Contrasting grain size and componentry in complex proximal deposits of the 1886 Tarawera basaltic Plinian eruption

The 1886 Plinian eruption of Tarawera, New Zealand, is a unique basaltic fissure-fed eruption with exceptionally well preserved fall deposits to within 200 meters of the source vents. These proximal deposits form a series of spatter/cinder half-cones along the northeastern 8-km-long segment of the 1886 fissure. Here we examine these deposits using grain size and clast componentry techniques. We contrast the products of the phreatomagmatic (phases I and III) and Plinian (Phase II) stages of the eruption and examine deposit variability as a function of contrasting eruptive intensity within the climactic phase (II) of the eruption. The opening phreatomagmatic phase I of the eruption involved gas-rich magma interacting with water and fragmenting at least 300 meters below the surface. The deposits of the climactic phase that followed have relatively uniform grain size but marked contrasts in the relative abundance of juvenile and wall rock (lithic) clasts. Deposits linked to vents associated with the high Plinian plume are more uniform than those characterized by a weaker cone-forming eruption style. During the third, and closing, phase of the eruption, magma withdrawal accompanied the onset of decoupling of the exsolved gas phase, leading to fragmentation at increasingly greater depths and significant wall rock collapse into the erupting vents. Variability in eruptive style during phase II along the fissure appears to be a function of shallow seated controls, in particular the variable extent of incorporation of lithic wall rock into the erupting jet, as a consequence of vent wall collapse. Widely dispersed beds centralized around Plinian sources along the fissure have very low lithic content; cone-forming beds at other craters that contain very high lithic contents. This incorporation led to a significant reduction of the velocity and stability of the jet at these latter steep-walled craters, and induced episodicity in the form of vent-clearing explosions. The result is a large reduction of the physical and thermal ability of these vents to contribute to a stable high eruptive plume. Instead large volumes of ejecta were sedimented prematurely from shallow heights at rates an order of magnitude greater than for historical Strombolian, Hawaiian and subPlinian eruptions. This study illustrates that sustained powerful Plinian eruptions can be accompanied by heterogeneities and instabilities of the eruptive jet. At Tarawera, the record of complex proximal transport and deposition processes in the eruptive jet cannot be inferred from the eruption products at distances greater than 400 m from the eruptive fissure. We suggest that study of ultraproximal deposits, as seen at Mt Tarawera, provides the only opportunity to document the complex, dynamic behavior of the jet region of Plinian eruptions.     

Use of terrain variables for mapping gully erosion susceptibility in Lebanon

This paper predicts the geographic distribution and size of gullies across central Lebanon using a geographic information system (GIS) and terrain analysis. Eleven primary (elevation; upslope contributing area; aspect; slope; plan, profile and tangential curvature; flow direction; flow width; flow path length; rate of change of specific catchment area along the direction of flow) and three secondary (steady‐state; quasi‐dynamic topographic wetness; sediment transport capacity) topographic variables were generated and used along with digital data collected from other sources (soil, geology) to statistically explain gully erosion field measurements. Three tree‐based regression models were developed using (1) all variables, (2) primary topographic variables only and (3) different pairs of variables. The best regression tree model combined the steady‐state topographic wetness and sediment transport capacity indices and explained 80% of the variability in field gully measurements. This model proved to be simple, quick, realistic and practical, and it can be applied to other areas of the Mediterranean region with similar environmental conditions, thereby providing a tool to help with the implementation of plans for soil conservation and sustainable management. Copyright © 2007 John Wiley & Sons, Ltd.     

Cyclic testing of unreinforced masonry walls in two‐way bending

An experimental programme was conducted in which eight full‐scale unreinforced masonry walls were subjected to cyclic face loading using a system of airbags. Of the eight walls, six contained a window opening and four were subjected to vertical pre‐compression. Combined supports at the vertical and horizontal edges ensured that under face loading the walls underwent two‐way bending. The test walls were found to possess good post‐peak strength and displacement capacity as well as reasonable energy dissipation characteristics. Significant strength and stiffness degradation and non‐symmetry of strength in the positive and negative displacement directions were also evident. Discussion of the causes of the aforementioned trends and their implications towards the seismic response of masonry walls is provided. Copyright © 2006 John Wiley & Sons, Ltd.     

The settling of consecutive spheres in viscoplastic fluids

One of the factors contributing to the uncertainties involved in the estimation of particle settling velocity in viscoplastic fluids is the time-dependent effect where the viscous parameters of the fluid change as a particle flows through and shears the medium. These changes, particularly at low shear Reynolds numbers, are reflected in the settling velocity of a following sphere that is released some time after an initial one, with the following sphere having a significantly greater velocity. This study found that changes in both fall velocity and equivalent viscosity can be correlated satisfactorily by a power law equation to the dimensionless form of the time interval between releases, and the rheogram shape factor for the fluid. A collision of particles occurs in cases where the time interval between releases is small, after which the particles combine and travel at a terminal velocity. A new variable, β, which takes into account the different surficial stress of the combined spheres, was introduced to the correlation of Wilson et al. [Wilson, K.C., Horsley, R.R., Kealy, T., Reizes, J.A., Horsley, M.R., 2003. Direct prediction of fall velocities in non-Newtonian materials. Int. J. Miner. Process. 71, 17–30] β was found to depend on the rheogram shape factor for the fluid and the shear Reynolds number for the particle. The validity of this approach was supported by experimental data.     

The use of tracers to assess leakage from the sequestration of CO2 in a depleted oil reservoir,New Mexico,USA

Geological sequestration of CO₂ in depleted oil reservoirs is a potentially useful strategy for greenhouse gas management and can be combined with enhanced oil recovery. Development of methods to estimate CO₂ leakage rates is essential to assure that storage objectives are being met at sequestration facilities. Perfluorocarbon tracers (PFTs) were added as three 12 h slugs at about one week intervals during the injection of 2090 tons of CO₂ into the West Pearl Queen (WPQ) depleted oil formation, sequestration pilot study site located in SE New Mexico. The CO₂ was injected into the Permian Queen Formation. Leakage was monitored in soil–gas using a matrix of 40 capillary adsorbent tubes (CATs) left in the soil for periods ranging from days to months. The tracers, perfluoro-1,2-dimethylcyclohexane (PDCH), perfluorotrimethylcyclohexane (PTCH) and perfluorodimethylcyclobutane (PDCB), were analyzed using thermal desorption, and gas chromatography with electron capture detection. Monitoring was designed to look for immediate leakage, such as at the injection well bore and at nearby wells, and to develop the technology to estimate overall CO₂ leak rates based on the use of PFTs. Tracers were detected in soil–gas at the monitoring sites 50 m from the injection well within days of injection. Tracers continued to escape over the following years. Leakage appears to have emanated from the vicinity of the injection well in a radial pattern to about 100 m and in directional patterns to 300 m. Leakage rates were estimated for the 3 tracers from each of the 4 sets of CATs in place following the start of CO₂ injection. Leakage was fairly uniform during this period. As a first approximation, the CO₂ leak rate was estimated at about 0.0085% of the total CO₂ sequestered per annum.     

Turbulent velocity distributions and implied trajectory models

Well-mixed, first-order Lagrangian stochastic (LS) particle trajectory models are derived from several idealized (“toy”) turbulent velocity distributions, and their performance is compared against the observations of Project Prairie Grass, i.e., the case of a continuous point source of tracer near the ground, in the horizontally homogeneous and neutrally stratified surface layer. Although in a context of limited information a Gaussian distribution is the preferred choice, and although the Gaussian corresponds to the simplest of this set of LS models (namely, the Langevin equation), models stemming from other velocity distributions give similar, albeit distinguishable, predictions.     

Predicting plume meandering and averaging time effects on mean and fluctuating concentrations in atmospheric dispersion simulated in a water channel

Plume meandering and averaging time effects were measured directly using a high spatial resolution, high frequency, linescan laser-induced fluorescence (LIF) technique for measuring scalar concentrations in a plume dispersing in a water channel. Post-processing of the collected data removed time dependent background dye levels and corrected for attenuation across the laser beam to produce accurate measurements over long sample times in both a rough surface boundary-layer shear flow and shear free grid-generated turbulent flow. The data were used to verify the applicability of a meandering plume model for predicting the properties of mean and fluctuating concentrations. The centroid position of the crosswind concentration profile was found to have a Gaussian probability density function and the instantaneous plume spread about the centroid fluctuated log-normally. A modified travel-time power law model for averaging time adjustment was developed and compared to the widely used, but much less accurate, 0.2 power-law model.