Managing biodiversity of floodplains in relation to climate change Open Access

The continuous supply and accumulation of nutrient‐rich sediments from upstream areas and lateral sources makes floodplains among the world's most biologically productive areas. The nutrient basis along with the high spatio‐temporal habitat heterogeneity due to the pulsing stream flow (Junk et al., 1989; Poff et al., 1997; Tockner et al., 2000), makes floodplain ecosystems highly rich in species (Tockner and Stanford, 2002).

Freshwater and wetlands/floodplains provide valuable ecosystems services, in most cases more than any other ecosystem type (Tockner, 2010). In spite of being an ecosystem of great importance and value for humanity through a variety of ecosystem services, floodplains are considered as among the most threatened ecosystems in the world. In Europe and North America, up to 90 per cent of floodplains are functionally extinct (Tockner and Stanford, 2002), mainly due to alteration of the natural flow regime, habitat alteration, species invasions, and pollution (Tockner and Stanford, 2002).

A wide variety of organisms have evolved adaptations of life history, behaviour and morphology to cope with alternating floods and droughts (Lytle and Poff, 2004). However, natural flow regimes (Poff et al., 1997) are heavily altered by human activities, especially by the construction of dams. By the end of the last century, over 45,000 large dams (>15 m high) were constructed worldwide. Dams reduce or disconnect the important longitudinal connectivity of rivers (Vannote et al., 1980), and lead to homogenisation of stream flow by reducing peak flows and increasing minimum flows (Moyle and Mount, 2007). The range of natural flow regimes may vary considerably between regions of the world (Poff et al., 2006), and the massive number of dams will not only homogenise the stream flow within rivers, but also between regions (Poff et al., 2007). In addition to the stream flow regulation by dams, channelisation and construction of levees, dikes, roads, and railroads close to the river channel may disconnect the river from its floodplain and reduce or even eliminate the lateral movement of water, alluvial sediments, nutrients, and biota (Ward and Stanford, 1995a, b; Blanton and Marcus, 2009). Thus, the fluvial dynamics that formed the floodplains in the first place have been severely altered (Ward et al., 1999).

Ward and Stanford (1995a) emphasise that the ecological integrity of floodplains is partly based on the diversity of water bodies (e.g. ponds and oxbow lakes), which differ in their degree of connectivity to the main river channel. During periods of low or no flow, these water bodies act like aquatic refugia (Sheldon et al., 2010). Further, the water bodies may serve as important habitats during different phases in life history for some species. The different water bodies also differ in their stages of succession, and form a highly diverse mosaic of habitat patches across the floodplain. The variation in design of the water bodies (Ward and Stanford, 1995a), along with the spatial and temporal heterogeneity from variation in hydrological connectivity makes riverine floodplains high in alpha, beta, and gamma diversity (Amoros and Bornette, 2002).

In periods with decreased magnitude, frequency, and duration of floods, Whited et al. (2007) found that the floodplain environment was dominated by late successional vegetation and exhibited low levels of physical restructuring. Following the construction of dikes and levees, the floodplain and its water bodies are often totally disconnected from the main river channel, and the forces resetting successional sequences from flood disturbance will no longer be present. In other words, the former diversity of water bodies along a gradient of different successional stages will probably be homogenised over time.

Different climate models point towards more frequent combined events of high temperatures and high precipitation (Benestad and Haugen, 2007), which in turn may increase the risk of spring time flooding. In addition, more frequent flash floods are expected throughout Europe (Alcamo et al., 2007). All together, more frequent and less predictable floods (more complex flooding pattern) are expected. In some settings, more frequent larger floods may work to re‐establish the connectivity (e.g. where it is disconnected by levees) between the river channel and the floodplains (Poff, 2002). However, larger floods might also result in the demand for construction of more extensive flood mitigation measures.

In this paper, we analyse species richness of crustacean zooplankton and water beetle from 20 water bodies on a riverine floodplain in Hedmark County with regards to varying degree of lateral connectivity and water body area. The findings were related to regional species diversity of crustaceans and water beetles. Further, the preservation of floodplain areas as centres of biodiversity are discussed in relation to probable scenarios of climate change.

This survey includes 20 water bodies in four different areas on the riverine floodplain at Flisa, Åsnes Municipality, Hedmark County (Figure 1, Table I). The floodplain surrounds the confluence area of the rivers Glomma and Flisa. Glomma is the largest river system in Norway with a total length and catchment area of 604 km and 42,000 km², respectively. The investigated area is dominated by agricultural land, some forest, built areas, and wetland.

The mean annual flood in the rivers Glomma and Flisa culminate at discharges of 1,520 and 71 m³ s⁻¹, respectively. The culmination discharges during 10, 50, 100, and 500 year floods are estimated to be 2,098, 2,721, 3,010, and 3,676 m³ s⁻¹, respectively, and the areas vulnerable for flooding during a 10, 50, 100, and 500 year flood are estimated to be 13.0, 16.6, 19.0, and 21.8 km², respectively. During 10, 50, 100, and 500 year floods, it is estimated that flood mitigation measures impede flooding of 66, 70, 53, and 0.2 per cent of the vulnerable area, respectively, (Haddeland, 2001).

Standard methods for qualitative samples of crustaceans and water beetles were carried out in September 2008 and 2009. The crustacean zooplankton samples were collected by a long‐shafted dip net (mesh size: 90 μm), which was operated both in free volume of water (circa 10 m) and in areas with water vegetation (circa 6 m). Water beetles were collected by the combined use of D‐framed pond net and a kitchen sieve in as many microhabitats as possible. Most water beetle species are found in the water vegetation along the shoreline, and the sampling effort was concentrated in these habitats. The collected material was transferred to a plastic tub, and individuals of water beetles were picked out manually until the sample included as few species represented by one individual as possible. Dependent on the size of the water body, the sampling of water beetles lasted from 1 to 2 hours per water body.

Land use information was available in digital format, AR5 – mapped in the scale of 1:5,000 (The Norwegian Forest and Landscape Institute, www.skogoglandskap.no). In an analysis of land‐cover changes over the past 40‐50 years, we have used this dataset as a background. The defined polygons have been redefined by interpretation of aerial photos from the early 1960s and 2000. Land‐cover/land use changes have been recorded in five classes: no change, regrowth, newly cultivated, newly built, and other technical encroachments in six specified parts of the study area (1‐6 in Figure 1).

All samples of crustaceans were analysed for presence/absence of species and dominance. Patterns in the distributions were summarised by a detrended correspondence analysis (DCA) (Hill, 1979; Hill and Gauch, 1980) using the program CANOCO (Ter Braak and Smilauer, 1998) with down weighting of rare species. By using pelagic and littoral samples as input for the zooplankton analysis, we expected the first axis to separate between the two main habitats.

A total of 57 species of crustaceans were found in the 20 investigated water bodies on the floodplain at Flisa in Hedmark County (Figure 2). The total recorded numbers of species of cladocerans and copepods were 36 and 21, respectively. Comparable studies in Hedmark County, including samples of littoral crustaceans, have not shown the same high number of species. A study which included ten large lakes documented 54 species, and the most intensively investigated lake in the county (Lake Atnsjøen) has a record of 37 species (Walseng, 1990; Halvorsen et al., 2004). The average number of zooplankton species per water body in the present study was 22.1 (SD=4.87; range: 14 (12 cladocerans and two copepods) to 32 (24 cladocerans and eight copepods)). A comparable study at the Ringebu floodplain in the river Gudbrandsdalslågen (Oppland County, 140 km NW) recorded an average of 15.6 species per water body (Schartau et al., 2005). The average number of species per locality in other studies in Hedmark (which also included littoral species) was 17.6. Ten new species for the county of Hedmark were recorded during the present study (Table II). These species are relatively infrequently found in Norway (in less than 5 per cent of investigated localities), but none of them are specified in the 2010 Norwegian Red List of Species (Table II). The most infrequently found species in Norway recorded in the present study was the copepod Paracyclops poppei.

Despite considerable variation amongst localities, a DCA ordination plot shows that the localities at the Flisa floodplain are relatively unique with regard to both the number and composition of species compared with other investigated localities in Hedmark County (n=83) (Figure 3). The plots representing the Flisa localities are distributed to the right in Figure 3, and the largest differences in the zooplankton communities were between pelagic samples in lakes and the Flisa localities (more than three units deviation on axis 1 signify no common species). This difference was probably partly due to different habitats sampled (pelagic vs pelagic/littoral samples), but also the comparable studies showed that the crustacean zooplankton community found on the Flisa floodplain represented something unique in the region of Hedmark County. Hence, at a regional scale, both the total number of species found within the floodplain at Flisa, and the average number of species per water body were high. Prior to the present study, the number of recorded zooplankton species in the county of Hedmark was 83 (based on data from 93 localities; Walseng, 1990; Walseng et al., 2011). Including the present study, a total of 93 zooplankton species are now recorded in Hedmark County (57 species of cladocerans and 36 species of copepods).

The different water bodies ranked from high to low by number of crustacean species showed that the most species rich water body had 55 per cent of the recorded species in this study, and that four and seven of the most species rich water bodies housed >80 per cent and >90 per cent of the recorded species (Figure 2).

There were no significant correlations between the number of crustacean species recorded and “area of water body” (Figure 4a, Pearson r=0,365, p=0.137) or “distance from main channel” (Figure 4b, Pearson r=−0.362, p=0.140). The average number of crustacean species recorded was higher in water bodies localised “behind flood defences” than in water bodies “not affected by flood defences” (Figure 4c, student t‐test: t=2,529, 16 df, p=0.022).

A total of 77 water beetle species was recorded during the present study. The average number of water beetles species per waterbody was 25 (SD=6.3; range: 16‐44), and the average number of individuals sampled per water body was 264 (SD=197; range: 66‐932). The following nine families were recorded: Gyrinidae (whirligig beetles, nine of 11 recorded species in Norway (rsN)), Haliplidae (crawling water beetles, nine of 13 rsN), Noteridae (burrowing water beetles, one of two rsN), Dytiscidae (predaceous diving beetles, 49 of 132 rsN), Hydraenidae (minute moss beetles, two of 13 rsN), Helophoridae (four of 21 rsN), Hydrochidae (one of three rsN), Hydrophilidae (water scavenger beetles, ten of 29 rsN) and Chrysomelidae (leaf beetles, two of 22 aquatic rsN). A total of 112 species of these nine families are previously described in this region (Dolmen and Aagaard, 1996). The present study recorded seven species on the Norwegian Red List of Species (Kålås et al., 2010) and eight first records of species in this region of Norway (Table II). The different water bodies ranked from high to low number of water beetle species showed that the most species rich water body had 57 per cent of the recorded species in this study, and that four and nine of the most species rich water bodies housed >80 per cent and >90 per cent of the recorded species, respectively, (Figure 2).

There were no significant correlations between the number of water beetle species recorded and “area of water body” (Figure 4d, Pearson r=−0.286, p=0.249) or “distance from main channel” (Figure 4e, Pearson r=−0.265, p=0.288). The average number water beetle species recorded in water bodies localised “behind flood defences” and in water bodies “not affected by flood defences” was not significantly different (Figure 4f, student t‐test: t=0.486, 16 df, p=0.631).

Most of the areas studied for land cover and land use changes have no major change recorded (Table III). The major change recorded is regrowth which affects some 10‐20 per cent of the area. Cultivation of new agricultural land covers about 10 per cent and newly built land somewhat more than 1 per cent of the area.

The high regional diversity of crustacean zooplankton and water beetles in water bodies at the riverine floodplain at Flisa, support the general consensus of floodplains being centres of biodiversity (Tockner and Stanford, 2002; Ward et al., 1999). The survey extended the list of recorded water beetle and crustacean zooplankton species in Hedmark County by eight and ten, respectively. Even though both zooplankton and especially water beetles are surveyed to a somewhat small extent in the region, both the total number, the number of “first records”, and the presence of several red listed species (Kålås et al., 2010) in this study imply that the Flisa floodplain fauna is of great importance regarding biodiversity. Further, our study shows that the water bodies on the Flisa floodplain are relatively unique with regard to composition of species compared with other investigated localities in Hedmark county.

Our results indicate that zooplankton species richness was higher in less flood‐affected areas (e.g. behind flood mitigation measures and road fillings). This is consistent with other studies finding higher zooplankton species richness with decreasing degree of flooding (Baranyi et al., 2002; Medley and Havel, 2007). Even though flooding and a high degree of connectivity may introduce new species, intense flooding can wash out entire populations (Medley and Havel, 2007). However, our analyses on the degree of flood affection is based on maps with areas that are subjected to flooding with ten year frequencies. Thus, there are uncertainties in the finer scale flood dynamics (shorter flood frequencies) and how efficient the flood mitigation measures really are. Baranyi et al. (2002) and Medley and Havel (2007) also found that flood frequency was a significant factor structuring floodplain zooplankton communities. This emphasises the importance of regarding floodplain biodiversity holistically. For zooplankton, the seven water bodies with the highest number of species accounted for 90 per cent of the total number of species. Corresponding numbers for water beetles were achieved in the nine most species rich water bodies. This implies that one must sample numerous floodplain ponds with varying characteristics (e.g. different degree of connectivity/flood frequency) to approach the total number of species.

Floodplain water bodies differ in their degree of connectivity to the main river channel, and in their successional stages (Ward and Stanford, 1995a). The variation in the design of the water bodies (Ward and Stanford, 1995a), along with the spatial and temporal heterogeneity from variation in hydrological connectivity makes riverine floodplains high in biological diversity (Amoros and Bornette, 2002). In an undeveloped state, the water bodies on the Flisa floodplain probably differed to a greater extent in their degree of connectivity to the main river channel. However, flood mitigation measures, roads, and other constructions have to a large extent altered the natural flood and disturbance regime, and it is estimated that flood mitigation measures impede flooding of 66 per cent of flood exposed areas during a ten‐year flood at Flisa (Haddeland, 2001). This disruption of lateral connectivity has probably reduced the diversity of hydrological disturbance regimes, which is of major importance in maximising and preserving the total biodiversity across a floodplain (Ward et al., 1999).

The Flisa floodplain is heavily influenced by flood mitigation measures and new flood defences were observed during the period from 1960 to 2000. The studied landscape has, however, not been under a heavy pressure of land use changes due to infrastructure and urbanisation. Changes in agricultural practices from a high level of husbandry (milk production) to extensive grain production has resulted in less grazing and overgrowing of former grazing lands. A major re‐growth of bushes and forest is seen along the edges of both the main river Glomma and the tributary river Flisa. This is probably partly due to less grazing, but may also be a result of the stop in the traditional use of the rivers for log‐driving, more stable water discharge, and/or a warmer climate. Despite the extensive flood mitigation measures and additional land use (e.g. agricultural development), the species richness of zooplankton and water beetles on Flisa's floodplains is regionally high. However, on a larger temporal scale, if the hydrological forces resetting successional sequences resume being absent, the diversity of both water body characteristics and biodiversity will probably homogenise and be reduced (Ward and Stanford, 1995a, b).

Scenarios on climate change point towards a more complex flooding pattern (Benestad and Haugen, 2007; Alcamo et al., 2007). The future climate may thus result in more frequent larger floods which may work to re‐establish the connectivity (e.g. where it is disconnected by levees) between the river channel and the floodplains (Poff, 2002). Larger floods might also result in construction of more extensive flood mitigation measures. Ironically, flood mitigation structures like flood walls and levees, tend to increase the magnitude of floods by decreasing the water storage capacity of rivers by preventing the lateral movement of water across floodplains and wetlands (Sparks, 1995).

Flooding may destroy property and threaten human life. Human use of floodplains has required flood mitigation measures, yet it is the dynamic nature of the floodplains that originally made these areas desirable, e.g. for agricultural exploitation, and at the same time created hotspots regarding biodiversity. These contrasting views confront local, regional, and national authorities when faced by the outlook for more frequent, less predictable and more extreme floods in the future.

Environmental adjustment of traditional flood protected river stretches and floodplains, and the use of more environmentally sensitive flood protection measures, have so far been applied to a small extent in Norway (Østdahl and Taugbøl, 1999). The implementation of the Water Framework Directive (European Parliament and Council of the European Communities, 2000) and the fact that floodplains are hotspots for preserving biodiversity call for a more holistic approach regarding watershed and thus floodplain management. Further, in connection to the recently implemented act relating to the management of biological, geological, and landscape diversity (Ministry of the Environment, 2009), prioritised ecosystems are to be selected. Since floodplains are threatened ecosystems in Norway (Johnsen et al., 2011), the future preservation of still intact and functional floodplains and floodplain areas would benefit from being among these prioritised ecosystems.

Reestablishment of the lateral connectivity of rivers by development of non‐structural flood management policies and/or by moving levees and dikes farther away from the main channel (retired levees and dikes) are imperative to preserve and restore intact floodplains in Norway.

Using multiple criteria analysis when evaluating trade‐offs between price and non‐priced hydro‐morphological impacts of flood mitigation projects may contribute to a more socially optimal approach to flood prevention (Barton and Dervo, 2009). Rauken and Kelman (2010) point out that the economic system in Norway leave municipalities with few economic incentives to leave flood zones undeveloped or restore the natural dynamic of floodplains. Expropriation of land and development of new insurance and compensation schemes may be important policy instruments to “give rivers room” in the future.

Poff (2002) argued that developing and implementing of non‐structural flood management policies based on ecological principles would benefit both riverine ecosystems as well as human society. Our study documented the Flisa floodplain as a regional hotspot for preserving biodiversity. Further research should emphasise the underlying mechanisms explaining the high number of species and the variation of species number amongst the different water bodies, to suggest alternative management and flood mitigation measures to restore the dynamic nature of floodplains. It may be unrealistic to restore or re‐establish the ecological functionality of entire floodplains. However, in addition to conserving functional floodplains, different areas within developed floodplains should be classified and prioritised to work out management strategies to partly restore the lateral connectivity.

Jon Museth is a Senior Research Scientist at the Norwegian Institute for Nature Research (NINA). His main research interests are human impact on aquatic ecosystems and fish migrations, e.g. hydropower development, mitigation measures, fishing, and introduced species. Jon Museth is the corresponding author and can be contacted at: jon.museth@nina.no

Stein I. Johnsen is a Research Scientist at NINA. He holds an MSc in Aquatic Ecology from the Norwegian University of Life Sciences (UMB). His main research interests are in aquatic ecology and management of freshwater fish and crayfish.

Bjørn Walseng is Research Scientist at NINA. His main research interest is crustacean ecology and environmental monitoring.

Oddvar Hanssen is Senior Technician at NINA. His special field of interest is entomology and taxonomy.

Lars Erikstad is a Senior Research Scientist at NINA and obtained a Candidatus realium in Physical Geography at the University of Oslo in 1979. He has interests in nature conservation with special emphasis on conservation of geological heritage, multidisciplinary landscape analysis, GIS modeling, and environmental impact assessments.