Independent. Built from Canal & River Trust's own published notices, with no affiliation to or endorsement by the Trust.
The archive / Method
Method and limits
How every figure here is produced, what it counts, and what it cannot tell you.
Where the figures come from
Every closure figure is read from Canal & River Trust's own published stoppage notices, collected nightly and archived. Only full navigation closures are counted — the thing that stops a boat.
Daily snapshots began on 18 September 2026; the archive has been observed on 5 days with 0 gaps. Notices before May 2025 are not in CRT's system at all: it migrated on 1 May 2025 and completed notices were not carried over.
Which closures are counted
Only full navigation closures — the thing that stops a boat. 18 notices that CRT cancelled are in no figure, and neither are 9 whose reopening date precedes their start, which cannot be right. A closure that has not started is scheduled: it is in no figure, and listed under Coming up. A notice that restricts passage rather than closing it — an overnight closure, a timed window — is marked restricted and kept out of every total. An open-ended closure is measured to today, never to a date nobody published.
Days are counted from 1 May 2025, when CRT's current system went live and this archive begins. A closure running from before then, such as Springs Branch, shows the years before the archive on its own notice page and is excluded from the totals: a total that mixes them is not about the period it claims to describe. Where several notices cover the same waterway at once, the days merge into unioned intervals for any figure of disruption, while the notice count is kept separately — five notices for one blocked flight are five notices and one shut canal.
Share of the season
The headline figure on each canal page is the share of the cruising season — 15 March to 31 October — on which at least one closure was in force somewhere on that canal. Its denominator is the calendar.
It carries a bias and we would rather say so than hide it: a closure somewhere on a long canal is likelier than on a short one, so this measures what a boater met rather than how well a canal is run. It does not mean the canal was impassable.
How rates are calculated
Closures per kilometre and per 100 locks use CRT's ArcGIS length and asset layers. Their licence is stated three incompatible ways — data.gov.uk lists them as Other (Non-Commercial), the portal metadata says CC-BY-SA, and an Ordnance Survey End User Licence is referenced — so they are published under the strictest reading: non-commercial use, attribution to Canal & River Trust, and no warranty. The denominators behind them are in the data exports, so any rate can be re-derived rather than taken on trust.
A rate ranks how well a canal is run. The season share above states what a boater met. They are not the same claim and should not be read as one, so the canal pages lead with the rate and keep the season share beneath it. A rate is shown only where the denominator is at least 10km; below that it says more about the denominator than the canal.
What this cannot tell you
- Whether the network is getting less reliable. Only 3 months allow a like-for-like year-on-year comparison, and the 2025 baseline may be under-recorded while CRT's system bedded in.
- Anything about closures that ended before 1 May 2025.
- Whether a route or ring is passable: other navigation authorities are not in this archive.
- How many canals have never been shut. The only source of waterway names is the notices themselves, so a canal that has never closed is invisible here.