Natural Flood Management: EA Research Deep Dive & Analysis

Natural flood management (NFM) does work, but the honest answer is more useful than a simple yes. The Environment Agency's latest evidence shows that measures which slow, store and filter water across a catchment can measurably lower flood peaks, especially for smaller and more frequent floods, while also improving water quality, habitats and carbon storage. What they cannot yet promise, on their own, is a guaranteed level of protection for a town during an extreme flood. This deep dive works through what the evidence actually says, technique by technique, and where the limits sit.

The reference point is the Environment Agency's Working with Natural Processes evidence directory, updated in February 2025 (research report FRS21232). It reviews around 800 studies covering 17 NFM measures, grouped into river and floodplain management, woodland, run-off, and coastal and estuary management. That structure gives us a clean way to assess each family of measures against the same question: how much flood risk does it actually reduce, and under what conditions?

The directory's headline finding is that during a one-in-25-year storm, a combination of NFM measures could cut flood flows by roughly half and lower peak water levels by 10 to 30 centimetres in urban areas downstream.

At Unda, we read these developments so that developers, landowners and planners do not have to. This article is the expert translation layer between the primary research and a decision on the ground. It sits alongside our shorter piece on how natural flood management is being used to combat flood risk in schemes like Calderdale and Northamptonshire, and goes further into the evidence base behind those schemes.

What the EA's 2024 evidence directory found

The 2024 evidence directory updates the Environment Agency's influential 2017 review, and it is deliberately cautious. It does not claim NFM is a universal fix. It grades the strength of evidence for each measure on a high, medium or low confidence scale, and it is candid where the science is thin or rests on modelling rather than monitored catchments. That caution is what makes it credible, and it is why the directory has become the backbone of national policy. It covers four families of measures:

• River and floodplain management — re-meandering, floodplain and wetland restoration, leaky barriers, beavers and offline storage. The largest evidence base, with more than 250 studies reviewed.
• Woodland management — catchment, cross-slope, floodplain and riparian planting to intercept rainfall, roughen the surface and improve infiltration.
• Run-off management — soil and land measures, run-off pathway features and retention in headwater and peatland systems.
• Coastal and estuary management — saltmarsh and mudflat restoration, beach nourishment, dune management, reefs and submerged aquatic vegetation.

The update looks in detail at 17 measures, three of them new since 2017: beavers, reefs and submerged aquatic vegetation. Sustainable drainage systems (SuDS) are deliberately kept out of scope, because they sit within development-scale drainage rather than catchment-scale NFM. Crucially, the report is explicit about the boundary of what NFM can promise on its own.

The directory is precise about scale, because scale is what decides whether a measure is likely to help. It uses a simple classification that is worth keeping in mind for the rest of this article:

The 2024 update also weighs the wider gains hard defences cannot deliver, valuing measures across eight benefit types, from flood risk reduction and water quality to biodiversity, carbon storage and community wellbeing.

That reframing matters for the economics. A scheme that only shaves a flood peak can look marginal on cost, while the same scheme valued for its habitat, carbon and water-quality gains often stacks up well. The report's wider list of benefits reads:

• Water and soil — reduced soil erosion and less sedimentation of lakes and rivers.
• Climate — increased carbon capture and storage.
• Water quality — cleaner water reaching watercourses.
• Habitat — rivers reconnected with species-rich floodplain wetlands and new habitat for biodiversity.
• People — more recreation and amenity space and improved wellbeing for local communities.

River and floodplain management

This is the most established family of NFM measures and the one with the clearest hydrological logic: give a river back its length, its floodplain and its roughness, and it holds water back rather than rushing it downstream. The EA holds medium confidence in the flood-risk benefit of river restoration and leaky barriers, and low to medium confidence for floodplain and wetland restoration, where most of the evidence is still modelled rather than monitored.

Following restoration works, a 25 km² catchment in the New Forest recorded a 21% reduction in its flood peak, one of the stronger monitored results in the UK evidence base.

River restoration

River restoration reinstates natural processes within a river system, usually through re-meandering, removing artificial structures and reintroducing braided channels. It lowers flood peaks through a few linked mechanisms:

• More channel length — re-meandering lengthens the river and lowers its gradient, which slows water velocities.
• Floodplain reconnection — letting water spill onto adjacent land adds storage and takes pressure off downstream areas.
• Changed sediment transport — slower flows and natural geomorphology reduce erosion, an effect the EA now rates with high confidence.

The size of the flood benefit depends on the length of restored reach relative to the whole catchment and on how well the river reconnects with its floodplain. Maintenance requirements depend on natural sediment and vegetation dynamics, and some restored channels need occasional intervention to prevent excessive sedimentation or erosion. The ecological gains are substantial, which is why river restoration is treated as a multi-functional measure rather than a single-purpose flood scheme.

Floodplain and wetland restoration

Floodplains modified for agriculture and development lose their ability to absorb excess water. Restoring them, and reconnecting them to river channels, rebuilds that capacity to attenuate flood peaks. Wetlands such as fens, reedbeds and wet woodlands act as natural sponges, storing floodwater and releasing it slowly. Their role is most pronounced in headwater catchments, where they moderate run-off before it reaches main rivers. The River Wensum in Norfolk is a worked example, where re-meandering, floodplain reconnection and wetland creation increased flood storage while improving biodiversity and filtering pollutants and sediment. The EA notes the flood evidence here is still largely modelled, so this remains an area where more monitored sites are needed.

Leaky barriers

Leaky barriers, also called leaky dams, are permeable timber or stone structures placed within watercourses to slow flow. Unlike a solid weir, they let water through while temporarily ponding it upstream, which flattens flood peaks and reduces downstream velocities. The EA is specific about where they help: they "generally show flood risk reduction for small floods in small catchments (under 10 km²) depending on design and placement". The "Slowing the Flow" project at Pickering in North Yorkshire, where hundreds of leaky barriers were installed through headwater streams, is the reference case for that finding. The flip side is that the storage they create is often too small to make much difference in higher return-period events, so they are a small-catchment, frequent-flood measure rather than a defence against extreme floods. We cover where they help and where they fall short in our explainer on leaky dams and flood risk.

Beavers as natural flood managers

Beavers (Castor fiber) are ecosystem engineers, and they are one of the three measures newly assessed in the 2024 update. Their dam networks create wetland environments that slow water, increase storage and recharge groundwater, which attenuates peak flows and lengthens the lag between rainfall and run-off. The EA reports that beaver activity "can significantly mitigate peak flows during large storms" and that beaver dams can lower flood levels even when the ground is already saturated, a notable finding given how most measures fade in wet conditions. The confidence rating is medium and growing, though understanding of their effect at larger catchment scales in the UK is still limited. The trade-off is management: localised flooding, agricultural conflict and infrastructure impacts all need handling through flow devices and buffer zones. Our article on whether beaver dams reduce flood risk sets out the evidence and the caveats in full.

Offline storage areas

Offline storage areas are purpose-built flood storage zones next to river channels, engineered with embankments or controlled spillways to take excess water during high flows. By diverting floodwater into designated storage, they cut peak discharges and downstream risk. Their effectiveness turns on volume capacity, inflow control and how well they fit into the wider flood management strategy. The Thames scheme at Oxford is the best-known example, using controlled floodplain inundation to protect the city; the recently approved Oxford Flood Alleviation Scheme shows how offline storage sits within a larger engineered project. These areas add habitat and recreation value, but they need land, which is the binding constraint in developed catchments.

None of these measures works in isolation. Their real strength shows when they are combined across a catchment: restoration and floodplain reconnection add storage, leaky barriers and beaver activity slow the headwaters, and offline storage catches what remains before it reaches the town. A distributed network keeps working even when one element reaches capacity, which is why a catchment-wide plan consistently outperforms scattered, opportunistic interventions.

Woodland management

Woodland influences flooding by intercepting rainfall, roughening the surface and improving infiltration. The effect is real but conditional, and the conditions matter more here than in any other family of measures.

Catchment woodland effects are most detectable where forest cover exceeds 15 to 20% of the area and in catchments below 100 km²; above that size, larger hydrological processes dominate and the woodland signal fades.

Catchment woodland

Catchment woodlands include all forested areas within a drainage basin. Tree canopies intercept rainfall and promote evaporation, an effect strongest in dense coniferous forest with year-round cover. Woodland soils tend to have higher organic content, which improves infiltration and storage, and the added surface roughness slows water across the landscape. The catch is that these effects diminish as flood magnitude rises: in a large flood, woodland soils saturate and interception becomes marginal against the volume of rainfall involved.

Cross-slope woodland

Cross-slope woodlands are planted perpendicular to the slope, acting as barriers to surface run-off. They intercept water before it reaches watercourses, improve soil structure through rooting, and trap sediment and nutrients. Modelling studies suggest a useful reduction in run-off velocity in small to medium catchments, though the effect depends on slope, soil and planting density. Much of the evidence here rests on models rather than monitored sites, so the directory treats it as promising but not proven.

Floodplain woodland

Floodplain woodlands sit within river floodplains and attenuate flows by increasing hydraulic roughness, encouraging sediment deposition and temporarily storing floodwater. There is a genuine trade-off to design around: densely planted floodplain woodland can locally raise flood levels by reducing conveyance, so siting and density need care. The measure attenuates moderate floods well; in extreme events its effect is less pronounced.

Riparian woodland

Riparian woodlands line riverbanks and deliver most of their value away from peak-flow reduction. Their roots bind bank soils, reducing erosion and channel widening in high flows, which maintains channel capacity and limits downstream siltation. They also buffer run-off, letting sediment and pollutants settle before water reaches the channel. Their direct effect on flood peaks is limited, but their contribution to channel stability supports the wider system.

Species choice and long-term upkeep

Two practical decisions shape whether woodland delivers. The first is species: native broadleaves with deep root systems tend to improve infiltration more effectively than shallow-rooted conifers, though conifers offer denser year-round interception. The second is management. Woodland is not a plant-and-forget measure; thinning, coppicing and controlling invasive species keep it hydrologically effective, and neglected planting can lose its benefit or, in a floodplain, impede storage. The scale of planting needed to shift flood peaks in a large catchment is also considerable, which is where land-use conflict with farming and development becomes the real constraint on woodland-led NFM.

Run-off management

Run-off management works on the land before water ever reaches a channel, by improving soil health, holding water in the landscape and lengthening the path it travels. It is the family with the most local, farm-scale evidence and the least certainty at large catchment scale.

• Soil and land management — aeration, subsoiling and reduced compaction raise infiltration; roughening the surface, disrupting tramlines and using low-pressure vehicles cut run-off from arable land.
• Grassland management — lower stocking density and strategic grazing keep vegetation cover intact, slowing water across the land and extending lag times.
• Flow pathway features — run-off attenuation features (RAFs), hedges and buffer strips store water temporarily, promote infiltration and trap sediment.
• Headwater and peatland retention — restoring vegetation and blocking gullies, grips and drains keeps water in peat systems for longer and buffers rapid run-off.

In Bishopdale, reducing grazing density across more than a quarter of the catchment produced a noticeable fall in peak run-off and a longer lag time during extreme weather.

Soil and land management

Managing run-off starts with the soil. Intensive arable practices, including root crops and late-harvested crops, worsen compaction and leave fields exposed to run-off over winter. Aeration, subsoiling and spiking relieve compaction and lift infiltration, and the evidence is clearer for grassland than for arable systems. Reducing stocking density has shown a consistent effect on slowing run-off, and converting arable land to pasture raises evapotranspiration, which lowers overall flood risk. In the Brimfield Brook catchment, subsoiling and reduced compaction on a small share of the land produced a minor reduction in flood peaks, which suggests wider uptake could deliver more.

Managing flow pathways

The second lever is the route water takes. Disconnecting and lengthening flow pathways, and adding temporary storage, slows flood peaks. RAFs hold water and release it gradually; hedges and buffer strips slow run-off, aid infiltration and trap sediment; well-designed farm tracks stop concentrated flow channels forming. These measures are most effective for smaller, more frequent floods, and their capacity is finite. In the Corney Fell catchment, two bunds produced a measurable reduction in stream flow that declined as floods grew more frequent, and at Rogerscale, hedges used as leaky dams significantly attenuated peak flows and reduced downstream water depth.

Retention in headwater and peatland systems

The third lever is holding water high in the catchment. Peatlands act as natural stores, and restoring them, by revegetating, blocking gullies and grips and obstructing artificial drainage, slows water and encourages infiltration. Pipe blocking prevents rapid drainage so water stays in the landscape longer. The empirical base is still limited, but the evidence points to restored peatlands buffering rapid run-off and contributing to flood risk reduction at catchment scale, alongside carbon and water-quality gains.

As with the river measures, run-off management pays off most when the techniques are layered across a landscape rather than tried one at a time. Better soil, a network of attenuation features and restored headwaters each slow a different part of the journey water takes from field to channel. The persistent limitation is scale: the farm-level evidence is encouraging, but proving a reliable effect on a large river remains difficult, which keeps monitoring central to any serious run-off programme.

Coastal and estuary management

On the coast, NFM works by putting natural buffers back in front of hard defences so that saltmarsh, dunes, beaches and reefs absorb wave and tidal energy before it reaches a wall or embankment. Reefs and submerged aquatic vegetation are the other two measures newly assessed in the 2024 update.

The Hesketh Out Marsh managed realignment covers 322 hectares and, alongside re-establishing priority saltmarsh, provides a 1-in-200-year standard of protection to 143 residential properties, three commercial buildings and 300 hectares of farmland.

Saltmarshes and mudflats attenuate wave energy through their vegetation, and denser, more flexible species diminish wave force most during storms, exactly when protection matters. Field measurements from the Scheldt Estuary recorded a 20% to 40% reduction in wave height over the first 12 metres of marsh, and modelling has indicated up to a 75% reduction in wave height for a 10-year return-period event where waves meet marsh vegetation. As with every other family of measures, that attenuation falls away at longer return periods, which is why coastal NFM is paired with, not substituted for, engineered defences.

Marshes are dynamic, not fixed. Wave energy during storms drives both erosion and sediment deposition, so their long-term stability depends on sediment supply and on understanding how a marsh evolves after it is created. Recent research has improved estimates of how long managed realignment schemes, where defences are deliberately breached to restore intertidal habitat, take to establish functional marsh and mudflat. The main coastal measures are:

• Beach nourishment — adding sediment restores a beach's capacity to dissipate wave energy; it needs regular topping up as natural processes redistribute the sediment.
• Sand dune management — stabilising dunes through planting and fencing creates a natural barrier and a sand reservoir; dune restoration formed part of the Medmerry managed realignment in West Sussex.
• Reefs — constructed or restored oyster and mussel beds dissipate wave action, reduce erosion and add habitat and water-quality benefits.
• Submerged vegetation and kelp — seagrass and kelp reduce water velocities, stabilise sediment and store carbon, complementing marsh and dune systems.

The Hesketh Out Marsh scheme shows the pattern at its best: 322 hectares of restored intertidal habitat that also carries a defined standard of flood protection, so the flood benefit and the environmental gain come from the same investment. Aligning these measures with the relevant shoreline management plans is now standard practice, and it is the clearest illustration of NFM's multi-benefit case anywhere in the evidence base.

Does natural flood management work in extreme floods?

This is the question that decides whether NFM protects a town or simply helps a catchment. The evidence is consistent: NFM reduces flooding well for frequent, moderate events, and its effect weakens as floods become larger and rarer. Woodland soils saturate, storage features fill, and marsh attenuation drops at longer return periods. For a genuinely extreme flood, NFM alone will rarely prevent serious flooding. The EA's own confidence ratings make the pattern plain across the best-studied family of measures.

The practical conclusion is a hybrid one: NFM is most reliable as a complement to engineered defences, adding resilience and reducing the load on hard infrastructure rather than replacing it.

There is a scale effect at work. A measure that trims a flood peak in a 10 km² headwater can be swamped in a 1,000 km² river basin, where the volumes dwarf the storage any realistic network of leaky barriers or wetlands can provide. This is why the strongest monitored results, the New Forest's 21% peak reduction or Pickering's leaky barriers, come from smaller catchments. Independent reviews reach the same view: the parliamentary POST briefing on nature-based flood and drought resilience flagged a persistent evidence gap on performance at scale and in extreme conditions. The useful question for a given site is not whether NFM works in the abstract, but how much of the flood risk it can realistically carry, and what has to sit behind it to cover the rest.

How is natural flood management funded and supported in 2026?

Policy has moved a long way since the first pilots. NFM is no longer a fringe experiment; it is being written into how national flood investment is allocated, which changes the calculation for landowners and developers weighing it up.

The Environment Agency updated its natural flood management guidance in February 2025 to encourage wider use, and new tools such as NFM benefit methods and heat maps now help build the economic case. A reformed flood and coastal funding policy takes effect from 1 April 2026, explicitly aiming to mainstream NFM and property flood resilience alongside traditional defences. The economic argument is strengthening too: research on Wildlife Trust schemes found around £10 of benefit for every £1 invested once wider gains are counted.

For anyone bringing forward development, this shift matters because it raises the profile of catchment-scale thinking in planning, and it sits close to the on-site drainage expectations that already apply. NFM operates at catchment scale; sustainable drainage operates at the development scale, and the two increasingly need to be considered together, as we explain in our piece on green SuDS and nature-based drainage in planning.

The challenges to scaling natural flood management

The EA's evidence is a strong endorsement of NFM, but scaling it into mainstream flood planning runs into real obstacles. These are the issues that have to be solved for NFM to move from promising to routine.

• Scalability — most monitored successes are in small catchments; replicating them across major river basins that affect thousands of properties needs comprehensive modelling, monitoring and coordination between many landowners.
• Performance in extreme floods — NFM reduces moderate flooding but its effect during high-magnitude events is uncertain, so it rarely stands alone against catastrophic flooding.
• Land-use and economic trade-offs — setting land aside for storage, planting or wetlands competes with farming and development, so uptake at scale needs funding, incentives and clear regulation.
• Maintenance and monitoring — leaky barriers degrade, wetlands evolve and restored channels need sediment management, so measures require long-term, adaptive upkeep to keep working.
• Integration with existing defences — much hard infrastructure is nearing the end of its design life, and the task is to weave NFM into broader strategies as a complement, not a replacement.

The challenge is no longer proving that natural flood management works; it is implementing it at scale, in the right places, and funding the long-term maintenance that keeps it effective.

Climate change sharpens all of this. The Environment Agency estimates around 6.3 million properties in England are at risk from rivers, the sea or surface water today, rising toward 8 million by mid-century, which is precisely why nature-based approaches are being taken seriously. Rising risk also feeds directly into how development is assessed, through tighter climate change allowances for flood risk assessment.

What this means for flood risk professionals

Natural flood management has earned its place in the UK's flood toolkit. The evidence supports it as an effective, multi-benefit approach that lowers flood peaks, delays run-off and strengthens landscape resilience, particularly when applied across a whole catchment and paired with engineered defences. It is not a silver bullet, and the EA is clear-eyed about where it reaches its limits, but the direction of policy and funding leaves little doubt that catchment-scale, nature-based thinking will feature more heavily in flood planning through the rest of this decade.

If you are planning a development, buying land, or assessing a site where flood risk and catchment measures interact, our chartered consultants can help you make sense of what the evidence means for your scheme. Speak to our team through our flood risk assessment service for advice grounded in the current research and policy.

Frequently asked questions