skip to primary navigation skip to content


Forests on floodplains: The European Challenge
FLOBAR2 workshop May 16th and 17th 2002

Workshop Report

A workshop was held in Madingley Hall, University of Cambridge, UK on the topic: 'Forests on floodplains:The European Challenge'

Participants are listed at the end of this report. They included members of the FLOBAR2 project who are primarily academics carrying out research in the fields of floodplain forest ecology, floodplain geomorphology and hydraulic modelling ; members of the FLOBAR2 UK advisory committee who are primarily practitioners in river management and forestry; and invited forestry and river management practitioners from continental Europe and the USA. We were very grateful to all people who could attend the workshop.

Workshop aims

The aims of the meeting were:

  • To promote communication with each other about our experiences in the restoration, conservation and management of floodplain forests and how they relate to river management practices in Europe.
  • To focus on 4 issues which we identified as being difficult or contentious areas. In particular we wanted inputs on how to think about these issues so that we could write about them in a document that the FLOBAR2 project is producing on floodplain forests in Europe. The four issues were:

1. How should we manage floodplain forests?

2. Is it possible in any European context to use flood pulses to manage or restore floodplain forests?

3. Can we use forests on floodplains in flood defence?

4. What institutional dynamics and constraints exist for the restoration or conservation of floodplain forests?

In the report below, we have given an introductory section on the nature of floodplain forests and how they function. This is followed by a resume of the discussions held at the workshop under the titles of the four main issues discussed.

Background to floodplain forests in Europe

What are floodplain forests?

  • Forest is the natural vegetation on many floodplains around the globe.
  • Floodplain forests are spatially complex mosaics of small scale patches of different forest types, often interspersed with wetland and grassland ecosystems. They are most extensive when associated with meandering or braided floodplain areas.
  • They show similar planform patterns the world over. The planform patterns are driven by the hydrological regime and the delivery of sediment to the floodplain, and these are similar in all rivers although species vary considerably with bioclimatic zone.
  • Flood patterns are very important controls on floodplain forests:
  • ANNUAL hydrographs control regeneration patterns and growth in any year
  • INTERANNUAL flood variability controls channel movement and the formation of regeneration sites
  • Construction of dams and embankments have shown the importance of:
  1. The connection between upstream areas which are sources of water and floodplains which are the recipients of floods and sediment and
  2. Local, lateral connections between the river and the floodplain because where embankments have been built to sever these connections, channel movement and regeneration of forests no longer occur.

What has happened to floodplain forests in Europe?

  • In Europe we have built many river structures which reduce or remove floods from floodplains. In addition there is severe competition for use of floodplain land, forestry, agriculture, urban spread, industrial complexes etc..
  • Maps show the huge decrease- The UNEP-World Conservation Monitoring Centre Forest Gap Analysis Report (2000) uses a forest classification from the Natural Vegetation Map of Europe (see Bohn 1994, Neuhausl 1990) to describe potential forest extent and loss. It includes data from 45 countries in Europe. Alluvial Forests (U) is used as a general category (which includes, U2-Boreal alluvial forests, U3-Alluvial and moist lowland forests in the nemoral zone, U4-Mediterranean wet lowland and alluvial forests and scrub, U5-Continental alluvial forests (Populus alba, P. nigra and Salix alba and Tamarisk alluvial scrub) and U11-Riverine forests). Alluvial forests are the forest category showing the greatest losses of all European forest types,(88.8% of potential area) and within that category, the smallest losses are in U2 (54.22%) and the highest in U5 (99.46%). 6% (by area) of European Alluvial forests are protected.

What are the challenges for restoration of floodplain forests?

  • There is a considerable and diverse challenge to restore it in different parts of Europe. The challenge is great because even if we take away competition for floodplain land and plant trees on it again, that does not give us a naturally regenerating floodplain forest ecosystem, it gives us an area of trees on a floodplain. Floodplain forests need to receive flooding which will create regeneration sites and replenish water tables and aquifers. In other words, hydrological connections between upper and lower parts of rivers and between rivers and their floodplains also need to be restored.
  • In most places, river corridors have multiple users so that restoration of naturally regenerating forests on floodplains will almost inevitably be possible only as part of a multi-functional scheme. This might, for example, include forestry for timber or flood holding options as part of a flood-defence scheme. In northern Europe and in some locations in eastern parts of Europe there may be opportunities to manipulate flows at a dam site to achieve downstream ecological goals. In addition regulation of abstraction from rivers affects volumes of water available through the year and needs to be carried out with floodplain as well as instream ecological targets in mind.

Workshop discussion topics

1. How should we manage floodplain forests?

(Introduced and chaired by Prof. Christer Nilsson.)

Floodplain forests are often portrayed as centres of biodiversity and as highly species-rich environments. On the other hand, it has also been suggested that floodplain forests are sinks for species that have their sources elsewhere and that we should be protecting those sources instead. This view suggests that floodplain forests are less important contributors to regional biodiversity than is sometimes argued. It becomes important to understand what the relative values are of natural functions of floodplain forest. In practical terms we need to ask questions about how to value these natural functions so that comparisons can be made between, for example, the value of an area of natural floodplain forest and the value of an area of forestry on a floodplain.

It is possible to identify 3 sets of values:

  1. The inherent natural values of floodplain forests: these include
  • hydrological values such as storing water;
  • organic productivity values such as producing organic matter, timber etc;
  • biotic values which includes species richness; and
  • biogeochemical values which are valuable in terms of nutrient cycling.
  1. Forestry values: these rely on organic productivity values.
  2. Other values: might include social values, landscape diversity, recreation.

Three key questions were posed in this discussion session:

  1. Are the natural functions of floodplain forests important enough in biodiversity terms to warrant their conservation or restoration?
  2. What are the relative values of the natural functions of floodplain forests and forestry on floodplains
  3. Is it possible to value the other functions that floodplain forests might naturally carry out and to what extent can the 3 different value sets co-exist?

Summarised discussion (Topic 1):

  1. The model of upstream areas acting as sources of biodiversity for downstream floodplain areas was thought to be an over simplification. There are obligate riparian species which are particularly associated with floodplain environments, implying that both their source and sink are in the floodplain. In addition, not all dispersal along river corridors takes place by hydrochory, wind dispersed species such as Populus nigra might well move in the direction of prevailing winds rather than from upstream to downstream areas. There is also an argument in large river basins with high elevational variation that some species can live in lower sections which are unable to tolerate conditions at higher altitude and vice versa.
  2. If floodplains are sinks, then protecting a floodplain forest would protect species from a diverse set of source areas as well as obligate floodplain species. Thus floodplain forests should be protected as centres of regional as well as local biodiversity as they are able to provide species for restoration initiatives elsewhere and as natural research laboratories. In Europe, the most diverse floodplain forests are in large river systems such as the Danube River. These should be protected so that they can act as sources for restoration elsewhere. It could be argued that some of the most species-rich floodplain forests are also very remote, frequently flooded and difficult of access and should be given high priority for protection. It is also true that protection at a site does not confer protection at all if the physical processes which maintain the forest are disrupted elsewhere in the catchment.
  3. Floodplain forests are very variable in their biodiversity depending on their age, origins and management. In Britain the few woodlands on floodplains are of recent origin and sources are more important to protect in this case. At sites where floodplain forests have existed for a long time there is a stronger case for their protection. The history of management (eg forestry and other floodplain land use) is also an important determinant of floodplain forest composition and species diversity.
  4. Species diversity in floodplain forests can be higher or lower than that of surrounding terrestrial ecosystems depending on the bioclimatic zone. Thus, in humid areas they are generally less diverse and in semi-arid areas they can be more diverse than surrounding areas. Of interest is how much overlap in species exists between the floodplain and surrounding areas. An example from Croatia showed very different species in the floodplain forests compared with surrounding forests.
  5. Definitions of biodiversity are very variable. However, one scheme differentiates between genetic diversity, species diversity and landscape diversity. In practical conservation terms, genetic diversity is largely left to look after itself. At the level of species diversity, an area is worthy of protection if the species assemblages are different from elsewhere even if the range of species is the same. If diversity of landscapes is important then protection of physical and biological processes that create different landscape units should be high on the agenda as floodplain ecosystems add to overall landscape diversity. The European Habitats Directive includes conservation of the structure and function of habitats. Since floodplain forests have different structures and functions to upland forests, it is important to have both forest types and to protect the hydrological and sedimentological processes that maintain them. In Europe there is very little naturally functioning floodplain forest, only forests with different degrees of cultural modification.
  6. Attitudes to what ecosystems should be found on a floodplains are also very different depending on cultural and environmental context. Thus in Sweden and the UK, people are more interested in restoring meadows of different types on floodplains than in restoring forests, while in Spain the concern is with soil erosion by the wind and many non-native tree species are planted on floodplains and elsewhere to reduce this.
  7. The impacts of climate change are difficult to assess in the context of floodplain forests. On the one hand these forests are naturally unstable in terms of their physical fabric and their species have high natural levels of adaptation to a disturbed environment through diverse regeneration strategies. This means that they may be better able to cope with the consequences of climate change than other ecosystems. On the other hand they are also highly specialised and might not cope with either higher water tables due to sea-level rise or reduced hydrological inputs.
  8. When we stop thinking about species diversity and consider whole communities we can begin to evaluate the differences between naturally functioning forests and forestry plantations on floodplains and whether they have totally different functions or overlapping functions. Again the history of forests is important in evaluating this question. In countries like the UK, The Netherlands and Denmark, there are no floodplain forests left so we can make a case for planting trees for forestry purposes since such forests will have a higher inherent natural value than the arable farming it replaces. In places where floodplain forests still exist, albeit usually with a reduced natural dynamism, such as in countries that border the Danube River or along the Oder River, then replacing such forest with plantations probably reduces the inherent natural value of the forests. In this latter situation there is perhaps a real clash between economic values and the preservation of biodiversity in all its definitions. Again, the situation is very variable in different parts of Europe. Thus an example from the Garonne River in France showed that the economic returns from poplar plantations on the floodplain are very low when calculated as an annual return per hectare. It suggests that in this location inherent natural values or other values such as amenity, recreation and flood control might provide better value than plantations and should drive decisions on what we do with floodplain forests.
  9. A naturally functioning floodplain forest is a mobile mosaic of forest communities, in some areas it is eroded and in other places it regenerates as the channel moves. Such a dynamic system is not compatible with plantations since these require stable sites for the length of time it takes to produce a timber crop. In many places, plantations have become possible because the floodplain is no longer mobile. Thus in most places we are not dealing with the replacement of fully natural floodplain forests with plantations but with already highly impacted forests in terms of their physical dynamics and species composition even if they still occupy the same area.
  10. Where plantations are being established on previously unforested floodplain lands it is probably best to sell the idea in terms of economic returns from forestry and best planting practices for creating highest species diversity. An example of the latter was given from the US where in floodplain plantations, native poplar species were interspersed with oaks. Farmers were able to cut down the poplars to get financial return for the timber and the oaks began to give the benefits of a natural forest. Once plantations are established, it is possible to keep improving their management to improve species diversity ie have compromises between timber returns and the availability of habitats for a range of species. Cutting practices are also very important, both their extent and timing . In some countries such as the UK, there are rules about how close to streams forests can be planted. In the UK a 10 meter wide strip has to be left for engineering purposes.
  11. The discussion on management practices in plantations on floodplains has to be broadened into a discussion on management of floodplains and the rivers on which they depend. It is much easier to recommend management practices for forest plantations than for management of floodplains as it moves our discussion from what we do at a particular site to what we do with a complex system at the catchment scale. There are numerous issues at this scale many of which are discussed under discussion topic 3.
  12. Not everyone would want to see floodplain forests returned to floodplains as they are considered a negative feature in flood control in some areas (friction and sedimentation) and in other places people would prefer to have open habitats for walking or wetlands for bird-watching. No-one knew of any studies which compared the value of wetland habitats with woodland habitats on floodplains. On the other hand, some studies in the USA have shown that forests on floodplains are greatly valued in some areas for their 'ecological services' (included in 'inherent natural values' in this discussion) such as flood control and nutrient filtering. These services would otherwise cost billions if they were provided in an instrumental way.
  13. It is very difficult to evaluate the amenity and recreational functions that floodplain forests might have. There are some methodologies such as using 'willingness to pay' criteria and 'travel costs' to evaluate the value put on them by people but these vary greatly depending on their location (eg near an urban centre or not). No one knew of any studies which evaluated people's preferences for different types of forests. Many studies have found that people like to have water as well as trees so floodplain forests would have some clear amenity advantages.
  14. Clearly 'value' is a relative term. An example from Hungary compared the value of timber production from a particular floodplain area along the Danube River with the value of electricity production from a HEP plant whose construction has adversely affected the forests in the same floodplain area. The value of the timber was valued at 1% of the value of the electricity produced. This assessment did not take into account the capital costs of constructing the power plant.
  15. There may be a value in the genetic diversity found on floodplains. An example was given of the elm trees which have suffered widespread death from Dutch Elm Disease in Europe except on floodplains where elms have survived.
  16. Forests on floodplains could contribute to carbon sequestration although this positive value will be very small compared to the emissions side of the equation.

2. Is it possible in any European context to use flood pulses to manage or restore floodplain forests?

(Introduced and chaired by Dr. Francine Hughes)

The best general restoration principle is to restore biological function by restoring physical processes. Using this principle there are two basic approaches to the restoration of floodplain forests, each operating at a different spatial scale:

a) At the river basin/regional scale, flow management can take place using an impoundment structure to release or retain water for downstream ecological targets. Water allocation practices and patterns of water abstraction also influence the quantity and timing of water delivery to downstream floodplain zones.

b) At more local scales, connections between the channel and the floodplain can be restored or improved by removing artificial embankments. Local water tables can also be manipulated using engineering structures.

In Europe we have followed the second approach because it is easier to do in a crowded landscape, but we should explore the possibility of using the first approach and consider how we might apply environmental flows to the restoration of floodplain forests.

What are environmental flows and how can they help floodplain forests?

  1. They are flows which are allocated to some environmental purpose such as maintaining a fish population or diluting pollution. In practice they are usually minimum flows that are allocated for instream environmental targets on a monthly basis.
  2. In Europe, environmental flows have not been allocated for 'out of stream' targets although the European Water Framework Directive now demands that stream flows are maintained for the benefit of 'dependent terrestrial ecosystems'.
  3. Outside Europe, a number of methodologies exist to allocate flows for floodplain ecosystems as well as instream ecosystems. They have been largely developed in countries which have semi-arid areas (South Africa, Australia and southern Canada) and where floodplain ecosystems are given a high priority. These methodologies include a number of 'expert panel' approaches or 'building-block' approaches which usually involve some form of consensual agreements about optimal flows. The latter approaches build up flows from a starting point of zero flows each month, adding in the needs of each part of the ecosystem through the year. In some cases, which have involved planned releases from an impoundment structure, a consensus has been arrived at on the timing and duration of a single artificial flood event through discussion between diverse downstream end-user groups.
  4. It is important to identify which kinds of flows do which kinds of jobs and for the present discussion, what kinds of flows are needed by floodplain forests? Regular, low to medium flows help to maintain floodplain water tables and are needed for the annual growth of established trees. Periodic high flows are needed for channel movement when erosion and sediment deposition create sites suitable for the regeneration of floodplain trees. Following these high flows, gradually tapering flows are required during the first growing season to ensure the establishment of young trees. Thus, as a general rule floodplain forests need varied inter-annual flows over the decadal time frame or, putting it another way, they can tolerate a wide range of flows as long as periodically, flows suitable for regeneration occur.

Three key questions were posed in this discussion session:

  1. At what scale can we consider using environmental flows for the restoration or conservation of floodplain forests.
  2. Can we use flows stored during flood periods for this purpose?
  3. Do forests on floodplains take up more water than other vegetation types?

Summarised discussion (Topic 2):

  1. No one knew of any examples in Europe where Building Block Methodologies or similar had been used in the determination of Environmental flows. The normal model appears to be a minimum flows model, loosely based around the needs of key fish populations and often set as a percentage of monthly flows (eg Q95 set by EA in the UK).
  2. There are also very few examples in Europe of the re-instatement of connections between rivers and their floodplains by removal of dykes or other flood defences. One example is the Regelsbrunner Au restoration site on the River Danube in Austria. At another location on the River Danube, just over the border from Slovakia in Hungary, there is a conflict of interests arising from the construction of the Gabcikovo Dam and power plant on the Slovakian side of the border. On the Slovakian side there is the possibility of planned releases from the dam but not on the Hungarian side. Problems lie not just with the quantity of water that arrives on the Hungarian side but also its timing. In different years the water needs of the floodplain forests in Hungary will also vary. A number of scenarios have been devised to allow different percentages of the water back onto the Hungarian floodplain zone but there is considerable conflict between different groups of people on how and when to do this.
  3. The issue of variable water amounts between years was identified as a difficult one in the context of the normal management/decision-making timeframes used by river management agencies. On the other hand, building such inter-annual variability into water allocations for floodplain forests would also allow no or little water to be allocated in years of low rainfall and low flows, thus fitting in with natural patterns rather than sticking to rigid annual water quantities. In all years hydrological patterns should attempt to mimic natural seasonal patterns. The importance of inter-annual variability was demonstrated in an example from Sweden. Here there are two types of river, regulated and free-flowing. Hydrological patterns since the 1970s show predictable, stable flows in regulated rivers, but also to some extent in free-flowing rivers. During the 1990s, regulated rivers experienced extreme floods which overtopped impoundments and free-flowing rivers had many flood peaks during the growing season. This was important in terms of ecosystem change as the increase in hydrological heterogeneity led to rejuvenated riparian ecosystems. This reinforces the need to include extreme hydrological events in planned flow regimes as well as average flows. Another example from Croatia showed that in the last ten years, 1 in 100 year events had been experienced several times, giving natural flood pulses despite river impoundment structures.
  4. Examples given from semi-arid regions of western Canada (Alberta) suggested that consideration of a longer time frame than the decadal one is needed for decisions on inter-annual flow quantities. Thus in some Canadian rivers, the 1 in 20 year flood promotes successful forest regeneration. Research in Canada is focusing on creating curves for a) vegetation and b) channel maintenance needs. Curves are more natural then threshold step intervals which have been used in the past and make it easier to reconcile conflicting uses and requirements. The extreme is capped to reduce the frequency of overbank full events.
  5. In Europe, flow records often go back 100 years or more. These can be used to study the variability in flows that has been experienced in the past, bearing in mind that most rivers have been severely impacted by human interference prior to and during this period. Can we use these records to say what the maximum variability has been during that period and then discuss what kinds of forest can be supported by this variability? An alternative approach would be to decide what kinds of forests are desirable and then specify what kinds of flows would be needed to promote them. This requires us to specify what kinds of flows support which kinds of forests. Our research shows that this is very variable between river systems. It also shows that there are considerable lags effects in floodplain forest ecosystems. Thus, mature forests that we see now, regenerated 150-200 years ago when climate conditions and hydrological patterns were different (often within the 'Little Ice Age' period) and land use and runoff patterns were also different. In addition, forest removal in the past has often led to river incision and a change of base levels (an example was given from the Garonne River in France). Therefore, to try and recreate forests like the ones that remain might not be appropriate as the preconditions for their creation have changed radically. This suggests that choice of appropriate restoration targets for floodplain forests is critical.
  6. It seems likely that in Europe there are relatively few places where hydrological processes can be manipulated at the catchment scale in order to promote the regeneration of floodplain forest because floodplains are too heavily developed to allow this. The only exceptions might be parts of Scandinavia. On the other hand, in The Netherlands, restoration of parts of the Rhine floodplain is being planned in the context of the whole of the Rhine from where it enters the country to the sea. In many places dykes are being removed in The Netherlands and the floodplain is being lowered in order to increase flood holding capacity. To some extent the approaches used are culturally determined. Thus in the Netherlands, much of the landscape has been created on land reclaimed from the sea. There is already a well-developed sense of artificially creating landscapes and they are less constrained by ideas of preserving historic landscapes than, for example, in the UK. This allows a more radical approach to floodplain restoration than in many places.
  7. Creating space for rivers is an important justification for restoration projects and is currently being considered in many places as part of improved flood defence strategies. The option of making deliberate flood pulses for floodplain maintenance is good for mitigation in the context of Environmental Impact Assessments. Flood control benefits could give a legal basis for the financing of infrastructural projects which also include a flood pulse component. On the other hand, meeting flood control objectives by setting back flood defences will not necessarily lead to the best ecological outcomes. The problem is likely to be too much water at the wrong time of the year for vegetation regeneration and growth. Thus soft engineering options might be quite saleable on the flood defence side but not the ecological side. As and when such schemes are created, monitoring of ecological outcomes is vital.
  8. The water use by floodplain forests is a complex issue and might in different situations have implications for water conservation, flood storage and water table levels. It depends on age, structure, length of growing season and species type. Thus, there is the potential for willows to decrease flows as they take up a lot of water. Other species like oak and beech use the same amount of water as grassland. In Australia work has shown that old forests reduce their water requirements, and that you get peaks of high water use when trees are young and growth is active. The interaction with the floodplain geomorphology is also important – if you create a multi-channel system, this will help subsequent low flows and compensate for high water use by trees. The more wetland storage features you have on the floodplain, the better it is, as there is more gradual water release in dry periods so you can enhance low flows. This is good for dry climates where water deficit is a problem. In wet places, they might have the benefit of reducing water you don’t want on the floodplain. An additional consideration is the scale of the river and the difference between low and high flows. Thus on a large river like the Danube where flows can vary from low flows of 2000cm/sec to high flows of 10,000cm/sec, the volume of water taken up by floodplain trees is proportionally much less significant than it would be in a much smaller river system.
  9. Uptake of water in a clump of trees is very difficult to measure and to scale up to the community level. During measurement it is important to consider not just the outer rings of the tree, but also the inner parts of the tree. Sub flow density changes with season (there is a lot less water use in winter as trees are leafless) and between years. So when you do simple or infrequent measurements, you may miss the real uptake figures of trees. You can’t come up with figures that say that 'x' hectares of forest take up 'x' amount of water, yet stakeholders and managers need such figures. It is problematic if you extrapolate from such uncertain figures in an uncontrolled way.
  10. It is also important that forests are viewed not just as assemblages of trees, but also as complex systems consisting of undergrowth, soil etc. This is often forgotten. Soils vary greatly with bioclimatic zone and some soils, like peat have massive water bearing capacities. Nevertheless, trees like poplar and willow do take up a lot more water than non-forest vegetation although density of planting is also an important factor. There is a danger here that people will expect high uptake trees to eliminate the risk of flooding but in a large flood event, the presence of trees is unlikely to have a significant effect.
  11. When scaling up information about water use by trees, it is also important to take into account the changes in evaporation rates at the soil surface as a result of changes in vegetation cover. In general, these rates are less under forests than under other vegetation types. It becomes important to balance the increased water use of phreatophytes with higher evaporation rates, loss of soil structure and increased runoff rates if trees are removed. It becomes clear that when looking at overall water balances it is difficult to draw boundaries around the system because of the interconnectedness of all parts.

3. Can we use forests on floodplains in flood defence?

(Introduced and chaired by Professor Keith Richards)

There are competing issues at stake when there is talk of reinstating forests on floodplains:

  1. There are the possible positive dimensions of increased biodiversity, improved landscape values and improved flood defence.
  2. There are the possible negative dimensions of large woody debris and the increase in flooding because the forest holds up flood waters.

We need to examine the possibility that the negative dimensions could be turned into benefits if these forests are judiciously located. Forests attenuate flood hydrographs which could be beneficial for locations downstream. Floodplain forests could therefore form part of flood defence strategies.

There are numerous questions about the effects of forests on flood hydrographs. For example, is the extra head of water upstream (the afflux) generated by an area of woodland on the floodplain sufficient to store enough flood water that it makes a significant difference during a flood event? Research in Grenoble has studied the relationship between river stage and discharge, and found that there can be a significant amount of extra storage in a reach of river upstream of woodland. However, the amount of extra storage is dependent on the geometry of the valley which varies locally and would need modelling to choose the most appropriate sites for afforestation. We also don’t know enough about the resistance effects – there is a need for research which will rely on modelling. But modelling outputs depend on how you represent those effects.

A negative aspect of forests on floodplains is the delivery of woody debris from the forest into the river and the potential for this debris to get stuck behind bridges and in culverts during floods. The strategy for many years has been to get wood out of rivers to increase efficiency. Ecological and water quality issues more recently have led to the reversal of this view, and woody material from floodplains is now more acceptable.

Another issue may be land use elsewhere in the catchment, and maybe that should be the real focus. It could be that woodlands on slopes in upper catchments are more effective in preventing runoff than floodplain woodlands are at managing excess runoff. So are we missing the point in advocating floodplain forests? This seems like a re-run of the 1940s debate in the US about upstream land-phase and downstream channel-phase management of flooding.

Also how would these things be financed? Can Cost-Benefit Analyses be carried out that demonstrate the gains of forests in terms of flood control benefits? Or would it have to rely on other demonstrable benefits such as carbon sequestration, to be viable? In the UK there is currently a consultation about floodplain developments becoming subject to charges and flood taxes. Could this favour the development of floodplain forests? Also in Britain the EA is responsible for flood defence, and DEFRA for financial support. Would the potential role of floodplain forests as flood defence instruments be something DEFRA could support financially under present funding arrangements?

Three key questions were posed in this discussion session:

  1. Do forests on floodplains cause flooding?
  2. Can they be used to hold floods in desirable locations?
  3. What are the research needs to make decisions in this area?

Summarized discussion (Topic 3):

  1. There seems to be a strong case for trees contributing to flood defence if they are strategically located in the right places and provided no habitation is put at risk. It is not known how much of an effect there is. It depends on many factors including the scale and size of floodplain, degree of channel movements, nature of the vegetation etc. In order for forests to have a significant effect, they would have to cover quite big areas (10% of the floodplain?) since just a few riparian trees wouldn't have much effect. The size of area needed to have a particular effect is largely a modelling issue, and would be very difficult to calculate – at the limits of our current modelling capability. Studies have been carried out by RIZA in The Netherlands that model increases in stage for given discharges and with varying percentages of willow cover on the floodplain. The nature of the vegetation is important because of the type of friction effects it has. Thus, trees create more of a barrier than bushes because the latter flatten during high flows whereas trees don't. The smoothness of trunks, presence of branches lower down on the trunk etc all have an effect and friction remains very difficult to measure. In this regard, the management of forests (for forestry or nature reserves) affects friction because of spacing of trees, levels of undergrowth etc. In the last 2 or 3 years these friction effects (flow resistance) have become the subject of more research.
  2. When the % of floodplain land with forests is plotted against water flow, the relationship is unlikely to be linear because of varying degrees of connectivity between areas of forest. The degree of connection between patches would also be a conservation issue.
  3. In many countries the priority would be creation of engineered flood storage structures not planting of forests. For example, the Environment Agency, which manages floods in the UK, would be most likely to identify locations where flood storage is needed, create engineered structures to hold flood waters and then ask if forests can also be accommodated within these structures. It would be assumed that if trees are to be planted within flood storage areas then these would have to be of a greater capacity to accommodate the trees. (However, in some storage schemes which involve banks built across valleys to slow down water eg. the Lincoln attenuation scheme in the UK, the presence of trees in the storage area doesn't really matter.) This is because it is important to build confidence in areas upstream of urban centres that have had bad flooding and the Environment Agency would prioritise engineered structures. On the other hand, from the point of view of DEFRA in the UK which tries to bring together environmental and rural aspects of river management, the wider question has to be asked...are floodplain forests a more sustainable solution than an engineering structure? Within the UK Catchment Management Plans there is scope for including forests because they have certainty of resistance and at present solutions are being sought to the management of higher level floods than just the 1 in 100 year.
  4. Floodplain forests should be viewed as being primarily useful in flood impedence. Forests are not complete barriers to floods so their role in storage involves holding up waters over a longer period of time than if they were not present, attenuating the flood hydrograph so that peaks are lower but of longer duration. A flood defence approach that uses forests is therefore different from one that involves engineered structures, more like having a sponge in the system than a water-tight flood holding area. Floodplain forests have naturally carried out this role in the past and their removal has probably contributed to an increase in flooding severity. There is therefore a strong case for research that quantifies their flood impedence effects.
  5. It is important to compare the costs of flood damage with the costs of different land uses in floodplain areas. In Hungary, a recent 1 in 500 year flood on the Tisza River caused 300 million Euros worth of damage. For that price it would be possible to plant 375,000 ha of forest on the floodplain. Construction of a large flood storage reservoir has been discussed but it might be better value to afforest the area and let the forest hold up floods.
  6. It is also important not to overstate the potential of floodplain forests to solve flooding problems. Floods cannot be stopped. It is a European priority to try and reinstate forests on floodplains. Because of resistance from regulators, this can only be carried out on low/no risk sites. Eventually they can become a more common flood defence tool once confidence in them increases. There is an overwhelming need for a demonstration floodplain forest site in a country where this habitat has largely disappeared from the landscape. In other countries where floodplain forests remain but are largely disconnected from their rivers by embankments, there is the potential for reconnection of floodplains as part of flood defence schemes that might set back flood defences and give more space to the river to flood. In the UK at present, DEFRA is carrying out research into the factors constraining flood defence realignment, trying to encourage integration of the need of all stakeholders rather than just engineering.
  7. The potential for instalment of floodplain forests is affected by land ownership issues not just in the forested area but also in the areas upstream which will experience higher water levels due to the barrier effect of the forest. There is a case for carrying out research on the impedence effects of narrow linear forests that follow the channel compared to blocks of forest because these would encounter fewer restrictions.
  8. In larger catchments, tributary streams often have their own floodplain areas. It might be easier to target restoration of floodplain forest on these thereby promoting flood attenuation in upper parts of the catchment and reducing the speed with which floods reach lower parts of the catchment. The risks of afforesting floodplains of smaller river are lower than those of large rivers and might at the present time be more acceptable. They might also provide the scope for empirical research that could be transferred to larger rivers. However, In large transboundary river systems such as the Danube, it is not possible to suggest such activities in upper parts of catchments since they are in different countries and integrated catchment management is much harder to achieve.
  9. It remains important to hold up runoff in upper catchments as much as possible. When it comes to flood defence, land-use issues throughout the catchment need as much discussion as what happens on the floodplain. At the European level, this requires integration of different policy environments (eg the CAP and the Water Framework Directive). In the UK, DEFRA is currently quantifying the effects of land use on runoff in the interests of flood management.
  10. Flood managers are powerful actors. So ignoring the flood defence issue in floodplain forests would be missing some central issues. There are wider floodplain restoration opportunities. Issues like sustainability, conservation can be drawn in on the back of thinking in other fields. Windows of opportunity are created where policy sectors can get benefits despite not being directly associated. Flood defence debates are therefore crucial. Thus we can use flood defence as a justification for catering for wider needs.
  11. When water is slowed down through a forest, it has longer opportunities to percolate through the soil and refill water tables and aquifers. This is a hidden benefit of forests on floodplains. Other benefits of forests on floodplains might be the improvement of water quality although this was considered to be contentious especially where land is underdrained and drains discharge directly into the river.

4. What institutional dynamics and constraints exist for the restoration or conservation of floodplain forests?

(Introduced and chaired by Dr. Tim Moss)

A key issue in this topic is connecting the policy and project levels of floodplain woodland restoration. There are a number of themes that are important to consider:

  1. Attitudes and mindsets

In many places, the impact of trees on floods is perceived as a potential problem and this attitude holds up change. Fear of the unknown is therefore a major factor prompting river managers, when in doubt, to err on the side of caution. Entrenched procedures and traditional practices can be a further obstacle to FPR. For instance, farmers who have spent their whole lives draining land are often very unwilling to let it flood in accordance with a shift in flood management policy. Compartmentalised thinking, or thinking “in boxes”, poses particular difficulties for FPR which relies so heavily on integrated, cross-sectoral management. Organisations and public agencies have their own rationales and ways of doing things which become entrenched over time and are often shaped by their members’ disciplinary backgrounds. For instance, whilst flood managers focus on peak flows, nature conservationists are often more concerned with the troughs, i.e. low water levels. The same applies to the funding instruments we use. Even within flood defence funding there exist distinct – and often uncoordinated – budgets earmarked for capital investments and for maintenance. Funding from each budget is bound by strict criteria limiting options for integrated, cross-sectoral interventions. Nevertheless, we can observe the emergence in recent years of much greater openness to more integrated approaches to flood management in policy-making circles. Civil engineers, for instance, are exploring a wider range of flood management options than previously as a way to increase efficiency and additionally as a way of cutting costs Nature conservationists are finding it more fruitful to identify policy shifts conducive to river restoration aims and exploit emerging communities of interest than to seek to force others to follow their policy demands as in the past.

2. Timescales – The high durability of trees is indicative of a problem of matching different time-scales in floodplain woodland restoration. Farmers, for instance, need to see trees as a long term investment but the funding incentives to get trees planted are not so durable. They are often not prepared to plant trees if the subsidies are only guaranteed for, say, the first 5 years.

Each of the actor groups involved in FWR thinks and acts around very different time-scales. Civil engineers think in terms of 100-year flooding events; farmers are bound by seasons and the duration of funding programmes; politicians have re-election timescales in mind which influences what they do. These differences need to be taken into account when devising programmes and projects for FWR.

Similarly, the relevant instruments of FWR have different time-scales, the most obvious being between capital funding and maintenance funding. Creating funding packages which transcend this divide would appear to be very important in terms of creating lasting opportunities for FWR. This is the case not just at the policy level, but also at the project level. Narrow projects can broaden out over time to encompass a wider view and scope than previously envisaged.

3. Spatial scales – In recent years an “extensification” of interests can be observed (Newson 1992), with river managers increasingly looking up the basin and conservationists looking beyond the channel. People are looking increasingly to influence developments in policy areas beyond their own sphere of action. Consequently, they are becoming more dependent on acquiring the support of others.

Recognition is growing that a catchment-based approach is the most effective and efficient way of managing rivers and their ecosystems. Catchment-scale management has far-reaching institutional implications, in terms of the number of policy fields involved, actor groups affected and funding options available. It also challenges traditional practices of river restoration. Restoration projects in the past have been very site-specific and are not set in a broader catchment context. As a rule a site – or some funding – has become available and a project is designed around the site. It is very rare for a restoration project to emerge from a catchment-based assessment identifying the most favourable location.

A further important spatial factor is the need to connect better the policy nd project levels of FWR. It appears the more integrated approach at the policy level mentioned above is often not filtering down to the operational level. As a result, we can observe frustration at the policy level at poor implementation, but also at the project level at the demands of integrated policy initiatives.

4. Policy delivery – This raises the question of how responsive policy makers are to implementation problems on the ground. Whilst policy is changing quite radically, the mechanisms for implementation have remained largely the same. A common complaint is that policy-makers need to take greater consideration of where those affected are starting from and what steps are feasible in practice.

Coping with complexity is a significant issue of policy delivery. Today restoration projects are becoming more complex to run, due to consultation requirements, co-funding, feasibility studies etc.. Each of these requirements places an additional demand on resources. They also often come with conditions attached, as with funding sources, pushing projects in sometimes divergent directions. Managing restoration projects increasingly requires professional help, as in the form of a full-time project coordinator. The sum effect of this growing complexity can be to actively discourage projects. There is considerable anecdotal evidence to suggest that many small scale and low budget projects are being shelved for this reason.

Low levels of policy delivery raises the question of the limitations to the consensus-building approaches on which much recent policy places trust. As the limited up-take of some of the more far-reaching agri-environmental schemes illustrates, simply creating opportunities and incentives for FWR is often not enough. Some form of coercion is in some cases beneficial. The ineffectiveness of land-use planning and building restrictions in preventing development on floodplains calls for greater attention to the implementation of existing regulations.

Finally, policy delivery would appear to suffer from inadequate monitoring on the effectiveness of existing instruments and how they can contribute to various objectives. The ex-post evaluation of programmes and projects is rare, yet it could offer important insights into context-sensitive improvements to programme design.

Two key questions were posed in this discussion session:

1. How can recent favourable policy shifts be translated into more floodplain restoration on the ground? How can we address the policy-delivery gap?

2. What does past experience with floodplain restoration projects tell us about the kind of institutional mechanisms needed for floodplain forest restoration?

Summarized discussion (Topic 4):

  1. The boundaries of influence of local institutions (such as local authorities in the UK) rarely match up with catchment boundaries. For example, in Croatia there are two main catchment areas, the Danube and a series of smaller catchments flowing into the Adriatic sea but the country is split into 32 regions which makes integrated water management difficult. It is difficult to create one scale identity for the purpose of driving though floodplain restoration projects and other river management decisions, as that identity probably won't mix with other policy fields. The more integrated decisions become, the more dependent they are on those other fields eg the co-operation of local planning authorities. There is no ideal spatial fit and there is also the added issue of accountability if a body is created for a catchment which is not democratically elected. Several examples were given of catchment-scale bodies being formed. In the Garonne River which has its catchment primarily in France but also in northern Spain, an organisation has been formed called EPTB Garonne. In the Garonne basin, decisions are made at the department and regional level by organisations that do not correspond spatially with the catchment though they overlap it. EPTB Garonne is composed of representative members of all stakeholder groups in the whole catchment (about 400 in total). It does not have regulatory powers but acts in a consultative capacity. Similar organisations exist for all river basins in France. In the Wye River in the UK, there is a discussion forum where different stakeholders in the catchment can come together for discussion. It is a democratic body in the sense that it is made up of local councillors, it does coincide with the catchment and is politically accountable but it too has no regulatory powers. On the other hand, in Florida, USA, there are powerful water management agencies which can tell landowners in the catchment what to do with their land.
  2. Use of the term 'restoration' is not always clear in either policy or practice. In most cases in Europe it is not possible to restore form or function to some kind of natural state and perhaps rehabilitation is a more appropriate term. In some countries the term 'renaturalisation' is commonly used. The use of the term sustainability by many water management agencies suggests they are concentrating on benefits rather than on restoration. Terminology is important because of messages it gives. For example, if we use the terms restoration or rehabilitation it gives a negative impression to farmers as they can feel criticised for past land use and this sets up an initial barrier. In countries without floodplain forest, restoration can mean planting forests whereas in countries which still have floodplain forests, restoration means something quite different and refers more to the management of the forest than a complete change in land use.
  3. Organisations which would like to bring about change in policy and practice in floodplain restoration in European countries are varied and operate at different scales. For example, in the UK, the EA is trying to bring about change at the local scale by encouraging farmers to modify their practices and change attitudes and at the national scale by being involved in the realignment of organisational objectives whereby the EA, DEFRA and Regional Assemblies would work together in setting targets for grant aid etc.... In Hungary, local level drivers tend to be most effective in bringing about change because of the large size of river basins. Thus at the local level, forest authorities might rehabilitate some floodplain forest areas by opening up side branches of the river and allowing more water onto the floodplain during times of flood. To have concerted action over a whole basin as large as the Danube would require such a high level of political coordination that it is unlikely to happen.
  4. The way that land is held is also an influential factor. For example, in Croatia, more than 80% of forests are owned by the state so it is easier to devise policies to manage them. On the other hand, state water management agencies have more power than forestry agencies, including the power to charge taxes for water supply and flood protection. This produces an inter-agency conflict of interest in the management of floodplain forests. In Hungary much of the land is now privately owned where previously it was state owned. This has led to a greater range of initiatives in floodplain areas.
  5. The nature of instruments that might drive floodplain restoration are also varied and frequently in conflict. For example, in France the New Water Act of 1992, encourages re-wetting of floodplain wetlands. Previously wetlands were prioritised for draining under both previous water laws and under agricultural laws. This means that money is available to both drain and re-wet the same piece of land. Historically, the mind-set is for draining wetlands and it has proved difficult to change this entrenched attitude. Also in France, there is currently discussion about water pricing for agricultural use and talk of a tax on the use of soil or water that modifies water flow patterns (eg. by gravel extraction, drainage off new roads and other developments). There is resistance to change and particularly to monies from such taxes being used for wetland creation. In the UK, discussion about taxes on floodplain development extends to consideration of all development in a catchment which might have an impact on runoff patterns. Imposition of taxes on developers in this regard skews the type of development to high value developments where the developer can afford the extra charges. In general, there is a large dependence on legislation for encouraging floodplain restoration. Legislation tends to make reference to protection and restoration of floodplain habitats and to be more effective in protecting existing habitat than in creating new habitat. Thus institutional constraints are greater where there are no forests at all than where pre-existing forest requires improvement. In the latter case there is no requirement for a change in land-use type. Many instruments are market instruments such as insurances.
  6. The policy-delivery gap is difficult to address. In many organisations such as the EA in the UK there is integrated thinking at the top levels of the organisation but different attitudes at the lower levels. This makes ratification of projects difficult and delays their delivery. Another problem is the difficulty of integrating all the environmental requirements into delivery of a product. These delays reduce public confidence and increase the gap between the policy environment and the people involved. In the US, lack of project delivery because of conflicting pressures ( 'policy gridlock' )is also linked to a high degree of risk aversion on the part of managers. An example from the UK shows problems with public confidence. In the town of Lewes, which has been severely flooded, a series of flood defence options were presented to the pubic by the EA. These were prepared by a civil engineering company working to environmental standards set by the EA and working to a brief which does not allow the company to impose its engineering bias. In addition the EA consulted the UK River Restoration Centre on the proposed options. Many of the options involve major engineering works to reroute flood water and there is a conflict between the public's concern over the (in)appropriateness of these in the particular rural environment around Lewes and fear of further flooding. There is also a fear that flood defence solutions are not being delivered rapidly enough.
  7. We should actively seek for windows of opportunity to promote floodplain forests by identifying changes in rules or grant support that would enable their establishment. Reform of the CAP gives some hope for floodplain forest restoration, as does the Water Framework Directive by forcing managers to think in terms of land-use reform.When several EU Directives impact on the same piece of land, it is not clear which Directive takes precedence but under UK law the first law would take precedence.

Discussion of the FLOBAR2 document format

A detailed outline of the proposed FLOBAR2 document format was sent to all workshop participants. During the workshop a special session was held to discuss the format for the document and its appropriateness for its intended user groups. Some mocked up pages showing possible page and graphic styles were also presented. It is currently the intention to produce a limited number of paper copies of the document and to make it primarily available through the web. If additional funding can be found it can be disseminated in more diverse ways and could also be developed into regional versions with more detailed local and specific information on appropriate species to plant etc.

Opinions on the format varied between workshop participants. The most strongly held views on the document's format and dissemination are summarised below:

  • On the whole the proposed order of information presented in the document was considered appropriate
  • The introductory sections should be more 'in your face' and spell out more clearly and right at the start why floodplain forests are important and what gains they might bring to a river manager.
  • Opinions were divided on the length of the document. It is proposed to be 60-80 sides of A4 long. Some people thought it should be much shorter, on the other hand at a recent meeting with practitioners in Sweden, Prof. Christer Nilsson noted a strong interest in a comprehensive volume on restoration.
  • It was suggested that whatever length it ended up being, each double page should be 'stand alone' in its information so that individual sheets or sets of sheets on particular topics could be sent to interested parties. This also reduces the opportunities for information to be cut and pasted from the document and used out of context.
  • The version put on the web should also be carefully compartmentalised so that each chapter and sub section is easy to reach from an initial directory/contents summary diagram. This allows people rapid access to the information that interests them most. Some suggestions were made for a strong hierarchical structure to the information presented so that people could choose to look at particular topics in more or less detail. Hypertext technology should be used.
  • The proposed liberal use of case studies was approved as the best way to give reassurance on approaches to floodplain restoration.
  • The document should primarily be addressed at multipliers.

Workshop participants

Europe and the USA

Mr. Hans H. Riber
COWI Consulting Engineers & Planners AS, Denmark
Mme Aline Chaumel
EPTB Garonne, Toulouse, France
Dr. Vladimir Kusan
Faculty of Forestry, University of Zagreb, Croatia
Mme Pascale Mahé
Nature Midi-Pyrénées, Toulouse, France
Dr.Zoltan Somogyi
Head Forest Research Institute, Budapest, Hungary
Professor John Stanturf
US Dept of Agriculture, Forest Service, USA

United Kingdom, Flobar2 advisory committee

Mr.Peter Allen-Williams
Department of the Environment, Food and Rural Affairs (DEFRA), UK
Dr. Mark Diamond
Environment Agency (EA), UK
Dr. Nigel Holmes
River Restoration Centre and Alconbury Environmental Consultants
Dr. Keith Kirby
English Nature (EN), UK
Dr. Tom Nisbet
Forestry Commission (FC), UK
Dr. George Peterken
Forest Ecologist
Dr.Christoph Zockler
World Conservation Monitoring Centre-UNEP, UK

FLOBAR2 members

Prof. Keith Richards
University of Cambridge, UK. FLOBAR2 coordinator
Dr. Francine Hughes
University of Cambridge, UK. FLOBAR2 coordinator
Mr. Adrian Hayes
University of Cambridge, UK
Sarah Damery (raporteur)
University of Cambridge, UK
Dr. Tim Moss
Institute for Regional Development and Structural Planning, Erkner nr. Berlin, Germany
Ms. Manuela Siener
Institute for Regional Development and Structural Planning, Erkner nr. Berlin, Germany
Dr. Jacky Girel
University Joseph Fourier, Grenoble, France
Ms. Lori Gom
University of Lethbridge/ Alberta Environment, Canada
Dr. Etienne Muller
CESAC-CNRS, Toulouse, France
Prof. Christer Nilsson
University of Umeaa, Sweden

Unable to attend

Mr. Russell Cryer
Royal Society for the Protection of Birds ,Wise Use of Floodplains Project, UK
Dr. Gary Kerr
Forestry Commission, UK
Dr. Jackie King
Department of Zoology, University of Capetown, SA
Min. Otto Malek
Bundesministerium für Umwelt, Naturschutz und Reaktorsicherheit, Germany
Dr. Hartmut Barth
European Commission, Brussels, Belgium
Prof. Herbert Hager
Institute of Forest Ecology, Vienna, Austria
M. Jean-Luc Roy
Ministere de l'Amenagement du Territoire et de l'Environment, France
Dr. Karl Gutzweiler
WWF Institute for floodplain Ecology, Germany
Dr. Edith Wenger
WWF Institute for floodplain Ecology, Germany
Martin Schneider-Jacoby
European Nature Heritage Fund, Germany.
Mr. Bart Fokkens
Director, Wetland Development and Restoration Department, RIZA, The Netherlands