Acqua alta: how Venice lives with flooding, and what the UK takes from it

Acqua alta is the tidal flooding that covers parts of Venice several dozen times a year, and the city stays open through almost all of it. The reason is not the barrier at the lagoon mouth. It is that the buildings, the pavements and the public services above them were all designed on the assumption that the water arrives.

Between 1870 and 1949 the tide at Venice passed 1.1 metres thirty times. It has done so seventy-six times in the last nine years.

Those counts come from research reported in Nature, and that shift is the whole subject. Venice has not been hit by a run of bad luck; the baseline it was built against has moved, and everything the city does now is a response to a different set of numbers than the ones its architecture assumed. Britain is at an earlier point on the same curve, with the Environment Agency putting around 6.3 million English properties in areas at risk of flooding today and roughly 8 million by mid-century, a picture Unda has set out in its guide to flood risk under NaFRA2. What follows is how Venice measures the problem, what its barrier does and does not do, and which parts of the answer transfer to a UK site.

What is acqua alta, and when does it happen?

Acqua alta is a tidal event, not a rainfall one. It happens when a high astronomical tide coincides with weather that pushes water up the Adriatic, and it is measured in centimetres above a datum fixed at Punta della Salute in 1897. Anything above about 80 centimetres begins to show in the lowest parts of the city; anything above 140 is classed by the city as exceptional. Venice’s flooding season runs from roughly October to March. November is the worst month.

Venice publishes the levels at which specific things happen, which is the part worth borrowing. There is no single “flooded” state. There is a ladder, and each rung has a defined consequence and a defined response.

The precision matters. A city that knows Piazza San Marco starts to flood at 82 centimetres and the station forecourt at 135 can tell a shopkeeper, six hours ahead, whether to fit the boards. A warning that says only “flooding expected” cannot.

Why Venice floods: tide, wind and a moving baseline

Four things combine. The astronomical tide sets the base level. A sustained sirocco from the south-east drives water up the length of the Adriatic and piles it against the lagoon. Low atmospheric pressure lifts the sea surface further. And the Adriatic’s long, narrow shape sets up a seiche, a slow oscillation that can add tens of centimetres depending on where the tide sits in the cycle. The result is a storm surge in a basin shaped to amplify it.

None of that is new. What has changed is the level all four are added to. Venice has lost roughly 25 centimetres of relative height since the late nineteenth century, part sea level rise and part subsidence from industrial groundwater abstraction that was largely stopped by the 1970s. In a lagoon this flat, a few centimetres of extra baseline converts an occasional event into a seasonal one.

This is the same arithmetic that governs a UK site. A flood level is not a property of a place; it is a modelled output for a given annual exceedance probability, at a given date, against a stated climate change allowance. Move the allowance and the level moves with it, which is why a flood risk assessment for planning is built around a design flood level for the development’s lifetime rather than around what the site has done historically. UK Climate Projections put sea levels around England on a rising path through the century, so the allowance itself is a moving quantity. Unda’s explainers on flood return periods and on which climate change allowance applies set out how those two inputs are chosen from the published allowances. Venice has a century of measurements at one gauge to argue from.

How the MOSE barrier works

MOSE is a set of 78 hinged steel gates lying flat in concrete housings on the seabed across the three inlets that connect the Venetian Lagoon to the Adriatic. When a high tide is forecast, compressed air displaces the water inside the gates, they become buoyant, and they rotate up to form a temporary wall. The lagoon is sealed until the surge passes, then the gates are flooded and settle back into the seabed. The operator’s own description sets out the sequence in full.

1. Forecast. The city’s tide centre predicts a peak above roughly 110 centimetres, several hours ahead.
2. Decision. Operators weigh the flooding avoided against shipping disruption and the ecological cost of sealing the lagoon.
3. Raise. Compressed air empties the gates of water; buoyancy rotates them up from the seabed in about half an hour.
4. Hold. The barrier stands while the peak passes, typically three to five hours, and shipping through Malamocco stops for as long as it is up.
5. Lower. The gates are flooded again and settle back into their housings, and tidal exchange resumes.

The gates are not a permanent seawall, and the distinction is deliberate. The lagoon is a brackish system that depends on tidal exchange for its sediment, its salinity gradient and its water quality, so every closure carries an ecological cost weighed against the flooding it prevents. That makes MOSE a flood defence with an operating policy rather than a fixed line on a map.

How often does MOSE close, and what it costs

MOSE became operational in October 2020 and was raised about a hundred times in its first four years, at an operating cost of roughly €20 million over that period. The frequency has climbed since. Researchers writing in Nature recorded 31 activations in the system’s fourth operational year against fewer than twenty in earlier cycles, and reported modelling that puts closures at around 260 days a year by the end of the century if emissions are not curbed.

The winter of 2026 gave a preview. Venice’s tide centre reported thirty separate closures in the twenty-three days to 19 February, at a cost of about €6 million, with no single tide exceeding 140 centimetres. The barrier was not holding back an exceptional event. It was holding back an ordinary acqua alta season that now sits higher than the trigger.

Each closure costs more than the electricity. Commercial shipping through Malamocco stops. Water exchange between the lagoon and the sea stops with it, and in parts of the lagoon the exchange time already runs past eleven days, long enough for nutrients to accumulate and algal blooms to follow. Sediment that would otherwise settle on the salt marshes and let them grow vertically against rising seas stays outside the gates, an effect now being measured experimentally in the lagoon.

What MOSE does not do

MOSE removes the catastrophic tail. It does not remove water from the city. Below the activation threshold the gates stay down and the lower districts flood as they always have, which is most of the acqua alta Venice actually experiences in a year. It was built to stop a 1966 or a 2019, and the paving in November was never the point.

A barrier that excludes the extreme event leaves every event below its trigger to the buildings.

That is the residual flood risk problem stated in its clearest form. In UK practice it is the same question asked of any defended site: the defence has a standard of protection, and everything above it, everything behind a breach, and everything that happens while it is not deployed remains for the development itself to handle. Unda’s explainer on how residual flood risk is assessed covers how overtopping and breach are modelled in an FRA. The difference in Venice is that the same question has been answered at building scale for four hundred years.

How Venetian buildings are designed to get wet

Venetian ground floors are built to be wetted and dried rather than kept dry. Thresholds sit 10 to 30 centimetres above the paving. Floors are stone or terrazzo laid on solid substrates rather than timber on joists. Walls are solid masonry finished in lime plaster, which absorbs water and gives it back without delaminating the way gypsum board does. Sockets and consumer units sit high, well clear of anything a winter tide is likely to reach. Sleeping accommodation is upstairs, and the ground floor is for storage, circulation and trade.

Why a board is enough here

At the depths a routine acqua alta produces, typically 20 to 60 centimetres, the hydrostatic load on a doorway is modest and a slot-in board is a reasonable engineering answer. The same board against a metre of standing water would be a structural question rather than a joinery one.

• Channels stay put. Guide channels are cut into the masonry reveals and left in place all year.
• Minutes, not hours. Boards go in on a forecast and take a few minutes per opening.
• Sealed at the base. A compression seal and a tight fit in the reveal do most of the work.
• Visible by design. Nobody hides the channels, so nobody has to find them in a hurry.

None of this is unavailable in Britain. The materials and the details are set out in CIRIA’s Code of Practice for property flood resilience, and national planning practice guidance supports resistant and resilient construction where development goes ahead in a flood risk area. Unda has written separately on how property flood resilience works in the UK planning system and on what recovery actually involves once water has been inside.

What differs is the default. A UK ground floor is normally specified on the assumption that it will never be wet, so when it is, the strip-out and drying run to months rather than days. Flood resistance keeps water out up to a point; recoverability decides what the months after that point cost.

Passerelle, sirens and a city that knows its own levels

Venice’s forecasting is municipal, public and specific. The Centro Previsioni e Segnalazioni Maree publishes tide forecasts hours ahead, and warnings go out by text message, app, public display and a graded siren whose tones correspond to forecast bands. Residents and shopkeepers read the number, not a colour.

When the forecast crosses roughly a metre, the passerelle go down along routes chosen from a century of observation of where the low points are, at centimetre resolution. Above about 120 centimetres they float and are withdrawn.

The point is not the walkways. It is that a public authority holds, publishes and acts on a level-by-level map of its own ground, and the population is fluent in it. UK flood warnings tell people that flooding is expected; they rarely tell an individual property what depth to expect and when. Unda has looked at how UK forecasting and warning is organised and at where the Environment Agency’s role begins and ends. For a single site, the flood risk map by postcode is the nearest domestic equivalent of reading the ladder.

St Mark’s Basilica: why the barrier is made of glass

The narthex of St Mark’s Basilica floods at about 85 centimetres above the 1897 datum, which makes it one of the most exposed significant structures in the city. Repeated saltwater ingress does more damage than a single deep flood, because salt crystallises inside the pores of marble and brick as the water evaporates and breaks the surface apart from within. The November 2019 event, which caused over €300 million of damage across Venice, was described by the basilica’s conservators as ageing the building twenty years in a day.

The panels first held water on 7 November 2022 and stopped a 185-centimetre surge later the same month, and they are openly described as an interim measure while the paving of Piazza San Marco is raised.

The choice of glass is the interesting part. An opaque wall would have protected the fabric equally well and destroyed the setting. The transparent barrier does the hydraulic job while leaving the building visible, and it is openly a flood defence rather than a disguised one. British heritage practice tends to prefer protection that can be taken away and not seen, which often means it is not there when it is needed. Unda has covered the same tension in its work on floodable voids and raised structures, where the mitigation has to be visible in the design to be credible in the assessment.

What this means for a UK flood risk assessment

The transferable lesson is not the barrier. It is the pairing. Venice runs strategic exclusion and building-scale tolerance at the same time, and the second is what carries the events the first was never designed to catch.

The National Planning Policy Framework requires development in a flood risk area to be safe for its lifetime without increasing flood risk elsewhere, and UK practice is strong on the first half of that test and thin on the second. A scheme demonstrates that finished floor levels sit above the design flood level with freeboard, that the sequential and exception tests are satisfied, that access and egress are safe, and that runoff is attenuated. Once that test is passed there is often no further consideration of what happens if the level is exceeded. That is a defensible position while the modelled event and the real one stay close together.

The Environment Agency now puts about 4.6 million English properties in areas at risk of surface water flooding, 43 per cent more than its previous assessment, and records an 88 per cent rise in the number of properties in the highest risk band.

Those figures come from the Environment Agency’s national assessment of flood and coastal erosion risk. The distance between the design event and the real one is not fixed, which is what makes the exceedance case worth stating rather than assuming. Handling it properly is site-specific work: establishing the levels, testing what happens beyond them, and specifying what the building does when the water reaches it. If you need that assessed for a development, Unda’s consultants can prepare a flood risk assessment that sets out both the design case and what sits beyond it, and can carry out site-specific flood modelling where the national mapping is too coarse to answer the question.

Frequently asked questions