Subaqueous pyroclastic flows and ignimbrites: an assessment

An assessment of the literature on subaqueous pyroclastic flows and their deposits shows that the term pyroclastic flow is frequently used loosely to describe primary, hot gas-rich pyroclastic flows, mass-flows which resulted from the transformation of gassupported flows into water-supported ones, and secondary mass-flows carrying redeposited pyroclastic debris. Based on subaerial pyroclastic flows, the term pyroclastic flow should be restricted to demonstrably hot, gas-rich mass-flows of pyroclastic debris. Using this definition, very few examples of subaqueous pyroclastic deposits with evidence for hot emplacement and of having been wholly submerged have been described. In the majority of these cases, the evidence for a hot state of emplacement and for the subaqueous nature of the host depositional environment is inadequate. The only unequivocal cases of hot pyroclastic flow deposits with adequate supporting evidence are the Ordovician nearshore, shallow marine ignimbrites of Ireland and Wales, and Miocene ignimbrites of southwest Japan, resulting from the passage of subaerially erupted pyroclastic flows into shallow water. Other possible examples are near-vent dense clast deposits in the Donzurobo Formation of Japan, possible submarine intra-caldera ponded ignimbrite successions in California and Wales, and near-vent pumiceous deposits of Ramsay Island, Wales. All other purported cases are either clearly the result of water-supported mass-flow transportation and deposition (debris avalanches, debris flows, turbidity currents), or lack adequate supporting evidence regarding the heat state or the palaeoenvironment. Only the shallow marine ignimbrites of Ireland and Wales show adequate evidence of welding, but even these could have been nearly wholly exposed above sea-level when welding occurred. We conclude that when pyroclastic flows enter water they are generally disrupted explosively and/or ingest water and transform into water-supported mass-flows, and we suggest the various scenarios in which this occurs. There is no evidence to suggest that welding in wholly subaqueous environments is common.     

Depositional processes and gas pore pressure in pyroclastic flows: an experimental perspective

The depositional processes and gas pore pressure in pyroclastic flows are investigated through scaled experiments on transient, initially fluidized granular flows. The flow structure consists of a sliding head whose basal velocity decreases backwards from the front velocity (U _f) until onset of deposition occurs, which marks transition to the flow body where the basal deposit grows continuously. The flows propagate in a fluid-inertial regime despite formation of the deposit. Their head generates underpressure proportional to U _f ² whereas their body generates overpressure whose values suggest that pore pressure diffuses during emplacement. Complementary experiments on defluidizing static columns prove that the concept of pore pressure diffusion is relevant for gas-particle mixtures and allow characterization of the diffusion timescale (t _d) as a function of the material properties. Initial material expansion increases the diffusion time compared with the nonexpanded state, suggesting that pore pressure is self-generated during compaction. Application to pyroclastic flows gives minimum diffusion timescales of seconds to tens of minutes, depending principally on the flow height and permeability. This study also helps to reconcile the concepts of en masse and progressive deposition of pyroclastic flow units or discrete pulses. Onset of deposition, whose causes deserve further investigation, is the most critical parameter for determining the structure of the deposits. Even if sedimentation is fundamentally continuous, it is proposed that late onset of deposition and rapid aggradation in relatively thin flows can generate deposits that are almost snapshots of the flow structure. In this context, deposition can be considered as occurring en masse, though not strictly instantaneously.     

Sole Marks at the base of the late Pleistocene Abrigo Ignimbrite,Tenerife: implications for transport and depositional processes at the base of pyroclastic flows

Sole marks, which are common in turbidites, have been observed as casts at the base of the Abrigo Ignimbrite on Tenerife, Canary Islands. They have been engraved by pebble to cobble-sized lithic tools in a soft, cohesive fine-grained substrate. The casts range from long, parallel groove marks, often with the tool embedded at their termination, to short, elongate impact marks and are useful as a flow-direction marker. They were formed from a highly energetic pyroclastic flow pulse and were almost immediately infilled with ash after rapid waning of flow. Large lithic tools, which formed groove marks, were held in place under high gas and grain dynamic pressures and moved forward by their own momentum and the drag force exerted by a highly concentrated granular flow. Impact marks were formed by smaller lithic tools, which had more freedom of movement within the agitated, chaotic flow. Scour structures on the lee side of stationary lithic tools may have formed by local turbulence in their wake.Editorial responsibility: T. Druitt     

Explosive activity and generation mechanisms of pyroclastic flows at Arenal volcano, Costa Rica between 1987 and 2001

Explosive activity at Arenal and associated tephra fall that has occurred over the 14-year period from 1987–2001 is described. Explosions have been notably variable in both frequency and size. A marked decrease in both frequency and quantity of tephra fallout occurred in early 1998 until the end of 2001. Grainsize distributions of cumulative tephra samples collected once a month are typically bimodal. Aggregation causing premature fallout of fine ash and possibly fallout from ash plumes produced by pyroclastic flows are considered responsible for the bimodality of fallout. Scanning electron microscopy of the glass component of tephra from single explosions show predominantly blocky and blocky/fluidal clast types, interpreted as being the product of vulcanian type explosions. Fragmentation of a mainly rigid, degassed magma body, and a minor molten component is inferred for these explosions. Pyroclastic flows were produced either associated with the larger explosions by a mechanism of column collapse (1987–1990), or unrelated to explosions by partial collapse of the crater wall (1993, 1998, 2000, 2001). Pyroclastic flow activity has migrated from west to north during the period reported. Pyroclastic flow deposits are variable in the quantity of juvenile material and any associated surge component. Large juvenile blocks were partially molten on emplacement and many have a typical cauliform texture. Blocks with both juvenile and lithic textures indicate that at the summit magma was in intimate contact with the pre-existing edifice, rather than as a simple open crater or lava pool. Crater wall collapse may have been promoted by the reduction in explosive activity, which has increased the lava accumulation at the summit and in turn increased instability of the summit region. Thus although explosive activity has waned, if the lava output is maintained, the hazard of pyroclastic flows is likely to continue.Editorial responsibility: R. Cioni     

Damage to structures by pyroclastic flows and surges, inferred from nuclear weapons effects

In order to define the risk from explosive eruptions, one must constrain both the probability of explosive events and the effects, or consequences, of those events. This paper focuses on the effects of pyroclastic flows and surges (here termed ‘pyroclastic density currents', or PDCs) on buildings, infrastructure elements, and to some extent on vehicles. PDCs impart a lateral force to such structures in the form of dynamic pressure, which depends on the bulk density of the PDC (which in turn depends mainly on particle concentration) and its velocity. For reasonable ranges of particle concentration (10⁻³ to 0.5) and velocities (10 to 300 m/s), dynamic pressure on the upstream face of a structure ranges from 0.1 kPa to 10⁴ kPa. Lateral loads ranging up to about 100 kPa were produced during nuclear weapons tests in the 1940s and 1950s that were designed to study the effects of such loading on a variety of structures for civil defense and emergency response purposes in the event of nuclear war. Although considerable simplifications are involved, the data from these weapon tests provide useful analog information for understanding the effects of PDCs. I reviewed data from the nuclear tests, describing the expected damage from different loadings. Tables are provided that define the response of different structural elements (e.g., windows, framing, walls) and whole structures to loading in probabilistic terms, which in principle account for variations in construction quality, orientation, and other factors. Finally, damage documented from historical eruptions at Mt. Lamington (1951), Herculaneum (AD 79 Vesuvius eruption), and St. Pierre (1902 Mt. Pelee eruption) is reviewed. Damage patterns, combined with estimates of velocity, provide an independent estimate of particle concentration in the PDCs. Details of structural damage should be recorded and mapped around future eruptions in order to help refine this aspect of consequence analysis. Another fruitful approach would be to combine numerical simulations of eruption scenarios, which can produce simulated maps of dynamic pressure, with GIS-based data on structures for a given region; the result would be predictions of consequences that could be used for planning and emergency response training.     

Evaluation of a gravel transport sensor for bed load measurements in natural flows

A recent acoustic instrument （Gravel Transport Sensor, GTS） was tested for predicting sediment transport rate （bed load rate） in gravel bed streams. The GTS operation is based on the particle collision theory of submerged obstacles in fluids. When particles collide with the GTS cylinder their momentum is recorded in the form of ping rates. The GTS is attractive for further consideration here because of its potential to provide continuous unattended local bed load measurements, especially in areas found in streams that access may be difficult under extreme conditions. Laboratory experiments coupled with numerical simulations for the same flow conditions were performed in order to determine the conditions under which particles of different size will hit the GTS cylinder and be able to register a ping rate. The GTS was able to detect the number of particles with diameter in the range of 15.9 to 25.4 mm, with reasonable accuracy, if the applied Shields effective stress τ＊e = τ＊ - τ＊cr was in the range of 0.006 to 0.015. A drawback of the tested prototype GTS, however, was that it exerted increased resistance on the incoming particles. The added drag effects increased the overall resistance that was exerted by the flow on particles and thus increased the likelihood that particles will rest in the ambient region of the cylinder instead of hitting it. Numerical simulation of the flow around the GTS cylinder revealed that changing the prototype geometry from cylindrical to ellipsoid or rhomboid will increase the likelihood of the particles hitting the instrument under the same flow conditions failed by the original tested GTS cylinder.     

Fluvial seed dispersal of riparian trees: transport and depositional processes

Fluvial seed dispersal considers both the transport and deposition of seeds where channel geomorphic structures, hydrology and seed dispersal traits contribute to transport times and depositional locations. This study examines the influence of stream flow patterns on fluvial seed dispersal of buoyant white alder (Alnus rhombifolia) seeds by applying a one‐dimensional transport model. Conceptually, the model separates the stream into two components: (i) the main channel where the seeds are transported downstream; and (ii) the transient storage zone where seeds are temporarily detained or deposited on the river bank. Transport processes are characterized by an advection–dispersion equation which is coupled to a transient storage model using an exponential decay term. The model parameters: longitudinal dispersion (D_L), exchange coefficient (α), main channel area (A) and storage zone (A_s) are estimated based on field experiments conducted in a confined, bedrock‐gravel bed river with pool‐riffle morphology located in coastal northern California. The riparian zone is inhabited by Alnus rhombifolia that disperse buoyant seeds in mid‐spring coinciding with the end of the wet, Mediterranean season. Artificial seeds, with similar traits of buoyancy and density to alder seeds, were used to quantify transport times and depositional locations. Preferential deposition resulted in stream reaches with larger A_s, high A_s/A ratios, and faster exchange coefficients corresponding to divergent stream flow (back‐eddies, re‐circulating flow, flow expansions) caused by geomorphic structures such as the ends of bar/riffle features and bends in the stream. The results demonstrate the importance of transient storage for seed transport and depositional processes. Morphological features that increase a channel's complexity create complex flow structures that detain seeds and provide a greater opportunity for deposition to occur. The model provides a simplification of river hydraulics to represent dispersal dynamics and lends itself to further understanding of hydrochory processes and associated population structure. Copyright © 2015 John Wiley & Sons, Ltd.     

Basaltic pyroclastic flows of Fuji volcano, Japan: characteristics of the deposits and their origin

Fuji volcano is the largest active volcano in Japan, and consists of Ko-Fuji and Shin-Fuji volcanoes. Although basaltic in composition, small-volume pyroclastic flows have been repeatedly generated during the Younger stage of Shin-Fuji volcano. Deposits of those pyroclastic flows have been identified along multiple drainage valleys on the western flanks between 1,300 and 2,000 m a.s.l., and have been stratigraphically divided into the Shin-Fuji Younger pyroclastic flows (SYP) 1 to 4. Downstream debris flow deposits are found which contain abundant material derived from the pyroclastic flow deposits. The new¹⁴C ages for SYP1 to SYP4 are 3.2, 3.0, 2.9, and 2.5 ka, respectively, and correspond to a period where explosive summit eruptions generated many scoria fall deposits mostly toward the east. The SYP1 to SYP4 deposits consist of two facies: the massive facies is about 2 m thick and contains basaltic bombs of less than 50 cm in size, scoria lapilli, and fresh lithic basalt fragments supported in an ash matrix; the surge facies is represented by beds 1 to 15 cm thick, consisting mainly of ash with minor amount of fine lapilli. The bombs and scoria are 15 to 30% in volume within the massive facies. The ashes within the SYP deposits consist largely of comminuted basalt lithics and crystals that are derived from the Middle-stage lava flows exposed at the western flanks. SYP1 to SYP4 were only dispersed down the western flanks. The reason for this one-sided distribution is the asymmetric topography of the edifice; the western slopes of the volcano are the steepest (over 34 degrees). Most pyroclastic materials cannot rest stably on the slopes steeper than 33 degrees. Therefore, ejecta from the explosive summit eruptions that fell on the steep slopes tumbled down the slopes and were remobilized as high-temperature granular flows. These flows consisted of large pyroclastics and moved as granular avalanches along the valley bottom. Furthermore, the avalanching flows increased in volume by abrasion from the edifice and generated abundant ashes by the collision of clasts. The large amount of the fine material was presumably available within the transport system as the basal avalanches propagated below the angle of repose. Taking the typical kinetic friction coefficient of small pyroclastic flows, such flows could descend the western flanks where scattered houses are below 1,000 m a.s.l. A similar type of pyroclastic flow could result if explosive summit eruptions occur in the future.Editorial responsibility: R Cioni     

Gas transport and bubble collapse in rhyolitic magma: an experimental approach

A series of experiments was conducted to test concepts of porous flow degassing of rhyolitic magma during ascent and of the subsequent collapse of vesicles in degassed magma to form obsidian. Dense, synthetically hydrated, natural glasses were pressurized under water-saturated conditions and then decompressed to achieve a range of porosities in the presence of a tracer vapor, D₂O. Rapid isotopic exchange indicative of vapor transport rather than of simple diffusion occurred at a porosity >60 vol.%, in accord with earlier gas permeability measurements on cold natural samples. In another series of experiments, natural and synthetic pumices, vesiculated by degassing to atmospheric pressure, rapidly collapsed to dense glass on repressurization to the modest pressures prevailing in lava flows. No relict bubble textures remained. These results support the hypothesis that effusive eruptions result from the syneruptive escape of gas from permeable magmatic foam, and that a process analogous to welding yields dense lavas when such foams are extruded.     

Gas release rates from pyroclastic flows: a assessment of the role of fluidisation in their emplacement

Estimates are presented for the rates of release of dissolved water from the particles in a pyroclastic flow by diffusion. Velocities of gas escaping from pyroclastic flows of different thicknesses are calculated. For quite small residual gas contents (0.2 to 0.8% H₂O), gas velocities of 10 to over 100 cm/sec during the first 10³ sec of release are estimated for flows of thickness 1 to 20 m, which experimental studies demonstrate are the velocities required to fluidise fine to medium ash. Flows with high residual gas contents or large volume (thick flows) are likely to be substantially fluidised by exsolving gas. 30% to over 60% of the particles in such flows are predicted to be fluidised. Fluidisation is thus believed to be an important mechanism in the flow and in determining the mobility of the large magnitude, prehistoric pyroclastic flows which formed extensive ignimbrite sheets. Small pyroclastic flows, however, of the magnitude observed in several historic eruptions are not believed to be fluidised, because of their low residual gas contents, small volume, and the substantial amount of cooling that occurs during their emplacement.     

The anisotropy of magnetic susceptibility of lava flows: an experimental approach

Measurements of the anisotropy of magnetic susceptibility (AMS) of natural lavas have shown that AMS varies with depth within a lava flow. We have investigated the reasons for such variation by studying the effects of temperature and strain rate on the AMS of recent lava in the laboratory. Samples of lava from Kilauea were melted and subjected to a range of strain rate and cooling histories. The results show that the degree of anisotropy is a function of both the thermal and shearing history of a sample. High degrees of anisotropy were found only in samples that were deformed at temperatures close to those encountered during eruption and then rapidly quenched. Lavas subjected to similar shear stresses at high temperatures had low degrees of anisotropy if allowed to cool down slowly without further deformation. Additionally, lava subjected to complex shearing yield a lower degree of anisotropy even when high strain rates were imposed on it. These results lead to the conclusion that only the last phase of deformation is detectable using AMS and that high strain rates will not result in high degrees of anisotropy if either deformation ends while lava is still fluid or if the orientation of the maximum shear stress varies with time. The relation between the orientation of the principal susceptibilities and that of shear is less sensitive to variation on shear with time. Consequently, flow directions can be inferred confidently with this type of measurements.     

Complex spatter- and pumice-rich pyroclastic deposits from an andesitic caldera-forming eruption: the Siwi pyroclastic sequence,Tanna, Vanuatu

The small- to moderate-volume, Quaternary, Siwi pyroclastic sequence was erupted during formation of a 4 km-wide caldera on the eastern margin of Tanna, an island arc volcano in southern Vanuatu. This high-potassium, andesitic eruption followed a period of effusive basaltic andesite volcanism and represents the most felsic magma erupted from the volcano. The sequence is up to 13 m thick and can be traced in near-continuous outcrop over 11 km. Facies grade laterally from lithic-rich, partly welded spatter agglomerate along the caldera rim to two medial, pumiceous, non-welded ignimbrites that are separated by a layer of lithic-rich, spatter agglomerate. Juvenile clasts comprise a wide range of densities and grain sizes. They vary between black, incipiently vesicular, highly elongate spatter clasts that have breadcrusted pumiceous rinds and reach several metres across to silky, grey pumice lapilli. The pumice lapilli range from highly vesicular clasts with tube or coalesced spherical vesicles to denser finely vesicular clasts that include lithic fragments.Textural and lithofacies characteristics of the Siwi pyroclastic sequence suggest that the first phase of the eruption produced a base surge deposit and spatter-poor pumiceous ignimbrite. A voluminous eruption of spatter and lithic pyroclasts coincided with a relatively deep withdrawal of magma presumably driven by a catastrophic collapse of the magma chamber roof. During this phase, spatter clasts rapidly accumulated in the proximal zone largely as fallout, creating a variably welded and lithic-rich agglomerate. This phase was followed by the eruption of moderately to highly vesiculated magma that generated the most widespread, upper pumiceous ignimbrite. The combination of spatter and pumice in pyroclastic deposits from a single eruption appears to be related to highly explosive, magmatic eruptions involving low-viscosity magmas. The combination also indicates the coexistence of a spatter fountain and explosive eruption plume for much of the eruption.Editorial responsibility: R. Cioni     

The role of fluidization in the emplacement of pyroclastic claws: An experimental approach

A series of experiments was carried out to review the process of fluidization for a number of particulate materials having various sorting and grain shape characteristics. The up (increasing gas velocity) curve on a gas velocity/bed-pressure drop plot for a poorly sorted mixture of irregularly shaped particles is divided into three sections; non-expanded, expanded, and segregating. These sections are used to define a threefold genetic classification of pyroclastic flows which can be directly linked to conditions within a semifluidized parent flow. Type 1 flows are ungraded and mostly result from hot-avalanche flows and pyroclastic flows formed of relatively dense material. Type 2 and type 3 flows are mainly pumiceous ignimbrites and are distinguished by expansion induced coarse-tail grading, coupled with segregation structures in the latter. The implications of this classification are discussed with reference to the flow regimes, deposition slope angles, crystal concentration and “fossil fumaroles” variously developed in the flows. The relevance of the source of fluidizing gas is discussed in relation to the zoning of flow types within a single flow unit and it is shown how this, and its attendant structures, can be used to estimate the relative importance of each gas source during and after flow emplacement.     

Flow and deposition of pyroclastic granular flows: A type example from the 1975 Ngauruhoe eruption,New Zealand

Small-volume pyroclastic density currents (PDCs) are generated frequently during explosive eruptions with little warning. Assessing their hazard requires a physical understanding of their transport and sedimentation processes which is best achieved by the testing of experimental and numerical models of geophysical mass flows against natural flows and/or deposits. To this end we report on one of the most detailed sedimentological studies ever carried out on a series of pristine small-volume PDC deposits from the 1975 eruption of Ngauruhoe volcano, whose emplacement were also witnessed during eruption. Using high-resolution GPS surveys, a series of lateral excavations across the deposits, and bulk sedimentological analysis we constrained the geomorphology, internal structure and texture of the deposits with respect to laterally varying modes of deposition.     

Mt. Pelée, martinique: may 8 and 20, 1902, pyroclastic flows and surges

The distribution, stratigraphic relationships and fragmental components of the May 8 and 20, 1902, pyroclastic flows from Mt. Pelée, Martinique, together with eyewitness accounts, suggest the following explanation for those eruptions. The eruptions were vertically directed magmatic (perhaps initiated phreatically), and contained abundant juvenile lithics from congealed magma of the dome and neck. This resulted in a two-part eruption column having (1) a dense, lithic-charged part which collapsed into the crater and flowed out of a pre-existing notch in its side, giving rise to pyrochlastic flows, and (2) a magmatically derived column containing gases, juvenile vitric material and crystals which largely by-passed the neck and dome and escaped into the atmosphere. All of the energy of the flows was apparently focused through the notch. They emerged fully turbulent and flowed down Rivière Blanche. Gravity segregation of large and abundant fragments soon resulted in a dense, high-concentration, poorly fluidized block-and-ash flow confined to the valley, leaving above a fully turbulent, high-energy ash-cloud surge. As the ash-cloud surge moved down the mountain, it continued to expand outward. The process of gravity segregation continued as the ash-cloud surge expanded, resulting in secondary block-and-ash underflows. Toward St. Pierre, the secondary block-and-ash flows developed on a gently sloping upland surface 100 m or more above the valley of Rivière Blanche. The turbulent, fragment-depleted surges above the secondary block-and-ash flows maintained sufficient energy to devastate the landscape outward to about 3000 m, including St. Pierre. The surges refracted around obstacles and in one place, moved up a small valley in a direction opposite to the main flows.     

Probabilistic approach to modeling of velocity distributions in fluid flows

This paper presents a synthesis of a probability-based approach and the underpinning, mathematical and philosophical foundations that have evolved to date, as well as applications in modeling of vertical and two-dimensional velocity distributions that have direct implications to measurements and estimation of transport of mass, momentum and energy in fluid flows. The approach draws inferences from a probability law identified by maximizing the information entropy under the constraints imposed by the available information. It gives predictions considered to be the most probable or objective on the basis of the available information. The probabilistic approach complements the deterministic approach of hydrodynamics. The difference in the point of view between the two approaches creates a different view about the available information. Some information, such as the location and magnitude of maximum velocity, the ratio of mean and maximum velocities, that may not appear to have direct use to the deterministic approach in flow predictions become important and useful to the probabilistic approach.     

Vent area and depositional mechanisms of the Upper Member of the Neapolitan Yellow Tuff (Campi Flegrei, Italy): new insights from directional fabric through image analysis

In order to provide new information about the source area and depositional mechanisms of the Upper Member of the Neapolitan Yellow Tuff (NYT), a prominent pyroclastic deposit of the Campi Flegrei Volcanic District (southern Italy), statistics on directional fabric, by means of computer-assisted image analysis on 32 rock samples, were compiled. Seventeen samples were collected along vertical direction on two selected exposures and fifteen were taken from outcrops widely distributed all around the Campi Flegrei Volcanic District. Fabric measurements within the investigated successions reveal a vertically homogeneous direction of the mean particle iso-orientation, with considerable variability in the strength of particle iso-orientation even at cm-scale. The existence of particle iso-orientation can be related to continuous sedimentation from a concentrated bedload region beneath suspension currents, producing massive or inversely graded beds by traction carpet sedimentation. The considerable vertical variability in the strength of iso-orientation is the result of very unstable flow regimes, up to the extreme condition of discrete depositional events, with a variable combination of traction carpet and/or direct suspension sedimentation. The vertical homogeneity in the mean orientation values, found in the investigated sections, may derive from the sequential deposition of laminae to thin beds, whose relatively flat upper surfaces were unable to significantly deflect the depositional system of the following currents. According to the observed homogeneous mean particle orientation values along the investigated vertical profiles, samples collected through areal distribution are considered representative of the local paleo-flow directions of the whole deposit. The mean directions of the samples collected areally show two different coherent patterns which point to the existence of two different source areas. The first, which includes all samples from the northern outcrops, appears to converge in a narrow area about 2 km NE of the town of Pozzuoli, largely in coincidence with the inferred area on the basis of the pumice fall distribution. The second, which includes samples from Capo Miseno and Posillipo areas, points to the central part of the Pozzuoli Bay, about 4 km offshore the town of Pozzuoli.     

Particle clusters in gravel‐bed rivers: an experimental morphological approach to bed material transport and stability concepts

Structured gravel river beds clearly exert a major influence on bed stability. Indexing structural stability by field measurements of bed strength neglects the processes operating to entrain and transport bed material in different parts of each structure. This study takes a morphological approach to interpreting the critical processes, using particle tracing to determine the movement of individual cluster particles over a range of flood event magnitudes and durations. The experiment was carried out on the River South Tyne, UK; it uses flow hydrographs measured nearby and also benefits from previous studies of historical development, channel morphology and sediment transport at the same site. More than 30 clusters were monitored over a seven‐month period during which clusters occupied 7–16 per cent of the bed. Threshold flows delimiting three apparently contrasting bed sediment process regimes for cluster particles are tentatively set at 100 m³ s^?1 and 183 m³ s^?1; durations of flow at these levels are critical for cluster development, rather than flow peak values. Wake particles are transported most easily. Flow straightening in the wandering channel planform reduces the stability of clusters, since mechanical strength is markedly reduced by this change of direction. The overall area covered by clusters between significant transport events varies little, implying a dynamic equilibrium condition. Copyright © 2005 John Wiley & Sons, Ltd.