (FenSeg). Revista de Seguros, CNseg, June 2023. Scope: electrical damage is a broader category than lightning, and the body says so.
Weather events accounted for 43% of the causes of transmission line outages in Brazil between 2014 and 2023. EPE, based on ONS data, 2025.
United Kingdom, May 28, 2026: lightning damaged signaling equipment at Weaver Junction on the West Coast Main Line; the operator ran one train per hour between Crewe and Liverpool during repairs. Travel And Tour World.
Unplanned downtime in industry costs an average of US$ 125,000 per hour; more than two-thirds of companies report at least one unplanned stoppage per month. ABB survey with Sapio Research, July 2023, 3,215 plant maintenance decision-makers, global and multi-sector scope. ABB, Oct 11, 2023.
NBER: the improvement in US hurricane forecast accuracy between 2007 and 2020 cut total cost by 23%, about US$ 2 billion per hurricane. Molina, R. and Rudik, I., "The Social Value of Hurricane Forecasts", NBER Working Paper 32548 (2024). nber.org/papers/w32548.
Source: i4sea, proprietary data. Numerical model at 1 to 3 km resolution across Latin America, calibrated with more than 10 years of climate history; more than one hundred AI scenarios on top of the model; discharge measurement layer via satellite and ground sensors, within radii of 10 to 100 km around the operation; public forecasting at roughly 25 km.
Source: i4sea, internal estimate. Cost of one unanticipated event by sector: mining US$ 150,000 to 500,000 per event; rail US$ 50,000 to 200,000 per day; power US$ 40,000 per day; large construction projects US$ 50,000 to 150,000 per day.
Source: i4sea, real case. Vattenfall, wind power: 27x ROI against the global ECMWF model.
Source: i4sea, real case. Puerto Mejillones, port terminal in Chile: 426 alerts in Q1 2026, annualized benefit of US$ 305,000, preventive preparation in 28 events, about 1 serious incident avoided per year, and about 40 hours recovered per year.