Insights · Evidence dashboard
Six months, announced. Day 577: completion not publicly confirmed.
The Umzimkhulu River Bridge repair on the R102 near Port Shepstone, source by source: what was promised, what the public record shows, and what it cannot see. One lane reopened on 7 December 2025; we found no public confirmation of the two-way reopening promised for 25 February 2026. Much of the story reaches the public only through press reports, which can be wrong.
How to read this page. It takes no position on who is responsible. It shows what the public record establishes, what is only reported, and what we could not see. Not public is not the same as not done. Engineering sign-offs, completion certificates, payment certificates and programmes are working records of the department, its engineers and the contractor, and are not normally published. Where this page says a record was not found, it means only that it is not in public view; the work or the record may well exist. Press reports are shown as reported, not as fact, and some may be inaccurate. The bridge repair and the R102 road works are different contracts, so they are never added together. Use the switches to watch the story thin out when only official sources are allowed.
The bridge, day by day
Status of the bridge from the 17 March 2025 closure to 15 October 2026. Hatched means no public source we found says; the bridge's status in that period is simply not visible to us.
All timeline events as a table
Every announced target date has passed
Each mark is a target: across is the day it was announced, up is the day it pointed to. Anything below the grey line was a promise about a date that has now gone by.
Targets as a table
Progress figures as a table
The money: what each number is, and is not
Dots sit on a log scale, so a R7 million package and a R775 million corridor can share a page. Nothing here is added up, on purpose.
What would prove it is finished
Lane reopening is not the same as practical completion or structural acceptance. This is the test, and what the public can see against it. Most of the proof below lives in records that are not normally published, so "not public" here is a statement about visibility, not about whether the work or the sign-off happened.
Can the public check this?
We looked for the records on the stakeholders' own websites on 15 October 2026. "Not found" means not found by us, not that a record does not exist. Still, when the picture has to be assembled from press reports, that is a finding about public communication in its own right.
In fairness. The department does publish dated statements, and its archive is a browsable, dated list. SANRAL's release is dated, and cites its gazette numbers. What is missing is project-level information: the contract, the programme, progress against it, payments, and notice of closures. In our search, the March 2026 closure and reopening plan reached the public through the press; we found no department statement on either.
Where the sources disagree
Documents that would settle this
None of these is in the research set. Each is a record the public body, its engineers or the contractor would normally hold, and most are not published as a matter of course. Their absence here says nothing about whether they exist.
Source register
Evidence class shown for each. Where a table row could not be tied to a single source, it carries a "source TBC" flag instead of a guess.
Method, reply and corrections
Compiled from the public sources above on 15 October 2026 and drafted with Claude. In the final round the department's own statements were downloaded from its archive and read in full, which replaced earlier reproductions as the source for those dates and figures. Source numbers beside each row were matched to the register by the author; where the match is an inference rather than a citation in the original notes, it is marked. Contested claims are attributed to the report that made them and are not stated as fact.
What this method cannot see. Everything here comes from what was published. That has two limits. First, published sources can be wrong: press reports in particular are snapshots, written fast, and this page has already found errors in some of them (see "Where the sources disagree"). Second, much of what proves a project is progressing properly, such as engineers' inspections and sign-offs, test results, payment certificates and site instructions, is never published. A tool, or a person, that only reads public sources cannot tell "did not happen" from "happened but was not published". This page tries to say "not found in public sources" wherever that is the honest position, and does not draw a conclusion from silence.
Right of reply. Anyone named here is invited to send a correction or response through the contact page; it will be published beside the entry it concerns.
Corrections log. 17 October 2026: wording revised after reader feedback. The first version said "no completion record" and counted "0 engineering sign-offs located", which could be read as saying those things did not happen. Engineering sign-offs and completion certificates are not normally published, so the page now says only that they are not in public view. Press-only events are now labelled as reported. No dates, figures or sources changed.
Want a dashboard that shows where each number comes from, built on your own data? Talk to us, or book a 30-minute call.