1Nature conservation and restoration traditionally focus on protecting individual sites. In parts of the world where the natural landscape has been severely altered for agricultural or urban use, individual patches are too small and isolated to ensure effective nature protection. Spatial processes, such as metapopulation dynamics, are disrupted and natural linkages in the landscape, such as water flows, are severed by modifications in the landscape and hydrology.
2We propose the identification of Operational Landscape Units (OLUs), which are defined as combinations of landscape patches with their hydrogeological and biotic connections, as a tool to facilitate wetland restoration in catchments with a high degree of fragmentation and strongly altered hydrology. The combined consideration of biotic (i.e. dispersal, transports of organisms) and hydrological connections (flooding events, groundwater flowpaths) is a new approach.
3We developed criteria for the delineation of OLUs in regional wetland restoration initiatives. The key elements for delineation are definition of the restoration objectives, identification of spatial landscape mechanisms and information on historic and present land uses and hydrologic management.
4As a case study, we delineated an OLU by applying these criteria in a restoration initiative for a large agricultural area that used to be a floodplain until the early 1950s in N.E. Twente, The Netherlands. The OLU encompassed not only the floodplain area to be restored but also a relatively remote nature reserve upstream as well as the stream connecting both areas. By restoring the historic water regime, flooding events would become a regular feature in the two areas and organisms including seeds would flow from the nature reserve to the restored floodplain.
5Synthesis and applications. Discussions of the proposed Operational Landscape Units with stakeholders (water authorities, nature conservation agencies, farmers) resulted in shared insights which will lead to modifications of the original management plan for the area. We believe the OLU approach will make natural resource managers aware of the importance of spatial processes and connectivity in landscapes and, if properly applied, will lead to more natural and more successful restoration projects.
Nature conservation aims to preserve the world's species diversity and ecosystem functions in an increasingly human-dominated context. In the past, the focus has been on the protection of selected species, mainly large vertebrates and birds (e.g. the EU Birds Directive 1979). More recently, the focus has widened to the preservation of areas characterized by rare species habitat, high biodiversity or valuable ecosystem functioning (e.g. the EU Habitats Directive 1992; the Convention on Biological Diversity 1992; Prins & Wind 1993). This approach has evolved to include spatial considerations in nature conservation strategies, initially aimed toward the preservation of large areas (Devries 1995; Schwartz 1999) but recently focusing on the preservation of networks of habitat areas, or ‘ecological networks’ (e.g. the EU Natura 2000 network; Bani et al. 2002; Bruinderink et al. 2003; Kati et al. 2004). In the last decade, this focus became so strong that Schwartz (1999, p. 98) noted that ‘it has been posited that the era of species conservation is over and that ecosystems integrated within landscapes will be the conservation unit of the future’.
In her review on the status of Landscape Ecology, Turner (2005) summarized the progress that has been made in identifying spatial patterns and processes in landscapes, more specifically addressing the role of disturbance, the use of landscape metrics and spatial statistics, and the interplay between living organisms and abiotic landscape components. The review paid little attention to lateral connections among ecosystems in fragmented landscapes. In this study, we emphasize the importance of these lateral connections and their pivotal role in maintaining biodiversity and ecological functioning at a much higher level than would be possible in separate ecosystems lacking these spatial interactions. These connections include not only corridors between landscape patches (e.g. riparian forests connecting extensive floodplain forests), but also connections via water flow or via wind. Conservation efforts should always look beyond the boundaries of the ecosystem to protect its connections to other systems in the landscape.
The need for this approach to nature conservation in densely populated regions arises from the increasing degree of landscape fragmentation, due to progressive urbanization and the associated intensification of agricultural activities. This fragmentation has been shown to have dramatic consequences for biodiversity and for environmental quality. Landscape fragmentation results in the transformation of large natural landscape patches connected by wide, continuous corridors into a large number of small patches, many of which are unconnected (Soons et al. 2005). This leads to an overall reduction in available habitat due to destruction (e.g. for housing or agriculture) and loss of connectivity because of destruction of corridors or the construction of barriers such as roads. As predicted by island biogeography theory, small patch size leads to high extinction and a high degree of isolation leads to low colonization, resulting in an overall loss of natural landscape and severe loss of species richness in patches (Hanski 2005).
Agricultural intensification leads to major changes in the hydrological functioning of catchments. Measures to drain agricultural land have led to lower groundwater tables, loss of groundwater discharge in (semi-)natural landscape patches, straightening of streams, dampening of stream water level fluctuations, loss of floodplain habitat, loss of meandering as a natural stream habitat process and a deterioration of stream water quality (Brierley & Stankoviansky 2002). These changes have led to drastic alterations in ecosystem functioning, often reducing habitat value across the landscape, even in protected reserves with appropriate internal management. It is evident that these landscape-scale modifications, including hydrological measures and land amelioration such as levelling of microtopography, have resulted in additional losses of biodiversity, which have probably been just as severe as those caused by fragmentation per se.
Although the consequences of fragmentation have been recognized previously, this is the first conceptual framework which considers them in combination. The issue of patch size and connectivity has been addressed in metapopulation theory, which has been shown to adequately predict the consequences of fragmentation for animal species with a reasonable dispersal ability (e.g. birds, mammals, flying insects, Kot, Lewis & van den Driessche 1996; Hanski 1999; Hanski & Ovaskainen 2000; Cabeza 2003). The importance of hydrological functioning of catchments for biodiversity and water quality has been recognized mostly in studies emphasizing vegetation or buffer zone functioning. The concept of homogeneous Hydrogeomorphic Units within catchments (HGMU, Brinson 1993; Maltby et al. 1994; Maltby, Hogan & McInnes 1996; Janssen et al. 2005) and the concept of Hydrogeologic Setting of wetlands (HGS, Godwin et al. 2002) have been proposed to emphasize the need for delineating functional units in the landscape, and the relation between vegetation and water and soil chemistry, respectively. These concepts do not address connectivity, either hydrological or biotic, but the HGS concept does recognize the importance of intact hydrogeological processes for plant diversity. In addition, some catchment-based scientific frameworks have been used to investigate the relation between stream flow, geochemical templates and biotic characteristics (e.g. Bohn & Kershner 2002; Petts, Morales & Sadler 2006). None of these approaches specifically focuses on spatial connectivity and restoration.
We propose the Operational Landscape Unit (OLU) concept, which combines biotic and hydrological connectivity, to analyze the best options for restoration of wetland ecosystem functioning and plant biodiversity in fragmented landscapes. This framework enables resource managers (i) to select the best locations for restoring biodiversity within a landscape or catchment; (ii) to identify the extent of the total area needed for restoration of target species and communities; and (iii) to determine which measures need to be taken to restore biotic and hydrological connections. Our aim is to provide guidelines for combining ecological knowledge on the spatial requirements of species with the spatial distributions and connections of ecosystem processes in order to develop more effective regional conservation strategies. The OLU approach aims to preserve and, where necessary, restore those landscape elements that key species and ecosystem functions require to operate successfully. Preservation and restoration of a carefully selected subset of small landscape elements in an area will often be less costly but at least as effective as conservation of a large area. Preserving or restoring corridors and hydrological pathways will maintain ecological networks. This will be more effective than the preservation and restoration of small isolated areas within the landscape. Success depends on a combination of knowledge from population biology, ecosystem science and hydrology.
The OLU concept integrates biotic and abiotic processes and connectivity at the landscape scale to analyze the best options for preservation and restoration. We stress that it is crucial that regional conservation strategies take into account the spatial scales of the ecological, hydrological and biogeochemical processes that are relevant to preserving the abiotic conditions and the flora and fauna of the ecosystem under consideration. It is also important to consider both habitat connectivity for key species and abiotic connections between ecosystems in the landscape.
Operational Landscape Units
We define Operational Landscape Units as ‘combinations of landscape patches with their hydrogeological and biotic connections’. An example of an OLU in a stream valley is (from high to low elevation) a combination of flat or sloping patches with groundwater recharge, the patches with groundwater discharge to which they are connected through local groundwater flow or overland run-off, the floodplain or stream bank patches and the stream itself. OLUs in stream catchments are characterized by connections through surface water flow in the stream or through regional groundwater flow (Fig. 1a). It is important to note that the connections that characterize OLUs are often disrupted in fragmented landscapes: groundwater recharge areas are often drained so that water short-circuits to the stream and groundwater discharge is strongly diminished; streams have been straightened to maximize hydraulic flow rates; floodplains have been lost because of flood control schemes such as dykes or water level control structures. This situation has a direct bearing on water-dispersed organisms as well as animal-dispersed organisms.
We stress the importance of restoration measures at a spatial scale that ensures self-sustaining restoration of ecosystem processes and functions. Depending on the group of organisms and the ecosystem function under threat, the extent of the OLU may differ. For example, the conservation of species-rich plant communities in stream valleys requires intact groundwater and surface water hydrology with lateral connections leading to groundwater discharge in species-rich meadows, in combination with connections between individual meadow patches through stream water flow and associated flooding. The size of the OLU is in the order of tens to hundreds of hectares in this case and consists of adjacent as well as disjunctive landscape components (Fig. 1b). Conservation of populations of small mammals [e.g. squirrels (Sciurus vulgaris L.)] requires a network of the order of hundreds of hectares consisting of forest patches, corridors and larger forest stretches (Verbeylen et al. 2003). Such OLUs are robust and will allow conservation in the future, even if environmental conditions will become much more dynamic as a result of climate change.
The OLU is situation- and target-specific. For each conservation or environmental management target, the combination of landscape components into OLUs will be different. If, for instance, the main target is the restoration of an ecosystem function (e.g. nutrient buffering by riparian wetland situated between agricultural fields and a stream), the main spatial features to be restored are the lateral run-off flows through the marginal zone bordering a stream. In many cases, these flows will have been diverted by construction of a ditch system bordering the fields, which bypasses the wetland and directs run-off straight to the stream, at the same time resulting in a much drier riparian zone. Removing this drainage system will restore the lateral surface and subsurface flows and will at the same time restore the wetland character of the riparian zone and its nutrient removal capability. The agricultural fields, riparian zone and stream, together with the lateral surface and subsurface run-off flows encompass the OLU in this example (cf. Fig. 1a). Another example would be the restoration of flooding and associated wetland vegetation in the catchment of a regulated stream where the temporal variation in stream discharge has been minimized and flooding events have become rare and extremely local. The OLU in this case should aim to restore the normal seasonal discharge pattern of the stream by connecting any available remnants of floodplain vegetation with the floodplain area to be restored. The floodplain remnants, the stream with seasonally varying discharge and the area to be restored constitute the OLU in this case.
The following three-step approach can be used to identify an OLU. First, define the restoration targets and then identify the type of terrain to be restored. Which plant or animal species or community need to be restored? Alternatively, which ecosystem function is to be restored (e.g. water quality enhancement or floodwater detention)? What are the main characteristics of the restoration area in terms of geomorphology, hydrology, land use and vegetation cover? The second step is to identify and analyze the spatial mechanisms in the landscape in terms of hydrology and dispersal that are crucial for the species, community or ecosystem to be restored. If spatial mechanisms are not relevant for the species or system to be restored, conservation or restoration of individual patches is sufficient and the OLU approach may not be appropriate. The third step is to define the extent of the OLU by identifying the landscape components which are necessary as building blocks for restoration of hydrological functioning and water regime, and dispersal of plant and/or animal species. This often will involve the identification of (i) ‘source areas’, intact, species-rich nature reserves, as well as (ii) ‘receptor areas’ which would be suitable for restoration and encompass the restoration target area, and (iii) connecting pathways such as streams or other hydrological flow paths. To identify the receptor areas and connecting pathways, historical maps and other records of present and past hydrological functioning are very useful. A necessary further step is the creation of a spatial overview of the area under consideration. The OLU map represents the combination of landscape elements that are required for regional survival of a species or a self-sustaining ecosystem function, or for creating target conditions for plant communities or ecosystem functions.
Case study: stream valleys in N.E. Twente, The Netherlands
Floodplains along rivers are much diminished in Europe because of habitat destruction associated with agriculture, urbanization and water management. The remaining floodplain wetlands are often degraded in terms of species diversity, extent, and natural functioning (Antheunisse et al. 2006). Further losses of biodiversity and ecosystem services still occur in these remaining wetlands as a result of progressive river and stream regulation, ever more intense agricultural use and fertilization (Nienhuis et al. 2002). Fragmentation of (semi-)natural areas and disruption of hydrological landscape connections have resulted in functionally degraded catchments in which small, isolated nature reserves are still losing species and where water quality problems occur regularly (Hefting et al. 2004, 2006; Verhoeven et al. 2006). In addition, predicted future climate change will modify the hydrological functioning of European catchments with more severe droughts as well as more extreme rainfall events; hence, management of hydrological processes needs re-evaluation (van Stokkom, Smits & Leuven 2005). All these factors have stimulated research on restoration of stream valleys.
In this case study, we propose some approaches for restoration of floodplains in the Ottershagen area in the Dinkel catchment in N.E. Twente, The Netherlands, by identifying the OLU that comprises the most important spatial relationships. These ideas have been presented to representatives from water authorities, nature protection agencies and farmers, and have helped to focus the discussion of options in land and water use.
Restoration of the Ottershagen floodplain area specifically requires an analysis of the hydrology and the current as well as historic land use of the Hollandsegraven subcatchment, which is part of the Dinkel catchment (Fig. 2a). The area is currently mostly in intensive agricultural use for dairy farming. In the southern part, there is a relatively small wetland reserve, the Agelerbroek, with wet species-rich grasslands (Caricion gracilis and Calthion communities, e.g. Ranunculo-Senecionetum, Caricetum vesicariae), and brook forest (Alnion glutinosae, e.g. Carici elongatae-Alnetum), and many waterfowl and amphibian species. The Ottershagen area in the north, now an extensive flat area with intensively fertilized and grazed meadows, once was a large, species-rich wetland prior to drainage and reclamation for agriculture in the 1950s. Since that time, flooding by the Hollandsegraven and Dinkel streams has occurred only exceptionally in the Ottershagen or in the Agelerbroek, as a result of canalization and bank reconstructions.
The delineation of the OLU involved (i) the definition of the targets for restoration; (ii) the identification of the key spatial mechanisms for the restoration target, and (iii) the identification of different landscape components to be connected and the connecting streams or flow paths involved. The targets are the restoration of the large floodplain area Ottershagen to increase regional wetland biodiversity, and the creation of a floodwater storage basin to cope with high river water discharges, which are expected to increase in frequency due to climate change. Hence, we seek floodwater retention measures that would result in regular inundations of the Ottershagen area with water from the Hollandsegraven stream. At the same time, the goal is to create opportunities for vegetation succession in the current agricultural pastures towards floodplain plant communities such as Caricion gracilis, Caricion elatae, Calthion, Phragmition and Alnion glutinosae.
The boundaries of the OLU in the Hollandsegraven subcatchment were drawn according to a combination of the extent of the wetland areas indicated on the historical map (1900) and the extent of the areas that were flooded during a short, extremely wet period in 1998 (Fig. 2b). The historical map shows that the Agelerbroek wetland area used to be more than twice as large as in 1900 and that Ottershagen used to be another large floodplain wetland in the northern half of the Hollandsegraven catchment. The OLU now encompasses the Agelerbroek as a source area for plant and animal diaspores, an extensive wetland area immediately to the north and east, the Tilligterbeek, Hamburgerbeek and Hollandsegraven streams (actually one continuous major stream in the centre of the OLU, see Fig. 2a) and the northern Ottershagen floodplain area. The OLU shows that the area suitable for floodplain restoration is not only the former floodplain at Ottershagen, but also the former wetlands adjacent to the present Agelerbroek. Both former wetland areas have the potential to be restored to regularly flooded marshes if the hydrological connections between the Agelerbroek and the Tilligterbeek were optimized and the stream discharge characteristics changed through hydrological measures. The plant communities in the Agelerbroek are typical for stream valleys with regular flooding and are suitable as a source of species for the restoration areas.
The hydrological and spatial mechanisms to be restored are (i) the natural flooding frequency of the Ottershagen area; and (ii) the hydrological connection of the Agelerbroek reserve with the Hollandsegraven stream network. The Ottershagen area can be turned into a floodplain again by a different regulation of stream discharges in the Hollandsegraven subcatchment, resulting in flooding events with highest frequency and duration in the winter season. The Agelerbroek nature reserve, which currently is only partly and rather indirectly connected to the Tilligterbeek–Hollandsegraven stream system, should be fully connected again, giving a direct connection between the Agelerbroek and the Ottershagen areas during periods of flooding. There is an elevation gradient sloping downward from the Agelerbroek northwards, so that floodwater from that area would flow into the Ottershagen area. By allowing flooding also to take place in the Agelerbroek, diaspores could be carried northward in the floodwater. These restoration measures will (i) create potential for development of floodplain areas typical for medium-sized (stream order 3–4) streams with their characteristic gradients in sediment deposition; and (ii) create opportunities for water resource managers to discharge water in these areas to prevent flooding in economically vital areas.
The OLU concept emphasizes the need to make use of spatial processes beyond individual sites to increase the chances of success in the restoration of ecosystem functioning and biodiversity. In fragmented landscapes, the lack of colonization or dispersal of organisms between sites may be rehabilitated through restoration of connectivity by water flow paths or other types of corridors. At the same time, landscapes with a strongly altered hydrology often lack spatial gradients in wetness or geochemistry, which are very important for plant species diversity. The most important new aspect of the OLU concept is that it addresses two types of spatial processes in a wetland restoration context (i.e. biological and hydrological flows). The practical application of OLUs in the planning process of restoration initiatives might result in OLUs of different sizes and scales, depending on the hydrogeomorphic landscape characteristics, the restoration targets and the spatial mechanisms involved. Nevertheless, through the combination of information on ‘source’ areas, ‘receptor’ areas, on historic and present hydrological flow paths and flooding regimes, a map can be produced with a delineation of the OLU and indications of the characteristics of the landscape patches after restoration. The use of GIS is instrumental for combining spatial information from different sources into the OLU map.
In the example we have given for N.E. Twente, historic maps and information on the spatial extent of a recent, extremely rare flooding event were vital sources of information during construction of the OLU. As an initial validation of the approach, we have compared the OLU plan for the Ottershagen area with the restoration plan that would have arisen without the explicit consideration of biotic and hydrological connectivity issues. The plan to use the former Ottershagen floodplain, now in agricultural use, for floodwater retention and to combine that with nature restoration, would most probably have resulted in the transfer of excess water from the Dinkel catchment at times of high river discharge (e.g. through a weir). Only the Otterhagen area itself would be the target area and without any connection to ‘source’ areas, the recolonization of the agricultural grasslands by wetland vegetation would be problematic. At the same time, the artificial flooding regime would fail to create the typical gradients in sediment deposition in the floodplain, which are favourable for floodplain biodiversity. In comparison with this ‘routine’ restoration plan, the OLU approach has (i) added the Ageler-broek as a source area of the plant and animal communities to be restored; (ii) increased the target area by including the former, much larger, Agelerbroek; (iii) added a connection of this source area to the Ottershagen area; and (iv) made flooding events more natural by restoring the stream flooding regime rather than creating artificial flooding from a canal.
Although the example illustrates many aspects of the OLU approach, there are some aspects which were of minor importance in the area studied but may be quite important in other wetland restoration incentives. In addition to the importance of surface water hydrology, groundwater discharge and recharge phenomena are often fundamental in creating specific site conditions for species-rich plant and animal communities. In the OLU approach, these groundwater dynamics should also be addressed where relevant. It may be that remnants of species-rich wetlands are still present at sites with groundwater discharge in a catchment with modified hydrology. Changing the hydrological functioning of such a catchment, even in a more natural direction, could result in loss of diversity if the groundwater discharge sites were flooded regularly with polluted river water. More generally, if species-rich, fen wetlands are part of the area to be restored, their hydrology should be studied carefully before they are considered as source areas. Tiny, isolated, groundwater-fed fens with a high botanical diversity may not be a good source of diaspores for floodplain restoration, and may at the same time lose many characteristic species if groundwater discharge ceases because of hydrological restoration measures. An ecohydrological study identifying the basic hydrological functioning of the catchment and the major relations between hydrology and plant community distribution is a prerequisite for a successful restoration plan and OLU delineation.
Although it will be inevitable that regional priorities in nature conservation, water resources management and land use planning will influence the implementation of OLUs, the recognition of spatial hydrological and biotic processes will make catchment-scale restoration plans more realistic and successful. The OLU approach can be a valuable tool in catchment restoration initiatives in areas where landscape fragmentation and hydrological alterations have dramatically reduced the surface area as well as connectivity of (semi-)natural ecosystems. In such areas, the OLU concept can optimize planning and implementation of restoration strategies.
The study was supported by the EUROLIMPACS project (EU-FP6 505540; http://www.eurolimpacs.ucl.ac.uk) on the effects of climate changes on aquatic ecosystems. We are indebted to Fons Eysink of Staatsbosbeheer Nederland, Menno Huge of Natuurmonumenten and Maarten Zonderwijk of Waterschap Regge en Dinkel for stimulating discussions.