clm-76deae4754d7f3f2
quantitative
Two steel doors, each 210 metres wide, lie in dry docks on the banks of the Nieuwe Waterweg at Hoek van Holland
Permanent project revision
Revision 2026-11-06-deltares-enw-additions · effective 2026-11-06
Current revision. This is the snapshot used by the live entity page.
Added four sources (Deltares 2024 remaining-life study, ENW 2019 background report, ENW 2024 winter high-water evaluation, Waterforum on the Delta Commissioner, November 2026): ENW's estimate of the water level the 2023 closure prevented, the arms' design basis of 15 storm closures and the earliest end of life in 2040. Nine new claims.
sha256:1c20e6286e3e35d1e31b4583702957b71bb501f506c08d984db56e63620fd5a7
replace /snapshot/projectFingerprint — Updated projectFingerprint to the current sourced project record.replace /snapshot/claims — Updated claims to the current sourced project record.replace /snapshot/sources — Updated sources to the current sourced project record.The text below is copied verbatim from the published record. Each claim names the source dossiers that carry it; caveats stay attached to the claim they qualify.
clm-76deae4754d7f3f2
quantitative
Two steel doors, each 210 metres wide, lie in dry docks on the banks of the Nieuwe Waterweg at Hoek van Holland
clm-d47a6969d47997b2
quantitative
a forecasting system called BOS decides to close when the predicted level threatens to reach 3 metres above NAP at Rotterdam or 2.90 metres at Dordrecht
clm-abae159afa4221ac
descriptive
because the designers' failure analysis concluded that manual control would undermine the barrier's reliability
clm-840c4d911ea4d635
descriptive
A team of Rijkswaterstaat experts stands by and can intervene.
clm-b0fab00436b966c9
quantitative
A second system, BESW, carries out BOS's commands; it was replaced in 2012, and in August 2026 the government's ICT review board called its state unacceptable for a safety-critical structure.
clm-0b1c7c2b69a03fbb
quantitative
Rijkswaterstaat counts one closure against a real storm since the barrier was built: 21 December 2023
clm-a6b291d68c604fd1
quantitative
when the forecast reached 3.10 m above NAP
clm-f31a28e397ab56eb
quantitative
the barrier shut around 22:00
clm-0af2f6c0f260c989
quantitative
the national flood-protection expertise network ENW estimates that the water at Rotterdam peaked at about NAP +2.4 m instead of about +3.0 m
clm-ce75c13bcccf7fe6
quantitative
Storm closures in 2007 and 2018 ran at a closing level lowered to 2.60 m to test the systems
clm-b7d6ae96ec094e4c
quantitative
every September the barrier makes a test closure, the 28th in 2025
clm-771994a647e7dc96
quantitative
a 2006 study found both Europoort barriers failing a requirement of at most one failure in a thousand closures
clm-631b9fdadd497c2f
quantitative
the legal requirement today is one failure in 100
clm-2b2f6aae6b607ebf
quantitative
In August 2026 the ICT review board counted almost 100 known defects in BESW, at least four manual interventions in 17 closures and three failed repair projects since 2012.
clm-84f64dc1e26fa5af
descriptive
Rijkswaterstaat replies that the barrier meets the legal requirement.
clm-e200e3aaadba25bf
quantitative
The decision layer, built with formal methods thirty years ago, is not where the 2026 review found the trouble; the execution layer replaced in 2012 is.
clm-19fbffc5b628f032
quantitative
Rijkswaterstaat aims to replace it before the 2030 storm season and calls the chance of missing that date real
clm-645e58f896e3591c
descriptive
a human team, which has had to step in, stays on duty
clm-fc1b0b0abb06fa59
quantitative
The barrier was being closed from 20:15 and was shut around 22:00
clm-336e4551bf47a74f
quantitative
For the first time, the barrier's own criterion, an expected water level of 3.00 m at Rotterdam, had been reached
clm-6ef16a0e85d3f2ae
descriptive
All six of Rijkswaterstaat's storm surge barriers closed that night, the Maeslant barrier automatically for the first time
clm-9de89c8ff72cfd51
quantitative
The Maeslant and Hartel barriers reopened at 04:45 the next morning
clm-b879008533d459ee
quantitative
with the barrier shut, the water at Rotterdam reached about NAP +2.4 m, against about NAP +3.0 m had it stayed open
clm-961042b8efe80af1
quantitative
the quays on the Noordereiland in Rotterdam lie at NAP +2.4 m at their lowest point
clm-55be3628dfa35d89
descriptive
The Maeslant barrier was not part of the original Delta Plan
clm-b147061792b9a96c
descriptive
Unlike the other Delta Works, it was not built by Rijkswaterstaat; a design competition was held, and the Bouwkombinatie Maeslantkering won it
clm-a195d00053807247
quantitative
Construction began in 1991, and six years later Queen Beatrix officially commissioned the barrier
clm-f238d5bfb7ad5913
quantitative
Each of its two doors is 210 metres wide and 22 metres high
clm-283593ba8d5cd919
descriptive
Rijkswaterstaat says no water barrier in the world has larger moving parts
clm-df00cf589404171c
quantitative
The barrier had to close at 3 metres above sea level, but no more than once or twice every ten years
clm-74058323ae3ebb06
descriptive
an unnecessary closure costs millions of guilders in restricted shipping, while a closed barrier can also flood the city through the Rhine if its water cannot flow out to sea
clm-2f0a13ec257ae3b0
descriptive
a careful failure analysis concluded that manual control would undermine that reliability
clm-a1544d8ec7759d97
quantitative
a normal human being has a failure probability of one in a thousand for a complex task like deciding when to close the barrier and then closing it
clm-750f8ef0efaa19cd
quantitative
Failing to close when needed was to be less likely than one in ten thousand, and failing to open when asked less likely than one in a hundred thousand
clm-78e84ee84a4d90f5
quantitative
BOS was classified at the highest Safety Integrity Level under IEC 61508
clm-dc2655d204b906a6
descriptive
was developed by CMG Den Haag, commissioned by Rijkswaterstaat
clm-6144f1c8a77392af
descriptive
the hydrological knowledge and model-based water-level predictions supplied by the Waterloopkundig Laboratorium
clm-2c1f9c4eac0f6c90
descriptive
weather forecasts, water-level measurements and the results of flood simulations
clm-471bbd1a9c9995bb
quantitative
the computer makes a calculation for the coming 24 hours every ten minutes
clm-c9f4876a24c96d11
quantitative
Development took three years and about 25 person-years, producing 20,000 lines of formal specification and 450,000 lines of C++
clm-28513f5d8e0d296c
quantitative
200,000 lines formed the operational system and 250,000 the simulators, test systems and supporting software
clm-a3024c689b4f4b0e
descriptive
Model checking with Promela and Spin found significant errors and omissions in early designs
clm-5692e71641c0c4cc
descriptive
not a single line of program code was completely proved correct
clm-eeeb348dcabdb5d7
quantitative
BOS was completed in October 1998, on time and within budget
clm-f650b3617bf9892b
descriptive
On BOS's command, a second system called BESW operates the barrier
clm-0502b6d13dd51b3a
descriptive
the closure starts by flooding the dry docks in which the walls lie, so that they float
clm-9e148957dd8d3684
quantitative
At the 2025 test, sailing out, closing and sinking the arms took just over two hours
clm-760a18dd9b75cf2a
descriptive
just dropping the structure into the river would send big waves into Rotterdam
clm-dde102264e63a9ab
descriptive
Rijkswaterstaat's storm surge warning service reported that the Hartel and Maeslant barriers had closed during a storm surge for the first time
clm-99dde53f52453dbd
quantitative
the closure on the evening of 8 November was the work of a computer
clm-71fb707638d87046
quantitative
On 3 January 2018 the barrier closed again, for a test, and would normally not have closed because the water was not high enough
clm-469424379792f82c
quantitative
as in 2007, the closing level had been lowered to 2.60 metres above NAP
clm-44dd132aecd54fdb
quantitative
a water level of 3 metres above NAP had never occurred at Rotterdam during a storm
clm-9e03d7280c0c49e8
quantitative
Storms in 2002, 2013 and 2014 nearly closed it, but the official closing level was just not reached
clm-52ea7ed2ef49316c
quantitative
In 2007 and 2018 the barrier also closed during storms, but under a temporarily lowered closing level in order to test the systems
clm-f53bc68dba304b75
quantitative
since it was built, the barrier has closed because of a storm only once, in December 2023
clm-cd2909767cecc3d0
quantitative
If seven years pass without a storm closure, Rijkswaterstaat lowers the official closing level once to 2.6 m at Rotterdam, instead of 3 m, to test the barrier for real
clm-7c38f5e87350349c
descriptive
every September the Hartel and Maeslant barriers make a test closure
clm-04bbc21c762f6516
quantitative
the 28th, on 13 September 2025, went smoothly
clm-9a37ae509cf8b447
quantitative
In a 2006 study Rijkswaterstaat found that both the Maeslant and Hartel barriers failed their technical reliability requirement
clm-cedb95055247f567
quantitative
the barriers fail to close when they should at most once in a thousand closures
clm-5f2f2c7cd6230a00
quantitative
the probability of not closing at up to 1 in 100 per closure demand
clm-abebcaf9c82a1cd1
descriptive
a realistic and achievable value
clm-340f2e7532c0c21e
quantitative
Today the legal requirement is a failure probability on closing of 1 in 100
clm-1e3d9e838062584a
quantitative
after the renovation, which started in 2009 and was delivered in 2012, it turned out in 2014 that its reliability could not be sufficiently demonstrated
clm-6c31d596a4487985
quantitative
the barrier would have a control system of demonstrable reliability before the 2021/2022 storm season at the latest
clm-6900b6720c6185c4
quantitative
mistakes were made and things were left undone in the 2012 replacement of BESW, which significantly raised the failure probability
clm-fb293284d13f0fcd
quantitative
the software contains almost 100 known defects that can lead to serious failure
clm-95632840db5f111c
quantitative
three projects after 2012 to repair the faults and omissions in BESW all failed
clm-1d9842771c6b174c
descriptive
BESW fails much more often in reality than Rijkswaterstaat's models predict
clm-4e2a04e7e5c04743
descriptive
the current state is unacceptable as the control of a safety-critical object like the Maeslant barrier
clm-5c90897f1258422f
descriptive
formal methods were already used thirty years ago to develop BOS, the system that commands BESW
clm-6d6491e92354160f
descriptive
The need for improvement does not mean the barrier is not working properly now
clm-1acc8bf6d58ebaaf
descriptive
concludes that the barrier works according to the legal safety requirements
clm-afdb6dd8fd60ba51
quantitative
endorses working to the IEC 61508 standard and will adapt the project accordingly
clm-657c5390d6577f68
descriptive
When high water is expected, a team of Rijkswaterstaat experts is always present to watch the automatic process and intervene if necessary
clm-a39f1cb9f20831bf
descriptive
it happens automatically, but if needed the team can switch to manual operation
clm-bd5fb1b80017a5f8
descriptive
In an emergency the barrier can be operated locally without BOS or BESW
clm-8efdea47983879ae
quantitative
in 17 closures there was manual intervention in the control system at least four times
clm-d83a45c18a9dfd04
quantitative
the annual September test closures had each been carried out successfully, including a planned manual closure in 2016
clm-301c8196584ea21a
quantitative
The new BESW is to last until at least 2045, and Rijkswaterstaat will own the source code
clm-5c72d87bd376e929
quantitative
before 2030, because support for the hardware then ends, and calls the chance of not succeeding real
clm-499bb92bea7c3516
descriptive
a formal specification of BESW in which all behaviour, including manual and testing modes, is specified and all fault situations are taken into account
clm-8dc759190f6fdb72
descriptive
the motivation was that the barrier's software control is up for replacement
clm-bee51fa091f217dd
descriptive
the precise formal model is confidential for reasons of national security
clm-499480813f155602
quantitative
the truss arms were originally designed on the assumption of 15 storm closures in total over the barrier's life, which runs to 2097
clm-9e259706f4dd59ff
quantitative
Without measures, the 15th closure would come in 2091, 2076 or 2065, depending on the KNMI'23 sea-level scenario
clm-020212c6259d1c16
quantitative
on current knowledge the barrier reaches the end of its life at the earliest in 2040, because of too great a risk of buckling in the tube piles of the combi-walls along the abutments
clm-e8542bb16e2666c8
quantitative
the barrier will be due for replacement around 2070, the first of the Dutch Delta Works
clm-be61351083531590
quantitative
From 60 cm of sea-level rise the current storm surge barriers will close more often to protect the hinterland
clm-0e693af841f40a86
quantitative
the ministry expects to replace the barrier between 2070 and 2100, a period that coincides with about 70 cm of sea-level rise
clm-a752b7f0922982ef
quantitative
The barrier was designed for at most 50 centimetres of sea-level rise
clm-128b718a322408c6
descriptive
because its failure probability is defined per closure request, more closures also mean more failures
This layer records what is known about each source, what is still missing, and the exact snapshot this page is rendering. A missing date, archive or hash stays visible as missing.
Awaiting operator and domain-expert review; publication of this revision implies neither.
sha256:1c20e6286e3e35d1e31b4583702957b71bb501f506c08d984db56e63620fd5a7
src-65db1b142fce4e27The operator's page on the barrier: built 1991-1997 at Hoek van Holland; two doors 210 m wide, 22 m high and 15 m deep; closing level 3 m above NAP at Rotterdam or 2.9 m at Dordrecht; closes fully automatically, with an expert team present to monitor and intervene; annual test closure every September; a verification closure at a lowered level after seven years without a storm closure.
src-a742723df11b7b07The 28th test closure of the Maeslant and Hartel barriers on 13 September 2025 went smoothly; sailing out, closing and sinking the arms took just over two hours.
src-5fc9008976a3624fLive blog of 21-22 December 2023: all six storm surge barriers closed, the Maeslant barrier automatically for the first time; the first time its criterion of an expected 3.00 m at Rotterdam was reached; the 2007 and 2018 storm closures ran at a temporarily lowered level to test the systems; reopening at 04:45.
src-55b089c8cc309f97Interview with the barrier manager: the team was alerted 24 hours before an expected 2.60 m, and the forecast then showed the water would reach 3.10 m above NAP. The article gives both 20:15 and 22:15 as times.
src-bd77893529b056aaRijkswaterstaat's storm surge report on the north-westerly storm of 8-9 November 2007: the Hartel and Maeslant barriers were closed during a storm surge for the first time.
src-558708a8fac3d548The barrier closed automatically because of high water for the first time; closing from 20:15, shut around 22:00.
src-11c825d021b903edStorm of 3 January 2018: the barrier closed for a test and would normally not have closed; storms in 2002, 2013 and 2014 nearly closed it.
src-c3fa74ef7aa77a7dAll five big storm surge barriers closed on 3 January 2018; the Maeslant closing level had been lowered to 2.60 m, as in 2007; a Rijkswaterstaat spokesperson says 3 m had never occurred at Rotterdam during a storm.
src-c10c1956d1834d30Letter to parliament: after the 2009-2012 renovation of the control system its reliability could not be demonstrated (2014); a redesign is to deliver a demonstrably reliable system before the 2021/2022 storm season; 3 January 2018 closure in storm conditions; annual test closures always successful, including a planned manual closure in 2016.
src-ca6d3f24bccce7c2Hinterland study: both movable parts of the Europoortkering failed their technical reliability requirement of at most one failure to close in a thousand closures; the latest failure analysis put the probability of not closing at up to 1 in 100, which an external review called realistic and achievable.
src-947e470313bb369bAdvice of the government's ICT review board on the project to replace BESW: the state of BESW has been unacceptable since the 2012 replacement; almost 100 known defects; at least four manual interventions in 17 closures; three failed repair projects; BESW fails more often than Rijkswaterstaat's models predict; one storm closure to date; legal requirement 1:100; replacement aimed for before the 2030 storm season.
clm-d47a6969d47997b2clm-b0fab00436b966c9clm-0b1c7c2b69a03fbbclm-631b9fdadd497c2fclm-2b2f6aae6b607ebfclm-e200e3aaadba25bfclm-19fbffc5b628f032clm-645e58f896e3591cclm-f650b3617bf9892bclm-0502b6d13dd51b3aclm-f53bc68dba304b75clm-340f2e7532c0c21eclm-6900b6720c6185c4clm-fb293284d13f0fcdclm-95632840db5f111cclm-1d9842771c6b174cclm-4e2a04e7e5c04743clm-5c90897f1258422fclm-bd5fb1b80017a5f8clm-8efdea47983879aeclm-301c8196584ea21asrc-cbd2addc0dbf8fabRijkswaterstaat's memo to the minister on the board's advice: improvement is needed but the barrier works according to the legal safety requirements; it rejects the board's 'false certainty' characterisation but adopts IEC 61508; hardware support ends in 2030 and missing that date is a real risk.
src-869e582e10a3cc93Letter to parliament on the Sea Level Rise Knowledge Programme: replacement of the Maeslant and Oosterschelde barriers in 2070-2100, driven by end of service life; from 60 cm of sea-level rise the barriers will close more often.
src-6108dd3774730e68The flood museum's account: not part of the original Delta Plan; a design competition won by Bouwkombinatie Maeslantkering; brief to close at 3 m but no more than once or twice every ten years; construction from 1991, commissioned six years later by Queen Beatrix; cost nearly half a billion euros.
src-65c8efcb099993caOn the 2023 test closure, quoting the operations spokesman: the structure is sunk slowly because dropping it would send big waves into Rotterdam; cost €450 million; the lowered level for 2007 and 2018 given as 2.65 m.
src-b89d0dd9cc87d31bAuthor preprint of the journal article in Formal Methods in System Design (2001), by researchers from the University of Twente and CMG: why the decision was automated, the software failure-probability requirements, SIL 4, code size, completion in October 1998.
src-d21665ba0120c0dcFM'99 conference paper by the developers: three years and about 25 man-years; 20,000 lines of formal specification and 450,000 lines of C++; Promela and Spin found significant errors in early designs; no line of code was completely proved correct.
src-b37b333886a31dd2Paper co-written with Rijkswaterstaat staff on lessons from the formal methods used for BOS: BOS closes the barrier based on weather forecasts, water-level measurements and flood-simulation results; expertise in the original methods has declined.
src-bc2221e4cf28c8c8Formal specification of BESW in mCRL2 by researchers from Eindhoven University of Technology and Rijkswaterstaat: all behaviour including manual and testing modes and all fault situations; verification limited to a restricted set of water levels and not all pump and valve failure combinations; the model is confidential.
src-4185f94afa6280f6The peer-reviewed version of the same work in the FMICS 2025 proceedings, Lecture Notes in Computer Science volume 16040.
src-458a2eeaeb96b4eaWater-sector magazine interview with Rijkswaterstaat staff after the November 2007 closure: BOS calls the team at a forecast of 2.60 m and closes at 3 m; the closure was the work of a computer; the team can switch to manual operation.
src-66a51d9f8f0a9941IT trade press visit to the barrier: every ten minutes the computer makes a calculation for the coming 24 hours.
src-c33a5771670cadf5Study commissioned by Rijkswaterstaat on the barrier's remaining life: design life 100 years, to 2097; the truss arms were designed for 15 storm closures in total, which would be reached in 2091, 2076 or 2065 under the KNMI'23 scenarios; earliest end of life 2040 because of buckling risk in the combi-wall tube piles along the abutments, otherwise 2065 from fatigue of the arms.
src-f7013fac9f6155aeThe Expertise Network for Flood Protection's evaluation of the 2023/24 winter high water: the barrier closed automatically for the first time in its history; with the closure the water at Rotterdam reached about NAP +2.4 m, against about NAP +3.0 m had it stayed open; the lowest quays on the Noordereiland lie at NAP +2.4 m.
src-a49278dcaebdd0abBackground report to the ENW's 2019 advice on accelerated sea-level rise: the barrier was designed for at most 50 cm of sea-level rise; at one metre it would close on average three times a year, more than it is expected to bear structurally; because the failure probability is defined per closure request, more closures mean more failures.
src-8274511f3d89dd21Trade press report of Delta Commissioner Verdaas in parliament: the Maeslant barrier is the first Delta Work due for replacement, around 2070, unless the Rhine-Meuse mouth is arranged differently.