Tomorrow, on November 1st, the Maxi Edmond de Rothschild will set sail in the legendary Route du Rhum, its first major race and transatlantic crossing (Gitana team)
Saturday, October 31, 2026
Friday, October 30, 2026
Closing the sea: lessons from ancient Mediterranean blockades
Naval battle at the harbor of Syracuse, illustration circa 1890 (artist unknown / public domain)From CIMSEC by Dr. Stephen DeCasien
Nearly a century ago, Rear Admiral William A. Moffett, U.S.N., wrote in the foreword to Arthur M.
Shepard’s seminal work, Sea Power in Ancient History, that “Readers of this work cannot but be impressed by the similarity between the problems of sea power in ancient times and problems of sea power which are engaging the attention of maritime nations today.”
Practitioners of naval power have long looked to the past for warnings, lessons, and insights that might inform the challenges of the present and future.
This post follows in that tradition by considering what antiquity might offer for understanding one of the most enduring applications of naval power, the blockade.
The British naval theorist Julian Corbett distinguished between commercial blockade, intended to restrict an enemy’s maritime trade, and naval blockade, directed against an enemy’s fleet.
The latter could itself serve different purposes, either keeping an enemy fleet confined to port or watching it closely while encouraging it to put to sea, where it could be brought to battle.
Although Corbett developed these categories from the experience of more recent naval warfare, their basic purposes would have been readily recognizable in antiquity.
Ancient fleets likewise sought to restrict maritime commerce and supplies, contain hostile naval forces, or compel an adversary’s fleet to come out and fight.
The relevance of these forms of blockade is hardly confined to the past.
At the time of writing, the United States is enforcing a blockade against Iran, restricting maritime traffic to and from Iranian ports as part of the ongoing conflict.
At the same time, across the Indo-Pacific, the possibility of a Chinese blockade of Taiwan remains a central concern in discussions of a potential cross-Strait conflict.
Chinese military planning considers blockade operations as one means of isolating and coercing Taiwan, alongside the considerably more demanding possibility of an amphibious invasion.
Despite enormous differences in political circumstances, technologies, and scale, these situations underscore a common problem.
Establishing control of the sea is one thing, but maintaining that control long enough to achieve a political or military objective is another.
There are, of course, many examples of naval blockade in the ancient Mediterranean, but four cases highlight particularly well the different challenges such operations presented: Samos in 440–439 BCE, Syracuse in 413 BCE, Rhodes in 305–304 BCE, and Roman operations around Lilybaeum and Drepana in 250–249 BCE.
Each exposes a different problem associated with blockade, including maintaining forces on station, preserving or contesting command of the sea, restricting or sustaining access to supplies, and protecting the forces involved.
Samos, 440–439 BCE: The Challenge of Maintaining a Blockade
The Athenian campaign against Samos in 440–439 BCE provides a clear example of both the effectiveness and fragility of naval blockade.
After an earlier Athenian intervention on the island, a group of Samian exiles returned, overthrew the Athenian-backed government, and revolted.
Athens responded by sending sixty ships under the command of Pericles and other generals.
After defeating a Samian fleet of seventy vessels near Tragia and receiving additional ships from Athens, Chios, and Lesbos, the Athenians landed troops and surrounded Samos with three walls while simultaneously blockading the city “from the sea at the same time” (?κ θαλ?σσης ?μα).
The blockade quickly revealed a problem that remains fundamental to such operations, as forces maintaining a blockade may also be required to respond to threats elsewhere.
Reports reached Pericles that a Phoenician fleet was approaching to assist the Samians.
He therefore took sixty ships from the force stationed around Samos and sailed toward Caunus and Caria to meet the anticipated threat.
His departure weakened the blockade at precisely the moment the defenders needed.
The Samians exploited the opportunity.
They attacked the inadequately protected Athenian naval station, destroyed the guard ships, and defeated the vessels sent against them.
For fourteen days they “controlled the sea in their area” (τ?ς θαλ?σσης τ?ς καθ? ?αυτο?ς ?κρ?τησαν) and, as Thucydides puts it, “brought in and carried out whatever they wished” (?σεκομ?σαντο κα? ?ξεκομ?σαντο ? ?βο?λοντο).
When Pericles returned, however, the Athenians reestablished the blockade by land and sea.
Further reinforcements arrived from Athens, Chios, and Lesbos, and after a nine-month siege the Samians surrendered, demolished their walls, surrendered their fleet and hostages, and agreed to repay the costs of the war.
For fourteen days, the diversion of part of the Athenian fleet was enough to reopen Samian access to the sea, demonstrating the importance of persistence in maintaining a blockade.
For those under blockade, even a temporary weakening of the forces on station could provide an opportunity to restore maritime access.
Syracuse, 413 BCE: The Reversal of a Blockade
When the Athenians arrived at Syracuse in 415 BCE, seeking to conquer the city and extend their power into Sicily, their overwhelming superiority at sea was one of the expedition’s greatest advantages.
Over the course of the campaign, however, they gradually lost that advantage and ultimately found themselves trapped in the Great Harbor.
By 414 BCE, the campaign had developed into an effort to isolate Syracuse by land and sea.
The Athenians occupied Plemmyrium at the entrance to the Great Harbor, using it to protect their ships and store equipment, while their army attempted to complete a wall of circumvallation around the city.
Thucydides emphasizes how far the Athenian works had progressed by the time Spartan support for Syracuse arrived under the command of Gylippus.
Much of the wall toward the Great Harbor was already complete, while materials had been prepared to extend the line toward the opposite coast.
Gylippus’ intervention changed the strategic situation.
Syracusan counter-walls prevented completion of the Athenian circumvallation, while the capture of Plemmyrium deprived the expedition of its principal position at the mouth of the Great Harbor and cost it substantial stores and naval equipment.
The Syracusans also refused to remain blockaded within the Great Harbor, continuing to contest Athenian control of its waters.
In 413 BCE, they modified their triremes for its confined waters, strengthening and shortening their prows for direct, head-on, prow-to-prow attacks (?ντιπρ?ροις), rather than the maneuvering attacks (?κ περ?πλου) for which the more lightly constructed Athenian prows were designed.
After defeating the Athenians at sea, the Syracusans closed the mouth of the Great Harbor with merchant vessels, warships, and other crafts moored across the entrance.
The fleet that had arrived with overwhelming maritime superiority now had to fight its way out of the harbor it had once helped control.
In the final naval battle, the Athenians attempted to break the Syracusan blockade, attacking the barrier across the harbor entrance, but the Syracusan and allied fleet prevented their escape.
The surviving Athenians abandoned their ships and attempted to retreat overland, ending in the destruction of the expedition.
Syracuse demonstrates that a blockade depends on continued control of the surrounding maritime space.
As Athenian control weakened through setbacks ashore, the loss of key positions, Syracusan tactical adaptation, and defeat at sea, the Athenians went from attempting to cut off Syracuse to being trapped themselves.
For a force under blockade, contesting control rather than simply accepting isolation can fundamentally alter the strategic situation.
Rhodes, 305–304 BCE: The Limits of a Blockade
Few commanders in antiquity possessed a reputation for naval siege warfare comparable to Demetrius Poliorcetes, yet his enormous effort against Rhodes in 305–304 BCE failed to isolate the city completely.
Maritime commerce was fundamental to Rhodian wealth and strategic power.
Diodorus Siculus describes Rhodes as possessing considerable naval strength (?σχ?ουσα ναυτικα?ς δυν?μεσι) and notes its particularly close economic relationship with Ptolemy I of Egypt.
When Antigonus and Demetrius attempted to force Rhodes away from its relationship with Ptolemy, separating the city from outside support became a central part of the campaign.
Demetrius arrived with a formidable fleet.
He attacked the Rhodian harbor defenses from the sea, deployed ship-mounted artillery and siege machinery, and later shifted much of his effort ashore.
Yet his fleet could not make the waters around Rhodes impermeable.
Rhodian warships continued to operate, and outside powers repeatedly sent assistance.
Ptolemy, Cassander, and Lysimachus supplied the city, with Diodorus describing major shipments of grain and other provisions.
Their importance was reflected in the honors Rhodes later granted its benefactors.
Every successful entry weakened the coercive purpose of Demetrius’ effort to isolate the city.
So long as Rhodes retained sufficient maritime access to receive supplies and outside assistance, time did not necessarily favor the attacker.
After roughly a year of fighting, Demetrius accepted a settlement.
Rhodes preserved its autonomy and its relationship with Ptolemy, while agreeing to support Antigonus except in conflicts against Ptolemy.
Rhodes illustrates the difficulty of maintaining a fully effective blockade.
A blockade does not need to stop all traffic to exert pressure, nor does every successful passage render it ineffective.
Its effectiveness ultimately depends on whether enough traffic is restricted to achieve its strategic purpose, while even limited access may be sufficient for the defender to sustain resistance.
Lilybaeum and Drepana, 250–249 BCE: The Costs of Maintaining a Blockade
By 250 BCE, during the First Punic War, Roman successes in Sicily had narrowed the Carthaginian position to a handful of strongholds in the west, but the effort to blockade Lilybaeum exposed the Roman fleet to risks that would ultimately threaten the operation itself.
Polybius notes that the Romans believed its capture would open the way for attacks against North Africa, while Carthage resolved to hold it at virtually any cost.
The Romans therefore invested Lilybaeum by land and sea.
The harbor proved exceptionally difficult to blockade effectively.
Carthage exploited local knowledge, favorable winds, and skilled seamanship to penetrate the Roman blockade.
A force under Hannibal, son of Hamilcar, carried approximately 10,000 troops and supplies into Lilybaeum despite the Roman presence.
Hannibal “the Rhodian” later repeatedly sailed between Carthaginian territory and Lilybaeum, running the Roman blockade until the Romans captured a sufficiently fast Carthaginian vessel to intercept him.
These episodes demonstrated the difficulty of transforming naval superiority into complete control of a contested area.
The blockade also imposed costs on the Romans.
Combat, siege operations, and disease depleted the fleet’s manpower, requiring 10,000 replacement sailors.
In 249 BCE, the Roman consul Publius Claudius Pulcher used the reconstituted fleet to strike the nearby Carthaginian naval base at Drepana, apparently hoping to surprise the Carthaginian commander Adherbal and eliminate the threat to Roman blockade operations around Lilybaeum.
Instead, Adherbal brought his ships into open water.
The Romans, caught close to the coast with less maneuverable ships and inexperienced replacement crews, suffered a devastating defeat.
Polybius reports that ninety-three Roman vessels were captured while Claudius escaped with about thirty.
Carthaginian forces subsequently returned to the vicinity of Lilybaeum and attacked Roman ships there.
Lilybaeum and Drepana reveal the sustained costs of blockade.
Maintaining pressure exposes the forces involved to attrition, crew exhaustion, enemy action, and competing operational demands.
As these pressures accumulate, they can create opportunities for the opposing force to challenge the blockade and reshape the balance at sea.
Ancient Blockades and the Problem of Sea Power
Ancient and modern blockades operate with radically different technologies.
Samos, Syracuse, Rhodes, and Lilybaeum offer no tactical prescriptions for fleets operating with submarines, aircraft, missiles, satellites, or unmanned systems.
Their relevance lies instead in the strategic problems they expose.
Taken together, the four cases suggest that blockade should be understood as a continuing contest rather than a condition established simply by achieving maritime superiority.
Samos demonstrates the demands of persistence, as even the temporary removal of part of the blockading force created an opportunity to reopen access to the sea.
Syracuse shows how control can be reversed, as the Athenians went from attempting to isolate the city to being trapped within the Great Harbor themselves.
Rhodes demonstrates that a blockade need not be broken completely to fail; sufficient access may allow a defender to continue resisting.
Lilybaeum and Drepana show how the accumulating costs and vulnerabilities of maintaining a blockade can create opportunities for it to be challenged.
These cases raise several enduring strategic questions about blockades.
How difficult is a blockade to impose, sustain, resist, or break?
This post follows in that tradition by considering what antiquity might offer for understanding one of the most enduring applications of naval power, the blockade.
The British naval theorist Julian Corbett distinguished between commercial blockade, intended to restrict an enemy’s maritime trade, and naval blockade, directed against an enemy’s fleet.
The latter could itself serve different purposes, either keeping an enemy fleet confined to port or watching it closely while encouraging it to put to sea, where it could be brought to battle.
Although Corbett developed these categories from the experience of more recent naval warfare, their basic purposes would have been readily recognizable in antiquity.
Ancient fleets likewise sought to restrict maritime commerce and supplies, contain hostile naval forces, or compel an adversary’s fleet to come out and fight.
The relevance of these forms of blockade is hardly confined to the past.
At the time of writing, the United States is enforcing a blockade against Iran, restricting maritime traffic to and from Iranian ports as part of the ongoing conflict.
At the same time, across the Indo-Pacific, the possibility of a Chinese blockade of Taiwan remains a central concern in discussions of a potential cross-Strait conflict.
Chinese military planning considers blockade operations as one means of isolating and coercing Taiwan, alongside the considerably more demanding possibility of an amphibious invasion.
Despite enormous differences in political circumstances, technologies, and scale, these situations underscore a common problem.
Establishing control of the sea is one thing, but maintaining that control long enough to achieve a political or military objective is another.
There are, of course, many examples of naval blockade in the ancient Mediterranean, but four cases highlight particularly well the different challenges such operations presented: Samos in 440–439 BCE, Syracuse in 413 BCE, Rhodes in 305–304 BCE, and Roman operations around Lilybaeum and Drepana in 250–249 BCE.
Each exposes a different problem associated with blockade, including maintaining forces on station, preserving or contesting command of the sea, restricting or sustaining access to supplies, and protecting the forces involved.
Samos, 440–439 BCE: The Challenge of Maintaining a Blockade
The Athenian campaign against Samos in 440–439 BCE provides a clear example of both the effectiveness and fragility of naval blockade.
After an earlier Athenian intervention on the island, a group of Samian exiles returned, overthrew the Athenian-backed government, and revolted.
Athens responded by sending sixty ships under the command of Pericles and other generals.
After defeating a Samian fleet of seventy vessels near Tragia and receiving additional ships from Athens, Chios, and Lesbos, the Athenians landed troops and surrounded Samos with three walls while simultaneously blockading the city “from the sea at the same time” (?κ θαλ?σσης ?μα).
The blockade quickly revealed a problem that remains fundamental to such operations, as forces maintaining a blockade may also be required to respond to threats elsewhere.
Reports reached Pericles that a Phoenician fleet was approaching to assist the Samians.
He therefore took sixty ships from the force stationed around Samos and sailed toward Caunus and Caria to meet the anticipated threat.
His departure weakened the blockade at precisely the moment the defenders needed.
The Samians exploited the opportunity.
They attacked the inadequately protected Athenian naval station, destroyed the guard ships, and defeated the vessels sent against them.
For fourteen days they “controlled the sea in their area” (τ?ς θαλ?σσης τ?ς καθ? ?αυτο?ς ?κρ?τησαν) and, as Thucydides puts it, “brought in and carried out whatever they wished” (?σεκομ?σαντο κα? ?ξεκομ?σαντο ? ?βο?λοντο).
When Pericles returned, however, the Athenians reestablished the blockade by land and sea.
Further reinforcements arrived from Athens, Chios, and Lesbos, and after a nine-month siege the Samians surrendered, demolished their walls, surrendered their fleet and hostages, and agreed to repay the costs of the war.
For fourteen days, the diversion of part of the Athenian fleet was enough to reopen Samian access to the sea, demonstrating the importance of persistence in maintaining a blockade.
For those under blockade, even a temporary weakening of the forces on station could provide an opportunity to restore maritime access.
Syracuse, 413 BCE: The Reversal of a Blockade
When the Athenians arrived at Syracuse in 415 BCE, seeking to conquer the city and extend their power into Sicily, their overwhelming superiority at sea was one of the expedition’s greatest advantages.
Over the course of the campaign, however, they gradually lost that advantage and ultimately found themselves trapped in the Great Harbor.
By 414 BCE, the campaign had developed into an effort to isolate Syracuse by land and sea.
The Athenians occupied Plemmyrium at the entrance to the Great Harbor, using it to protect their ships and store equipment, while their army attempted to complete a wall of circumvallation around the city.
Thucydides emphasizes how far the Athenian works had progressed by the time Spartan support for Syracuse arrived under the command of Gylippus.
Much of the wall toward the Great Harbor was already complete, while materials had been prepared to extend the line toward the opposite coast.
Gylippus’ intervention changed the strategic situation.
Syracusan counter-walls prevented completion of the Athenian circumvallation, while the capture of Plemmyrium deprived the expedition of its principal position at the mouth of the Great Harbor and cost it substantial stores and naval equipment.
The Syracusans also refused to remain blockaded within the Great Harbor, continuing to contest Athenian control of its waters.
In 413 BCE, they modified their triremes for its confined waters, strengthening and shortening their prows for direct, head-on, prow-to-prow attacks (?ντιπρ?ροις), rather than the maneuvering attacks (?κ περ?πλου) for which the more lightly constructed Athenian prows were designed.
After defeating the Athenians at sea, the Syracusans closed the mouth of the Great Harbor with merchant vessels, warships, and other crafts moored across the entrance.
The fleet that had arrived with overwhelming maritime superiority now had to fight its way out of the harbor it had once helped control.
In the final naval battle, the Athenians attempted to break the Syracusan blockade, attacking the barrier across the harbor entrance, but the Syracusan and allied fleet prevented their escape.
The surviving Athenians abandoned their ships and attempted to retreat overland, ending in the destruction of the expedition.
Syracuse demonstrates that a blockade depends on continued control of the surrounding maritime space.
As Athenian control weakened through setbacks ashore, the loss of key positions, Syracusan tactical adaptation, and defeat at sea, the Athenians went from attempting to cut off Syracuse to being trapped themselves.
For a force under blockade, contesting control rather than simply accepting isolation can fundamentally alter the strategic situation.
Rhodes, 305–304 BCE: The Limits of a Blockade
Few commanders in antiquity possessed a reputation for naval siege warfare comparable to Demetrius Poliorcetes, yet his enormous effort against Rhodes in 305–304 BCE failed to isolate the city completely.
Maritime commerce was fundamental to Rhodian wealth and strategic power.
Diodorus Siculus describes Rhodes as possessing considerable naval strength (?σχ?ουσα ναυτικα?ς δυν?μεσι) and notes its particularly close economic relationship with Ptolemy I of Egypt.
When Antigonus and Demetrius attempted to force Rhodes away from its relationship with Ptolemy, separating the city from outside support became a central part of the campaign.
Demetrius arrived with a formidable fleet.
He attacked the Rhodian harbor defenses from the sea, deployed ship-mounted artillery and siege machinery, and later shifted much of his effort ashore.
Yet his fleet could not make the waters around Rhodes impermeable.
Rhodian warships continued to operate, and outside powers repeatedly sent assistance.
Ptolemy, Cassander, and Lysimachus supplied the city, with Diodorus describing major shipments of grain and other provisions.
Their importance was reflected in the honors Rhodes later granted its benefactors.
Every successful entry weakened the coercive purpose of Demetrius’ effort to isolate the city.
So long as Rhodes retained sufficient maritime access to receive supplies and outside assistance, time did not necessarily favor the attacker.
After roughly a year of fighting, Demetrius accepted a settlement.
Rhodes preserved its autonomy and its relationship with Ptolemy, while agreeing to support Antigonus except in conflicts against Ptolemy.
Rhodes illustrates the difficulty of maintaining a fully effective blockade.
A blockade does not need to stop all traffic to exert pressure, nor does every successful passage render it ineffective.
Its effectiveness ultimately depends on whether enough traffic is restricted to achieve its strategic purpose, while even limited access may be sufficient for the defender to sustain resistance.
Lilybaeum and Drepana, 250–249 BCE: The Costs of Maintaining a Blockade
By 250 BCE, during the First Punic War, Roman successes in Sicily had narrowed the Carthaginian position to a handful of strongholds in the west, but the effort to blockade Lilybaeum exposed the Roman fleet to risks that would ultimately threaten the operation itself.
Polybius notes that the Romans believed its capture would open the way for attacks against North Africa, while Carthage resolved to hold it at virtually any cost.
The Romans therefore invested Lilybaeum by land and sea.
The harbor proved exceptionally difficult to blockade effectively.
Carthage exploited local knowledge, favorable winds, and skilled seamanship to penetrate the Roman blockade.
A force under Hannibal, son of Hamilcar, carried approximately 10,000 troops and supplies into Lilybaeum despite the Roman presence.
Hannibal “the Rhodian” later repeatedly sailed between Carthaginian territory and Lilybaeum, running the Roman blockade until the Romans captured a sufficiently fast Carthaginian vessel to intercept him.
These episodes demonstrated the difficulty of transforming naval superiority into complete control of a contested area.
The blockade also imposed costs on the Romans.
Combat, siege operations, and disease depleted the fleet’s manpower, requiring 10,000 replacement sailors.
In 249 BCE, the Roman consul Publius Claudius Pulcher used the reconstituted fleet to strike the nearby Carthaginian naval base at Drepana, apparently hoping to surprise the Carthaginian commander Adherbal and eliminate the threat to Roman blockade operations around Lilybaeum.
Instead, Adherbal brought his ships into open water.
The Romans, caught close to the coast with less maneuverable ships and inexperienced replacement crews, suffered a devastating defeat.
Polybius reports that ninety-three Roman vessels were captured while Claudius escaped with about thirty.
Carthaginian forces subsequently returned to the vicinity of Lilybaeum and attacked Roman ships there.
Lilybaeum and Drepana reveal the sustained costs of blockade.
Maintaining pressure exposes the forces involved to attrition, crew exhaustion, enemy action, and competing operational demands.
As these pressures accumulate, they can create opportunities for the opposing force to challenge the blockade and reshape the balance at sea.
Ancient Blockades and the Problem of Sea Power
Ancient and modern blockades operate with radically different technologies.
Samos, Syracuse, Rhodes, and Lilybaeum offer no tactical prescriptions for fleets operating with submarines, aircraft, missiles, satellites, or unmanned systems.
Their relevance lies instead in the strategic problems they expose.
Taken together, the four cases suggest that blockade should be understood as a continuing contest rather than a condition established simply by achieving maritime superiority.
Samos demonstrates the demands of persistence, as even the temporary removal of part of the blockading force created an opportunity to reopen access to the sea.
Syracuse shows how control can be reversed, as the Athenians went from attempting to isolate the city to being trapped within the Great Harbor themselves.
Rhodes demonstrates that a blockade need not be broken completely to fail; sufficient access may allow a defender to continue resisting.
Lilybaeum and Drepana show how the accumulating costs and vulnerabilities of maintaining a blockade can create opportunities for it to be challenged.
These cases raise several enduring strategic questions about blockades.
How difficult is a blockade to impose, sustain, resist, or break?
How much force does it require, and for how long?
How complete must it be to achieve its political or military objectives?
How much traffic can continue to pass through before its effectiveness is undermined?
What outside powers might intervene or provide support?
And what forces, bases, or geographic features might strengthen, weaken, or threaten a blockade?
The central question is therefore not simply whether a blockade can be imposed, but whether it can be maintained long enough to achieve its purpose in the face of resistance, changing circumstances, and the costs of sustaining it.
That broader lesson returns us to Moffett’s observation nearly a century ago.
The instruments of sea power have changed beyond anything an ancient commander could have imagined.
The underlying strategic problem has not changed.
Control of the sea is only valuable insofar as it can be translated into lasting political or military results.
The central question is therefore not simply whether a blockade can be imposed, but whether it can be maintained long enough to achieve its purpose in the face of resistance, changing circumstances, and the costs of sustaining it.
That broader lesson returns us to Moffett’s observation nearly a century ago.
The instruments of sea power have changed beyond anything an ancient commander could have imagined.
The underlying strategic problem has not changed.
Control of the sea is only valuable insofar as it can be translated into lasting political or military results.
Thursday, October 29, 2026
How AI is transforming hurricane forecasts
A satellite image provided by the National Oceanic and Atmospheric Administration (NOAA) shows Hurricane Isaias approaching the Gulf Coast of Alabama and Florida on Friday, Oct. 9 2026.NOAA/NOAA
From NPR by Jessica Meszaros
Hurricane forecasters are increasingly using artificial intelligence to make predictions — and they say the technology is rapidly transforming the field.
When the first hurricane of the Atlantic season formed this month, an AI model beat out traditional models at predicting where it would go.
In early October, Hurricane Isaias made landfall near Destin, Fla., as a Category 2 storm.
Conventional forecasting models initially predicted it would head further west, toward southeastern Louisiana.
But Google DeepMind's artificial intelligence model called it right out of the gate: the hurricane was going to hit the western Florida Panhandle.
"Google DeepMind, in particular, has really proven itself the last few years as the gold standard model," says Michael Lowry, a longtime meteorologist and hurricane expert in Miami, who has previously advised both the Federal Emergency Management Agency and the National Hurricane Center (NHC).
"That model has been, by far and away, the best performing computer model that we've ever seen. Period."
But Google DeepMind's artificial intelligence model called it right out of the gate: the hurricane was going to hit the western Florida Panhandle.
"Google DeepMind, in particular, has really proven itself the last few years as the gold standard model," says Michael Lowry, a longtime meteorologist and hurricane expert in Miami, who has previously advised both the Federal Emergency Management Agency and the National Hurricane Center (NHC).
"That model has been, by far and away, the best performing computer model that we've ever seen. Period."
A different kind of model
The National Hurricane Center first started incorporating the Google DeepMind model into its official forecasts last year.
By the end of the 2025 Atlantic hurricane season, it was already one of the best at predicting a storm's path and strength, says Brian McNoldy, a hurricane researcher for the University of Miami.
Usually, new models take years to refine after first being introduced.
"It's something we're not used to – to have a model, the new kid on the block, very quickly becoming the best model out there," McNoldy says.
"This pretty much came out of the gate in first place, and that's a pretty impressive feat."
Traditional physics-based hurricane models use real-time atmospheric data, like temperature, humidity, pressure and wind shear, as measured by satellites and aircraft.
Traditional physics-based hurricane models use real-time atmospheric data, like temperature, humidity, pressure and wind shear, as measured by satellites and aircraft.
Then they use complicated equations to predict what path a storm might travel and how strong it could get.
An AI model operates very differently: it relies more on history.
An AI model operates very differently: it relies more on history.
It identifies patterns by analyzing thousands of past storms along with atmospheric data.
"The AI models don't really know anything about the atmosphere other than the current picture," Lowry says.
"The AI models don't really know anything about the atmosphere other than the current picture," Lowry says.
"Beyond that, it's not leaning on the governing equations of the atmosphere, the physics, to make predictions."
Most of the people designing and developing these models aren't even meteorologists, Lowry notes: "They're software engineers, and they're developing neural networks."
A "neural network" is a set of computer algorithms modeled on the human brain and trained to find patterns in massive amounts of data.
As a result, the models can seem like a "black box," Lowry says: "We don't know why they might get a forecast right. We also don't know why they might get a forecast wrong."
It can be unsettling for meteorologists to depend on a system they don't totally understand – especially during storms when millions of people might be relying on their forecasts.
"It makes us a little uncomfortable, because we like to know why you're getting the right answer," McNoldy says.
Most of the people designing and developing these models aren't even meteorologists, Lowry notes: "They're software engineers, and they're developing neural networks."
A "neural network" is a set of computer algorithms modeled on the human brain and trained to find patterns in massive amounts of data.
As a result, the models can seem like a "black box," Lowry says: "We don't know why they might get a forecast right. We also don't know why they might get a forecast wrong."
It can be unsettling for meteorologists to depend on a system they don't totally understand – especially during storms when millions of people might be relying on their forecasts.
"It makes us a little uncomfortable, because we like to know why you're getting the right answer," McNoldy says.
"And I feel like we don't – or at least I don't know why."
Not all AI models have performed well, though.
Not all AI models have performed well, though.
There are lots of unproven weather forecasting AI models out there, Lowry says.
The two that have so far proven to be the most accurate, despite some pitfalls, are the models developed by Google DeepMind and by the U.K.-based European Centre for Medium-Range Weather Forecasts (ECMWF).
The two that have so far proven to be the most accurate, despite some pitfalls, are the models developed by Google DeepMind and by the U.K.-based European Centre for Medium-Range Weather Forecasts (ECMWF).
People retrieve belongings from a collapsed boat house in the aftermath of Hurricane Isaias in Destin, Fla., Saturday, Oct. 10, 2026. An AI model from Google Deepmind beat out traditional physics-based models at predicting the path of the storm, though it didn't do as well in forecasting the storm's strength.
Gerald Herbert/AP
Relying on a "blend" of models
Forecasters at the National Hurricane Center use several proven models – both AI and traditional – when putting together their official forecasts.
"Every storm has different forecast challenges," says NHC Director Michael Brennan.
"That's why we take that consensus approach and use that blend of the best models."
And it's still humans who have to interpret the results of those models before deciding what to tell the public.
The NHC provides updates every six hours during active hurricanes.
And it's still humans who have to interpret the results of those models before deciding what to tell the public.
The NHC provides updates every six hours during active hurricanes.
The models by themselves can sometimes show drastic changes from one update to the next. Forecasters have to translate those shifting predictions into clear guidance that people can act on.
"That's one thing that the human forecaster provides a lot of, is continuity," Brennan says.
"That's one thing that the human forecaster provides a lot of, is continuity," Brennan says.
"We're using our expertise and our understanding to translate that hurricane forecast into the hazard space, and that's what people really need to pay attention to."
Scientists expect rapid advances
The Google DeepMind model still needs some work, Brennan says.
While it topped other models in tracking the path of Isaias, it didn't do as well as the NHC's blended forecast in determining the intensity, or strength, of the storm.
Brennan's team gives regular feedback to Google DeepMind to help improve the model.
Brennan's team gives regular feedback to Google DeepMind to help improve the model.
Google didn't respond to questions from NPR for this story.
The "future of forecasting has always been changing," Brennan says.
The "future of forecasting has always been changing," Brennan says.
"It's never just one thing that's driving those improvements."
It wasn't that long ago when forecasters couldn't predict how strong a storm would be even within a couple of days, he notes.
It wasn't that long ago when forecasters couldn't predict how strong a storm would be even within a couple of days, he notes.
That changed, because of advances in prediction models and better data collected by satellites and planes.
"Errors are about half of what they were, say, at the time of [Hurricane] Katrina, in terms of intensity forecasting," Brennan says.
The exciting thing about AI, says meteorologist Michael Lowry, is how quickly it's improving.
"Errors are about half of what they were, say, at the time of [Hurricane] Katrina, in terms of intensity forecasting," Brennan says.
The exciting thing about AI, says meteorologist Michael Lowry, is how quickly it's improving.
AI updates take only months or even weeks, unlike traditional models, which are updated annually at best.
"They are making advancements in skill and hurricane predictions over the course of a year that would have taken us 10 years, a decade or more, to make," Lowry says.
But even with AI coming in fast and strong, Lowry says he's not worried about artificial intelligence replacing people.
"For me, I don't feel some sort of existential crisis with AI," Lowry says.
"They are making advancements in skill and hurricane predictions over the course of a year that would have taken us 10 years, a decade or more, to make," Lowry says.
But even with AI coming in fast and strong, Lowry says he's not worried about artificial intelligence replacing people.
"For me, I don't feel some sort of existential crisis with AI," Lowry says.
"If anything, I think it is a boon for our business."
Links :
Links :
Subscribe to:
Posts (Atom)