Freshwater Biology (2002) 47, 517–539

Riverine landscape diversity

J.V.WARD,K.TOCKNER,D.B.ARSCOTTandC.CLARET Department of Limnology, EAWAG/ETH, Duebendorf, Switzerland

SUMMARY 1. This review is presented as a broad synthesis of riverine landscape diversity, beginning with an account of the variety of landscape elements contained within river corridors. Landscape dynamics within river corridors are then examined in the context of landscape evolution, ecological succession and turnover rates of landscape elements. This is followed by an overview of the role of connectivity and ends with a riverine landscape perspective of biodiversity. 2. River corridors in the natural state are characterised by a diverse array of landscape elements, including surface waters (a gradient of lotic and lentic waterbodies), the fluvial stygoscape (alluvial aquifers), riparian systems (alluvial forests, marshes, meadows) and geomorphic features (bars and islands, ridges and swales, levees and terraces, fans and deltas, fringing floodplains, wood debris deposits and channel networks). 3. Fluvial action (erosion, transport, deposition) is the predominant agent of landscape evolution and also constitutes the natural disturbance regime primarily responsible for sustaining a high level of landscape diversity in river corridors. Although individual landscape features may exhibit high turnover, largely as a function of the interactions between fluvial dynamics and successional phenomena, their relative abundance in the river corridor tends to remain constant over ecological time. 4. Hydrological connectivity, the exchange of matter, energy and biota via the aqueous medium, plays a major though poorly understood role in sustaining riverine landscape diversity. Rigorous investigations of connectivity in diverse river systems should provide considerable insight into landscape-level functional processes. 5. The species pool in riverine landscapes is derived from terrestrial and aquatic communities inhabiting diverse lotic, lentic, riparian and groundwater habitats arrayed across spatio-temporal gradients. Natural disturbance regimes are responsible for both expanding the resource gradient in riverine landscapes as well as for constraining competitive exclusion. 6. Riverine landscapes provide an ideal setting for investigating how complex interactions between disturbance and productivity structure species diversity patterns.

Keywords: biodiversity, connectivity, floodplains, landscape ecology, natural disturbance

framework of landscape ecology, defined by Turner Introduction (1998) as the study of interactions between spatial River corridors consist of a dynamic mosaic of spatial patterns and ecological processes in the context of elements and ecological processes arrayed hierar- spatial heterogeneity across a range of scales. This chically. As such, they fit comfortably within the review of riverine landscape diversity focuses on pattern and process at habitat, floodplain, and corri- Correspondence: Prof J. V. Ward, Department of Limnology, dor spatial scales across seasonal and successional EAWAG/ETH, Ueberlandstrasse 133, CH-8600 Duebendorf, time scales. River corridors are addressed as integra- Switzerland. E-mail: [email protected] ted ecological systems, as opposed to the traditional

Ó 2002 Blackwell Science Ltd 517 518 J.V. Ward et al. view of corridors as conduits between similar even the most important work. Detailed treatments of elements within the landscape. the fauna and flora of riverine landscapes are presen- Consideration is primarily restricted to natural ted elsewhere (Ward et al., 2000a; Adis & Junk, 2002; conditions. It is our contention that the remarkable Amoros, 2002; Robinson, Tockner & Ward, 2002). degree of spatio-temporal heterogeneity characteris- ing riverine landscapes has been masked in much of the world by a long history of river engineering (see Landscape elements of river corridors Petts, Moller & Roux, 1989; Benke, 1990; Dynesius & The river corridor Nilsson, 1994). Floodplain reaches, which exhibit the highest heterogeneity in the natural state, have been River corridors are linear features of the landscape, the most severely altered (e.g. Ward & Stanford, 1995), structured along ribbons of alluvium extending from resulting in a distorted perception of patterns and the headwaters to the sea. Narrow canyon-con- processes in riverine landscapes. An accurate com- strained reaches typically alternate with alluvial prehension of spatio-temporal heterogeneity in the floodplains, like ‘beads on a string’ (Fig. 1; Stanford unaltered state is crucial for a holistic understanding & Ward, 1993). In constrained reaches, the hillslopes of the structure and function of river ecosystems and descend abruptly to a single-thread channel bordered is essential for successful protection and restoration. by a narrow band of riparian vegetation. Alluvial As stated by Bornette et al. (1998a), ‘The sustainable deposits consist of only a thin layer of sediment conservation of diversity in riverine wetlands covering the bedrock and hydrological exchange is implies knowledge of the basic geomorphological predominantly unidirectional (downstream). and ecological processes that interplay at the land- The floodplain reaches (the ‘beads’) of river corri- scape scale.’ dors, in contrast, are expansive, with multiple chan- In the natural state, riverine landscapes exemplify nels and deep alluvial deposits. Alluvium is the the ‘new paradigm in ecology’ (sensu Talbot, 1996), in matrix upon and within which the landscape elements which ecological systems are widely recognised as are embedded. Hydrological exchange occurs along non-deterministic, open systems in continual states of longitudinal, lateral and vertical dimensions. Terraces flux, rather than internally regulated, homeostatic are former floodplain surfaces that were formed when systems exhibiting equilibrium conditions. Yet, the river was flowing at a higher level. Nanson & despite their highly dynamic nature, riverine land- Croke (1992) developed a typology, in an attempt to scapes provide predictable ecological conditions. relate the complex landforms of floodplains to two Although individual landscape features exhibit high key variables: stream power (the ability of a river to turnover, largely as a function of interactions between entrain and transport sediment) and sediment cohe- fluvial dynamics and successional phenomena, their siveness (the erosional resistance of the alluvium). relative abundances in the river corridor tend to Three basic types of floodplains were identified: (1) remain constant over ecological time. Disequilibrium floodplains (high energy, non-cohe- This review begins with an account of the diverse sive) which erode in response to extreme episodic array of landscape elements encompassing surface flow events. Such floodplains tend to be located in waters, alluvial aquifers, riparian systems and steep headwaters where channel migration is con- geomorphic features. Then landscape dynamics with- strained by coarse substratum and narrow valleys. in river corridors are examined in the context of Floodplain construction is mainly by vertical accre- landscape evolution, ecological succession and turn- tion. Dominant landforms include boulder levees, over rates of landscape elements. This is followed by sand and gravel splays, back channels and scour analyses of connectivity between landscape elements. holes. (2) Equilibrium floodplains (medium energy, Finally, a riverine landscape perspective of biodiver- non-cohesive) are formed by regular flow events in sity is presented. Special attention is given to flood- broad valleys. Such floodplains are thought to be in plains throughout the paper because of the high level dynamic equilibrium with the flow regime. Fluvial of landscape diversity they exhibit. Riverine land- energy from extreme floods is dissipated as flood- scape diversity is a broad topic with a vast literature; waters overtop the channel banks and disperse across in this review it is possible to include only a portion of an expansive surface. Floodplain construction

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 517–539 Riverine landscape diversity 519

Fig. 1 Idealised configuration of a river corridor as an alternating sequence of constrained and floodplain reaches. Pre- dominant hydrological exchange path- channel ways are indicated for longitudinal (horizontal arrows), lateral (oblique arrows) and vertical (vertical arrows) dimensions. involves lateral point bar accretion or braid channel deal with their permanence, connectivity and succes- accretion. Channels may be braided, anastomosed or sional trajectories. meandering. Characteristic landforms include aban- A diversity of lotic, semi-lotic and lentic waterbod- doned channels, bars and islands, oxbows, meander ies occur within river corridors. In dynamic river scrolls and backswamps. (3) Low-gradient floodplains ecosystems these three classes of surface waters (low energy, cohesive) are also formed by regular flow merely represent different positions along an inunda- events in broad valleys, but with channels laterally tion continuum and they are typically interconnected stabilised by erosion-resistant banks of fine cohesive during floods. During the dry phase, however, each alluvium. Floodplain construction involves mainly has distinctive attributes, as described below and vertical accretion of fine sediment deposits and illustrated in Fig. 3. occasional channel avulsion. Channels may be brai- Lotic waters flow within the main channel and the ded, anastomosed or meandering. Landforms include side arms having both upstream and downstream levees, islands, splays and backswamps. connections to the main channel. In addition, spring- Ideally, these three classes of floodplains occur brooks originate as upwelling ground water from the sequentially from the headwaters to the lower reaches alluvial aquifer beneath the river corridor (alluvial of a given river corridor, corresponding in general to springs) and from hillslope aquifers that emerge along the sediment production, transfer and storage zones the edge of the river corridor (hillslope springs). Surface of Schumm’s (1977) ‘fluvial system’. Many complica- drainage from the hillslope forms tributaries which ting factors, however, such as glacial activity, may may flow some distance within the corridor before disrupt the idealised sequence. forming a confluence or sinking into the alluvium. Semi-lotic waters include abandoned braided chan- nel segments and dead arms that retain a connection Surface waterbodies to the main channel only at their downstream ends. The major aquatic, semi-aquatic and terrestrial ele- Both typically require only minor flooding to recon- ments of riverine landscapes may be grouped within nect with the main channel, thereby frequently alter- four interactive categories (Fig. 2), one of which is nating between lentic and lotic conditions. Even when surface waterbodies. Here we address the diversity of disconnected, inflowing ground water may sustain a surface waters in river corridors; subsequent sections slight current within these waterbodies.

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 517–539 520 J.V. Ward et al.

Fig. 2 Major spatial elements of riverine landscapes.

Lentic waters within the river corridor are of both fluviatile and non-fluviatile origins. Lakes of non- fluvial origin that may be associated with riverine landscapes form by different mechanisms, including damming stream/river courses by lava flows, land- slides, moraines or beavers (see Hutchinson, 1957; for a detailed account of the origin of lake basins). The fluviatile action of running water forms lakes (and other lentic waterbodies) in various ways. Lateral lakes form when entering tributaries are dammed by sediment deposited as levees along the main channel during floods. Tributaries may also be blocked by wood debris to form riparian lakes (Triska, 1984). Alluvial fans from steep mountain tributaries may Fig. 3 Surface waterbodies and basic geomorphic features of an idealised river corridor in a braided-to-meandering transition dam the main channel to form lakes, although most zone. L ¼ lateral or riparian lake; BA ¼ bar; IS ¼ island; are soon drained by erosion of the fan. Fluviatile plesio ¼ plesiopotamal/plesiorhithral (abandoned braids); action originating within the floodplain also forms palaeo ¼ palaeopotamal/palaeorhithral (abandoned meanders); lentic waterbodies. As the main channel migrates para ¼ parapotamal/pararhithral (dead arms). across the river corridor meander loops are aban- doned to form oxbow lakes. Unlike abandoned braids, terrace, where erosion scours the floodplain surface which are smaller, shallower and without surround- during floods. ing levees, oxbow lakes are less frequently reconnect- A precise quantification of surface waters was ed to the main channel, thereby developing a more conducted in six geomorphic reaches along the cor- truly lentic character. Ponds and marshes occur in the ridor of an Alpine river, the Fiume Tagliamento, Italy depressions between successive meander scrolls, (Arscott, Tockner & Ward, 2000). Using digitised where relict levees, abandoned channels and point maps and aerial photographs in concert with field bars form a ridge and swale topography on the surveys, four major surface water habitats were floodplain (Nanson & Beach, 1977). Marshes and distinguished: surface-connected channels (SC), allu- swamps often occur adjacent to the hillslope or vial channels (AC), tributary channels (TC) and

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 517–539 Riverine landscape diversity 521 isolated standing waters (ISO). The first three habitats headwaters. The Shannon diversity index and Simp- included primary, secondary and tertiary branches son’s index of dominance were employed to examine plus backwaters (e.g. SC1, SC2, SC3, SCB), resulting in aquatic habitat heterogeneity along the river corridor a total of 13 habitat types (Fig. 4). The total area of (Fig. 4b), with the habitat types representing ‘species’ surface waters per river km exhibited a unimodal and the relative abundance of habitats representing pattern, with the highest value in the island- ‘species abundance’ values. Shannon diversity exhi- braided lower reach and the lowest in the constrained bited a unimodal pattern with maximum diversity in

Fig. 4 Quantification of surface waters in six geomorphic reaches along the corridor of the Fiume Tagliamento, Italy. Aquatic habitat area (a) for 13 habitat types described in the text; diversity/dominance of surface waterbodies (b); and a map of aquatic habitats in a 3-km long segment of the island-braided mid reach (c). Habitat abbreviations: SC ¼ surface-connected channels; AC ¼ alluvial channels; TC ¼ tributary channels; ISO ¼ isolated standing waters. The first three habitats included primary, secondary, tertiary and backwater (B) subtypes (modified from Arscott et al., 2000).

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 517–539 522 J.V. Ward et al. the island-braided lower reach; dominance exhibited where rivers empty into a lake or the sea, the channels the inverse pattern as expected. The remarkable forming a distributive network in which the streams degree of habitat heterogeneity along the Fiume divide and become increasingly smaller. Tagiamento is attributed to its morphologically intact Deposits of large woody debris (LWD) are common river corridor, forested catchment and natural flood landscape elements in natural rivers with forested regime (Ward et al., 1999). catchments. In a survey of LWD retention in a river corridor, Gurnell et al. (2000) estimated wood storage values of c.1tha–1 for open gravel reaches with Geomorphic features single-thread channels, 6 t ha–1 for open gravel rea- The geomorphic features of floodplains reflect com- ches with multiple-thread channels and 80 t ha–1 on plex interactions between climate, catchment geology, islands. Wood debris plays major geomorphologic topographic relief and fluvial dynamics, mediated by and ecological roles in riverine landscapes (Keller & vegetation. In this section, we present geomorphic Swanson, 1979; Maser & Sedell, 1994). Moreover, features largely as spatial landscape elements, reser- LWD may serve as nuclei for the development of ving consideration of the dynamics for a later section. vegetated islands (Abbe & Montgomery, 1996), the Alluvial channel networks have been traditionally presence of which is believed to contribute directly classified into four types (straight, meandering, brai- and indirectly to landscape heterogeneity (Gurnell ded and anastomosed), with recognition that these et al., 2001; Ward et al., 2000b). types are really part of a continuum. Straight channels have a sinuous thalweg and alternate bars that slowly Fluvial stygoscape move downstream. Meandering rivers have sinuous single-thread channels that erode the concave banks, The term fluvial stygoscape refers to the subterranean causing the lateral and downstream migration of the extension of the riverine landscape (Ward, 1997). channel network. Point bars that form on the convex Aquifer systems (water-saturated alluvium) associ- bends of meanders build the floodplain by lateral ated with rivers, traditionally regarded as rather accretion. Braided rivers consist of multiple shifting homogeneous, are now known to exhibit heterogen- channels that are highly unstable if separated by eity at multiple scales and may be highly interactive sand/gravel bars devoid of vegetation (bar braided), with surface waters (Gibert et al., 1994). A highly but less so if the bars have been stabilised by specialised fauna inhabits the water-filled interstices vegetation (island braided). Anastomosed rivers may between sediment particles (Ward et al., 2000a). The be regarded as an extreme form of island-braiding, the fluvial stygoscape is a three-dimensional mosaic, the prominent landform consisting of large, relatively spatial structure of which is a function of sediment permanent vegetated islands. particle size, particle size heterogeneity, pore size, Other geomorphic features of riverine landscapes porosity, pore geometry and hydraulic conductivity, include levees, crevasse splays, alluvial fans and and the distribution of preferential flow paths, buried deltas (Allen, 1965). Levees are ridges of sediment organic matter, and chemical and thermal gradients along channels formed when a river overtops its within the alluvial aquifer. Ground-penetrating radar banks and deposits the coarsest part of its load as may be used to identify subterranean structures such water velocity abruptly decreases. In large rivers as fills of former channels, ancient bars and interca- levees may be more than 5 m above the floodplain lations of buried organic debris (Fig. 5). surface. Crevasse splays are sediment deposits that The hyporheic zone, the ecotone between surface form when sediment-laden water breaches the levee waters and the alluvial ground waters of the phreatic and spreads over the floodplain dropping its load. zone, is a dynamic boundary zone characterised by Alluvial fans are depositional structures that form nutrient transformations, steep gradients (physical/ where steep tributaries enter a river corridor. They chemical/biological) and hydrologic exchange commonly occur in arid climates where intermittent (Castany, 1985; Brunke & Gonser, 1997; Boulton et al., tributaries carry large sediment loads associated with 1998). As the name implies, hyporheic zones occur violent rainstorms or in mountainous topography beneath running water channels; however, the hypo- where tributaries carry snowmelt runoff. Deltas form rheic zone may also extend some distance (up to

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 517–539 Riverine landscape diversity 523

Fig. 5 Fluvial stygoscape, as determined from ground penetrating radar, of the alluvial corridor of the lower Necker River, Switzerland (modified from Nae- geli, Huggenburger & Uehlinger, 1996). 1 ¼ Water surface; 2 ¼ a channel point bar; 3 ¼ point bar lateral accretion; 4 ¼ sand and intercalation of organic debris; 5 ¼ boulders forming the core of an ancient bar; 6 ¼ fills of former channel; 7 ¼ bedrock; 8 ¼ hillslope. kilometres) beyond the channel margin beneath the in the upper reaches of Central European river floodplain (Stanford & Ward, 1988) or be missing corridors (Pinay et al., 1990; Statzner & Kohmann, altogether under special conditions, such as where a 1995; Schnitzler, 1997). As rivers enter the foothills headwater stream flows directly on bedrock. and develop a wider floodplain, alders begin to decline in relative abundance, willows increase, pop- lars (Populus) appear, and an alluvial hardwood forest Riparian systems dominated by oak (Quercus) and elm (Ulmus) occupies The vegetation of river corridors forms a complex the less frequently inundated outer portions of the mosaic in response to gradients of climate, inundat- corridor. In the lowlands, willows and poplar pre- ion/soil moisture, disturbance and nutrients (Gregory dominate on frequently inundated sites, with an et al., 1991; Malanson, 1993; De´camps, 1996; Naiman alluvial forest dominated by elm and oak occupying &De´camps, 1997). Precipitation during the dry phase the remainder of the expansive floodplain. Different is a critical determinant of broad scale patterns of portions of the altitudinal gradient are characterised floodplain forest development. As stated by Junk, by different species of both Salix and Alnus. Riparian Bayley & Sparks (1989), ‘when local precipitation at species differ markedly in the ability to sprout low water is high, floodplains are forested, e.g. in the vegetatively from pieces of live wood deposited on middle and upper Amazon, Zaire and Mississippi sediments following flood recession (Gurnell et al., rivers. Conversely, when local precipitation is low, 2001). Because vegetative sprouting is one mechanism savannas with gallery forest develop, e.g. in the by which alluvial islands form, the altitudinal distri- floodplains of the lower Nile, Zambezi, and Volta bution of woody species influences landscape-level rivers.’ Within a catchment, elevation exerts primary processes. influence on controlling variables, and therefore on Along the lateral gradient from the channel to the riparian vegetation composition and structure, at uplands, river corridors with expansive floodplains three scales: (1) as altitude decreases along the river’s may contain a series of riparian zones, primarily course from mountain headwaters to sea level, (2) as reflecting species-specific responses to soil moisture/ elevation increases along the lateral gradient from the oxygenation, sediment deposition, the frequency and main channel to the uplands, and (3) as a function of duration of inundation, and the erosive action of local changes in elevation reflecting topographic flooding. An example of woody vegetation zones features on the floodplain surface, such as levees, along a lateral floodplain gradient is illustrated in islands, ridges and swales. Fig. 6. Pautou (1984) developed a typology for alluvial Under natural conditions, alders (Alnus) and wil- forests of the Upper Rhone based on their position lows (Salix) dominate the woody riparian vegetation along three axes: (1) site elevation above water level,

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 517–539 524 J.V. Ward et al.

Fig. 6 Schematic lateral transect across a braided reach of the alluvial floodplain of the upper French Rhone, showing the zonation of the dominant woody veget- ation (modified from Pautou, 1984).

(2) sediment particle size composition/concentration cannot withstand even brief periods of inundation of organic matter in the A1 horizon, and (3) the and are confined to high terraces. seasonal regime, including precipitation patterns, Riparian vegetation moderates myriad ecological which determines the time of year a given floodplain processes in river corridors by influencing tempera- unit is inundated. ture and light regimes; producing organic detritus The idealised lateral gradient at the floodplain scale (leaf litter, woody debris); by routing water and is, however, complicated by local topographic fea- sediment; by structuring the physical habitat at tures (Brinson, 1990). High levees, composed of several scales; by providing a substrate for biological coarse-grained well-drained soils, often support activity and habitat/cover for aquatic, amphibious riparian plants unable to tolerate conditions at lower and terrestrial animals. Riparian plants serve as levels on the floodplain. The ridge and swale topo- roughness elements during periods of inundation, graphy formed by relict point bars, levees and thereby influencing the hydrogeomorphological pro- abandoned channels, produces an undulating surface cesses that structure the riverine landscape (Johnson, across the floodplain. Depressions are colonised by 2000). species adapted to long hydroperiods, whereas the levees and ridges may contain species that also occur Landscape dynamics in river corridors in mesic upland sites. Woody alluvial species collec- tively occupy a broad range of flood tolerance Contemporary riverine landscapes reflect processes (Schnitzler, 1997). Salix alba L. tolerates up to 300 days occurring over a wide range of time scales (Table 1). of inundation; Quercus robur L. and Ulmus minor Mill. The following brief discussion of landscape evolution tolerate up to 151 days; Fraxinus excelsior L. 102 days; is intended as a prelude to the treatment of dynamic Acer campestre L. and Tilia cordata Mill. 13 days; patterns and processes occurring in ecological time whereas Acer pseudoplatanus L. and Fagus sylvatica L. (seasonal to centennial).

Table 1 Time scales of some major phenomena that structure patterns and processes in riverine landscapes

Time scale Phenomenon

Seasonal Spates, flow pulses, channel expansion/contraction Annual Flood pulse, seedling establishment, animal migration, reproduction, shallow ground water exchange Decadal Drought cycles, episodic events (extreme floods, debris flows), lateral channel migration, channel avulsion, island formation, channel abandonment Centennial Floodplain formation, hydrosere and riparian succession, deep ground water exchange Millennial Terrace formation, glaciation, climate change, sea level fluctuation, orogeny

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 517–539 Riverine landscape diversity 525 in a youthful stage characterised by numerous Landscape evolution on-channel lakes, falls and rapids (Fig. 7a). Eventually The action of running waters (erosion, transport, vertical erosion drains the lakes and reduces the falls. deposition) is the predominant agent of landscape Moreover, the transport capacity of the high-gradient evolution. Hydraulic action in concert with corrosion stream exceeds the available sediment load, resulting and corrasion erodes mineral matter from the banks in additional bed degradation and the creation of and bed. These materials are transported by the deep gorges (Fig. 7b). As weathering and mass wast- current and deposited at downstream locations (with- ing act on the canyon walls, the corridor begins to in the channel, on the floodplain, in standing water- widen (Fig. 7c). As tributary corridors develop in the bodies). same manner, the sediment supply to the main On a newly formed landmass not previously corridor increases. As the sediment load increases subjected to fluvial denudation, the river corridor is and the gradient decreases, a time is reached when the

Fig. 7 Landscape evolution of a river corridor (modified from Strahler, 1963, Copyright Ó Arthur N. Strahler, Used by permission.). (a) (youthful stage) to (e) (advanced stage) represent the ontogenesis of a river corridor (see text).

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 517–539 526 J.V. Ward et al. sediment supply matches the transport capacity, 1996), which undoubtedly reflects, at least in part, resulting in a graded channel. This ends the period of anthropogenic suppression of natural dynamics. rapid downcutting and initiates the formation of the Avulsion refers to the process by which a channel floodplain, as the channel undergoes lateral migra- segment is abandoned by an abrupt change in the tion. In the advanced stage of landscape evolution water’s course, such as when a meander loop is cut off (Fig. 7e), the river corridor consists of a complex low- to form an oxbow lake. In braided river networks gradient floodplain containing an expansive alluvial avulsion is a major process of channel dynamics, aquifer. The idealised landscape ontogenesis por- involving infilling of active channels by sediment and trayed over geological time in Fig. 7 is analogous to subsequent diversion of the main flow path (Leddy, the spatial landscape transformations occurring along Ashworth & Best, 1993). Avulsion and lateral channel river corridors arising in youthful mountains. migration play major roles in the riverine landscape by structuring topographic features of the floodplain, influencing successional trajectories, and determining Ecological time scales the turnover rates of landscape elements. Rivers are interactive four-dimensional systems (Ward, 1989) characterised by high levels of natural Turnover. Despite high levels of disturbance and disturbance. The disturbance regime of riverine land- corresponding turnover rates, the relative abundance scapes encompasses debris flows, fire, disease, and of landscape elements in a natural river corridor may animal activities among other factors. Here we focus remain relatively constant over ecological time (e.g. on fluvial action as the primary driver of landscape Kollmann et al., 1999), corresponding to the ‘shifting- dynamics. Fluvial action influences the riverine land- mosaic steady state’ model of Bormann & Likens scape directly, by structuring channel networks and (1979). For example, a landscape feature destroyed by other geomorphic features, and indirectly, by influen- localised shear forces is typically re-formed by cing successional phenomena. Although landscape aggradation processes elsewhere in the corridor. To dynamics within river corridors are more intense be more specific, floods eliminate islands at some during overbank flood events (the ‘flood pulse’ of locations while initiating the formation of new Junk et al., 1989), flow pulses below bankfull dis- islands at other locations. Therefore, a landscape charge also have landscape-level implications (Tock- element that appears to be increasing (or decreasing) ner, Malard & Ward, 2000). at a fine scale of resolution, may in fact be in a steady state if viewed at a broader scale. Fig. 8 illustrates Channel change. Channel migration and avulsion are aquatic habitat turnover during a late summer– two phenomena responsible for inducing channel autumn period in the bar-braided geomorphic reach change. Along meandering reaches lateral migration of an Alpine river with a natural flood regime. The of channels tends to be unidirectional, as is apparent high rate of turnover (53%) that occurred during an from the zonal array of riparian vegetation with average flood season elucidates the high level of increasing distance from the active channel (Kalliola landscape dynamics that must have characterised et al., 1992). In contrast, the migration pattern in Alpine rivers in the pristine state. Of course different anastomosed reaches is patchy rather than unidirec- geomorphic reaches responded differently to this tional and the alluvial forest exhibits a mosaic pattern. flood season. An even higher turnover (62%) was Annual channel migration rates ranging from 25 to documented in the island-braided headwater reach; 400 m have been reported from large rivers in the the lowest turnover of a floodplain segment occurred Upper Amazon basin (Terborgh & Petren, 1991; in the meandering reach (22%). Turnover in the Kalliola et al., 1992), with an extreme rate of constrained headwater reach was less than 10%. 115 m day–1 reported by Puhakka et al. (1992). Based Avulsion was the mechanism of channel change in on the lateral erosion rate of 25 m year–1 determined the island-braided headwaters and lateral channel by Terborgh & Petren (1991), the Manu River, Peru, migration the dominant mechanism in the meander- would require 240 years to traverse the 6-km wide ing reach, whereas both avulsion and lateral channel floodplain. Much lower lateral migration rates are migration contributed to turnover in the other reported for temperate rivers (Gilvear & Bravard, geomorphic reaches.

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 517–539 Riverine landscape diversity 527

Fig. 8 Aquatic habitat turnover occurring from August to November 1999 in the bar- braided reach of the Fiume Tagliamento, Italy. A ¼ Preflood configuration of sur- face waters; B ¼ postflood configuration; C ¼ August–November overlay (D.B. Arscott, unpublished data).

Employing GIS-based analysis of aerial photo- corridors (Stanley, Fisher & Grimm, 1997; Tockner graphs from 1984, 1986, and 1991, Kollmann et al. et al., 2000), has received little detailed study. An (1999) investigated turnover of landscape elements in exception is the detailed analysis of the channel the island-braided lower reach of the Tagliamento. network in the Val Roseg, a glacial stream with a Although erosive floods caused dramatic changes in predictable expansion/contraction cycle (Tockner landscape configuration, the relative proportion of et al., 1997; Malard, Tockner & Ward, 1999; Ward river channels, gravel bars and vegetated islands did et al., 1999). Total channel length within the 2.6 km not change significantly. The turnover rate between long braided floodplain varied from 5.9 to 20.9 km, 1984 and 1986 ranged from 15% for established the highest values occurring during peak glacial melt islands to 83% for pioneer islands/large woody in summer and the lowest values in winter. The entire debris. We believe that high turnover is a general floodplain shifts from dominance by glacial meltwa- feature of unimpaired riverine landscapes. ter during the expansion phase in summer to a ground water-controlled system during the contrac- Expansion/contraction cycles. The seasonal expansion tion phase in winter. The aquatic habitats of the and contraction of channel networks without over- glacial floodplain were by no means uniformly bank flooding, a common phenomenon in river cold and turbid. Rather, a remarkable degree of

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 517–539 528 J.V. Ward et al. spatio-temporal heterogeneity was documented. Sea- Succession. Over longer ecological time scales, the sonal changes in discharge, the spatial array of the erosive action of flooding is countered by ecological channel network, and landscape patterns in water succession. In constrained reaches succession is main- chemistry within the floodplain were linked to the tained in early seral stages by the concentration of shifting dominance of different hydrological reser- fluvial energy in a single-thread channel, whereas a voirs (englacial and subglacial water, hillslope and diversity of seral stages is apparent in floodplain alluvial aquifers) and flow paths (surface and sub- reaches. Therefore, at the river corridor scale, surface) within the catchment. Six lotic habitat types concomitant with the alternation of constrained and were identified within the floodplain, based on water floodplain reaches (see Fig. 1), there is a downstream source, flow paths and hydrochemical indicators: (1) spatial pattern of alternating successional trajectories. main channel, (2) side channels, (3) intermittently The following discussion focuses on the different connected channels, (4) mixed channels, (5) entering types of succession occurring in alluvial floodplains: tributaries and (6) groundwater channels (three sub- as some floodplain waterbodies slowly fill and types). Although individual groundwater channels undergo terrestialisation, other waterbodies are newly were characterised by temporally stable environmen- formed or rejuvenated; as portions of floodplain tal conditions, as a group groundwater channels forests are undercut by channel migration, seedlings exhibited higher spatial habitat heterogeneity than are established on newly deposited alluvium; as any other lotic habitat type, reflecting a diversity of mature islands are eroded by a flood, LWD deposited water sources (hillslope aquifer, deep and shallow on gravel bars serves as nuclei for island formation. alluvial aquifers). The main channel and intermit- These three types of succession – hydrarch succession, tently connected channels, in contrast, exhibited high floodplain forest succession and alluvial island temporal but low spatial heterogeneity. The other succession – are briefly addressed below. First, aquatic habitat types were characterised by interme- however, we will examine the major successional diate levels of temporal and spatial heterogeneity. trends for key components of the riparian flora, The high level of collective habitat heterogeneity aquatic vegetation and aquatic fauna associated with ensured that clear-water refugia were available in the the rapid terrestrialisation of a meander that in floodplain year round, even when highly turbid 1956–57 was cut off from the Ain River in France glacial meltwater dominated the system. Thermal (Fig. 9). Willows (Salix eleagnos L.) occurring on the heterogeneity between aquatic habitats was also highest alluvial surfaces were the only woody riparian higher than expected for a glacial system (U. Uehlin- plants in 1965. Subsequently, as the floodplain surface ger, unpublished data). was elevated relative to the water level, poplar The influence of the expansion/contraction cycle on (Populus nigra L.) and ash (Fraxinus) initially displaced riverine landscapes is expected to reflect geomorphic S. eleagnos to lower elevation sites (1980). Terrestriali- style (Ward & Tockner, 2001). Overall habitat diver- sation was predicted to eliminate S. eleagnos by 2000, sity was highest near maximum expansion in the at which time Populus and Fraxinus would co-domin- glacial floodplain of Val Roseg, where heterogeneity is ate, hawthorn (Crataegus) would become abundant, largely a function of diversification of water sources and elm (Ulmus) and the willow S. purpurea would and flow paths, rather than autogenic processes occupy sediment deposits in filled channel segments. within aquatic habitats. In contrast, the Fiume Taglia- The open water habitats with coarse substratum and mento, a flashy Mediterranean river characterised by some hydrophytes (not shown) in 1965 were partially extensive island-braided reaches, attains the highest filled by fine organic sediment and densely colonised landscape-level diversity at intermediate discharge. by hydrophytes and helophytes by 1980 and predicted At higher levels inundation of the entire Tagliamento to become temporary waters by the year 2000. floodplain has a homogenising effect, whereas the glacial floodplain sustains a distinct network of running water channels even during major spates. Fig. 9 Graphical depiction of successional development across an idealised cross-section of a former meander (Les Brotteaux) Expansion/contraction cycles are intimately related to on the Ain River floodplain, France, during 1965, 1980 and connectivity, a topic addressed in a later section of this projected to the year 2000 (modified from Amoros et al., 1986). paper. HW ¼ High water; LW ¼ low water.

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 517–539 Riverine landscape diversity 529

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 517–539 530 J.V. Ward et al. Salmonids (Thymallus thymallus L.), which occurred in concave banks of meander bends, primary succession the abandoned meander in 1965, were replaced by is initiated on fresh alluvium deposited as point bars pike (Esox lucius L.) by 1980; the fish fauna would on convex bends. As water levels decline during the disappear completely by 2000 concomitant with the dry season, the woody composite Tessaria integrifolia disappearance of permanent water. Gammarus,a Ruiz & Pav. germinates on the point bars and grows predominant member of the zoobenthos in 1965 gave to heights up to 2 m before the next flood season. The way to Asellus, anisopteran and coleopteran larvae as next flood removes many Tessaria plants, but some eutrophication proceeded. Benthic chydorids were stems are flattened and buried within the new layer of also abundant in 1980. sediment. The buried stems sprout vegetatively and Hydrarch (hydrosere) succession refers to the Tessaria attains heights of 3–4 m during the next dry changes that occur over time (eutrophication/terres- season. Vertical accretion on the floodplain surface trialisation) following the initial formation of lentic and a greater distance from the migrating river allows waterbodies. In river corridors successional trajector- upright plants to withstand the next flood season and ies are confounded by floods, which not only initiate grow to full size (8–10 m) within 3–5 years. The cane hydrarch succession by forming new waterbodies by Gynarium invades and eventually overtops Tessaria, channel abandonment, but also reset successional causing its demise. Cane stands provide suitable sequences through rejuvenation of existing waterbod- conditions (adequate light penetration, reduced flood ies. Waterbodies that are situated far from the active scour) for seedling establishment by many tree spe- channel or those otherwise shielded from fluvial cies. Subsequent invasion of the cane by Cecropia,a action by their position in the riverine landscape, fast-growing, short-lived tree, forms a Gynaria–Cecr- undergo different sucessional trajectories than water- opia association that lasts several years, but is even- bodies subjected to frequent flooding. This was well tually replaced by taller, slower-growing trees. A illustrated by a study of six braided channel segments transitional mixed forest stand persists for several isolated from the main course of the Rhone at about decades during which a dense herbaceous understory the same time in the late 19th century (Bornette, suppresses further tree recruitment. Two slow-grow- Amoros & Chessel, 1994). Only two of the six channels ing trees (Ficus and Cedrella) that persist for about a had progressed to an advanced stage of hydrarch century and attain heights of 35–40 m are co-domin- succession, characterised by eutrophic species and the ant in the late successional stage of the floodplain accumulation of fine organic sediment. Those aban- forest. The mature floodplain forest has five vertically doned braids are protected from flood scouring by the superimosed strata (the highest exceeding 50 m) with forest vegetation growing on the alluvial plug. over 200 tree species per hectare. This mature stage Conversely, the four other channels of the same age requires 300–500 years to develop and is therefore exhibited much retarded successional development, limited to areas of the floodplain with lower than as indicated by oligo-mesotrophic species assemblag- average rates of channel migration. The spatially es. Two of these had coarse alluvial plugs without variable rates of channel migration, in concert with woody vegetation and lacked protection from flood the temporal sequence of seral stages, each character- scouring. The other two were located far from the ised by distinctive three-dimensional structure, results active channel and had forested alluvial plugs, but in a diversity of forest patches across the dynamic were fed by cold, nutrient-poor ground water, that riverine landscape. slowed the rate of succession. Forest succession on temperate zone floodplains, Floodplain forest succession is initiated on freshly although involving fewer species and lower structural deposited alluvium, a major landscape element of diversity, exhibits generally similar trajectories to that natural rivers. For example, recent alluvial deposits described above for Amazonia. In the north temperate were estimated to occupy about 26% of the Amazo- zone, the pioneer stage consists of patches of shrubs nian lowlands of Peru (Salo et al., 1986). Terborgh & and tree saplings of Populus and Salix, the predomin- Petren (1991) provide a useful summary of alluvial ant woody plants on fresh alluvial deposits (Schnit- forest succession along the Rio Manu, a meandering zler, 1997). Like Tessaria of Amazonia, they are ruderal Peruvian river with a 6-km wide floodplain. During strategists (sensu Grime, 1977), enabling them to annual floods, as the forest is undercut along the colonise frequently disturbed environments. These

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 517–539 Riverine landscape diversity 531 pioneer softwoods are tolerant of fluvial action, burial burying the trunk and branches. Shoots sprout from and submersion, but are vulnerable to shade and the buried wood of Salix and Populus during the first drought. They are weak competitors that are replaced growing season. This complex of LWD, sediment, by other species as landforms stabilise. The transi- entrapped fine organic matter and regrowth from tional or pre-equilibrium forest of European flood- buried wood constitutes the incipient stage of island plains (defined by canopy closure of the Salix/Populus formation, termed phase I islands by Edwards et al. association) is characterised by two strata, a canopy c. (1999). If suitable conditions prevail, phase I islands 20 m high and the grass layer of 2 m height (Schnit- develop into pioneer islands (phase II) with young zler, 1997). Canopy closure occurs within one or two Salix and Populus trees several years old forming a decades. These softwoods are gradually replaced by canopy up to 4 m high. The woody vegetation of hardwoods emerging from the understory. The ‘cli- established islands (phase III) is dominated by pat- max’ (terminal successional stage) of the floodplain ches of shrubs consisting of Salix spp. and a tree forest, characterised by the hardwoods Ulmus, Fraxi- canopy up to 20 m high formed by Populus, Alnus and nus and Quercus, consists of five strata with a canopy S. alba. Established islands have a distinctive commu- reaching 30–35 m in height. It takes two to three nity composition, including an assemblage of species centuries to reach the terminal hardwood stage of less tolerant of inundation, but more tolerant of alluvial forest succession. However, given that the drought and low nutrient levels (Kollmann et al., rivers of Europe have been regulated for centuries, it 1999). Succession from gravel bars to phase III islands is possible that the hardwood association may not requires 10–20 years, during which 1–2 m of fine represent the primeval floodplain forest of the active sediment is deposited on top of the underlying gravel floodplain; prior to flow regulation natural fluvial bar. Although the probability of an island being dynamics may have prevented alluvial forests from washed away decreases as islands develop, islands of developing beyond the softwood successional stage. all phases may succumb to erosive forces, their Islands, a dominant landform of pristine alluvial accumulated materials contributing to island forma- rivers, are simultaneously forming, developing and tion downstream. Nonetheless, an important effect of eroding by processes that operate continuously as island succession is the creation of relatively stable well as by episodic flood events (Harwood & Brown, habitats in the active floodplains of highly dynamic 1993; Abbe & Montgomery, 1996; Osterkamp, 1998; rivers. In this process of island succession, the plants Kollmann et al., 1999; Gurnell et al., 2001; Ward et al., serve as autogenic ecosystem engineers (sensu Jones, 2000b). Because islands integrate hydrologic, morpho- Lawton & Shackak, 1994), trapping sediment and logic and vegetation attributes, the presence of organic matter, thereby structuring the riverine land- various successional stages of islands provides a scape. landscape-level indicator of the condition of river It is clear that phenomena associated with hydrarch, corridors. There are two principal modes of island alluvial forest and island succession greatly contribute formation, vegetation establishment on bars within to riverine landscape diversity. The broad range of the active river corridor and dissection of the lateral seral stages occurring simultaneously in natural river floodplain forest by channel avulsion (Gurnell et al., corridors results in landscape patches differing mark- 2001). Here we briefly address island formation and edly in successional trajectories, geomorphic features, succession on gravel bars, based on studies conducted environmental conditions and biotic structure. How- on a braided Alpine river (Edwards et al., 1999; ever, if unchecked, succession eventually leads to Kollmann et al., 1999; Gurnell et al., 2001). reduced environmental heterogeneity: waterbodies Island formation is typically initiated when LWD, undergo rapid terrestrialisation; islands merge with often an uprooted tree that is being transported the floodplain, forming a single-thread channel; allu- downstream, lodges on the crest of a gravel bar on vial forests senesce. A high degree of natural distur- the declining limb of a flood hydrograph. The root bance is necessary to counter such successional plate of the tree usually faces upstream with the trunk trends. Connectivity, induced by the kinetic energy and branches trailing downstream. A plume of fine of flooding (fluvial dynamics), is the countervailing sediment accumulates immediately downstream mechanism that sustains heterogeneity across the from the LWD accumulation, partially or completely riverine landscape.

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 517–539 532 J.V. Ward et al. mechanism, rejuvenating abandoned waters through Connectivity between landscape elements scouring action that removes accumulations of fine The concept of connectivity, initially applied to gene sediments, organic detritus and organisms. Other flow between subpopulations of a metapopulation materials, including propagules, are imported during (Merriam, 1984), here refers to hydrological connec- reconnection. tivity, the exchange of matter, energy and biota Connectivity in alluvial rivers is, however, a more between different elements of the riverine landscape complex phenomenon than implied by simple gradi- via the aqueous medium (Amoros & Roux, 1988). ents. The example presented earlier demonstrated Connectivity operates at different scales, as illustrated how hydrological connectivity of abandoned braids by the following examples of the distribution of may be reduced by vegetation growing on alluvial animals in subsurface waters. At a fine scale of plugs or increased by upwelling of nutrient-poor resolution (centimetres to meters), different faunal ground waters, in both cases altering hydrarch elements exhibit different degrees of connectivity successional trajectories. Some additional manifesta- between surface waters and ground waters (Gibert tions of connectivity, based on examples from Euro- et al., 1990). The subterranean amphipod Niphargus pean rivers, follow. freely traverses the ecotone (hyporheic zone) between Fig. 10 compares shoreline ecotone length as a ground water and surface water, whereas the hypo- function of water level in floodplains of two rivers, rheic zone is a barrier for the strictly subterranean both of which have dynamic flow regimes but differ in amphipod Salentinella. The surface-dwelling amphi- connectivity. Low connectivity in the Danube is pod Gammarus and some aquatic insects freely colo- engendered by artificial levees that limit surface nise the hyporheic zone, whereas other epigean forms connectivity between the river and the floodplain, are confined to surface waters. Floods alter connec- whereas the Tagliamento is unconstrained. The tivity within the alluvium, with corresponding fine- Tagliamento had an ecotone length 2–3 times greater scale shifts in faunal distribution patterns within the that that of the Danube during the study period. In sediment interstices (Marmonier & Dole, 1986). In an addition, ecotone length remained constant over a expansive alluvial aquifer characterised by high wide range of discharge and exhibited high resilience hydraulic conductivity, specialised stoneflies migrate to major floods. Despite dramatic expansion/contrac- up to 2 km from the river beneath the floodplain tion cycles, in the intact floodplain suitable shoreline during their 2-year larval stage, returning to surface habitats for edge species were available in similar waters of the river to emerge, mate and oviposit quantities throughout the year (K. Tockner, unpub- (Stanford & Ward, 1988). True groundwater crusta- lished data). ceans, in contrast, are confined to the alluvial aquifer. Landscape indices (Gustafson, 1998) were used to Hydrological connectivity varies greatly between dif- investigate seasonal dynamics in the spatial configur- ferent sedimentary units within the fluvial stygoscape. ation of the braided channel network of Val Roseg, a Preferential flow paths (buried former river channels) glacial floodplain in the Swiss Alps (Malard et al., transmit a disproportionate amount of water through 2000). Connectivity was expressed as the relative alluvial aquifers and provide especially favourable proportion of the channel network length having an environmental conditions for groundwater fauna. upstream surface connection with the main river Fluvial dynamics, including the expansion/contrac- channel. When surface connectivity was plotted tion of surface waters, is the primary driving force against total channel length, two distinct phases were that sustains connectivity in alluvial rivers. Different apparent. During the first expansion phase channel types of floodplain waters may be arrayed along a length more than doubled without a substantial gradient based on their degree of surface connectivity increase in surface connectivity. During this period with the main channel. For example, abandoned snow melt recharged hillslope and floodplain aquifers braids lack surface connectivity during base flow, via subsurface pathways; increased surface flow was but are reconnected frequently by relatively small largely restricted to groundwater-fed tributaries and increases in water level. In contrast, abandoned alluvial channels. Only when total channel length meanders may require exceptional floods to re-estab- exceeded a threshold (15 km in this 3-km long lish connectivity. Reconnection serves as a reset floodplain) was there a corresponding increase in

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 517–539 Riverine landscape diversity 533

Fig. 10 Contrasting relationship between shoreline ecotone length and discharge in dynamic segments of two river floodplains, one with low connectivity and one with high connectivity between surface waters (modified from Ward et al., 2001).

surface connectivity. These relationships have import- These few examples demonstrate that connectivity ant implications, not only for groundwater–surface plays a major role in riverine landscapes (see also water interactions, but for colonisation pathways and Amoros, 2002), although this phenomenon has not refuge use by the aquatic fauna in this harsh alpine been afforded the attention it deserves. Detailed landscape. analysis of connectivity in diverse river systems Water temperature was recorded continuously for should provide considerable insight into structural 1 year in 18 waterbodies along a connectivity gradient and functional attributes of riverine landscapes, in the Alluvial Zone National Park, Austrian Danube including a greater understanding of the factors (Fig. 11). Habitats ranged from the main river channel structuring biodiversity patterns. to an isolated pond at the edge of the floodplain. During the winter months there was a more or less Biodiversity in riverine landscapes homothermous spatial pattern. However, for much of the year, a remarkable degree of thermal heterogen- Biodiversity (sensu Noss, 1990; Ward & Tockner, 2001) eity is apparent across the riverine landscape, caused includes structural diversity and functional diversity, in part by strong upwelling of cooler ground waters at thereby encompassing much of what has already been some locations. During the summer differences in presented in this paper. In this final section, we daily mean temperatures between the warmest and explore the relationship between spatio-temporal the coolest habitats were as much as 16 °C, greater heterogeneity and species diversity in a landscape than that which occurs along the entire length of the context. main channel. Therefore, the habitats within this As stated by Levine (2000a), ‘The study of bio- single floodplain collectively provide a wide range diversity is the study of how competitive exclusion is of temperature regimes, meeting the diverse thermal foiled – through the exploitation of heterogeneity and requirements of different aquatic species. pattern in the environment….’ The biota inhabiting

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 517–539 534 J.V. Ward et al.

Fig. 11 Colour plot of daily mean tem- peratures during 1999 from 18 waterbod- ies along a connectivity gradient from the main channel to an isolated pond on the edge of the floodplain, Alluvial Zone National Park, Austrian Danube (data from K. Tockner and C. Baumgartner). Locations with strong upwelling of cooler groundwaters are indicated by horizontal streaks of blue and green. riverine landscapes are indeed adapted to exploit the Carbiener & Tre´molie`res (1992) emphasise the role of spatio-temporal heterogeneity engendered by natural fine-scale environmental gradients in the ecological disturbance regimes (Junk et al., 1989; Petts & Amo- segregation of closely related plant species in alluvial ros, 1996; Naiman & De´camps, 1997; Schnitzler, 1997). forests. For example, nine congeneric species of Maintenance of community diversity is a composite willows coexist in close proximity in the upper Rhine, function of the range of resources available and the segregated along gradients of soil acidity, sediment degree to which species segregate those resources. In grain size, habitat stability and groundwater level, this context, landscape spatial heterogeneity expands among other factors. the resource gradient (e.g. Fig. 4) and temporal Dole (1983) examined interstitial faunal diversity heterogeneity (e.g. Fig. 11) increases the potential for within alluvial sediments (– 0.5 m depth) in three niche overlap. floodplain habitats representing a disturbance gradi- Natural disturbances are responsible both for struc- ent on a Rhone River floodplain. A total of 112, 157 turing spatial heterogeneity and for creating condi- and 107 species were identified, respectively, from the tions under which niche overlap can occur. For active channel (high disturbance), side arm (moderate example, Schnitzler (1994) considers disturbance by disturbance) and an isolated meander arm (low fluvial processes largely responsible for the fact that disturbance). The active channel contained a species the alluvial hardwood association (Querco–Ulmetum) assemblage resistant to disturbance, whereas an is one of the most productive and speciose forest assemblage sensitive to disturbance occurred under ecosystems of the temperate zone. In the Upper Rhine, the stable conditions characterising the abandoned the Querco–Ulmetum association, with 25 species of meander. The proportion of hypogean forms was trees and 23 species of shrubs, consisted of five forest strongly correlated with the disturbance gradient, units (phytosociological sub-associations). Three of constituting 8, 20 and 47% of the total fauna from the these were arrayed along a flood-risk gradient on the active arm to the meander. active floodplain. The fourth was associated with The intermediate disturbance hypothesis (IDH; terraces and the fifth forest unit occurred in a part of Connell, 1978) predicts low community diversity in the floodplain influenced by a major tributary carry- environments exposed to high disturbance levels ing more acidic water and suspended clay. Schnitzler, where only highly tolerant or fugitive species can

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 517–539 Riverine landscape diversity 535 occur, as well as in low disturbance environments the connectivity gradient (Tockner, Schiemer & Ward, where superior competitors monopolise resources. 1998). Fish diversity, for example, may peak in highly According to the IDH, species diversity is highest in connected habitats, whereas amphibian diversity situations experiencing intermediate levels of distur- tends to be highest in habitats with low connectivity. bance, which constrains competitive exclusion, there- Other groups attain maximum species richness by allowing the coexistence of taxa with divergent between these two extremes. The resulting pattern is species traits and adaptive strategies. Salo et al. (1986) a series of overlapping species diversity peaks along the proposed that river dynamics, by constraining com- connectivity gradient, suggesting that habitats collec- petitive exclusion, is responsible in part for the high tively traversing a broad range of connectivity will biodiversity of the upper Amazon basin. Pollock, optimise community diversity in riverine landscapes. Naiman & Hadley (1998) reported highest species The dynamic equilibrium model of Huston (1979, richness in wetland plant communities subjected to 1994) adds a resource dimension (productivity) along intermediate flood frequencies. However, only limited the disturbance axis of the IDH. This integrated model attempts have been made to empirically test the IDH predicts that the level of disturbance at which species in complex riverine landscapes. diversity is maximised varies as a function of the The only investigation known to us designed resource base, because greater disturbance is needed specifically to test the IDH in riverine floodplains to prevent competitive exclusion as productivity examined the relationship between connectivity and increases. Alluvial floodplains provide an ideal set- aquatic plant diversity in 12 waterbodies that were ting for investigating how the balance between pro- inundated by the river from 0 to 37 days year–1 ductivity and disturbance influences biodiversity (Bornette et al., 1998b). The authors postulated that patterns. Apart from Pollock et al. (1998), whose work high connectivity would constrain diversity by exces- on plant species richness in wetlands, including some sive flood scouring; that low connectivity would floodplain sites, provides partial support for the constrain diversity through competitive exclusion; dynamic equilibrium model, we are not aware of whereas intermediate connectivity would result in research that has been conducted in this context. the highest species richness. No clear relationship However, Yachi & Loreau (1999) developed a general between connectivity and species diversity was stochastic dynamic model to examine the relationship demonstrated; however, a complex of interacting between biodiversity and productivity in fluctuating factors structuring aquatic plant diversity in flood- environments. The model output predicts increased plain waterbodies was identified and incorporated ecosystem productivity and reduced variance in into a predictive model (Amoros & Bornette, 1999). productivity in more speciose systems. Tilman They found that the scouring effect is determined not (1999) predicts that the more complete utilisation of only by frequency of inundation, but also by the limiting resources by diverse communities reduces configuration of the waterbody in the landscape; that their susceptibility to invasion by exotic species. groundwater connectivity may influence aquatic plant Results from experimental manipulations of riparian diversity; that nutrient levels, shading by phytoplank- plant tussocks, however, suggest that the intrinsic ton or turbidity may override other factors; that invasion resistance of diverse local assemblages may submerged and floating macrophytes respond differ- not be realised at larger scales (Levine, 2000b). ently to connectivity; that silting and scouring are There is a great need for rigorous empirical studies both important; that degree of connectivity deter- that examine the relationship between landscape mines whether sexual or vegetative reproduction diversity and species diversity in alluvial river corri- predominate and whether antiherbivore defences are dors. The species pool in riverine landscapes is needed; and that retention structures determine whe- derived from terrestrial and aquatic communities ther or not entering propagules become established. inhabiting diverse lotic, lentic, riparian and ground- The relationship between species richness and water habitats arrayed across spatio-temporal gradi- connectivity is therefore determined by complex ents (Fig. 12). Understanding the complex interactions relationships among several interacting variables. In between disturbance regimes, spatial heterogeneity addition, species richness maxima for different faunal and biodiversity in riverine landscapes provides a and floral elements occur at different positions along major challenge for the future.

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 517–539 536 J.V. Ward et al. level of a process (connectivity for example). A more balanced approach is to assure that a given process varies sufficiently among individual habitat patches to sustain a broad range of that process at the landscape scale.

Acknowledgments

Our recent research has been supported by grants from the Swiss National Science Foundation (SNF grants 21-49243.96; 31-50440.97; 31-50444.97) and the ETH Forschungskommission (0-20572-98).

References

Abbe T.B. & Montgomery D.R. (1996) Large woody debris jams, channel hydraulics and habitat formation in large rivers. Regulated Rivers, 12, 201–221. Adis J. & Junk W. (2002) Terrestrial invertebrates inhabiting lowland river floodplains of Central Amazonia and Central Europe: a review. Freshwater Biology, 47, 711–732. Allen J.R.L. (1965) A review of the origin and character- istics of recent alluvial sediments. Sedimentology, 5, 89– 191. Fig. 12 The diverse patterns and processes that contribute to the species pool of riverine landscapes. Amoros C. (2002) Connectivity and biocomplexity in waterbodies riverine floodplains. Freshwater Biology, Conclusions 47, 761–776. Amoros C. & Bornette G. (1999) Antagonistic and Riverine landscapes in the natural state exhibit high cumulative effects of connectivity: a predictive model levels of complexity across a range of scales. The based on aquatic vegetation in riverine wetlands. expansive perspective provided by landscape ecology Archiv fu¨r Hydrobiologie (Supplementband), 115, 311–327. holds the potential for developing a truly holistic Amoros C. & Roux A.-L. (1988) Interactions between perspective of river corridors, one that rigorously water bodies within the floodplains of large rivers: function and development of connectivity. In: Connec- integrates structure, dynamics, and function. In addi- tivity in Landscape Ecology (Ed. K.-F. Schreiber), tion, landscape ecology offers an effective framework pp. 125–130. Mu¨ nstersche Geographische Arbeiten, to integrate pattern and process in river corridors, to Mu¨ nster, Germany. examine environmental dynamics and interactive Amoros C., Bravard J.-P., Castella C. et al. (1986) pathways between landscape elements, to link Recherches interdisciplinaires sur les e´cosyste`mes de research with management, and to develop viable la basse-plaine de l’Ain (France). Documents de Carto- strategies for river conservation. Remote sensing graphie Ecologique, 29, 1–166. provides a variety of techniques for analysing land- Arscott D.B., Tockner K. & Ward J.V. (2000) Aquatic scape patterns and dynamics in river corridors, with habitat diversity along the corridor of an alpine promises of even greater utility in the future (Mertes, floodplain river (Fiume Tagliamento, Italy). Archiv fu¨r 2002). A more complete understanding of ecological Hydrobiologie, 149, 679–704. processes from a landscape perspective, will not only Benke A.C. (1990) A perspective on America’s vanishing streams. Journal of the North American Benthological advance the discipline of river ecology, but will also Society, 9, 77–88. enhance the effectiveness of conservation and restor- Bormann F.H. & Likens G.E. (1979) Pattern and Process in ation initiatives. However, present understanding a Forested Ecosystem. Springer-Verlag, New York. indicates the futility of managing for the optimal

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 517–539 Riverine landscape diversity 537

Bornette G., Amoros C. & Chessel D. (1994) Effect of Petts & C. Amoros), pp. 68–97. Chapman & Hall, allogenic processes on successional rates in former London. river channels. Journal of Vegetation Science, 5, 237–246. Gregory S.V., Swanson F.J., McKee W.A. & Cummins Bornette G., Amoros C., Piegay H., Tachet J. & Hein T. K.W. (1991) An ecosystem perspective of riparian (1998a) Ecological complexity of wetlands within a zones. Bioscience, 41, 540–551. river landscape. Biological Conservation, 85, 35–45. Grime J.P. (1977) Evidence for the existence of three Bornette G., Amoros C. & Lamouroux N. (1998b) Aquatic primary strategies in plants and its relevance to plant diversity in riverine wetlands: the role of ecological and evolutionary theory. American Natural- connectivity. Freshwater Biology, 39, 267–283. ist, 111, 1169–1194. Boulton A.J., Findlay S., Marmonier P., Stanley E.H. & Gurnell A.M., Petts G.E., Harris N., Ward J.V., Tockner Valett H.M. (1998) The functional significance of the K., Edwards P.J. & Kollmann J. (2000) Large wood hyporheic zone in streams and rivers. Annual Review of retention in river channels: the case of the Fiume Ecology and Systematics, 29, 59–81. Tagliamento, Italy. Earth Surface Processes and Land- Brinson M.M. (1990) Riverine forests. In: Forested forms, 25, 255–275. Wetlands (Eds A.E. Lugo, M.M. Brinson & S.L. Brown), Gurnell A.M., Petts G.E., Hannah D.M., Smith B.P.G., pp. 87–141. Elsevier, Amsterdam. Edwards P.J., Kollmann J., Ward J.V. & Tockner K. Brunke M. & Gonser T. (1997) The ecological significance (2001) Riparian vegetation and island formation along of exchange processes between rivers and groundwa- the gravel-bed Fiume Tagliamento, Italy. Earth Surface ter. Freshwater Biology, 37, 1–33. Processes and Landforms, 26, 31–62. Castany G. (1985) Liaisons hydrauliques entre les aqui- Gustafson E.J. (1998) Quantifying landscape spatial fe`res et les cours d’eau. Stygologia, 1, 1–25. patterns: what is the state of the art? Ecosystems, 1, Connell J. (1978) Diversity in tropical rainforests and 143–156. coral reefs. Science, 199, 1302–1310. Harwood K. & Brown A.G. (1993) Fluvial processes in a De´camps H. (1996) The renewal of floodplain forests forested anastomosing river: flood partitioning and along rivers: a landscape perspective. Verhandlungen changing flow patterns. Earth Surface Processes and der Internationalen Vereinigung fu¨r Theoretische und Landforms, 18, 741–748. Angewandte Limnologie, 26, 35–59. Huston M. (1979) A general hypothesis of species Dole M.-J. (1983) Le domaine aquatique souterrain de la diversity. American Naturalist, 113, 81–101. plaine alluviale du Rhoˆne a` l’est de Lyon. 1. Diversite´ Huston M.A. (1994) Biological Diversity: The Coexistence of hydrologique et bioce´notique de trois stations repre´- Species on Changing Landscapes. Cambridge University sentatives de la dynamique fluviale. Vie et Milieu, 33, Press, New York. 219–229. Hutchinson G.E. (1957) A Treatise on Limnology, Vol. I. Dynesius M. & Nilsson C. (1994) Fragmentation and flow Wiley, New York. regulation of river systems in the northern third of the Johnson W.C. (2000) Tree recruitment and survival in world. Science, 266, 753–762. rivers: influence of hydrological processes. Hydrological Edwards P.J., Kollmann J., Gurnell A.M., Petts G.E., Processes, 14, 3051–3074. Tockner K. & Ward J.V. (1999) A conceptual model Jones C.J., Lawton J.H. & Shackak M. (1994) Organisms of vegetation dynamics on gravel bars of a large as ecological engineers. Oikos, 69, 373–386. Alpine river. Wetlands Ecology and Management, 7, Junk W.J., Bayley P.B. & Sparks R.E. (1989) The flood 141–153. pulse concept in river-floodplain systems. Canadian Gibert G., Dole-Olivier M.-J., Marmonier P. & Vervier P. Special Publication of Fisheries and Aquatic Sciences, 106, (1990) Surface water-groundwater ecotones. In: The 110–127. Ecology and Management of Aquatic-Terrestrial Ecotones Kalliola R., Salo J., Puhakka M., Rajasilta M., Haeme T., (Eds R.J. Naiman & H. De´camps), pp. 199–216. Neller R.J., Raesaenen M.E. & Danjoy Arias W.A. UNESCO MAB and the Parthenon Publishing Group, (1992) Upper Amazon channel migration. Naturwis- Paris and Carnforth. senschaften, 79, 75–79. Gibert J., Stanford J.A., Dole-Olivier M.-J. & Ward J.V. Keller E.A. & Swanson F.J. (1979) Effects of large organic (1994) Basic attributes of groundwater ecosystems and material on channel form and fluvial processes. Earth prospects for research. In: Groundwater Ecology (Eds Surface Processes, 4, 361–380. J. Gibert, D.L. Danielopol & J.A. Stanford), pp. 7–40. Kollmann J., Vieli M., Edwards P.J., Tockner K. & Academic Press, San Diego, CA. Ward J.V. (1999) Interactions between vegetation Gilvear D. & Bravard J.-P. (1996) Geomorphology of development and island formation in the Alpine river temperate rivers. In: Fluvial Hydrosystems (Eds G.E. Tagliamento. Applied Vegetation Science, 2, 25–36.

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 517–539 538 J.V. Ward et al.

Leddy J., Ashworth P. & Best J. (1993) Mechanisms of Petts G.E. & Amoros C. (Eds) (1996) Fluvial Hydrosystems. anabranch avulsion within gravel-bed braided rivers: Chapman & Hall, London. observations from a scaled physical model. In: Braided Petts G.E., Moller H. & Roux A.-L. (Eds) (1989) Historical Rivers (Eds J. Best & C.S. Bristow), pp. 88–104. Changes of Large Alluvial Rivers, Western Europe. Wiley, Geological Society Special Publication 75, London. Chichester. Levine S.A. (2000a) Multiple scales and the maintenance Pinay G., De´camps H., Chauvet E. & Fustec E. (1990) of biodiversity. Ecosystems, 3, 498–506. Functions of ecotones in fluvial systems. In: The Ecology Levine J.M. (2000b) Species diversity and biological and Management of Aquatic-Terrestrial Ecotones (Eds R.J. invasions: relating local process to community pattern. Naiman & H. De´camps), pp. 141–169. Parthenon Science, 288, 852–854. Publishers, Casterton Hall, Carnforth. Malanson G.P. (1993) Riparian Landscapes. Cambridge Pollock M.M., Naiman R.J. & Hadley T.A. (1998) Plant University Press, Cambridge. species richness in riparian wetlands – a test of Malard F., Tockner K. & Ward J.V. (1999) Shifting biodiversity theory. Ecology, 79, 94–105. dominance of subcatchment water sources and flow Puhakka M., Kalliola R., Rajasilta M. & Salo J. (1992) paths in a glacial floodplain, Val Roseg, Switzerland. River types, site evolution and successional vegetation Arctic, Antarctic and Alpine Research, 31, 135–150. patterns in Peruvian Amazonia. Journal of Biogeography, Malard F., Tockner K. & Ward J.V. (2000) Physico- 19, 651–665. chemical heterogeneity in a glacial riverscape. Land- Robinson C.T., Tockner K. & Ward J.V. (2002) The fauna scape Ecology, 15, 679–695. of dynamic riverine landscapes. Freshwater Biology, 47, Marmonier P. & Dole M.-J. (1986) Les Amphipodes des 661–677. se´diments d’un bras court-circuite´ du Rhoˆne: logique Salo J., Kalliola R., Ha¨kkinen I., Ma¨kinen Y., Niemela¨ P., de re´partition et re´action aux crues. Sciences de l’Eau, 5, Puhakka M. & Coley P.D. (1986) River dynamics and 461–486. the diversity of Amazon lowland forests. Nature, 322, Maser C. & Sedell J.R. (1994) From the Forest to the Sea: The 254–258. Ecology of Wood in Streams, Rivers, Estuaries and Oceans. Schnitzler A. (1994) European alluvial hardwood forests St Lucie Press, Delray Beach, FL. of large floodplains. Journal of Biogeography, 21, 605–623. Merriam G. (1984) Connectivity: a fundamental ecolo- Schnitzler A. (1997) River dynamics as a forest process: gical characteristic of landscapes. Proceedings of the interactions between fluvial systems and alluvial for- International Association for Landscape Ecology, 1, 5–15. ests in large European river plains. Botanical Review, 63, Mertes L.A.K. (2002) Remote sensing of riverine land- 40–64. scapes. Freshwater Biology, 47, 799–816. Schnitzler A., Carbiener R. & Tre´molie`res M. (1992) Naegeli M.W., Huggenburger P. & Uehlinger U. (1996) Ecological segregation between closely related species Ground penetrating radar for assessing sediment struc- in the flooded forests on the upper Rhine plain. New tures in the hyporheic zone of a prealpine river. Journal Phytologist, 121, 293–301. of the North American Benthological Society, 15, 353–366. Schumm S.A. (1977) The Fluvial System. Wiley, New York. Naiman R.J. & De´camps H. (1997) The ecology of Stanford J.A. & Ward J.V. (1988) The hyporheic habitat of interfaces: riparian zones. Annual Review of Ecology river ecosystems. Nature, 335, 64–66. and Systematics, 28, 621–658. Stanford J.A. & Ward J.V. (1993) An ecosystem perspec- Nanson G.C. & Beach H.F. (1977) Forest succession and tive of alluvial rivers: connectivity and the hyporheic sedimentation on a meandering-river floodplain, corridor. Journal of the North American Benthological northeast British Columbia, Canada. Journal of Biogeo- Society, 12, 48–60. graphy, 4, 229–251. Stanley E.H., Fisher S.G. & Grimm N.B. (1997) Ecosystem Nanson G.C. & Croke J.C. (1992) A genetic classification expansion and contraction in streams. Bioscience, 47, of floodplains. Geomorphology, 4, 459–486. 427–435. Noss R.F. (1990) Indicators for monitoring biodiversity: a Statzner B. & Kohmann F. (1995) River and stream hierarchical approach. Conservation Biology, 4, 355–364. ecosystems in Austria, Germany and Switzerland. In: Osterkamp W.R. (1998) Processes of fluvial island River and Stream Ecosystems (Eds C.E. Cushing, K.W. formation, with examples from Plum Creek, Colorado Cummins & G.W. Minshall), pp. 439–478. Elsevier, and Snake River, Idaho. Wetlands, 18, 530–545. Amsterdam. Pautou G. (1984) L’organisation des foreˆts alluviales dans Strahler A.N. (1963) The Earth Sciences. Harper & Row, l’axe Rhodanien entre Gene`ve et Lyon; comparaison New York. avec d’autres syste`mes fluviaux. Documents de Carto- Talbot L.M. (1996) The linkage between ecology and graphie Ecologique Grenoble, 27, 43–64. conservation policy. In: The Ecological Basis of Conser-

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 517–539 Riverine landscape diversity 539

vation (Eds S.T.A. Pickett, R.S. Ostfeld, M. Shachak & Ward J.V. (1997) An expansive perspective of riverine G.E. Likens), pp. 368–378. Chapman & Hall, New landscapes: pattern and process across scales. GAIA, 6, York. 52–60. Terborgh J. & Petren K. (1991) Development of habitat Ward J.V. & Stanford J.A. (1995) Ecological connectivity structure through succession in an Amazonian flood- in alluvial river ecosystems and its disruption by flow plain forest. In: Habitat Structure (Eds S.S. Bell, E.D. regulation. Regulated Rivers: Research and Management, McCoy & H.R. Mushinsky), pp. 28–46. Chapman & 11, 105–119. Hall, London. Ward J.V. & Tockner K. (2001) Biodiversity: towards a Tilman D. (1999) The ecological consequences of changes unifying theme for river ecology. Freshwater Biology, 46, in biodiversity: a search for general principles. Ecology, 807–819. 80, 1455–1474. Ward J.V., Tockner K., Edwards P.J., Kollmann J., Tockner K., Malard F., Burgherr P., Robinson C.T., Bretschko G., Gurnell A.M., Petts G.E. & Rossaro B. Uehlinger U., Zah R. & Ward J.V. (1997) Physico- (1999) A reference river for the Alps: the ‘Fiume chemical characterization of channel types in a glacial Tagliamento’. Regulated Rivers: Research and Manage- floodplain ecosystem (Val Roseg, Switzerland). Archiv ment, 15, 63–75. fu¨r Hydrobiologie, 140, 433–463. Ward J.V., Malard F., Stanford J.A. & Gonser T. (2000a) Tockner K., Schiemer F. & Ward J.V. (1998) Conservation Interstitial aquatic fauna of shallow unconsolidated by restoration: the management concept for a river- sediments, particularly hyporheic biotopes. In: Ecosys- floodplain system on the Danube River in Austria. tems of the World. Subterranean Ecosystems (Eds H. Aquatic Conservation, 8, 71–86. Wilkens, D.C. Culver & W.F. Humphreys), pp. 41–58. Tockner K., Malard F. & Ward J.V. (2000) An extension of Elsevier, Amsterdam. the flood pulse concept. Hydrological Processes, 14, Ward J.V., Tockner K., Edwards P.J., Kollmann J., 2861–2883. Gurnell A.M., Petts G.E., Bretschko G. & Rossaro B. Triska F.J. (1984) Role of wood debris in modifying (2000b) Potential role of island dynamics in river eco- channel geomorphology and riparian areas of a large systems. Verhandlungen der internationale Vereinigung fu¨r lowland river under pristine conditions: a historical case theoretische und angewandte Limnologie, 27, 2582–2585. study. Verhandlungen der Internationale Vereinigung fu¨r Ward J.V., Tockner K., Uehlinger U. & Malard F. (2001) Theoretische und Angewandte Limnologie, 22, 1876–1892. Understanding natural patterns and processes in river Turner M.G. (1998) Landscape ecology, living in a corridors as the basis for effective river restoration. mosaic. In: Ecology (Eds S.I. Dodson, T.F.H. Allen, Regulated Rivers: Research and Management, 17, 311–323. S.R. Carpenter, A.R. Ives, R.J. Jeanne, J.F. Kitchell, N.E. Yachi S. & Loreau M. (1999) Biodiversity and ecosystem Langston & M.G. Turner), pp. 78–122. Oxford Univer- productivity in a fluctuating environment: the insur- sity Press, New York. ance hypothesis. Proceedings of the National Academy of Ward J.V. (1989) The four-dimensional nature of lotic Sciences, 96, 1463–1468. ecosystems. Journal of the North American Benthological Society, 8, 2–8. (Manuscript accepted 18 July 2001)

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 517–539