From 6-Hour Warning to Spade in Ground: What SWFFIP Means for Planning

The press release reads as an emergency-response story: better warnings, faster action. That is the obvious headline. There is a less obvious one underneath, about what happens to planning evidence when the forecasting layer matures, especially in light of work being done by SWFFIP.

What SWFFIP delivers

SWFFIP is the umbrella programme. Two models sit underneath it, doing different jobs.

SWFHIM has been running at the Flood Forecasting Centre since April 2020 (technical overview, FFC 2022). It forecasts surface water flood impact from six hours to three days out, using a 24-member Met Office ensemble at 2.2 km resolution, the UK Centre for Ecology and Hydrology’s Grid-to-Grid hydrological model, and an HSE-built impact library that translates modelled runoff into estimated impacts at 1 km resolution. The framework underpinning it, the Natural Hazards Partnership’s Hazard Impact Framework, was peer-reviewed at development stage by Halcrow Group and JBA Consulting. The output is probabilistic: it tells you the likelihood of a given impact severity at county level, across an ensemble of rainfall scenarios.

N-FOREWARNS handles shorter lead times. It is a 1–6 hour nowcast development of the parent FOREWARNS daily forecasting tool, published in Meteorological Applications in early 2026 by Smith and colleagues at Leeds, Oxford and the FFC. The peer-reviewed evaluation shows useful skill out to roughly three hours, with performance falling off sharply beyond that. That predictability ceiling is worth holding onto.

What SWFFIP brings is mostly operational maturation: a tighter verification regime, the Met Office’s next-generation rainfall nowcast from 2027, and full national rollout by 2029.

Why this matters for planning

Forecasting and assessment look like different jobs. One looks at the next 72 hours, the other at the next century. But they draw on the same underlying inputs: rainfall datasets, NaFRA maps, impact catalogues, LiDAR. When the forecasting layer is upgraded, those inputs get used hundreds of times a day. Errors that would sit quietly in an FRA for years get found and corrected within a forecast cycle.

That has a planning shadow. The trial’s finding that “improvements could have been made on nearly 50% of occasions” is more than an emergency-response statistic. It is an evidence-quality statistic. The same datasets that make a forecast better also make an FRA better, because both lean on the same source material. By the time SWFFIP is fully operational in 2029, the data underpinning a planning-stage FRA will not look the same as the data consultants are working with today.

Three things that change for FRAs

The first is how residual risk gets described. SWFHIM does not produce a binary zone. It produces a distribution: likelihood of impact at a given location across an ensemble of rainfall scenarios. FRAs that still describe surface water risk in deterministic, zone-based language will start to look thin against that evidence base. Consultee challenges will follow.

The second is how impact gets characterised. SWFHIM’s impact library is built from the Risk of Flooding from Surface Water maps at 2 m resolution, run through the Jflow+ hydraulic model for nine effective rainfall scenarios. It categorises consequences into four impact types: danger to life, damage to buildings, disruption of key sites and infrastructure, and disruption of transport. The thresholds are explicit. Water depth above 0.3 m counts as damage to a residential property; hazard rating of 1.25 or above counts as danger to life; a road or railway flooded to 0.15 m over 10 m is classed as transport disruption. Where FRAs have leaned heavily on depth alone, expect more weight on velocity, hazard rating and receptor type. The new National Standards for SuDS and Schedule 3 are pushing in the same direction.

The third is source convergence. When the EA, the Met Office, UKCEH, the HSE and academic groups are pointing at the same map, an EA Flood Map Challenge needs more than a LiDAR overlay and a hopeful narrative. The “we used the published map” defence weakens with every iteration of NaFRA2 and SWFHIM. Any consultancy still treating mapping as the FRA, rather than the start of the FRA, will feel that.

What SWFHIM means for the Sequential Test

Surface water flooding sits inside the FRA, not the Sequential Test, under NPPF. But Mead, Yatton and Faversham, the appeals that have shaped the last two years of practice, all turned partly on how rigorously the residual-risk story was told once a site had cleared the Sequential Test.

Better forecasting and richer impact data give inspectors more granular evidence about consequences, even where the headline probability zone is unchanged. An Exception Test argument that calls residual risk “manageable” will face a higher evidentiary bar in 2029 than it does today. The September 2025 PPG update already pushed in that direction. SWFFIP gives consultees the data to enforce it.

For a planning consultant, that has lead-time implications. Sites at the marginal end of surface water risk, particularly those approaching reserved matters in the late 2020s, should be designed against the data they will be assessed by, not the data published when the application went in. Otherwise a refusal can be queued up by the calendar alone.

The forecast is sharper. The argument has not changed.

There is a temptation to read SWFFIP as a story about better emergency response and stop there. The press release encourages that reading. But forecasting is a response tool, and planning is a prevention tool. The published verification on N-FOREWARNS makes the distinction sharper: at one hour out, the nowcast captures roughly a third of observed surface water flood events; at four hours, it captures almost none. SWFHIM picks up more but generates more false alarms. SWFFIP does not replace good drainage design. It makes the consequences of bad drainage design easier to see and harder to argue away.

That cuts in two directions. Developers carrying strong SuDS strategies and well-evidenced FRAs into 2029 will find consultees better equipped to recognise the work. Those carrying weak schemes will find the evidence base has quietly tightened around them. The middle ground, the marginal site that scrapes through on a generous reading of the published map, is the ground that shrinks fastest.

Related surface water guidance for planning

Surface water is now the most widespread source of flood risk in England, and it runs through several strands of planning practice. These Unda guides cover the parts that most often affect development:

• The May 2026 Flood Map for Planning update — the new surface-water and climate-change layers, and what they change for site appraisal.
• What pluvial (surface water) flooding is — causes, risks and planning implications.
• Urban flood risk in England is now dominated by surface water — the wider context.
• The 2025–30 London surface water strategy — what it means for planning.
• The 2025 national standards for SuDS — designing surface-water drainage to the expected standard.
• The Sequential and Exception Tests and the September 2025 PPG update — how surface water feeds the policy tests.