Formation of high-grade ignimbrites Part II. A pyroclastic suspension current model with implications also for low-grade ignimbrites

 Analogue experiments in part I led to the conclusion that pyroclastic flows depositing very high-grade ignimbrite move as dilute suspension currents. In the thermo–fluid–dynamical model developed, the degree of cooling of expanded turbulent pyroclastic flows dynamically evolves in response to entrainment of air and mass loss to sedimentation. Initial conditions of the currents are derived from column-collapse modeling for magmas with an initial H₂O content of 1–3 wt.% erupting through circular vents and caldera ring-fissures. The flows spread either longitudinally or radially from source up to a runout distance that increases with higher mass flux but decreases with higher gas content, temperature, bottom slope and coarser initial grain size. Progressive dilution by entrainment and sedimentation causes pyroclastic currents to transform into buoyant ash plumes at the runout distance. The ash plumes reach stratospheric heights and distribute 30–80% of the erupted material as widespread co-ignimbrite ash. Pyroclastic suspension currents with initial mass fluxes of 10⁷-10¹² kg/s can spread for tens of kilometers with only limited cooling, although they move as supercritical, strongly entraining currents for the eruption conditions considered here. With increasing eruption mass flux, cooling during passage through the fountain diminishes while cooling during flow transport increases. The net effect is that eruption temperature exerts the prime control on emplacement temperature. Pyroclastic suspension currents can form welded ignimbrite across their entire extent if eruption temperature is T_o>1.3^.T_mw, the minimum welding temperature. High eruption rates, a large fraction of fine ash, and a ring-fissure vent favor the formation of extensive high-grade ignimbrite. For very hot eruptions producing sticky, partially molten pyroclasts, analysis of particle aggregation systematics shows that factors favoring longer runout also favor more efficient aggregation, which reduces runout. As a result, very high-grade ignimbrites cannot spread more than a few tens of kilometers from their source. In cooler pyroclastic currents, particles do not aggregate, and the sedimentation process may involve re-entrainment of particles, which potentially leads to more extensive cooling and longer runout; such effects, however, are only significant when net erosion of substrate occurs. Model results can be employed to estimate mass flux and duration of ignimbrite eruptions from measured ignimbrite masses and aspect ratios. The model also provides an alternative explanation of the observed decrease in H/Lratios with ignimbrite mass. Received: 10 May 1998 / Accepted: 21 October 1998     

The Afragola settlement near Vesuvius,Italy: The destruction and abandonment of a Bronze Age village revealed by archaeology,volcanology and rock-magnetism

Public works in progress in the Campanian plain north of Somma-Vesuvius recently encountered the remains of a prehistoric settlement close to the town of Afragola. Rescue excavations brought to light a Bronze Age village partially destroyed and buried by pyroclastic density currents (PDCs) of the Vesuvian Pomici di Avellino eruption (3.8 ¹⁴C ka BP) and subsequently sealed by alluvial deposits. Volcanological and rock-magnetic investigations supplemented the excavations.Careful comparison between volcanological and archaeological stratigraphies led to an understanding of the timing of the damage the buildings suffered when they were struck by a series of PDCs. The first engulfed the village, located some 14 km to the north of the inferred vent, and penetrated into the dwellings without causing major damage. The buildings were able to withstand the weak dynamic pressure of the currents and deviate their path, as shown by the magnetic fabric analyses. Some later collapsed under the load of the deposits piled up by successive currents. Stepwise demagnetization of the thermal remanent magnetization (TRM) carried by potsherds embedded in the deposits yields deposition temperatures in the order of 260–320 °C, fully consistent with those derived from pottery and lithic fragments from other distal and proximal sites. The fairly uniform temperature of the deposits is here ascribed to the lack of pervasive air entrainment into the currents. This, in turn, resulted from the lack of major topographical obstacles along the flat plain.The coupling of structural damage and sedimentological analyses indicates that the currents were not destructive in the Afragola area, but TRM data indicate they were still hot enough to cause death or severe injury to humans and animals. The successful escape of the entire population is apparent from the lack of human remains and from thousands of human footprints on the surface of the deposits left by the first PDCs. People were thus able to walk barefoot across the already emplaced deposits and escape the subsequent PDCs. The rapid cooling of the deposits was probably due to both their thinness and heat dissipation due to condensation of water vapour released in the mixture by magma–water interaction.     

Gravity currents down canyons: effects of rotation

The flow of dense water in a V-shaped laboratory-scale canyon is investigated by using a non-hydrostatic numerical ocean model with focus on the effects of rotation. By using a high-resolution model, a more detailed analysis of plumes investigated in the laboratory (Deep-Sea Res I 55:1021–1034 2008) for laminar flow is facilitated. The inflow rates are also increased to investigate plume structure for higher Reynolds numbers. With rotation, the plumes will lean to the side of the canyon, and there will be cross-canyon geostrophic currents and Ekman transports. In the present study, it is found that the cross-canyon velocities are approximately 5 % of the down-canyon velocities over the main body of the plume for the rotational case. With rotation, the flow of dense water through the body of the plume and into the plume head is reduced. The plume head becomes less developed, and the speed of advance of the head is reduced. Fluid parcels near the top of the plume will to a larger extent be left behind the faster flowing dense core of the plume in a rotating system. Near the top of the plume, the cross-canyon velocities change direction. Inside the plume, the cross-flow is up the side of the canyon, and above the interface to the ambient there is a compensating cross-flow down the side of the canyon. This means that parcels of fluid around the interface become separated. Parcels of fluid around the interface with small down-canyon velocity components and relative large cross-canyon components will follow a long helix-like path down the canyon. It is found that the entrainment coefficients often are larger in the rotational experiments than in corresponding experiments without rotation. The effects of rotation and higher inflow rates on the areal patterns of entrainment velocities are demonstrated. In particular, there are bands of higher entrainment velocities along the lateral edges of the plumes in the rotational cases.     

Numerical study of pyroclastic surges

The dynamics of pyroclastic surges accompanied by co-ignimbrite plumes is investigated numerically. The numerical simulations are performed with a newly developed numerical model, which is based on the Navier–Stokes equations for time-dependent flows of a compressible fluid in two-dimensional Cartesian coordinates. We regard pyroclastic surges as dilute turbulent suspensions in which hot gases and fine solid particles are homogeneously mixed owing to vigorous turbulence. In other words, the gas–particle mixture is treated as a single-phase fluid whose bulk density is represented by averaging the density of each component in the numerical model. We focus on the effect of buoyancy forces generated by the thermal expansion of the air mixed into pyroclastic surges from the calm surroundings. For our purpose, the numerical model is designed to simulate relatively simple flows spreading over a horizontal flat surface. Topographic irregularity and the sedimentation process of solid particles are neglected in the present simulations. The motion of pyroclastic surges is generated by the instantaneous release of a gas–particle mixture whose density is initially larger than the ambient air density and changes nonlinearly with the temperature and concentration of suspended solid particles. Turbulent mixing is evaluated by adopting the Smagorinsky model. By employing cubic interpolated pseudo-particle (CIP) method and C-CUP method, we obtain the fine structure of flows. The behavior of calculated flows agrees fairly well with observed pyroclastic surges in nature. The current head, which remains hot and dense, keeps spreading over a horizontal surface at a speed of about 20 m s⁻¹. The spreading speed is of the order of the speed of a gravity current that excludes the influence of thermal expansion. Besides, turbulent mixing between the basal dense layer and the ambient air is enhanced by the successive development of an interfacial less-dense layer. This results in the formation of a number of buoyant plumes rising above a horizontally spreading current. Consequently, the tails of the current thickens as time progresses. A parametric study shows that the initial temperature of a gas–particle mixture should be higher than about 600 K when buoyant plumes occur owing to the thermal expansion of mixed air. The result is quantitatively interpreted by introducing a diagram that describes the relationship among the bulk density, temperature and concentration of solid particles suspended in pyroclastic surges.     

Dilute gravity current and rain-flushed ash deposits in the 1.8 ka Hatepe Plinian deposit,Taupo, New Zealand

Two groups of poorly sorted ash-rich beds, previously interpreted as rain-flushed ashes, occur in the ca. AD 180 Hatepe Plinian pumice fall deposit at Taupo volcano, New Zealand. Two ash beds with similar dispersal patterns and an aggregate thickness of up to 13 cm make up the lowermost group (A). Group A beds extend 45 km north-east of the vent and cover 290 km². In the southern part of the group A distribution area, a coarse ash to lapilli-size Plinian pumice bed (deposit B) separates the two group A beds. The scarcity of lapilli (material seen elsewhere from the still-depositing pumice fall) in group A beds indicates that they were rapidly transported and deposited. However, this rapid transportation and deposition did not produce cross-bedding, nor did it erode the underlying deposits. It is proposed that thick (>600 m) but dilute gravity currents generated from the collapsing outer margin of the otherwise buoyant Hatepe Plinian eruption column deposited the group A beds. The upper ash beds (group C) consist of one to seven layers, attain an aggregate thickness of 35 cm, and vary considerably in thickness and number of beds with respect to distance from vent. Group C beds contain variable amounts of ash mixed with angular Plinian pumices and are genuine rain-flushed ashes. Several recent eruptions at other volcanoes (Ukinrek Maars, Vulcan, Rabaul, La Soufrère de Guadeloupe and Soufrière, St Vincent) have produced gravity currents similar in style, but much smaller than those envisaged for group A deposits. The overloaded margins of otherwise buoyant eruption plumes generated these gravity currents. Laboratory studies have produced experimental gravity current analogues. Hazards from dilute gravity currents are considerable but often overlooked, thus the recognition of gravity current deposits will contribute to more thorough volcanic hazard assessment of prehistoric eruption sequences.     

Sensitivity of the dynamics of volcanic eruption columns to their shape

This paper presents a one-dimensional steady-state model to investigate the sensitivity of the dynamics of sustained eruption columns to radius variations with height due to thermal expansion of the entrained air, and decreases in atmospheric pressure with height. In contrast to a number of previous models using an equation known as the entrainment assumption, the new model is based on similarity arguments to derive an equation set equivalent to the model proposed by Woods [Bull Volcanol 50:169–193, 1988]. This approach allows investigation of the effect of gas compressibility on the entrainment rate of ambient air, which has been little examined for a system in which a decrease in pressure significantly affects the density stratification of a compressible fluid. The new model provides results that include two end members: one in which the volume change within the eruption columns affects only the radial expansion without changing the vertical motion, and the other is the converse. The Woods [Bull Volcanol 50:169–193, 1988] model can be regarded as being between those two end members. The range of uncertainty arises because the extremely high temperature of discharged materials from a volcanic vent, and the exceptional terminal height of the eruption columns, allow significant expansion of the gas component in the eruption columns, making them behave differently from common turbulent plumes. This study indicates that the maximum height of the eruption columns is affected considerably by this uncertainty, particularly when the eruption columns extend above a height of 10 km, at which the pressure is about one-fourth the pressure at the ground surface. Column collapse may also be suppressed in wider parameter ranges than previously estimated. However, the uncertainty can be reduced by measuring column radii through a vertical profile during actual volcanic eruptions. Accordingly, this paper suggests that appropriate observation of eruption column shapes is essential for improving our understanding of the dynamics of eruption columns.     

Deposition temperature of the AD 472 Pollena pyroclastic density current deposits, Somma-Vesuvius, Italy

The deposition temperature of the pyroclastic density current (PDC) deposits emplaced during the AD 472 Pollena eruption (Somma-Vesuvius, Italy) has been investigated using the thermal analysis of the magnetic remanence carried by lithic clasts embedded within the deposits. A total of 310 lithic clasts were collected from all the PDC units in the Pollena stratigraphic succession, at different distances from the inferred vent and at different stratigraphic levels. The temperature reached by each individual clast during residence in the PDC was estimated through stepwise thermal demagnetization, with the values from all clasts collected at each site being used to infer the deposition temperature (T _dep). Although the sedimentological features of these PDC deposits show some variation, the deposition temperature typically falls in the range 300 to 320°C, with a maximum range of 260 to 360°C. The fairly uniform temperature observed in both the dune bedded and massive deposits points to homogeneity in attainment of T _dep for the different deposits and suggests similarity in the depositional regime of the different PDCs and/or in heat transfer to lithic fragments. Similarity in depositional regime was also favoured by the limited control exerted by topography on the distribution of these PDCs, with the northern wall of the Somma caldera that did not act as a morphological barrier. As a result the currents were capable of moving away from the vent, without topographic disturbances and, thus, significant variations in the cooling regime. Because the Pollena eruption is considered similar to the maximum expected event at Somma-Vesuvius, the characteristics of its deposits best simulate the likely maximum hazard for the Vesuvius region. In this regard, Pollena produced hot, dilute PDCs which were able to travel up to 12 km from the vent maintaining high temperatures across this distance.     

Pyroclastic flow hazard assessment at Somma–Vesuvius based on the geological record

During the past 22 ka of activity at Somma–Vesuvius, catastrophic pyroclastic density currents (PDCs) have been generated repeatedly. Examples are those that destroyed the towns of Pompeii and Ercolano in AD 79, as well as Torre del Greco and several circum-Vesuvian villages in AD 1631. Using new field data and data available from the literature, we delineate the area impacted by PDCs at Somma–Vesuvius to improve the related hazard assessment. We mainly focus on the dispersal, thickness, and extent of the PDC deposits generated during seven plinian and sub-plinian eruptions, namely, the Pomici di Base, Greenish Pumice, Pomici di Mercato, Pomici di Avellino, Pompeii Pumice, AD 472 Pollena, and AD 1631 eruptions. We present maps of the total thickness of the PDC deposits for each eruption. Five out of seven eruptions dispersed PDCs radially, sometimes showing a preferred direction controlled by the position of the vent and the paleotopography. Only the PDCs from AD 1631 eruption were influenced by the presence of the Mt Somma caldera wall which stopped their advance in a northerly direction. Most PDC deposits are located downslope of the pronounced break-in slope that marks the base of the Somma–Vesuvius cone. PDCs from the Pomici di Avellino and Pompeii Pumice eruptions have the most dispersed deposits (extending more than 20 km from the inferred vent). These deposits are relatively thin, normally graded, and stratified. In contrast, thick, massive, lithic-rich deposits are only dispersed within 7 to 8 km of the vent. Isopach maps and the deposit features reveal that PDC dispersal was strongly controlled by the intensity of the eruption (in terms of magma discharge rate), the position of the vent area with respect to the Mt Somma caldera wall, and the pre-existing topography. Facies characteristics of the PDC deposits appear to correlate with dispersal; the stratified facies are consistently dispersed more widely than the massive facies.     

The impacts of pyroclastic surges on buildings at the eruption of the Soufrière Hills volcano, Montserrat

We investigated the impacts on buildings of three pyroclastic surges that struck three separate villages on 25 June, 21 September and 26 December, 1997, during the course of the andesitic dome building eruption of the Soufrière Hills Volcano, Montserrat, which began on 18 July, 1995. A detailed analysis of the building damage of the 26 December event was used to compare the findings on the flow and behaviour of dilute pyroclastic density currents (PDCs) with the classical reports of PDCs from historical eruptions of similar size. The main characteristics of the PDC, as inferred from the building damage, were the lateral loading and directionality of the current; the impacts corresponded to the dynamic pressure of the PDC, with a relatively slow rate of rise and without the peak overpressure or a shock front associated with explosive blast; and the entrainment of missiles and ground materials which greatly added to the destructiveness of the PDC. The high temperature of the ash, causing the rapid ignition of furniture and other combustibles, was a major cause of damage even where the dynamic pressure was low at the periphery of the current. The vulnerability of buildings lay in the openings, mainly windows, which allowed the current to enter the building envelope, and in the flammable contents, as well as the lack of resistance to the intense heat and dynamic pressure of some types of vernacular building construction, such as wooden chattel houses, rubble masonry walls and galvanised steel-sheet roofs. Marked variability in the level of damage due to dynamic pressure (in a range 1–5 kPa, or more) was evident throughout most of the impact area, except for the zone of total loss, and this was attributable to the effects of topography and sheltering, and projectiles, and probably localised variations in current velocity and density. A marked velocity gradient existed from the outer part to the central axis of the PDC, where buildings and vegetation were razed to the ground. The gradient correlated with the impacts due to lateral loading and heat transfer, as well as the size of the projectiles, whilst the temperature of the ash in the undiluted PDC was probably uniform across the impact area. The main hazard characteristics of the PDCs were very consistent with those described by other authors in the classic eruptions of Pelée (1902), Lamington (1951) and St Helens (1980), despite differences in the eruptive styles and scales. We devised for the first time a building damage scale for dynamic pressure which can be used in research and in future volcanic emergencies for modelling PDCs and making informed judgements on their potential impacts. Editorial responsibility: T. Druitt     

The gas content and buoyancy of strombolian ash plumes

Plinian plumes erupt with a bulk density greater than that of air, and depend upon air entrainment during their gas-thrust phase to become buoyant; if entrainment is insufficient, the column collapses into a potentially deadly pyroclastic flow. This study shows that strombolian ash plumes can be erupted in an initially buoyant state due to their extremely high initial gas content, and in such cases are thus impervious to column collapse. The high gas content is a consequence of decoupled gas rise in the conduit, in which particles are ultimately incidental. The relations between conduit gas flow, eruption style and plume density are explored here for strombolian scenarios and contrasted with conventional wisdom derived from plinian eruptions. Considering the inherent relation between gas content and initial plume density together with detailed measurements of plume velocities can help unravel ambiguities surrounding conduit processes, eruption styles and hazards at poorly understood volcanoes. Analysis of plume dynamics at Santiaguito volcano, Guatemala adds further support for a model involving decoupled gas rise in the conduit.     

The Pomici di Avellino eruption of Somma–Vesuvius (3.9 ka BP). Part II: sedimentology and physical volcanology of pyroclastic density current deposits

Pyroclastic density currents (PDCs) generated during the Plinian eruption of the Pomici di Avellino (PdA) of Somma–Vesuvius were investigated through field and laboratory studies, which allowed the detailed reconstruction of their eruptive and transportation dynamics and the calculation of key physical parameters of the currents. PDCs were generated during all the three phases that characterised the eruption, with eruptive dynamics driven by both magmatic and phreatomagmatic fragmentation. Flows generated during phases 1 and 2 (EU1 and EU3pf, magmatic fragmentation) have small dispersal areas and affected only part of the volcano slopes. Lithofacies analysis demonstrates that the flow-boundary zones were dominated by granular-flow regimes, which sometimes show transitions to traction regimes. PDCs generated during eruptive phase 3 (EU5, phreatomagmatic fragmentation) were the most voluminous and widespread in the whole of Somma–Vesuvius’ eruptive history, and affected a wide area around the volcano with deposit thicknesses of a few centimetres up to more than 25 km from source. Lithofacies analysis shows that the flow-boundary zones of EU5 PDCs were dominated by granular flows and traction regimes. Deposits of EU5 PDC show strong lithofacies variation northwards, from proximally thick, massive to stratified beds towards dominantly alternating beds of coarse and fine ash in distal reaches. The EU5 lithofacies also show strong lateral variability in proximal areas, passing from the western and northern to the eastern and southern volcano slopes, where the deposits are stacked beds of massive, accretionary lapilli-bearing fine ash. The sedimentological model developed for the PDCs of the PdA eruption explains these strong lithofacies variations in the light of the volcano’s morphology at the time of the eruption. In particular, the EU5 PDCs survived to pass over the break in slope between the volcano sides and the surrounding volcaniclastic apron–alluvial plain, with development of new flows from the previously suspended load. Pulses were developed within individual currents, leading to stepwise deposition on both the volcano slopes and the surrounding volcaniclastic apron and alluvial plain. Physical parameters including velocity, density and concentration profile with height were calculated for a flow of the phreatomagmatic phase of the eruption by applying a sedimentological method, and the values of the dynamic pressure were derived. Some hazard considerations are summarised on the assumption that, although not very probable, similar PDCs could develop during future eruptions of Somma–Vesuvius.     

Nature of sediment dispersal off the Sepik River, Papua New Guinea: preliminary sediment budget and implications for margin processes

A sediment budget is constructed for the slope and narrow continental shelf off the Sepik River in order to estimate the relative importance of turbid plumes versus bottom gravity transport through a near-shore submarine canyon in the dispersal of sediment across this collision margin. ²¹⁰Pb geochronology and inventories of Kasten cores are consistent with the northwestward dispersal of sediment from the river mouth via hypopycnal and possible isopycnal plumes. Sediment accumulation rates are 5 cm yr⁻¹ on the upper slope just off of the Sepik mouth, decreasing gradually to 1 cm yr⁻¹ toward the northwest, and decreasing abruptly offshore (<0.2 cm yr⁻¹ at 1200 m water depth). A sediment budget indicates that only about 7–15% of the Sepik River sediment discharge accumulates on the adjacent open shelf and slope. The remainder presumably escapes offshore via gravity flows through a submarine canyon, the head of which extends into the river mouth. The divergent sediment pathways observed off the Sepik River (i.e., surface and subsurface plumes versus sediment gravity flows through a canyon) may be common along high-yield collision margins of the Indo–Pacific archipelago, and perhaps are analogous to most margins during Late Quaternary low sea-level conditions.     

Long-term multi-hazard assessment for El Misti volcano (Peru)

We propose a long-term probabilistic multi-hazard assessment for El Misti Volcano, a composite cone located <20 km from Arequipa. The second largest Peruvian city is a rapidly expanding economic centre and is classified by UNESCO as World Heritage. We apply the Bayesian Event Tree code for Volcanic Hazard (BET_VH) to produce probabilistic hazard maps for the predominant volcanic phenomena that may affect c.900,000 people living around the volcano. The methodology accounts for the natural variability displayed by volcanoes in their eruptive behaviour, such as different types/sizes of eruptions and possible vent locations. For this purpose, we treat probabilistically several model runs for some of the main hazardous phenomena (lahars, pyroclastic density currents (PDCs), tephra fall and ballistic ejecta) and data from past eruptions at El Misti (tephra fall, PDCs and lahars) and at other volcanoes (PDCs). The hazard maps, although neglecting possible interactions among phenomena or cascade effects, have been produced with a homogeneous method and refer to a common time window of 1 year. The probability maps reveal that only the north and east suburbs of Arequipa are exposed to all volcanic threats except for ballistic ejecta, which are limited to the uninhabited but touristic summit cone. The probability for pyroclastic density currents reaching recently expanding urban areas and the city along ravines is around 0.05 %/year, similar to the probability obtained for roof-critical tephra loading during the rainy season. Lahars represent by far the most probable threat (around 10 %/year) because at least four radial drainage channels can convey them approximately 20 km away from the volcano across the entire city area in heavy rain episodes, even without eruption. The Río Chili Valley represents the major concern to city safety owing to the probable cascading effect of combined threats: PDCs and rockslides, dammed lake break-outs and subsequent lahars or floods. Although this study does not intend to replace the current El Misti hazard map, the quantitative results of this probabilistic multi-hazard assessment can be incorporated into a multi-risk analysis, to support decision makers in any future improvement of the current hazard evaluation, such as further land-use planning and possible emergency management.     

How to anchor hotspots in a convecting mantle?

Laboratory experiments were performed to study the influence of density and viscosity layering on the formation and stability of plumes. Viscosity ratios ranged from 0.1 to 6400 for buoyancy ratios between 0.3 and 20, and Rayleigh numbers between 10⁵ and 2.10⁸. The presence of a chemically stratified boundary layer generates long-lived thermochemical plumes. These plumes first develop from the interface as classical thermal boundary layer instabilities. As they rise, they entrain by viscous coupling a thin film of the other layer and locally deform the interface into cusps. The interfacial topography and the entrainment act to further anchor the plumes, which persist until the chemical stratification disappears through entrainment, even for Rayleigh numbers around 10⁸. The pattern of thermochemical plumes remains the same during an experiment, drifting only slowly through the tank. Scaled to an Earth’s mantle without plate tectonics, our results show that: (1) thermochemical plumes are expected to exist in the mantle, (2) they could easily survive hundreds of millions of years, depending on the size and magnitude of the chemical heterogeneity on which they are anchored, and (3) their drift velocity would be at most 1-2 mm/yr. They would therefore produce long-lived and relatively fixed hotspots on the lithosphere. However, the thermochemical plumes would follow any large scale motion imposed on the chemical layer. Therefore, the chemical heterogeneity acts more as a ‘floating anchor’ than as an absolute one.     

On the dynamics of 3-D single thermal plumes at various Prandtl numbers and Rayleigh numbers

Three-dimensional (3-D) numerical simulations of single turbulent thermal plumes in the Boussinesq approximation are used to understand more deeply the interaction of a plume with itself and its environment. In order to do so, we varied the Rayleigh and Prandtl numbers from Ra?～?10⁵ to Ra?～?10⁸ and from Pr?～?0.025 to Pr?～?70. We found that thermal dissipation takes place mostly on the border of the plume. Moreover, the rate of energy dissipation per unit mass ε _T has a critical point around Pr?～?0.7. The reason is that at Pr greater than ～0.7, buoyancy dominates inertia and thermal advection dominates wave formation whereas this trend is reversed at Pr less than ～0.7. We also found that for large enough Prandtl number (Pr?～?70), the velocity field is mostly poloidal although this result was known for Rayleigh–Bénard convection (see Schmalzl et al. [On the validity of two-dimensional numerical approaches to time-dependent thermal convection. Europhys. Lett. 2004, 67, 390--396]). On the other hand, at small Prandtl numbers, the plume has a large helicity at large scale and a non-negligible toroidal part. Finally, as observed recently in details in weakly compressible turbulent thermal plume at Pr?=?0.7 (see Plourde et al. [Direct numerical simulations of a rapidly expanding thermal plume: structure and entrainment interaction. J. Fluid Mech. 2008, 604, 99--123]), we also noticed a two-time cycle in which there is entrainment of some of the external fluid to the plume, this process being most pronounced at the base of the plume. We explain this as a consequence of calculated Richardson number being unity at Pr?=?0.7 when buoyancy balance inertia.     

DISCONTINUOUS FLOW OF TURBID DENSITY CURRENTS I. CHANNEL EXPANSION AND CONTRACTION

Laboratory experiments on turbid density currents were conducted to observe the flow features of these currents with abrupt contracted and expanded reaches. Experimental data were used to determine water entrainment coefficients for both channel expansion and contraction. Expressions for turbid density currents with water entrainment coefficients in abrupt contracted and expanded reaches were derived, and compared with experimental data.     

Small scale turbulent diffusion determined with underwater flow visualization

Dye plumes were generated at three depths in the seasonal thermocline between 7 and 11 m, 22 km south of Key West on 21 August 1980 and photographed at about 10 second intervals with an underwater camera system. Eleven pairs of consecutive pictures are analyzed to determine the mean current vertical shear and the width of the plumes by positioning reference points relative to the rod attached to the camera system. The relative distances of reference points are calibrated with the stereophotogrammetric method for one pair. The eddy diffusivity is calculated by use of a model of turbulent diffusion developed byTaylor (1921). Its values range from 5 to 25 cm²s^–1 for the plume widths ranging from 33 to 132 cm. The Richardson number is calculated for each pair of pictures with the vertical density gradient estimated from temperature profiles. Its values are higher than the critical value of 0.25 except for one case. The diffusivity was higher by orders of magnitude than the molecular one and indicates the presence of turbulence together with billow like features of the plumes in spite of high Richardson numbers. This suggests that the billow turbulence might be caused by effects of surface gravity waves and not by the Kelvin-Helmholtz instability.     

Gravity currents over a rigid and emergent vegetated slope

Lock-exchange experiments are conducted to investigate the effects of emergent vegetation on gravity currents flowing down a slope. Rigid and emergent cylinders are used to represent vegetation such as reeds in aquatic environments. The results show that the head of gravity currents without cylinders forms a semi-elliptic shape, similar to the flat bed case. For the slope-induced gravity, however, the head of gravity currents gradually grows and accelerates. A steeper slope without cylinders causes the evident entrainment and subsequent energy dissipation of gravity currents with ambient fluid. When the cylinder population becomes denser, developments of the head become slower, and the semi-elliptic head can be only seen at the very front end. As the density reaches 6.9%, the interface between the saltwater and freshwater performs an inclined straight line, corresponding to the pervious theoretical derivation. The gravity current event could go through two different routes under the similar toe speeds. The current head without cylinders grows along the downslope path, and the enrollment and mixing with ambient fluid could occur at the edge of the head; while the currents within dense cylinders (∼6.9%) has a nearly constant thickness and a smaller and more streamlined head at the very front end. As the cylinder density increases, the front speed of gravity currents transforms from acceleration to deceleration phases. The experimental results reveal that the head would accelerate over the downslope course if the cylinder density is less than 2%.     

A laboratory simulation of pyroclastic flows down slopes

Laboratory experiments are described which explore the dynamical consequences of buoyant convective upflow observed above hot pyroclastic flows. In nature, the convection is produced by the hot ash particles exchanging heat with air mixed into the front and top of the pyroclastic flow. This effect on the buoyancy due to the mixing of air and ash has been modelled in the laboratory using mixtures of methanol and ethylene glycol (MEG), which have a nonlinear density behaviour when mixed with water. Intermediate mixtures of these fluids can be denser than either initial component, and so the laboratory experiments were inverted models of the natural situation. We studied MEG flowing up under a sloping roof in a tank filled with water. The experiments were performed both in a narrow channel and on a laterally unconfined slope. The flow patterns were also compared with those of conventional gravity currents formed using fresh and salt water. The presence of the region of reversed buoyancy outside the layer flowing along the slope had two significant effects. First, it periodically protected the flow from direct mixing with the environment, resulting in pulses of relatively undiluted fluid moving out intermittently ahead of the main flow. Second, it produced a lateral inflow towards the axis of the current which kept the current confined to a narrow tongue, even on a wide slope.In pyroclastic flows the basal avalanche portion has a much larger density contrast with its surroundings than the laboratory flows. Calculations show that mixing of air into the dense part of a pyroclastic flow cannot generate a mixture that is buoyant in the atmosphere. However, the overlying dilute ash cloud can behave as a gravity current comparable in density contrast to the laboratory flows and can become buoyant, depending on the temperature and ash content. In the August 7th pyroclastic flow of Mount St. Helens, Hoblitt (1986) describes pulsations in the flow front, which are reminiscent of those observed in the experiments. As proposed by Hoblitt, the pulsations are caused by the ash cloud accelerating away from the front of the dense avalanche as a density current. The ash cloud then mixes with more air, becomes buoyant and lifts off the ground, allowing the avalanche to catch up with and move ahead of the cloud. The pulsing behaviour at the fronts of pyroclastic flows could account for the occurrence of cross-bedded layer 1 deposits which occur beneath layer 2 deposits in many sequences.