Thomas Dunne

The Network Dynamics Hypothesis: How Channel Networks Structure Riverine Habitats

Abstract Hierarchical and branching river networks interact with dynamic watershed disturbances, such as fires, storms, and floods, to impose a spatial and temporal organization on the nonuniform distribution of riverine habitats, with consequences for biological diversity and productivity. Abrupt changes in water and sediment flux occur at channel confluences in river networks and trigger changes in channel and floodplain morphology. This observation, when taken in the context of a river network as a population of channels and their confluences, allows the development of testable predictions about how basin size, basin shape, drainage density, and network geometry interact to regulate the spatial distribution of physical diversity in channel and riparian attributes throughout a river basin. The spatial structure of river networks also regulates how stochastic watershed disturbances influence the morphology and ages of fluvial features found at confluences.

Effects of Rainfall, Vegetation, and Microtopography on Infiltration and Runoff

Apparent, or effective, infiltration rates on grassland hillslopes vary with rainfall intensity and flow depth because of the interaction between rainfall, runoff, and vegetated microtopography. The higher parts of the microtopography are occupied by greater densities of macropores and therefore have much greater hydraulic conductivities than the intervening microdepressions. On short hillslopes and plots the apparent infiltration rate is simply the spatial average of the saturated and unsaturated conductivities of this surface. The proportion of the surface which is saturated and the value to which the unsaturated conductivity is raised depends on the rainfall intensity. On longer hillslopes the downslope increase in flow depth in microtopographic depressions progressively inundates more permeable, vegetated mounds so that the hydraulic conductivity of a greater proportion of the surface is raised to its saturated value. For this reason the apparent infiltration rate increases downslope, even in the absence of spatial trends in any of the surface characteristics that affect infiltration. Apparent, or effective, infiltration rate depends on hillslope length. Consequently, steady state discharge does not increase linearly with distance downslope. These two fundamental relationships between infiltration, rainfall intensity, and runoff are analyzed on the basis of sprinkling-infiltrometer measurements and a mathematical model.

Stochastic forcing of sediment supply to channel networks from landsliding and debris flow

Sediment influx to channel networks is stochastically driven by rainstorms and other perturbations, which are discrete in time and space and which occur on a landscape with its own spatial variability in topography, colluvium properties, and state of recovery from previous disturbances. The resulting stochastic field of sediment supply interacts with the topology of the channel network and with transport processes to generate spatial and temporal patterns of flux and storage that characterize the sedimentation regime of a drainage basin. The regime varies systematically with basin area. We describe how the stochastic sediment supply is generated by climatic, topographic, geotechnical, and biotic controls that vary between regions. The general principle is illustrated through application to a landscape where sediment is supplied by mass wasting, and the forcing variables are deterministic thickening of colluvium, random sequences of root-destroying wildfires, and random sequences of rainstorms that trigger failure in a population of landslide source areas with spatial variance in topography and colluvium strength. Landslides stop in channels or convert to scouring debris flows, depending on the nature of the low-order channel network. Sediment accumulates within these channels for centuries before being transferred downstream by debris flows. Time series of sediment supply, transport, and storage vary with basin scale for any combination of climatic, topographic, and geotechnical controls. In a companion paper (Benda and Dunne, this issue) we use simulations of timing, volumes, and locations of mass wasting to study the interaction between a stochastically forced sediment supply and systematic changes of storage and flux through channel networks.

Storage and Remobilization of Suspended Sediment in the Lower Amazon River of Brazil

In the lower Amazon River, suspended sediment is stored during rising stages of the river and resuspended during falling river stages. The storage and resuspension in the reach are related to the mean slope of the flood wave on the river surface; this slope is smaller during rising river stages than during falling stages. The pattern of storage and resuspension damps out the extreme values of high and low sediment discharge and tends to keep them near the mean value between 3.0 x 106 and 3.5 x 106 metric tons per day. Mean annual discharge of suspended sediment in the lower Amazon is between 1.1 x 109 and 1.3 x 109 metric tons per year.

Exchanges of sediment between the flood plain and channel of the Amazon River in Brazil

Sediment transport through the Brazilian sector of the Amazon River valley, a distance of 2010 km, involves exchanges between the channel and the flood plain that in each direction exceed the annual flux of sediment out of the river at Obidos (∼1200 Mt yr −1 ). The exchanges occur through bank erosion, bar deposition, settling from diffuse overbank flow, and sedimentation in flood-plain channels. We estimated the magnitude of these exchanges for each of 10 reaches of the valley, and combined them with calculations of sediment transport into and out of the reaches based on sediment sampling and flow records to define a sediment budget for each reach. Residuals in the sediment budget of a reach include errors of estimation and erosion or deposition within the channel. The annual supply of sediment entering the channel from bank erosion was estimated to average 1570 Mt yr −1 (1.3 × the Obidos flux) and the amount transferred from channel transport to the bars (380 Mt yr −1 ) and the flood plain (460 Mt yr −1 in channelized flow; 1230 Mt yr −1 in diffuse overbank flow) totaled 2070 Mt yr −1 (1.7 × the Obidos flux). Thus, deposition on the bars and flood plain exceeded bank erosion by 500 Mt yr −1 over a 10–16 yr period. Sampling and calculation of sediment loads in the channel indicate a net accumulation in the valley floor of approximately 200 Mt yr −1 over 16 yr, crudely validating the process-based calculations of the sediment budget, which in turn illuminate the physical controls on each exchange process. Another 300–400 Mt yr −1 are deposited in a delta plain downstream of Obidos. The components of the sediment budget reflect hydrologic characteristics of the valley floor and geomorphic characteristics of the channel and flood plain, which in turn are influenced by tectonic features of the Amazon structural trough.

STOCHASTIC FORCING OF SEDIMENT ROUTING AND STORAGE IN CHANNEL NETWORKS

The stochastic field of sediment supply to the channel network of a drainage basin depends on the large-scale interactions among climatically driven processes such as forest fire and rainstorms, topography, channel network topology, and basin scale. During infrequent periods of intense erosion, large volumes of colluvium are concentrated in parts of a channel network, particularly near tributary junctions. The rivers carry bed material and wash load downstream from these storage sites at different rates. The bed material travels slowly, creating transient patterns of sediment transport, sediment storage, and channel morphology along the channel network. As the concentrations of bed material migrate along the network their waveforms can undergo changes by diffusion, interference at tributary junctions, and loss of mass through temporary sediment storage in fans and terraces and through particle abrasion, which converts bed material to wash load. We investigated how these processes might influence the sediment mass balance in channels of third and higher order in a 215-km 2 drainage basin within the Oregon Coast Range over a simulated 3000-year period with a climate typical of the late Holocene. We used field measurements and a simulation model to illustrate interactions between the major controls on large-scale processes functioning over long periods of time in complex drainage basins.

Interferometric radar measurements of water level changes on the Amazon flood plain

Measurements of water levels in the main channels of rivers, upland tributaries and floodplain lakes are necessary for understanding flooding hazards, methane production, sediment transport and nutrient exchange. But most remote river basins have only a few gauging stations and these tend to be restricted to large river channels. Although radar remote sensing techniques using interferometric phase measurements have the potential to greatly improve spatial sampling, the phase is temporally incoherent over open water and has therefore not been used to determine water levels. Here we use interferometric synthetic aperture radar (SAR) data, acquired over the central Amazon by the Space Shuttle imaging radar mission, to measure subtle water level changes in an area of flooded vegetation on the Amazon flood plain. The technique makes use of the fact that flooded forests and floodplain lakes with emergent shrubs permit radar double-bounce returns from water and vegetation surfaces, thus allowing coherence to be maintained. Our interferometric phase observations show decreases in water levels of 7–11 cm per day for tributaries and lakes within ∼20 km of a main channel and 2–5 cm per day at distances of ∼80 km. Proximal floodplain observations are in close agreement with main-channel gauge records, indicating a rapid response of the flood plain to decreases in river stage. With additional data from future satellite missions, the technique described here should provide direct observations important for understanding flood dynamics and hydrologic exchange between rivers and flood plains.

Episodic sediment accumulation on Amazonian flood plains influenced by El Niño/Southern Oscillation

Continental-scale rivers with a sandy bed sequester a significant proportion of their sediment load in flood plains. The spatial extent and depths of such deposits have been described, and flood-plain accumulation has been determined at decadal timescales, but it has not been possible to identify discrete events or to resolve deposition on near-annual timescales. Here we analyse 210Pb activity profiles from sediment cores taken in the pristine Beni and Mamore river basins, which together comprise 720,000 km2 of the Amazon basin, to investigate sediment accumulation patterns in the Andean–Amazonian foreland. We find that in most locations, sediment stratigraphy is dominated by discrete packages of sediments of uniform age, which are typically 20–80 cm thick, with system-wide recurrence intervals of about 8 yr, indicating relatively rare episodic deposition events. Ocean temperature and stream flow records link these episodic events to rapidly rising floods associated with La Niña events, which debouch extraordinary volumes of sediments from the Andes. We conclude that transient processes driven by the El Niño/Southern Oscillation cycle control the formation of the Bolivian flood plains and modulate downstream delivery of sediments as well as associated carbon, nutrients and pollutants to the Amazon main stem.

Channel-floodplain geomorphology along the Solimões-Amazon River, Brazil

Across the cratonic landscape of Brazil the Solimoes-Amazon River transports to its delta plain 1240 Mt of suspended sediment derived from Andean erosion and reworks another 3200 Mt of floodplain sediments. Distribution of these sediments has resulted in a variable along-stream pattern of geomorphology. The upstream reaches are characterized by sediment erosion in the main channel and deposition in floodplain channels that are an order of magnitude smaller in discharge than the main channel. Sediment deposition in and migration of the floodplain channels erases oxbow lakes of the main channel and yields an intricate scroll-bar topography that forms the boundaries of hundreds of long, narrow lakes. In contrast, downstream reaches are characterized by channels restricted by stabilizing, long-term, levee building and floodplain construction dominated by overbank deposition. Overbank deposition buries the scroll-bar topography, resulting in a flat floodplain covered by a patchwork of large, more equant, shallow lakes. On the basis of estimated rates of recycling of floodplain sediments, the modern floodplain of the Brazilian Amazon could have been recycled in <5000 yr, and is recycled more rapidly in the upstream than the downstream reaches. The cratonic interior is interrupted by structural arches that bound intracratonic basins. Four of these arches cross the valley of the main river system at intervals of several hundred kilometres and impart a tectonic imprint on the channel-floodplain geomorphology at this spatial scale. Structural arches appear to exert a primary influence by promoting entrenchment of the river as it passes through zones of deformation, thus restricting channel movement. For example, as the river crosses the Purus arch, the valley narrows to <20 km compared to an average of ∪45 km, the water-surface gradient decreases, sediment is deposited, and yet the rate of channel migration is negligible. Hence, the effect of the arches is to create a landscape where, on the spatial scale of hundreds of kilometres, the river is confined and entrenched in its valley, is straight, and is relatively immobile. Local valley tilting apparently unrelated to the arch structures also imprints the geomorphology. In particular, a tilted valley in the upstream reaches appears to have caused avulsions which have left behind the only large-scale, oxbow-type features on the Brazilian Amazon River floodplain.

Relation of field studies and modeling in the prediction of storm runoff

Abstract Routine methods of predicting storm hydrographs do not require accurate conceptual models of storm runoff processes. Application of hydrology to the analysis of landform evolution and other scientific and land-management problems requires realistic concepts of runoff processes and their variation within drainage basins. These concepts need to be refined, developed, and formalized through more vigorous combination of rigorously designed field experiments and realistic physically-based mathematical models.