David M. Merritt

Homogenization of regional river dynamics by dams and global biodiversity implications.

Global biodiversity in river and riparian ecosystems is generated and maintained by geographic variation in stream processes and fluvial disturbance regimes, which largely reflect regional differences in climate and geology. Extensive construction of dams by humans has greatly dampened the seasonal and interannual streamflow variability of rivers, thereby altering natural dynamics in ecologically important flows on continental to global scales. The cumulative effects of modification to regional-scale environmental templates caused by dams is largely unexplored but of critical conservation importance. Here, we use 186 long-term streamflow records on intermediate-sized rivers across the continental United States to show that dams have homogenized the flow regimes on third- through seventh-order rivers in 16 historically distinctive hydrologic regions over the course of the 20th century. This regional homogenization occurs chiefly through modification of the magnitude and timing of ecologically critical high and low flows. For 317 undammed reference rivers, no evidence for homogenization was found, despite documented changes in regional precipitation over this period. With an estimated average density of one dam every 48 km of third- through seventh-order river channel in the United States, dams arguably have a continental scale effect of homogenizing regionally distinct environmental templates, thereby creating conditions that favor the spread of cosmopolitan, nonindigenous species at the expense of locally adapted native biota. Quantitative analyses such as ours provide the basis for conservation and management actions aimed at restoring and maintaining native biodiversity and ecosystem function and resilience for regionally distinct ecosystems at continental to global scales.

Riparian vegetation and channel change in response to river regulation : a comparative study of regulated and unregulated streams in the Green River Basin, USA

The effects of river damming on geomorphic processes and riparian vegetation were evaluated through field studies along the regulated Green River and the free-flowing Yampa River in northwestern Colorado, USA. GIS analysis of historical photographs, hydrologic and sediment records, and measurement of channel planform indicate that fluvial processes and riparian vegetation of the two meandering stream reaches examined were similar prior to regulation which began in 1962. Riparian plant species composition and canopy coverage were measured during 1994 in 36, 0.01 ha plots along each the Green River in Browns Park and the Yampa River in Deerlodge Park. Detrended correspondence analysis (DCA) of the vegetation data indicates distinctive vegetation differences between Browns Park and Deerlodge Park. Canonical correspondence analysis (CCA) indicates that plant community composition is controlled largely by fluvial processes at Deerlodge Park, but that soil chemical rather than flow related factors play a more important role in structuring plant communities in Browns Park. Vegetation patterns reflect a dichotomy in moisture conditions across the floodplain on the Green River in Browns Park: marshes with anaerobic soils supporting wetland species (Salix exigua, Eleocharis palustris, Schoenoplectus pungens, and Juncus nodosus) and terraces having xeric soil conditions and supporting communities dominated by desert species (Seriphidium tridentatum, Sarcobatus vermiculatus, and Sporobolus airoides). In contrast, vegetation along the Yampa River is characterized by a continuum of species distributed along a gradual environmental gradient from the active channel (ruderal species such as Xanthium struminarium and early successional species such as S. exigua, Populus deltoides subsp. wislizenii, and Tamarix ramossissima) to high floodplain surfaces characterized by Populus forests and meadow communities. GIS analyses indicate that the channel form at Browns Park has undergone a complex series of morphologic changes since regulation began, while the channel at Deerlodge Park has remained in a state of relative quasi-equilibrium with discharge and sediment regimes. The Green River has undergone three stages of channel change which have involved the transformation of the historically deep, meandering Green River to a shallow, braided channel over the 37 years since construction of Flaming Gorge Dam. The probable long-term effects of channel and hydrologic changes at Browns Park include the eventual replacement of Populus-dominated riparian forest by drought tolerant desert shrublands, and the enlargement of in-channel fluvial marshes. Copyright

The role of hydrochory in structuring riparian and wetland vegetation

Hydrochory, or the passive dispersal of organisms by water, is an important means of propagule transport, especially for plants. During recent years, knowledge about hydrochory and its ecological consequences has increased considerably and a substantial body of literature has been produced. Here, we review this literature and define the state of the art of the discipline. A substantial proportion of species growing in or near water have propagules (fruits, seeds or vegetative units) able to disperse by water, either floating, submerged in flowing water, or with the help of floating vessels. Hydrochory can enable plants to colonize sites out of reach with other dispersal vectors, but the timing of dispersal and mechanisms of establishment are important for successful establishment. At the population level, hydrochory may increase the effective size and longevity of populations, and control their spatial configuration. Hydrochory is also an important source of species colonizing recruitment‐limited riparian and wetland communities, contributing to maintenance of community species richness. Dispersal by water may even influence community composition in different landscape elements, resulting in landscape‐level patterns. Genetically, hydrochory may reduce spatial aggregation of genetically related individuals, lead to high gene flow among populations, and increase genetic diversity in populations receiving many propagules. Humans have impacted hydrochory in many ways. For example, dams affect hydrochory by reducing peak flows and hence dispersal capacity, altering the timing of dispersal, and by presenting physical barriers to dispersal, with consequences for riverine plant communities. Hydrochory has been inferred to be an important vector for the spread of many invasive species, but there is also the potential for enhancing ecosystem restoration by improving or restoring water dispersal pathways. Climate change may alter the role of hydrochory by modifying the hydrology of water‐bodies as well as conditions for propagule release and plant colonization.

PROCESSES GOVERNING HYDROCHORY ALONG RIVERS: HYDRAULICS, HYDROLOGY, AND DISPERSAL PHENOLOGY

Rivers are important corridors for movement, migration, and dispersal of aquatic organisms as well as for dispersal of the seeds and vegetative propagules of riparian plants. In this investigation, the relationships between flow regime, channel morphology, dispersal phenology, and seed deposition patterns were evaluated using experimentation in a flume. A channel with geomorphic features common to a wide range of stream morphologies was constructed in a 1.8 × 20 m experimental flume through which three hydrologic regimes (one natural and two typical of dam releases) were routed in replicated trials. Relationships between dispersal phenology and hydrologic regime were examined using color-coded Betula fontinalis seeds released over each 10-min trial. Spatial patterns of seed deposition along stream margins were then compared to determine the individual and combined effects of flow regime, fluvial feature, and timing of seed release. Reynolds number, Froude number, Weber number, flow velocity, and a dimensi...

HYDROLOGIC REGIMES AND RIPARIAN FORESTS: A STRUCTURED POPULATION MODEL FOR COTTONWOOD

Riparian cottonwood (Populus deltoides) forests form the one of the most extensive deciduous forest ecosystems in arid regions of the western United States. However, cottonwood populations are threatened by flow alteration and channel degradation caused by dams, water diversions, and groundwater pumping. We developed a stochastic, density- dependent, population model to (1) consolidate information concerning cottonwood pop- ulation dynamics in a conceptual and analytical framework, (2) determine whether complex forest stand dynamics can be predicted from basic cottonwood vital rates and river hy- drology, and (3) aid in planning prescribed floods by projecting how altered flow regimes might affect populations. The model describes how annual variation in the hydrograph affects cottonwood mortality (via floods and droughts) and recruitment (via scouring of new habitat and seedling establishment). Using parameter values for the undammed Yampa River in Colorado, we found that abundances of seedlings and younger trees followed a boom-bust cycle driven by high flood mortalities while reproductive adult abundance fol- lowed a less erratic 5-15-yr periodicity driven by multiyear sequences of flows favorable to stand recruitment. Conversely, chance occurrences of multiple drought years eliminated cottonwood from up to 50% of available habitat, providing opportunities for competing plant species to establish. By simulating flow alterations on the Yampa ranging from chan- nelization (many floods/droughts) to damming (few floods/droughts), the model suggested that mature cottonwood forest should be most abundant near the observed natural flow regime. Model analysis also suggested that flow regimes with high flood frequencies result in stable (albeit small) population sizes, while stable flows result in highly variable pop- ulation sizes prone to local extinction.

Factors controlling the establishment of Fremont cottonwood seedlings on the Upper Green River, USA

Declines in cottonwood (Populus spp.) recruitment along alluvial reaches of large rivers in arid regions of the western United States have been attributed to modified flow regimes, lack of suitable substrate, insufficient seed rain, and increased interspecific competition. We evaluated whether and how these factors were operating during 1993–1996 to influence demographics of Fremont cottonwood (P. deltoides Marshall subsp. wislizenii (Watson) Eckenwalder) along reaches of the Green and Yampa Rivers near their confluence in northwestern Colorado. We examined seedling establishment, defined as survival through three growing seasons, at three alluvial reaches that differed primarily in the level of flow regulation: a site on the unregulated Yampa, an upper Green River site regulated by Flaming Gorge Dam, and a lower Green River site below the Green–Yampa confluence. Seed rain was abundant in all sites, and led to large numbers of germinants (first-year seedlings) appearing each year at all sites. The regulated flow in the upper Green River reach restricted germination to islands and cut banks that were later inundated or eroded; no seedlings survived there. Mortality at the lower Green River site was due largely to desiccation or substrate erosion; 23% of 1993 germinants survived their first growing season, but at most 2% survived through their second. At the Yampa River site, germinants appeared on vegetated and unvegetated surfaces up to 2.5 m above base flow stage, but survived to autumn only on bare surfaces at least 1.25 m above base flow stage, and where at least 10 of the upper 40 cm of the alluvium was fine-textured. Our studies of rooting depths and the stable isotopic composition of xylem water showed that seedlings in the most favorable locations for establishment at the Yampa site do not become phreatophytic until their third or fourth growing season. Further, the results of experimental field studies examining effects of shade and competition supported the hypothesis that insufficient soil moisture, possibly in combination with insufficient light, restricts establishment to unvegetated sites. Collectively, the demographic and experimental studies suggest that, in arid regions, soil water availability is at least as important as light level in limiting establishment of Fremont cottonwood seedlings. We hypothesize that in cases where arid land rivers experience large spring stage changes, recruitment is further constrained within bare areas to those sites that contain sufficient fine-textured alluvium, saturated during the spring flood, to provide the flood-derived soil moisture normally necessary for late-summer seedling survival. Copyright

Shifting dominance of riparian Populus and Tamarix along gradients of flow alteration in western North American rivers

Tamarix ramosissima is a naturalized, nonnative plant species which has become widespread along riparian corridors throughout the western United States. We test the hypothesis that the distribution and success of Tamarix result from human modification of river-flow regimes. We conducted a natural experiment in eight ecoregions in arid and semiarid portions of the western United States, measuring Tamarix and native Populus recruitment and abundance at 64 sites along 13 perennial rivers spanning a range of altered flow regimes. We quantified biologically relevant attributes of flow alteration as an integrated measure (the index of flow modification, IFM), which was then used to explain between-site variation in abundance and recruitment of native and nonnative riparian plant species. We found the likelihood of successful recruitment of Tamarix to be highest along unregulated river reaches and to remain high across a gradient of regulated flows. Recruitment probability for Populus, in contrast, was highest under free-flowing conditions and declined abruptly under even slight flow modification (IFM > 0.1). Adult Tamarix was most abundant at intermediate levels of IFM. Populus abundance declined sharply with modest flow regulation (IFM > 0.2) and was not present at the most flow-regulated sites. Dominance of Tamarix was highest along rivers with the most altered flow regimes. At the 16 least regulated sites, Tamarix and Populus were equally abundant. Given observed patterns of Tamarix recruitment and abundance, we infer that Tamarix would likely have naturalized, spread, and established widely in riparian communities in the absence of dam construction, diversions, and flow regulation in western North America. However, Tamarix dominance over native species would likely be less extensive in the absence of human alteration of river-flow regimes. Restoration that combines active mechanical removal of established stands of Tamarix with a program of flow releases conducive to native species establishment and persistence is hypothesized to facilitate the codominance of Populus in reaches where it has become rare. Our findings have implications for planning flow-related stream restoration, for developing realistic expectations for yield on investment in prescribed flow releases, and for planning flow-related interventions that might be possible if control and management of invasive plant species along rivers is a goal.

Bedrock channel morphology

Analyses of 41 bedrock channel reaches indicate quantifiable relationships between bedrock channel morphology and reach- scale hydraulic and substrate variables. Discriminant analysis was used to develop a discriminant criterion based on reach-averaged channel gradient, substrate heterogeneity, and Selby rock-mass strength. This criterion correctly classified 70% of the observations into one of five channel morphologic types. Channels formed at higher gradients have a morphology that effectively maximizes the erosional force, whereas a morphology that evenly distributes flow energy or dissipates flow energy internally is associated with lower gradients. These results suggest that bedrock channel morphology, like alluvial channel morphology, reflects a quantifiable balance between hydraulic driving and substrate resisting forces.

Downstream hydraulic geometry and channel adjustment during a flood along an ephemeral, arid-region drainage

In September 1997, a dissipating tropical storm caused a flood with an estimated maximum discharge of 240 m 3 /s along Yuma Wash, an ephemeral braided system draining 186 km 2 in southwest Arizona. Older high-water marks that record a flood peak of 1280 m 3 /s provide a reasonable estimate for the probable maximum flood along the wash. Detailed channel crosssectional surveys during 1995 and again in 1998, <6 months after the 1997 flood, facilitated examination of downstream hydraulic geometry and channel adjustment during the flood. Channel width increased substantially downstream (exponent of 0.78), presumably because of low bank resistance, whereas hydraulic depth and velocity had modest downstream increases (0.15 and 0.14, respectively). Channel aggradation generally occurred along wider, braided reaches; moreover, degradation occurred in narrow reaches with fewer channels. Aggradation and degradation also appeared to be governed by a threshold relationship between flow depth and vegetated bars. Degradation occurred where flow was confined within a channel or subchannel. At discharges sufficiently high to overtop vegetated bars, greater roughness facilitated sediment deposition and channels aggraded. A discriminant function correctly classified nearly 90% of the cross-sections as scoured or filled using a single hydraulic variable, maximum depth of flow during the dissipating tropical storm. D 2002 Elsevier Science B.V. All rights reserved.

Consequences of propagule dispersal and river fragmentation for riparian plant community diversity and turnover

The spatial distribution and temporal availability of propagules fundamentally constrain plant community development. This study experimentally tested several hypotheses about the relative roles of wind and water dispersal in colonization and development of riparian communities along rivers. Through controlling the source of propagules (dispersed by wind, water, or both) reaching newly created, bare river margin sites, we isolated the relative roles of dispersal and other factors in plant community development over five years. Replicated treatments were established at 12 sites spanning 400 km along two adjacent rivers in northern Sweden, one fragmented by a series of dams, the other free-flowing. Bare river margins receiving only water-dispersed propagules had significantly higher species richness compared to plots receiving only wind-dispersed propagules during the initial two years of colonization. Species richness increased annually throughout the study along tranquil and turbulent reaches of the free-flowing river but reached an asymptote at comparatively low richness after a single year on the impounded river. Propagule source strongly influenced species richness during the initial establishment along both rivers, with richness being significantly higher in plots receiving water-dispersed seeds. This strong treatment effect continued to be important through time along the regulated river but diminished in importance along the free-flowing river where other factors such as soil moisture, light availability, and exposure of sites to fluvial disturbance overshadowed the influence of dispersal pathway in mediating species richness. This suggests that hydrochory (plant dispersal by water) may be more important for maintenance of diversity in regulated systems where long-distance dispersal is absent or negligible, but that the rich local propagule source along free-flowing rivers supports high species richness. The number of unique species was higher in water-dispersal plots along both the regulated and free-flowing rivers. This result suggests that hydrochory may contribute to temporal variability of sites, may enhance richness over time, and may have an important role in meta-population and meta-community dynamics of plant communities through long-distance (and local) dispersal and chance colonization. Our findings provide experimental evidence that water dispersal of plant propagules influences colonization dynamics and is important for long-term community development in riparian zones.