Scottish Government

Natural Flood Storage and Extreme Flood Events Final Report

2 ROLE OF FLOODPLAIN ATTENUATION IN SUSTAINABLE FLOOD MANAGEMENT

2.1 Schemes designed to encourage better floodplain utilisation for flood storage

A review of available literature on either existing or proposed flood management schemes located in rural areas has identified a number of examples in England and Wales that incorporate more natural flood storage options within their designs.

A joint statement prepared by English Nature (EN), the Environment Agency (EA), the Department for Environment, Food and Rural Affairs (Defra) and the Forestry Commission (FC) (EN/EA/Defra/FC, 2003) highlighting the potential for sustainable flood management included a number of examples of the use of more natural storage for flood management, namely the Lincoln washlands, the Severn/Vyrnwy washlands to protect riverside properties in Shrewsbury and the Aire washlands to provide some of the flood protection for Leeds.

In 2001 English Nature produced a research report, which sets out the case for washlands as a viable flood defence option (Risk & Policy Analysts Ltd, 2001). A 'washland' is defined in this report as "an area of land adjoining a river or stream that floods from the positive act of directing floodwaters onto it as part of a flood defence measure (such as a flood storage reservoir)". It should be noted that some washlands do comprise natural unmanaged areas, which have not been subject to large-scale engineering works but which nevertheless provide similar functions to those of a managed one.

The report included a short review of the approach to the economic value of such sites and highlighted the problem that funding for washlands in England and Wales usually require agreements from two very different funding streams (both divisions of Defra) - flood defence for the construction and agri-environment payments for the long term management of these features. The report recommended the development of a new washland agri-environment scheme to maximise all the potential benefits of washlands as a flood defence option. A number of case studies were also described indicating that washlands can provide an environmentally sustainable, technically feasible and economically robust method to manage flood risk. These include the River Calder washlands in Yorkshire, River Witham washlands in Lincolnshire, Melton Mowbrey washlands in Leicestershire and the River Trent washlands in Lincolnshire.

The methods for compensating the landowner for flooding land varied considerably between these washland schemes. On the Trent, the land was purchased by the River Board (now EA) in the 1930's and 1940's and has been rented out to farmers ever since on the agreement that there would be no compensation for flooding; so the farmer carries the risk. At Melton Mowbray all the affected landowners are being paid, with the level of payment being related to future agricultural losses from increased flooding. On the Witham, the EA (J.Oldfield, EA, pers. comm.) has negotiated compensation packages with the individual landowners based on a sliding scale, depending on the flood return period which causes a particular land area to flood. The compensation payment is based on the current land value, associated with its current use and the local land quality, and includes some business reorganisation costs, if appropriate. Compensation is paid in one go, rather than as and when the flooding occurs. This agreement of payment levels required a lengthy period of negotiation between the regional EA land agent and the land agents working on behalf of the landowner. Other examples of studies investigating natural flood storage include the River Severn Strategy (EA, 2002) and the Parrett Catchment Project in Somerset (FWAG, 2001).

The Curry Report on the future of farming and food in England (Cabinet Office, 2002) highlighted the role that farming could play in sustainable flood management. The report recommended that the Government should take action to encourage farm flood management schemes and, in particular, future environmental schemes should include water management as an option for support. In the same way, the Government should ensure that land management responses to flooding are eligible for funding from flood management budgets alongside more traditional methods of flood defence. The use of land management options will however require a shift to whole catchment planning, such as the Catchment Flood Management Plans (CFMPs) currently being developed by the EA. The report also suggests that the farming industry should look to embrace water management as a viable 'alternative crop'. The recommendation to explore the use agri-environment measures to achieve flood benefits was taken forward into the Defra Strategy for Sustainable Farming and Food (Defra, 2002).

2.2 Floodplain schemes in Scotland

In Scotland, the Rural Stewardship Scheme (RSS) is an Agri-environment Scheme designed to encourage farmers, crofters and Common Grazings Committees to adopt environmentally friendly practices and to maintain and enhance particular habitats and landscape features (Scottish Executive, 2004). Within the "Prescriptions for Wetland Features" section of the booklet is an option entitled "Management of Flood Plain". The aim of this option is to create and maintain a mosaic of washlands and dry lands by allowing the watercourse to overflow on to its natural floodplain.

The management payment is made where the whole floodplain is allowed to flood and the watercourse must be allowed to flow naturally with any impediments to natural flooding removed. This payment will be in addition to any other payments for the management of other specific habitats within the floodplain. Anyone applying for this payment must ensure that the interests of adjoining landowners are not prejudiced by any change in management. In reality this will generally mean that all landowners in the affected floodplain would need to apply for this option before any payments are made. The exception to this would be if a single landowner owned the whole of the affected floodplain. Data provided by the Scottish Executive indicates that the take up of the floodplain option within RSS is still very limited (Table 2-1).

Supplied by Scottish Executive. 2004 data correct as of 7 February 2005

Just 1,140ha of land is currently managed under this option on the agreement that flooding will not be prevented, with a management fee of £25/ha per year to the landowners. This area of land currently under RSS floodplain agreement represents a very small proportion (about 0.4%) of the estimated 2,950km ² of inland floodplain in Scotland (Werritty et al, 2002). Recent increases in the amount of money made available by the Scottish Executive for the RSS may encourage more landowners to apply for the floodplain option, though joint applications from landowners covering coherent floodplain areas should be more actively encouraged wherever possible. Some Non-Governmental Organisations (NGOs), such as the RSPB and EN, suggest that the payment levels for flood alleviation measures on agricultural land should be raised to more adequately compensate the farmers for loss of income or to permit a change in the land management, within the current economic climate of the farming industry (RSPB Scotland, 2004; Risk & Policy Analysts Ltd, 2001).

In Scotland, to date, the controlled use of natural floodplain storage to reduce the risk of flooding downstream is very limited and hard-engineered protection schemes (e.g. flood walls) are widespread. The Insh Marshes floodplain in Strathspey is often cited as the classic Scottish example of naturally functioning floodplain that prevents extensive flooding to properties and farming downstream, including the town of Aviemore (RSPB Scotland, 2004). The floodplain here has a very high conservation value with many national and international environmental designations. In contrast, in many rural floodplain areas there is a legacy of low level embankments along watercourses which would typically be overtopped by 1 in 2-5 year flood events and therefore exert no control during more extreme events. Other Scottish schemes using more 'natural' storage include Linlithgow and Conon Bridge using playing fields, Fraser Road in Aberdeen and Moredun in Renfrewshire.

There are examples that some Scottish Local Authorities are now actively exploring sustainable flood management and soft engineering solutions during the assessment and appraisal of options for new flood prevention schemes (e.g. Renfrewshire Council at Lochwinnoch and Moray Council at Forres and Elgin), though whether these get taken forward to the final scheme are still to be determined. Heavily engineered washlands with high embankments and large control structures, similar to the Ouse Washes in eastern England, are not common. However, one of the biggest schemes in Scotland is an area of washlands upstream of Kilmarnock, on the Irvine near Galston, and on Cessnock Water. The proposed highly embanked areas on the White Cart and its tributaries to store flood water in the rural environment and therefore reduce the flood risk in south Glasgow is a relatively new venture for flood prevention schemes in Scotland.

2.3 Opportunities for biodiversity improvements through natural floodplain storage

Shifts in flood management philosophy, to some extent driven by predictions of greater flooding potential through future climate change, combined with changes in the way the rural economy is viewed, has resulted in a growing awareness of the potential for re-introduction of natural floodplain storage to UK river systems. This change has been recognized and encouraged by the nature conservation lobby, which has welcomed the opportunity to embrace the dual benefit of flood alleviation and biodiversity gain.

Three types of washland are generally considered:

• Flood management washlands - primarily for flood management with biodiversity a lesser consideration
• Integrated washlands - flood management and biodiversity are given equal weight
• Conservation wetlands - where biodiversity is the main concern

Recent research by English Nature has investigated the potential to embrace biodiversity benefits within the flood defence management framework (Morris et al, 2004). Findings of the study suggest that where washlands are currently used for arable production (requiring infrequent flooding and fast drainage) there are more limited opportunities, with greater scope for existing grassland or woodland areas (more frequent flooding and wetter ground conditions). In the context of the test catchments for this study, many of the floodplains are grassland based, with relatively little arable or horticultural production except in the South Esk (see Table 1-1) that would support the potential for floodplain storage.

The English Nature study concludes that there is the possibility of synergistic uses, provided the fundamental conflicts between existing land uses (mainly agricultural) and future aspirations for biodiversity gains can be reconciled - a key consideration would however be the management of water levels within the flood areas to maximize benefit during non-flood event periods.

2.4 Washland flooding requirements to enhance biodiversity

Three water-related characteristics of washlands determine their vegetation habitat potential, namely duration of flooding, seasonality of flooding and soil water regime (Morris et al, 2004):

• Duration of flooding (short/medium/long). The presence of surface water following a flood event is important from a vegetation perspective because it will limit soil aeration and may also prevent plants of short stature obtaining oxygen from the atmosphere.
• Seasonality of flooding (winter only or year round). This is relevant because many plant communities are able to tolerate flooding and waterlogged soils in winter but not summer.
• Soil water regime as determined by the drainage characteristics of the soil profile. This is important because, following the recession of the flood, some washlands may drain freely, re-aerating their soils and allowing non-wetland specialist vegetation to persist. Soils that have no subsurface drainage or drainage management may only support species adapted to anoxic and prolonged wet rooting environments.

The washland classification focuses on habitats and vegetation types whose composition is largely determined by the prevailing water regime. The assumption within the classification is that the flood frequency of sites is greater than once every three years. The vegetation of sites with lower flood frequencies is unlikely to be primarily determined by the flooding regime. The only arable land use is that associated with short duration flooding in winter on soils with rapid drainage, and this would be confined to extensive arable such as cereals, possibly spring sown.

Land that floods only rarely is likely to be committed to relatively high value cropping, and there are likely to be measures in place to evacuate water quickly in order to minimise the duration of inundation and waterlogging. Thus, sites subject to medium and long duration flooding are those that are likely to experience relatively frequent flooding, of at least once every three years or so. Of course, measures which otherwise are used to evacuate flood water to avoid long duration flooding can be modified or immobilised to help create a desired washland habitat.

The three components listed above (duration, seasonality and soil water regime) were chosen to form the basis of a washland habitat classification matrix (Table 2-2) because they can be readily estimated for an existing or potential washland. They also summarise the flooding and soil water regime requirements of a habitat in a way that is clear to both flood and environmental managers. The degree to which they are determined by flood management, other sources of water such as that draining from higher land or drainage infrastructure, will be a site-specific issue. It is important to distinguish between a flood event and water level management beyond a flood event - the latter relying on stored floodwater or another source of water where one exists. It is the management of field water levels that arguably will have the greatest effect on the water related biodiversity interest of a washland.

2.5 Flood Duration and Seasonality

Table 2-2 contains the washland habitat classification matrix, which describes washlands by flood duration, seasonality of flooding and soil water regime. The rows of the matrix classify washlands according to the typical duration of flood events, namely short (less than 3 days), medium (3 days to 14 days) and long term (more than 14 days). The matrix is divided vertically into two sections that denote the seasonality of flood occurrence, namely; winter flooding only, and winter and summer flooding.

Note: Soil drainage is a function both of soil conductivity and drainage infrastructure

Rapid soil drainage = Following inundation, water table typically falls by > 30 cm in < 10 days in winter

Moderate soil drainage = Following inundation, water table typically falls by > 30 cm in < 30 days in winter

Slow soil drainage = Water table does not fall below 30 cm following an inundation event in winter until late April

Short duration of surface water: typically 3 days per event.

Medium: typically less than 2 weeks per event.

Long: typically more than two weeks per event

These seasonality categories are further classified according to the rapidity of soil drainage after the flood event, namely rapid, medium and slow soil drainage. The body of the habitat matrix contains cells that denote the habitats associated with given flood duration, seasonality and soil wetness regimes. It is noted that the wetness of soils in the period following a flood event is a key determinant of habitat potential.

Each cell in the matrix can be described in terms of detailed habitat types reflecting variation in other site factors such as soils, topography and habitat management practices such as grazing or hay making. Decision trees can be developed for each cell in the matrix to indicate which National Vegetation Classification (NVC) type is compatible with the given washland characteristics and management regimes. The current matrix is illustrative rather than definitive in terms of its assignment of NVC types to particular cells. The majority of the cells in the matrix have more than one vegetation type. The communities listed represent the vegetation that could develop on the site over a long period of consistent management. Such communities may not be achievable in the short (1-10 years) or even medium term (10-50 years), but they may be used to represent either future goals or as a guide to the appropriate management of the land, even though it may be recognised that the full community is unlikely to assemble at a site within the time-frame of a specific project.

The consequence of the findings of the Morris et al (2004) review for the present study are significant in terms of any aspirations for biodiversity gain through adoption of integrated natural floodplain storage (i.e. schemes to benefit both flooding and biodiversity). In essence, any natural flood storage area that is inundated at a frequency of less than approximately 1 in 3 to 5 years, and certainly at the 1 in 100 to 200 years frequencies, will not increase significantly the inherent quality of biodiversity within the area.

This could be a significant restriction to aspirations for large-scale schemes capable of mitigating floods with very infrequent return period (1 in 100-200 years). Simply put, the large areas involved and the continued management required to benefit biodiversity (more frequent flooding), which is likely to lead to a change in land use, would not be practicable at these scales, if biodiversity gains are a key driver.

However, planning for attenuation of more extreme flood events should not preclude consideration of conjunctive use options for smaller or confined areas of the larger flood storage footprint which could be engineered to provide biodiversity gain. There would be considerable potential for biodiversity gains from sub-schemes that attenuate smaller return period floods (say 1 in 3-10 years) in certain sub catchments or on the main stem of the river. The flood management purpose of these integrated washlands would be two-fold: a) to provide some attenuation that contributes to reduction in the immediate overall flood volume b) to potentially provide attenuation to individual sub-catchment contributions, thereby reducing the combined flood peak by de-synchronising peak propagation downstream (i.e. holding back and de-coupling peak sub-catchment flows).

The current study is focused on extreme flood events (1 in 100-200 years) and is not considering in detail the utility of integrated flood and biodiversity flood management for higher return period events. Much of the biodiversity benefit that can be achieved in integrated flood management of washlands is not likely to be relevant to the discussions later in the report, but is reported here for completeness.

2.6 Examples of schemes that provide flood attenuation and biodiversity gains

There are an increasing number of sites in the UK that provide flood attenuation through washland creation with integrated benefits for biodiversity (Morris et al, 2004). These can either be online (controlled or uncontrolled) or off-line (usually controlled) systems. In the majority of cases the sites are inundated at least yearly, and in some cases much more frequently: the Ouse washes is flooded 30 to 40 times per year and the Nene washes approximately 10 times per year. The duration of flooding is often in the region of two to five days per event, but in some circumstances can be much longer.

The examples of washland schemes in operation reinforce the message that to introduce biodiversity benefit to a flood management scheme usually requires the managed area to be inundated far more frequently than is being considered for the attenuation of the extreme floods in our study.