Sixteen flood-prone homes on Clydach Terrace in Ynysybwl, Rhondda Cynon Taf, are set to be purchased and demolished by the local authority, bringing an end to years of repeated flooding, fear and uncertainty for the people who live there.
Clydach Terrace sits in a narrow, constrained valley in south Wales, where the Nant Clydach runs immediately alongside a row of terraced houses built on the natural floodplain. The physical setting leaves little room for floodwaters to dissipate during intense rainfall, concentrating risk directly against the properties.
The location has a long history of flooding, but Storm Dennis in 2020 exposed the full severity of the hazard. Natural Resources Wales’ project documentation describes flooding that was “extreme”, with “rapid onset” and “internal property flood depths above 6ft at the lowest lying houses”. Post-event surveys quantified this further, recording a “Storm Dennis Observed Depth – 1.96m”.
NRW does not present this data neutrally. The project brief states that these observations “serve to illustrate the significant danger posed by the flooding during Storm Dennis”. The risk is framed explicitly in terms of safety and danger to life, not simply property damage.
Crucially, the same document confirms that this danger is ongoing. “Without any flood changes to risk management,”NRW concludes, “Clydach Terrace will remain at high risk of flooding”. The problem was therefore well understood, formally evidenced and accepted within the flood risk system.
Managed retreat without a national policy
Despite this, there is no formal inland managed retreat policy for flood risk in England or Wales. Unlike coastal erosion, there is no established framework for relocating communities exposed to river flooding, no clear decision pathway, and no dedicated funding mechanism.
These included raised or reconstructed flood walls, channel maintenance and blockage removal, natural flood management, flood warning systems, property-level resilience measures, and the removal of people and property from the highest-risk area. The issue was not a lack of technical options. It was how those options were judged.
When flood defences fail the viability test
In NRW’s appraisal framework, options are assessed against Critical Success Factors (CSFs) – the criteria used to decide whether a flood risk management scheme is allowed to progress through the business case process. An option must meet all CSFs to remain viable; failure on any single criterion rules it out entirely.
Economic performance is central to this assessment. A viable option must be “an economically viable option to manage flood risk, with a Benefit Cost Ratio (BCR), a standard Green Book measure comparing scheme benefits to costs, greater than one”. The threshold is explicit and absolute. As the project brief explains, “An option will not be considered viable if it can meet only some or only partially meet a CSF.”
In practice, this logic narrows the field quickly. In the options appraisal tables, conclusions are often reduced to a single line: “This does not meet the CSFs.”
Raised flood defences were explored, but when the council later summarised NRW’s findings it recorded that a raised flood wall was “not an economically viable option under the UK and Welsh Government funding rules for flood risk management purposes”. The danger remained, but the flood defence system could not justify progressing a scheme.
Why the council bought flood-prone homes instead
The risk at Clydach Terrace did not disappear when flood defence options stalled. Instead, the problem moved institutions.
The council was not overriding NRW’s conclusions, but acting in a different legal capacity, using powers that sit outside the flood risk management funding system. By the time the issue reached cabinet, the framing had shifted decisively. The report refers to “the significant risk to health presented and danger to life at this locality” and acknowledges that “there is no viable proposal available to protect their homes”.
Rather than attempting to reopen the flood defence case, the council relied on its general power of competence. The report notes that the authority “has a statutory general power of competence”, allowing it “to do anything that individuals generally may do”. This is a broad enabling power, not one specific to flood risk, and the report explicitly links its use to flooding and climate impacts. It states that the power would include “protecting communities against the threat of climate change and from a high risk of flooding, through the purchasing of property”.
The practical outcome is described plainly: the council would “demolish the properties thereby removing an ongoing flooding threat to residents”.
This is managed retreat in effect, even if it is never labelled as such.
Managed retreat by default, not by design
Large-scale public buyouts of flood-prone homes are not unknown in the UK, but they are overwhelmingly associated with coastal erosion, where specific policy mechanisms exist to support them. Inland fluvial flooding has no equivalent framework.
Clydach Terrace therefore appears exceptional not because the circumstances are unique, but because the system offers so few routes to resolution. Retreat emerges only once flood risk management rules have been exhausted and local government intervenes using powers never designed as a substitute for national flood policy.
This is not an argument that retreat should be the default response to flood risk. It is an observation of how retreat emerges when existing systems cannot accommodate extreme cases. Seen in full, this is not retreat undertaken early or strategically. It is retreat that arrives late, after years of exposure, once engineering solutions fail economic tests and the human consequences become unavoidable.
No document ever states that Clydach Terrace should be abandoned. Instead, the decision emerges incrementally – through appraisal thresholds, stalled business cases and, finally, a council stepping in to remove people from harm.
The question this leaves behind is not whether other inland floodplain communities will face similar choices. It is whether they too will have to wait until managed retreat becomes the only option left.
