MOSE Closures: How Often the Venice Flood Gates Close

How often does MOSE close now?

The lagoon authority publishes every activation, numbered in sequence since the system entered service. The last entry on the log read in October 2026 is number 156, dated 15 May 2026, so the gates have gone up 156 times in five and a half years (Autorità per la Laguna di Venezia, 2026a).

By calendar year the counts run 13 in 2020 (from October alone), 20 in 2021, 11 in 2022, 25 in 2023, 28 in 2024, 27 in 2025, and 32 in 2026 up to the middle of May. By season, counted from autumn to spring the way the log groups them, the figures are 20, 13, 19, 32, 23 and then 49 for 2025/2026 (Autorità per la Laguna di Venezia, 2026a).

Closures are short. Across 2020 to 2023 they averaged 8.6 hours per activation and 146.3 hours a year, with a standard deviation of 6.4 hours on the first figure and 48.5 on the second (Michielotto et al., 2026, p. 850). The log measures duration from the start of lifting to the end of ballasting once the gates are down again, and notes that the hydraulic effect lasts about 25 minutes less than the published figure (Autorità per la Laguna di Venezia, 2026a).

Most activations fall in the autumn and winter storm surge season, though there are exceptions: the gates were raised in May and late August in 2023 (Michielotto et al., 2026, p. 848), and twice in mid-May 2026 (Autorità per la Laguna di Venezia, 2026a).

When does MOSE close, and why it closes below 110 cm

The operating rule is to close when a tidal peak above 1.10 m is forecast, hold the water inside the lagoon below 0.90 m, and reopen as soon as the ebbing tide drops back under 0.90 m (Melaku Canu et al., 2026, p. 15). Those heights are measured against the Zero Mareografico di Punta della Salute, the datum Venetian tide bulletins use, which now sits about 32 cm below mean sea level. At 1.10 m on that datum roughly 12 per cent of the city floods, an extent handled with footbridges and raised pavements (Michielotto et al., 2026, p. 848).

In practice the gates often go up for tides that end up short of the trigger: the official log records activations where the highest sea level measured outside reached 84, 95 or 96 cm (Autorità per la Laguna di Venezia, 2026a). The decision rests on the forecast rather than the outcome. Predicted maxima from the municipal alert system have often exceeded what was then observed in the city, which points to precautionary management rather than changed weather as the main driver of closure frequency (Michielotto et al., 2026, p. 848).

That has a cost in closure time. Closures ran on average 107.2 per cent longer than flood protection required, and an adaptive rule tested on the same four years could have cut their total number by about 30 per cent while keeping levels below the safety thresholds set for Venice, Chioggia and Burano (Michielotto et al., 2026, pp. 848, 851).

How often will MOSE close by 2050?

Two independent modelling studies have put numbers on the next twenty-five years, and they agree on the order of magnitude.

The first simulated the whole lagoon for 2020 as a control and for a set of 2050 scenarios. Against 16 acqua alta events and 357 hours of closure in the control year, 3 per cent of the year, a continuation of the observed trend gives 34 events and 917 hours (7 per cent), a moderate emission scenario 42 events and 1,097 hours (9 per cent), and a high emission scenario 54 events and 1,324 hours (11 per cent) (Ferrarin et al., 2024, p. 10).

The second counted closure days rather than hours, for the five years 2045 to 2049, and found 17 days a year under the moderate scenario and 26 under the high one, which it puts at 5 and 8 per cent of annual time (Melaku Canu et al., 2026, pp. 4, 6). The two count different things, hours of gate closure against days on which a closure happens, so the figures do not line up directly. Read together they bracket mid-century between 5 and 11 per cent of the year, against 1 to 2 per cent now.

The driver is the same in both: relative sea level at Venice rose 4.9 mm a year over 1991 to 2020, and a comparison with the Trieste record, which is considered free of vertical land movement, attributes part of that to subsidence (Ferrarin et al., 2024, p. 7).

How often will MOSE close by 2100?

For 2095 to 2099 the projections stop being incremental. The gates would close 138 days a year under a moderate emission scenario, 40 per cent of annual time, and 248 days a year under a high one, 70 per cent of the year (Melaku Canu et al., 2026, pp. 4, 6). Those runs assume the present operating rule, with sea-level forcing of plus 0.52 m by 2100 under the moderate scenario and plus 0.71 m under the high (Melaku Canu et al., 2026, p. 15).

An earlier assessment framed the question as thresholds. Closing the inlets for three weeks a year was judged unlikely before the 2040s and virtually certain before the end of this century even under a low emission scenario. Two months a year was judged unlikely before the late 2050s even under a high-end scenario, and virtually certain by the late 2080s under high emissions. Six months a year, the criterion those authors use for considering the present defence strategy inadequate, is likely before 2100 under a high emission scenario (Lionello et al., 2021, p. 2638).

None of this says the barrier stops working. It says it keeps working by being shut. The projected end-century rise in mean lagoon water level falls from 0.74 m to 0.51 m under the high scenario once the gates are operated, and from 0.55 m to 0.44 m under the moderate one, so the system reduces the rise inside the lagoon without eliminating it (Melaku Canu et al., 2026, p. 4).

What more closures would cost the lagoon

Water renewal time is the days a basin needs to replace its water, and it governs nutrient cycling, oxygen dynamics and pollutant dispersal (Melaku Canu et al., 2026, pp. 3, 11). The lagoon mean today is about 9.5 days, with under 1.5 per cent of the modelled area over 20 days (pp. 6, 7).

The projections separate what the sea does from what the gates do, because every scenario was run twice, with the barrier operated and with none at all. By the end of the century under the high emission scenario, sea-level rise on its own takes the mean renewal time to 11.9 days. Operating the barrier takes it to 17.7. The share of the lagoon needing more than 20 days goes from 1.5 to 33.9 per cent, and the share needing more than 30 days from 0.4 to 11.2 per cent (Melaku Canu et al., 2026, p. 6).

The mechanism is straightforward: a higher mean sea level means a larger volume to renew through inlets that cannot pass any more of it, and each closure suspends the exchange altogether (Melaku Canu et al., 2026, p. 11). On this measure the moderate scenario offers less comfort than elsewhere: with the barrier it produces renewal distributions comparable to the high scenario (p. 8). The authors' summary is that the system can buffer sea-level rise and cut flood exposure for Venice, though not the effects of warming, and that the protection costs substantially reduced water renewal (p. 13).

What a closure means if you are in Venice

Almost nothing, which is the point. On an ordinary day the gates lie flat on the bed of the inlets, full of water, and boats pass over them. During a closure the water in the canals simply stops rising, held below 0.90 m while the tide outside goes higher. A visitor on the ground sees no barrier and no works, and most closures are over inside a working day.

The planning consequence is about tides rather than gates. Flooding has not stopped: at the trigger level about 12 per cent of the city still goes under, which is why footbridges appear across the lowest squares (Michielotto et al., 2026, p. 848). What has changed is that the deep events which used to shut shops for a day or more are now headed off, at the price of a system in near-constant use through the autumn and winter.

The early months of 2026 showed what a dense run of closures looks like. The lagoon authority recorded 30 activations between 28 January and 20 February 2026, called the frequency exceptional, and commissioned a reanalysis of the meteorological and tide-gauge data to improve future management of the barriers, along with a simulation of where the city would have flooded had the gates stayed down (Autorità per la Laguna di Venezia, 2026b). That 24-day stretch is a preview of the closure frequency the end-century projections describe as normal.