Extreme Rain: How to Avoid Unplanned Downtime

In May 2024, the Rio Grande do Sul flood in Brazil shut down highways and railways at the same time: the Federal Highway Police recorded 159 highway closures in the state from landslides and flooding, and Brazil's rail regulator, ANTT, counted 46 closures across seven railways, all caused by the rain.

It was also the flood event with the largest insured loss ever recorded in Brazil, US$1 billion in claims paid by insurers, according to Swiss Re. That figure measures only the insured share of the damage: using the DaLA methodology, the World Bank, the IDB, and CEPAL estimated the total damage from the same disaster at R$88.9 billion, most of it with no insurance coverage at all. None of this required a dam to break: the rain simply found the points that already had nowhere left to drain.

Extreme rain is the cause. What it produces has at least two different names, and the difference matters to whoever operates the asset. When local rain exceeds a point's drainage capacity, urban drainage, a yard, an internal haul road, the result is urban flooding, with no river involved at all. When a river receives more water than its channel can carry, often from rain that fell hours earlier and kilometers away, at the headwaters of the basin, the result is river flooding. Brazil's official civil defense classification (COBRADE) separates the two exactly this way: extreme rain is a meteorological event; urban flooding and river flooding are distinct hydrological consequences of it.

Without enough warning of which point will flood and in what window, extreme rain charges on three fronts at once:

Life: whoever is in the low point when the water rises. The field crew at the access road, the drivers, the team at the foundation front.
Operations: whatever depends on that point stops working. The highway is closed, the rail corridor backs up train formations, the job site sits idle.
Budget: every hour of that stoppage enters the month's operating cost, spread across several different lines.

Climate disasters cost Latin America US$11.6 billion in 2024, with only US$1.5 billion (13%) covered by insurance. Urban flooding, river flooding, and flash floods fall into that count within the regional aggregate, with no line item of their own. In the first quarter of 2026, floods, landslides, and convective storms caused US$260 million in losses in Brazil, and the Colombia flood was the region's costliest event in the period, US$2.2 billion, also in a grouped category.

Across the 2022-2024 period, Brazil alone accounted for R$184 billion in climate-related losses, and 91% of that value had no insurance protection. In the same survey, extreme rain and flooding appear as the most frequent events of the period.

Extreme rain hits multiple points of the same operation at once, from ports to mining, from highways to construction sites, and it tends to stay off the risk spreadsheet of most companies. It's the invisible cost of climate that shows up as force majeure, hits the whole company vertically, and needs to be analyzed.

It stops the gate access, the mine's internal haul road, the low-lying kilometer of the concession, the rail head, which is the running surface of the track, the substation yard, and the job site front. Each of these stoppages lands in a different record (productivity, maintenance, contractual penalty, force majeure), so the real pain stays scattered across several cost lines and rarely shows up consolidated in a single report with the word "rain" underneath it.

In this article

The danger of extreme rain: what sets it apart from urban and river flooding, and how it forms in Mexico, Europe, Brazil, and Chile.
Turning the danger of extreme rain into risk calculated by asset.
What the same rain impacts in each sector, starting with the table that sums it all up.
- Ports and terminals: the same rain costs differently in bulk, container, and berth.
- Mining: the haul road decides whether the mine stops with a plan or on the fly.
- Highways: the kilometer that closes has a probable hour.
- Railways: a few centimeters over the rail head lock the whole corridor.
- Power transmission and distribution: the water reaches the substation before the crew.
- Heavy civil construction: the rain already stops the work front; the record that backs the schedule is missing.
The return on investment from anticipating the hazard.
The method: three steps to apply starting tomorrow.
Where this becomes a daily decision.

The danger of extreme rain: what sets urban flooding apart from river flooding

Extreme rain is the cause. Under Brazil's official Disaster Classification and Coding system (COBRADE), run by the National Civil Defense, intense rain is an event in the meteorological group, and what it triggers on the ground and in rivers falls into three different categories within the hydrological group: urban flooding, river flooding, and flash flooding. All three share the same atmospheric cause, but not the same mechanism, and it's the mechanism that decides which point of your asset will stop.

Urban flooding is drainage capacity being overwhelmed: too much local rain for that point's storm drain, culvert, or channel, with no river involved at all. River flooding is a watercourse overflowing beyond its normal limits, usually gradually, fed by rain accumulated across the entire basin, including upstream, far from where the water finally rises. Flash flooding is sudden, fast surface runoff, typical of a small, dry basin on steeper terrain, that overflows within hours with major destructive force. A port with no river nearby can flood just because its yard can't drain the day's rain. A city kilometers away from any rain can flood because it rained at the headwaters of the river running through it. These are different problems, and each needs a different read.

The mechanism changes from continent to continent, and even changes within the same country.

Mexico: river flooding too, at a distance that proves the point. On October 10, 2025, intense rain associated with tropical storm Raymond fell in northern Veracruz and reached Poza Rica as a flood wave: the Cazones River rose more than 3 meters above the alert level at the Cazones bridge, and the city woke up underwater without the heaviest rain having fallen there. Thirty-eight municipalities were affected in the state and 35 communities were isolated, with 37 deaths in Veracruz and 83 across the five affected states combined.

Europe: here the mechanism is the third one, flash flooding. In the October 29, 2024 DANA (an isolated low-pressure system at high altitude) in Spain, the Turís station in Valencia recorded 772 mm in 24 hours, and the 185 mm in a single hour is the Spanish national record, according to AEMET's report on the event. The Poyo ravine and other dry channels, normally empty, concentrated that water and carried it into urban areas within two to three hours of the rain, too fast to be classic river flooding. The official toll is 237 deaths, 229 of them in the Comunitat Valenciana, with about 90% of victims in the Júcar basin concentrated in the Poyo and Saleta ravine catchments.

Brazil: the two mechanisms show up side by side. On February 23-24, 2026, the Paraibuna River rose over Juiz de Fora and Ubá after a supercell with a cloud top at 16 km, over a region that had already accumulated 589.6 mm that month: that's river flooding, a river overflowing from rain accumulated across the basin. On May 4-5, 2026, more than five hours of rain flooded historically vulnerable points in Porto Velho (Rondônia), and the local infrastructure department itself treated the episode as a recurring drainage problem: that's urban flooding, with no river involved.

Chile: the dominant pattern is river flooding. In the August 1-3, 2026 frontal system, Constitución (Maule region) recorded 45 to 60 mm in under an hour, and the Mataquito, Cautín, Imperial, Bueno, and Rahue rivers, among others, overflowed. The official toll reached 1,142 people displaced, with more than 14,000 affected in La Araucanía alone and two deaths in Panguipulli. On July 28, Chile's emergency agency Senapred had already issued a triple alert for overflow risk in the Los Ríos region, where residents along the Llollelhue River were evacuated three times.

Four countries, one shared cause, and three different mechanisms: urban flooding in Porto Velho, river flooding in Mexico, Chile, and Brazil, flash flooding in Spain. That's why the climate hazard of extreme rain applies to any operation with a low point, an access road, a drain, or a watercourse in its path, in any of these markets, and that's why the right response depends on knowing which of the three mechanisms is at play for that specific asset.

Turning the danger of extreme rain into risk calculated by asset

Two operations receive the same rain. The one that knows which of its points will flood, and in what window, clears the drainage beforehand, reroutes its fleet, and keeps operating. The other one pays for improvisation. In practice, the difference shows up as drainage cleared the day before instead of a pump rented in a rush, a fleet rerouted on a planned shift instead of a truck stuck on the haul road, and concrete pours rescheduled from Thursday to Saturday without losing the ready-mix already ordered.

Urban flooding is the hazard. Loss is the risk. Between the two sit factors that multiply or reduce the impact, and they're what decides whether the same rain drains without incident over a flat, inactive area, or shuts down the only kilometer of access to a highway concession for six hours. i4sea uses three layers to make that conversion.

Layer 1, the data. The foundation is a proprietary numerical model with 1-3 km resolution across all of Latin America, calibrated with more than 10 years of real climate history. Where public forecasting operates at roughly 25 km resolution and speaks for the region, i4sea already starts from a reading sized to the gate access, the mine's haul road, the low-lying kilometer, or the substation yard.

Layer 2, the interaction with the terrain and the business. The same extreme rain doesn't produce the same result at two different points, and the first factor that decides this is which of the two mechanisms is at play. The reading crosses the hazard with:

Mechanism, urban or river flooding: a point with no river nearby only floods when local rain overwhelms the drainage. A point on the banks of a watercourse can also flood from rain that fell hours earlier and kilometers away, at the basin's headwaters, as happened in Poza Rica. These are two different readings, and the second requires monitoring the entire basin, not just the sky over the asset.
Elevation, slope, and land use: the real geography of that point decides where the water goes before it decides whether it rises.
Drainage capacity and condition: storm drains, culverts, and channels sized for another era, or without recent maintenance, cross the threshold with far less rain.
Surrounding impermeabilization: a paved yard, a storage area, and a service road reduce infiltration and concentrate flow at the lowest point.
Prior soil saturation: soil already soaked by last week's rain sends almost all of the new rain straight back to the surface.
What's in the water's path: a flooded area with no human or economic activity generates no business risk. The same water covering an access road, a rail line, or a yard with people working does.
What depends on the affected point: the more people, assets, and revenue that pass through that specific point, the higher the risk, even with an identical climate hazard.

Layer 3, the risk matrix. i4sea connects that threat's index, intensity, and timing window at the exact point with the impact that causes to that specific business. That connection is what turns "it's going to rain hard in the region" into a risk reading calculated by asset, ready for a decision.

These three layers are what support the four gains an operation feels: lower cost, less exposure of lives, crews positioned where the risk is, and planning locked in before the event.

What the same rain impacts in each sector

The physics are the same. The point that floods, the metric that suffers, and who signs off on it change completely.

Table 1 · What extreme rain costs by sector

Sector	What the water interrupts	Where the cost shows up
Ports and terminals	Gate access, yard, exposed reefer containers, low productivity	Demurrage, truck queues, disorganized line-up, cargo loss
Mining	Haul road, stockpile drainage, yard, outbound access	Unplanned downtime, production off pace
Highways	Low point closed, alternative route	Closure, concession SLA, missed ETA
Railways	Rail head covered, low-lying stretch	Downtime hour, cascading network effect
Power Transmission and Distribution	Access to switching point, substation yard	Restoration time, reliability indicators
Civil construction	Foundation front, earthworks, concrete pours	Day of delay, schedule change order, material loss

Ports and terminals: the same rain costs differently in bulk, container, and berth

Start with the outcome. A terminal that knows on Tuesday that its gate access has a high probability of flooding Thursday morning reschedules cargo intake, notifies the trucking company, and keeps the yard sequence intact. A terminal that finds out on Thursday absorbs the entire queue, receives cargo out of sequence, and loses sync between the berth and dispatch.

UNCTAD found that 70% of the world's ports confirm impact from climate events. A risk that hits seven out of every ten ports on the planet is already a condition of operation.

MarineTraffic estimates the cost of waiting time from poor planning at congested ports at US$18 billion a year. Brazilian ports rack up US$2.3 billion in demurrage in a single year.

Treating a port as a single block hides the fact that the same rain hits three operations in different ways.

Bulk: rain soaks the cargo before it ever closes the terminal. Ore with a moisture content higher than specified loses loading productivity and can have the batch's quality disputed in the lab report, on top of entire days without working the stockpile. The queue of vessels grows the same way it does for containers, but without a weather demurrage clause the cost doesn't show up on that line: it shows up as a vessel held up, a disorganized line-up, and a quality dispute with the buyer. An i4sea case that cross-referenced weather-related operational stoppage records with vessel-swap and stoppage decisions made using i4cast®, across three solid bulk export terminals, measured a 12x ROI; client names are not disclosed publicly.

Container: the yard floods and the problem goes beyond the queue. Access: the gate floods, and the queue that would have entered that window enters later. Yard: cargo arrives out of its planned sequence, and equipment productivity drops. Window: the day's line-up falls behind, and the cost migrates to demurrage and to the carrier relationship. This sequence charges without generating a single damage record, with one costly exception: refrigerated containers depend on power being connected at all times, and if the yard floods enough to compromise that connection or the crew's access to the generator, the perishable cargo loses its cold chain, and the loss becomes the cargo's, not just the schedule's.

River-influenced terminals: here the mechanism shifts from urban to river flooding, the same distinction from the previous section. The port can close for days, not hours, from currents stronger than normal in the access channel. The risk stops being just a delay: a moored vessel under a strong current can break loose or collide, and the berth itself can suffer structural damage, a cost that never shows up on any demurrage line.

Per event, the order of magnitude for containers is smaller than for other sectors, and frequency is what makes the math work: an unanticipated port event costs, on the demurrage line, between US$10,000 and US$50,000. That range doesn't cover the quality loss in bulk cargo or the structural damage at a river terminal, which weigh more per event and are rarer.

The question that changes the conversation in the boardroom changes with the type of terminal: for containers, what accumulation, over how many hours, floods the gate access to the point of suspending cargo intake. For bulk, what's the best day to resume operations. For river terminals, what current speed requires suspending mooring. In all three, who gets notified when the probability crosses the threshold.

Mining: the haul road decides whether the mine stops with a plan or on the fly

At a port, water slows the pace. At a mine, it interrupts the entire chain, and interrupting without a plan costs more than interrupting with one.

An unanticipated stoppage in mining costs between US$150,000 and US$500,000 per event. The number is high because the chain is long: it stops the mining face, interrupts the plant's feed, eats into the next shift's maintenance window, delays shipment, and reschedules a vessel already en route. A day of a stalled mining face at a large-scale operation runs in the tens to hundreds of thousands of dollars.

The points that flood first are known to the crew:

The haul road connecting the pit to the crusher.
The drainage serving the stockpile.
The storage yard.
The access road the fleet uses to move material out.

What's usually missing is the trigger with a timestamp. Picture persistent rain saturating the haul road's drainage overnight: at the start of the morning shift, the road is impassable for off-road trucks, the operation finds out in the field, reroutes the fleet on the fly, and burns through the afternoon's maintenance window. With 24 hours of warning about that specific point, the same rain turns into drainage cleared ahead of time, a planned reroute, and a safe, scheduled stoppage.

Torrential rain also raises the risk to containment structures, and that increased risk was recorded for tailings dams in 2024. The gain from anticipation lies in seeing the accumulation that requires inspection and protocol ahead of time, and logging that decision. Predicting a structural failure falls outside the scope.

Calling this force majeure hands over to the weather a decision that belonged to the operation, and closes the door on learning from it.

Highways: the kilometer that closes has a probable hour

At a mine, the asset is concentrated in one area. In linear transport, it stretches across the entire network, and the pain changes shape: water covers a low point and the road closes.

Highway concessionaires already have a contingency plan, a field crew, and a signage protocol. The plan says what to do once the stretch closes. Anticipation says which kilometer has a high probability of closing tomorrow afternoon, which lets you position crews, notify the shipper, and recalculate the ETA while an alternative still exists.

In Rio Grande do Sul alone, in May 2024, the Federal Highway Police recorded 159 highway closures from landslides and flooding. Brazil's national transport confederation, CNT, estimated the recovery of the state's highway network at R$18.9 billion. Of what was actually allocated to highways hit by the rain in the state, the recorded figure is R$250 million. On the federal network, climate-related closures jumped from 61 in 2018 to 486 in 2022, roughly eightfold.

The cost doesn't stay on the road. CNT's Climate Resilience Survey, from November 2025, found that 70.6% of trucking companies had losses from climate events in the preceding five years. Based on i4sea's read of highway operations, the choice of an alternative route typically happens within the first four hours after the event, once the cost is already running.

On highways, anticipation changes three concrete decisions. Variable message sign: activated before the point closes. Preventive restriction: applied to known low-lying stretches. Field crew: pre-positioned outside the flooding area, not inside it.

Linear assets are hit hardest by generic forecasting. A long concession crosses several basins, and the point that matters floods within a handful of kilometers of it.

Railways: a few centimeters over the rail head lock the whole corridor

On highways, a truck reroutes. On railways, the train waits for the stretch to clear, and that's what makes the corridor more exposed than the road.

The trigger is direct: accumulated water only needs to rise a few centimeters above the rail head, the running surface, for safe circulation to stop existing and the line to be closed. A flash flood doesn't need to wash away the embankment.

In May 2024, ANTT counted 46 closures across seven railways in Rio Grande do Sul because of rain, and the year saw three shutdowns in four months across railways from climate-related events.

A case with a different cause sizes what a stopped corridor is worth: five days of blockage on the Estrada de Ferro Vitória-Minas railway added up to R$645 million in losses. That blockage was caused by a protest, not weather, and it serves as an order-of-magnitude reference for the cost of an interrupted line, whatever the cause.

A single flooded low point locks the entire corridor, backs up train formations, eats into network capacity, and cascades into a missed SLA for the shipper.

The differentiator is forecasting accumulation by track section, with a probable hour. The section with poor drainage that floods every year at the same kilometer is already known to the Operations Control Center. What it usually lacks is the accumulation forecast for that kilometer over the next 36 hours, with the threshold that typically covers the rail head and a speed-restriction recommendation before the water rises.

Power transmission and distribution: the water reaches the substation before the crew

On the road, water blocks passage. On the grid, it blackouts entire neighborhoods and still keeps the crew from reaching the site to reconnect it.

Distribution and transmission utilities already live with continuity indicators on the table and a formal contingency plan. Anticipation delivers the gain in access: knowing which bases and routes will be compromised lets you pre-position crews outside the flooding area, ahead of the peak, and choose the switching window with the information in hand. An unanticipated event in the sector costs around US$40,000 per day.

ANEEL's Technical Note No. 90/2024 records that the October 11, 2024 storm in São Paulo, with strong winds, lightning, and torrential rain, left about 3.1 million Enel SP customers without power. It was a combined wind-and-rain event, and what it measures is the scale an outage can reach, not an isolated urban flooding case.

Picture a feeder serving a low-lying area that trips offline during the storm. The crew is dispatched and takes hours to arrive because the access to the switching point flooded along with it. The restoration clock starts running at the flooded access point, before the crew ever reaches the switching point.

Heavy civil construction: the rain already stops the work front, the record that backs the schedule is missing

In several sectors, the threshold still needs to be written down. On a job site, the field team already knows rain stops concrete pours, earthworks, and lifting operations, and the surprise happens anyway.

The cost of the delay is well known: a day stalled on large job sites costs between US$50,000 and US$150,000.

Picture a concrete pour scheduled for Thursday morning. The night before's rain floods the foundation area and saturates the access road for the mixer truck, the work front stops, the ready-mix already ordered turns into a loss or a rebooking, the crane sits idle, and the sequence of following fronts slips. With 48 hours of warning, the same rain turns into a rescheduled front, protected materials, and crews reassigned to covered tasks.

There's a second gain that job sites underuse and legal teams value. When the team logs every decision to stop, protect, or reschedule with the forecast, the time, who made the call, and the action taken, the project ends up with documented evidence to support a schedule change-order claim, contest a contractual penalty, and file an insurance claim. Without that record, the same rain turns into a dispute over whose version of events is right between the builder and the client.

The rain crosses the threshold and the work front stops. What anticipation eliminates is improvisation: the stoppage stops being a race against the clock and becomes a rescheduled calendar entry instead.

The return on investment from anticipating the hazard

No system stops the rain. What changes is the cost of dealing with it. Two cases show the size of that gap, each with its scope stated openly.

Metropolitan rail operation. The operation received 6 hours of advance warning about a high probability of flooding on the stretch and activated a speed-reduction protocol before the water rose, instead of reacting once the rail was already covered. The operator's name is not disclosed publicly.

Solid bulk export terminals. Cross-referencing weather-related operational stoppage records with vessel-swap and stoppage decisions made using i4cast®, the analysis of three terminals measured a 12x ROI. Client names are not disclosed publicly.

The logic behind both cases has independent academic validation. An NBER working paper found that improved hurricane forecast accuracy in the United States between 2007 and 2020 saved around US$5 billion per hurricane in damages and emergency spending, a 19% reduction in total cost. The hurricane kept happening. What changed was the quality of the preparation.

It's the same principle behind extreme rain, on a smaller scale and with far greater frequency: the rain will fall regardless; what varies is how much you pay for not having prepared for it.

The method: three steps to apply starting tomorrow

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

1. List the five points that flood first: write down each one's threshold. Pick five: gate access, haul road, low-lying kilometer, substation yard, job site front. For each one, log what accumulation and over how many hours typically crosses the threshold that stops the operation, and the estimated cost of one hour stopped there. Use the history your team already carries in their heads. If the number doesn't exist, that's the first gap to close, and it's internal to the operation.

2. Turn the threshold into a decision: define the window, the owner, and the action. A threshold with no owner never becomes a decision. Define who receives the alert, on which channel, how far in advance, and what they do when they receive it. "Accumulation above X mm in 6 hours, forecast for Thursday morning" needs to become "So-and-so suspends cargo intake Wednesday at 10am." Treat the lead time probabilistically: accumulation typically opens a window of 24 to 72 hours, a severe storm 1 to 6 hours, and rain already in motion 15 minutes to 2 hours.

3. Log the forecast, the action, and the outcome: a simple record with the date, the predicted risk, the decision made, and what actually happened. Within three months you'll have your own hit rate, know which thresholds are miscalibrated, and have evidence for contract negotiations, schedule claims, insurance filings, and board conversations. That trail is what separates an operation that learns from one that repeats the same cost every rainy season.

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

Where this becomes a daily decision

A rain threshold written into a procedure only protects the operation when the team gets the alert in time, on the channel it already uses, with the recommended action attached.

The AI Climate Agent delivers the team a ready-made decision before the event: which point, which window, which action. It already knows your operation's thresholds, protocols, and history, reaches the channel your team already uses, WhatsApp, Teams, or email, and logs the source of the data behind every response it gives.

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

If you'd rather start with a diagnostic, we map the climate hazards on your assets and show you what the system will see, today and going forward, at your critical points, with real data.

→ Request a free climate exposure diagnostic by asset

The warning hours your operation had, and what it did with them, is what decides the cost of the next extreme rain event.

Sources

COBRADE (Brazil's Disaster Classification and Coding system): intense rain is an event in the Meteorological group; river flooding, flash flooding, and urban flooding are events in the Hydrological group. COBRADE, Goiás Military Fire Department.
Cemaden: definitions of river and urban flooding. River flooding is the gradual overflow of a watercourse from rain accumulated across the basin. Urban flooding is urban drainage capacity being overwhelmed. Cemaden, Inundação.
Mexico, October 2025: the Cazones River rose more than 3 meters above alert level and flooded Poza Rica (Veracruz), kilometers from the point of heaviest rain; 37 deaths in Veracruz, 83 total across five states. El Informador; UDG TV.
Spain, DANA of October 29, 2024: 772 mm in 24 hours at Turís (Valencia); dry channels concentrated the water and carried it into urban areas within 2-3 hours, a flash-flood mechanism; 237 deaths total. AEMET; MITECO.
Brazil, Juiz de Fora and Ubá (MG), February 2026: 209.4 mm in a few hours, monthly accumulation of 589.6 mm; 65 deaths, 3,000 displaced; the Paraibuna River overflowed. O Tempo; Correio Braziliense.
Brazil, Porto Velho (RO), May 2026: more than five hours of rain flooded historically vulnerable points, a recurring drainage problem, with no river involved. Rondônia Dinâmica.
Chile, August 2026: frontal system with overflow of the Mataquito, Cautín, Imperial, Bueno, and Rahue rivers; 1,142 people displaced, 2 deaths in Panguipulli. Cooperativa.cl.
Climate disasters cost Latin America US$11.6 billion in 2024, with US$1.5 billion insured (13%); the RS flood is the flood event with the largest insured loss ever recorded in Brazil, US$1 billion in claims paid. Swiss Re, sigma NatCat 2025. That figure measures the insured loss, not the total damage.
Total damage from the May 2024 RS floods: R$88.9 billion (69% productive sector, 21% social sectors, 8% infrastructure, 1.8% environment). Damage and Loss Assessment (DaLA methodology), World Bank, IDB, and CEPAL, Nov/2024. World Bank; CEPAL.
Floods, landslides, and convective storms caused US$260 million in losses in Brazil in Q1 2026; the Colombia flood was the region's costliest event, US$2.2 billion. Aon, via Revista Cobertura.
R$184 billion in climate-related losses in Brazil between 2022 and 2024, 91% with no insurance protection; extreme rain and flooding as the most frequent events. Climate Events and Insurance Radar for Brazil, CNseg/EY, Nov/2025.
70% of the world's ports confirm impact from climate events. UNCTAD.
US$18 billion a year in waiting-time costs from poor planning at congested ports; US$2.3 billion in demurrage at Brazilian ports in a single year. MarineTraffic.
159 highway closures in RS in May 2024 from landslides and flooding; recovery of the state's highway network estimated at R$18.9 billion; R$250 million actually allocated; climate-related closures on federal highways up from 61 (2018) to 486 (2022). Brazil's Federal Highway Police (PRF); National Transport Confederation (CNT).
70.6% of trucking companies with losses from climate events over five years. Climate Resilience Survey, CNT, Nov/2025.
46 closures across seven railways in RS in May 2024 due to rain; three shutdowns in four months that year. ANTT / Revista Ferroviária.
R$645 million in losses from a five-day blockage on the Estrada de Ferro Vitória-Minas railway (protest-caused blockage, not weather-related; used only as an order-of-magnitude reference).
The October 11, 2024 storm in São Paulo left about 3.1 million Enel SP customers without power. ANEEL, Technical Note No. 90/2024.
NBER: improved hurricane forecast accuracy between 2007 and 2020 saved around US$5 billion per hurricane (a 19% reduction). Molina, R. and Rudik, I., "The Social Value of Hurricane Forecasts," NBER Working Paper 32548 (2024). nber.org/papers/w32548.