Le volcanisme de point chaud est l'une des activités les plus intrigantes de la Terre. L'origine de ce type de volcanisme est encore débattue mais il est largement admis que les panaches sont formés de matériel chaud qui peut potentiellement entrainer avec lui des matériaux recyclés. La plupart des études se concentrent sur les panaches puissants montrant une forte production magmatique comme Hawaï ou la Réunion, mais les panaches montrant une plus faible production magmatique comme la Polynésie (Pacifique sud), peuvent apporter des informations complémentaires.Ce travail consiste en une étude pétrologique et géochimique de la chaine de Pitcairn-Gambier, située au sud-est du panache Polynésien. Ce travail s'organise autour de mesures en éléments majeurs et en trace, de compositions isotopiques en Pb, Sr, Nd et Hf sur une centaine d'échantillons, et autour d'analyses ponctuelles sur olivines et sulfures contenus dans certains échantillons.Les compositions isotopiques et élémentaires montrent une évolution systématique avec le temps (de Mururoa, île la plus vieille, aux monts sous-marins de Pitcairn les plus jeunes), depuis des compositions isotopiques « HIMU jeunes » avec une forte anomalie positive en Nb (1.8) jusqu'à des compositions isotopique EMI typiques avec une faible anomalie négative en Nb (0.8). Les différents modèles développés dans cette thèse ont permis de dater, contraindre la lithologie et la proportion des composants recyclés dans la source de ces îles. Les îles de Mururoa, Fangataufa et Gambier possèdent dans leur source de la croûte basaltique recyclée de 1.5 Ga ainsi que des sédiments âgés de 1-1.5 Ga (<0.36%). L'étude ciblée de l'île de Gambier met en évidence une source plutôt froide (~1450°C) comparée aux autres panaches comme Hawaï (1550°C). La source de l'île de Pitcairn et de ses monts sous-marins se révèle extrêmement différente car elle contient non seulement de la croûte basaltique recyclée (1.5 Ga) mais également jusqu'à ~18% de cherts archéens (~2.5 Ga). La présence inattendue de cherts archéens dans la source des îles de Pitcairn qui ont les valeurs EM1 les plus prononcées des OIB, remet en cause les autres théories développées pour tenter de contraindre ce composant.Les anomalies positives en Nb indiquent la présence dans la source de croûte basaltique recyclée. La diminution de cette anomalie au cours du temps le long de la chaîne est expliquée par l'augmentation progressive du taux de sédiments dans la source du panache, les sédiments possédant une anomalie négative en Nb.Les hétérogénéités isotopiques et élémentaires mises en évidence dans cette étude sont expliquées par fusion, sous chaque île, de matériel de composition et d'âge différents. Un modèle possible pour expliquer la structure de l'ensemble du panache polynésien, consiste en de «petits panaches» générés à partir d'un super-panache. Ces petits panaches produisent alors de petits alignements, et échantillonnent de manière imprévisible les différents matériels présents dans le super-panache. L'ensemble des données et modèles proposés dans cette thèse offre un nouvel angle pour apprécier la variabilité géochimique du panache Polynésien et à plus grande échelle de nouvelles perspectives pour comprendre les processus de recyclage au sein du manteau terrestre. / Plume volcanism is one of the most puzzling features of present-day activities of the Earth. The origin of this type of volcanism remains a matter of debate, but it is generally agreed that its source is hot material containing some potentially recycled material. Most studies have focused on ‘strong' plumes with high magmatic production, e.g. Hawaii or Réunion, but weaker plumes such as Polynesia (South Pacific) may provide useful complementary information.This work presents a petrological and geochemical study of the Pitcairn-Gambier chain, situated at the South East of the Polynesian plume. I report major and trace element concentrations and Pb, Sr, Nd and Hf isotopic measurements of ~100 samples as well as probe analyses on olivine and sulphides in a few selected samples.Both isotopes and trace elements show a strong correlation with the age of the volcanoes (from the oldest island of Mururoa to the younger Pitcairn Seamounts), from « Young HIMU » isotopic compositions associated with high positive Nb anomalies (1.8), to typical EMI compositions with a slight negative anomaly (0.8). Geochemical modeling enables to constrain the age, the nature and the proportion of the recycled components in the source. Mururoa, Fangataufa and Gambier show in their source, the presence of 1.5 Gy recycled oceanic crust, along with 1-1.5 Gy sediments (<0.36%). The source of Gambier Island basalts is ‘cold' (~1450°C) compared to that of other plumes like Hawaii (1550°C). In contrast, the Pitcairn Island and Seamounts source is different as it contains 1.5 Gy recycled basaltic oceanic crust, as well as up to ~18% Archean (~2.5 Ga) cherts. The unexpected discovery of Archean cherts in the source of Pitcairn islands (which have the most extreme EMI compositions) is at odd with previous models on the origin of the EMI component.Positive Nb anomalies indicate the presence of recycled oceanic crust in the source, and this anomaly decreases with time along the chain. This feature can be explained by the increasing contribution of the sediment input in the source of the plume, since sediments display a negative Nb anomaly.The isotopic heterogeneities evidenced in this study are best explained by melting, beneath the islands, of different materials with various ages and compositions. The general structure of the Polynesian plume can be explained by the presence of small plumes generated from a superplume. These small plumes create small alignments, and they randomly sample various materials present in the plume. Overall the data and models in this study give new insights into the geochemical variations observed across the Polynesian plume, and at a larger scale they provide new perspectives to better understand crustal recycling processes and the evolution of the mantle through time.

abstract: Assessments for the threats posed by volcanic eruptions rely in large part on the accurate prediction of volcanic plume motion over time. That predictive capacity is currently hindered by a limited understanding of volcanic plume dynamics. While eruption rate is considered a dominant control on volcanic plume dynamics, the effects of variable eruption rates on plume rise and evolution are not well understood. To address this aspect of plume dynamics, I conducted an experimental investigation wherein I quantified the relationship between laboratory jet development and highly-variable discharge rates under conditions analogous to those which may prevail in unsteady, short-lived explosive eruptions. I created turbulent jets in the laboratory by releasing pressurized water into a tank of still water. I then measured the resultant jet growth over time using simple video images and particle image velocimetry (PIV). I investigated jet behavior over a range of jet Reynolds numbers which overlaps with estimates of Reynolds numbers for short-duration volcanic plumes. By analysis of the jet boundary and velocity field evolution, I discovered a direct relationship between changes in vent conditions and jet evolution. Jet behavior evolved through a sequence of three stages - jet-like, transitional, and puff-like - that correlate with three main injection phases - acceleration, deceleration and off. While the source was off, jets were characterized by relatively constant internal velocity distributions and flow propagation followed that of a classical puff. However, while the source was on, the flow properties - both in the flows themselves and in the induced ambient flow - changed abruptly with changes at the source. On the basis of my findings for unsteady laboratory jets, I conclude that variable eruption rates with characteristic time scales close to eruption duration have first-order control over volcanic plume evolution. Prior to my study, the significance of this variation was largely uncharacterized as the volcanology community predominately uses steady eruption models for interpretation and prediction of activity. My results suggest that unsteady models are necessary to accurately interpret behavior and assess threats from unsteady, short-lived eruptions. / Dissertation/Thesis / Ph.D. Geological Sciences 2012

A biodiversidade e a estrutura da comunidade fitoplanctônica na Zona Costeira Amazônica Brasileira foi avaliada com base em 272 publicações, relativas ao período entre 1861 e 2016, e a partir de oito cruzeiros oceanográficos realizados em diferentes fases do ciclo do Rio Amazonas entre os anos de 2013 a 2015. A seleção das publicações restringiu-se as que apresentaram identificação taxonômica, em diferentes níveis, e/ou quantificação, expressa em densidade celular, obtidas exclusivamente através de microscopia ótica. A composição específica e a densidade celular do fitoplâncton foram avaliadas através da metodologia Utermöhl (1958). As informações disponíveis, principalmente sob a forma de resumos em congressos (67%), monografias (13%) e artigos publicados em periódicos (11%), foram resgatadas de estudos desenvolvidos no litoral dos Estados do Pará (131), Maranhão (105) e Amapá (3), assim como na plataforma continental (36). Entre os ecossistemas mais bem estudados, estão os estuários (135 referências), plataforma continental (36 referências) e as praias (30), sendo o enfoque ecológico o principal objetivo das investigações científicas (96%). Atualmente, o inventário de espécies inclui 1157 táxons distribuídos em 612 diatomáceas, 252 dinoflagelados, 199 clorofíceas, 56 cianobactérias, 19 cocolitoforídeos, 19 euglenofíceas e uma criptofícea. Com base nesta revisão e nas amostragens realizadas neste estudo, a análise da estrutura da comunidade fitoplanctônica na Plataforma Continental Amazônica é característica de ambientes costeiros sob influência de plumas fluviais, com dois cenários ecológicos relacionados aos períodos de vazão do Rio Amazonas: um de maior influência, nos meses de abril e julho, e outro oposto, nos meses de outubro e janeiro. Observou-se uma distribuição em faixas, de acordo com as características ambientais e a estrutura da comunidade fitoplanctônica, divididas em plataforma interna sob influência da pluma, zona de transição e plataforma externa com influência oceânica. As Diatomáceas são responsáveis pelas altas densidades (105 cel.L-1) na região costeira, especialmente sob influência da pluma, contribuindo para o incremento de Clorofila a (>10 &mu;g.L-1). Na porção oceânica, sem influência da pluma, as concentrações de Cla são baixas (<2 &mu;g.L-1) e predominam (104 cel.L-1) as cianobactérias filamentosas e cocolitoforídeos. A variabilidade temporal da descarga do Rio Amazonas e, consequentemente, da dinâmica espacial da pluma, é o principal fator responsável por mudanças na salinidade e disponibilidade de nutrientes, determinando a estrutura da comunidade fitoplanctônica. Este estudo reforça a importância da manutenção de séries temporais e o levantamento da biodiversidade para a compreensão da dinâmica da comunidade fitoplanctônica em ambientes de alta complexidade, como a Plataforma Continental Amazônica. / Phytoplankton biodiversity and structure in the Brazilian Amazon Coastal Zone was evaluated based on 272 publications from the period between 1861 and 2016 and from eight oceanographic cruises carried out in different phases of the Amazon river hydrograph between the years 2013 to 2015. Selected publications were restricted to those which presented taxonomic identification at different levels, and / or quantification in terms of cell density, obtained exclusively by optical microscopy. Specific composition and phytoplankton cell density was evaluated by Utermöhl methodology (1958). Available information in the form of conference abstracts (67%), monographs (13%) and journal articles (11%) were rescued from studies developed on the coast of Pará (131), Maranhão (105) and Amapá (3), as well as on the continental shelf (36). Within these, the best studied ecosystems are estuaries (135 references), the continental shelf (36 references) and the beaches (30), with an ecological approach as the main research objective (96%). Currently, species inventories include 1157 taxa, distributed in 612 diatoms, 252 dinoflagellates, 199 chlorophytes, 56 cyanobacteria, 19 coccolithophorids, 19 euglenoids and one criptoficean. Based on this review and on the results obtained in this study, the structure of the phytoplankton community in the Amazon Continental Shelf is characteristic of coastal environments under the influence of river plumes, with two ecological scenarios related to the Amazon river discharge, one with greater influence in April and July and another with less influence in October and January. The distribution occurs in bands, according to the environmental characteristics and the structure of the phytoplankton community, divided into internal continental shelf under the influence of the plume, transition zone, and external continental shelf with oceanic influence. Diatoms are responsible for the high densities (105 cells.L-1) in the coastal region, especially under the influence of the plume, contributing with the increase of chlorophyll a (> 10 &mu;g.L-1). At the oceanic portion without influence of the plume, Cla concentrations are low (<2 &mu;g.L-1) and filamentous cyanobacteria (104 cells.L-1) and coccolithophorids dominate. The temporal variability of the discharge of the Amazon river and, consequently, the spatial dynamics of the plume, is the main factor responsible for changes in salinity and nutrient availability, determining the structure of the phytoplankton community. This study reinforces the importance of maintaining time series and biodiversity surveys to understand phytoplankton community dynamics in highly complex environments such as the Continental Amazon Shelf.

Magister Scientiae - MSc / Nuclear reactors release small amounts of radioactivity during their normal operations. The most common method of calculating the dose to the public that results from such releases uses Gaussian Plume models. We are investigating these methods using CAP88-PC, a computer code developed for the Environmental Protection Agency (EPA) in the USA that calculates the concentration of radionuclides released from a stack using Pasquill stability classification. A buoyant or momentum driven part is also included. The uptake of the released radionuclide by plants, animals and humans, directly and indirectly, is then calculated to obtain the doses to the public. This method is well established but is known to suffer from many approximations and does not give answers that are accurate to be better than 50% in many cases. More accurate, though much more computer-intensive methods have been developed to calculate the movement of gases using fluid dynamic models. Such a model, using the code FLUENT can model complex terrains and will also be investigated in this work. This work is a preliminary study to compare the results of the traditional Gaussian plume model and a fluid dynamic model for a simplified case. The results indicate that Computational Fluid Dynamics calculations give qualitatively similar results with the possibility of including much more effects than the simple Gaussian plume model. / South Africa

Les travaux de recherche présentés dans ce manuscrit portent sur le mécanisme de remplissage et de vidange simultanés d’un local ventilé naturellement, pouvant éventuellement être rencontré dans des situations d’incendie. L’objectif de ce travail et d'améliorer la compréhension des phénomènes physiques dominants ce mécanisme, notamment, une fois que le régime stationnaire est établi à partir d’études théoriques et expérimentales. Le travail est divisé en deux parties. Dans la première partie, nous reformulons dans un premier temps le modèle théorique de Linden et al. (1990) dans le cadre de l’approximation de Boussinesq basé sur la modélisation du panache turbulent proposé par Morton et al. (1956). Ce modèle est par la suite étendue au cadre général non-Boussineq. Dans un second temps, nous proposons un nouveau modèle basé sur les solutions exactes du panache turbulent à partir des travaux de Michaux & Vauquelin (2008). Une campagne d’essais densimétriques réalisée à échelle réduite permet ensuite d’éprouver et valider ces modèles. Dans la deuxième partie de ce mémoire, nous abordons la problématique de remplissage et de vidange dans le cadre plus spécifique de l'ingénieur de la sécurité incendie (ISI) appliquée au désenfumage. Nous commençons par présenter quelques modèles de flammes panache. Ensuite, ces modèles de flammes panache sont implémenté dans un modèle de remplissage vidange et comparés aux modèles présentés dans la première partie du manuscrit. La dernière partie porte sur une campagne d'essais thermiques à l'échelle du laboratoire. Le premier modèle théorique présenté dans la première partie est confronté aux différents résultats expérimentaux. / Research works presented in this thesis deals with the filling and simultaneous emptying of a naturally ventilated room subject to a continuous source of buoyancy, for example, the problem of natural ventilation of a room containing a fire. The aim of this work is to improve from a theoretical and experimental point of view the understanding of the dominant physical phenomena of this mechanism, especially once the stationary state is reached. This work is divided into two parts. In the first "academic" part, we revisit the theoretical model developed by Linden et al. (1990) under the Boussinesq approximation, based on Morton et al. (1956) turbulent plume assumption. This model is then extended in the general non-Boussinesq case and a parametric study highlights the influence of the governing parameters on which it depends. Secondly, we propose a new model based on the exact turbulent plume solutions proposed by Michaux & Vauquelin (2008). Laboratory experiments were also conducted using a light gas air-helium mixture in order to test and validate these models. In the second part of this work, we address the problem of filling and simultaneous emptying in the more specific context of fire safety engineering applied to smoke management. We begin by presenting some engineering relations for fire plumes, which we implement in a filling emptying model. A comparison is then made between these models and those presented in the first part of the manuscript. The last part deals with a fire test campaign at the laboratory scale and on full-scale. The first theoretical model presented in the first part is confronted with different experimental results.

No description available.

Doctor en Ciencias, Mención Geología / La Dorsal de Juan Fernández (JFR) es una cadena volcánica (~ 800 km) de intraplaca emplazada sobre la Placa de Nazca en el Pacífico SE alejada de márgenes activos. Mediante datos de geoquímica (roca total y mineral), isotópicos (Sr-Nd-Pb) y geocronológicos (40Ar/39Ar) se busca comprender los procesos petrogenéticos implicados en la generación del volcanismo y en la evolución magmática de JFR. Se determina que los 4 edificios volcánicos más volumétricos de JFR: O Higgins (~ 9.26 8.41 Ma), Alpha (~ 4.63 4.58 Ma), Robinson Crusoe (~ 4.10 3.40 Ma) y Alejandro Selkirk (~ 0.94 0.83 Ma); satisfacen una progresión de edades más joven hacia el W coherente con la teoría de plumas mantélicas. La fase de construcción del escudo representa casi la totalidad del volumen de los edificios estudiados, se compone principalmente de basaltos con signatura geoquímica (e.g., alto contenido de TiO2, alto FC3MS y anomalía TITAN) e isotópica (FOZO-A con participación adicional de DM) que sugiere la presencia de piroxenita (formada a partir de corteza oceánica reciclada) como heterogeneidad en una fuente mantélica peridotítica. Dicha presencia es confirmada mediante un modelo petrogénetico para la pluma que indica una baja temperatura potencial (rango de 1290 1322 °C para Robinson Crusoe vs. 1312 1362 °C en Alejandro Selkirk), presión de término de fusión (2.34 2.54 vs. 2.24 2.52 GPa) probablemente relacionado al límite litósfera-astenosfera, y una similar participación en el melt final de fundidos provenientes de piroxenita (38.6 56.4 vs. 35.8 55.6 wt%) pese a su baja presencia en la pluma (4 8 vs. 6 12 wt%). Las variaciones composicionales internas se explican por fraccionamiento de olivino + clinopiroxeno ± plagioclasa, mezcla y/o recarga magmática y acumulación de cristales de olivino en una cámara magmática somera (~ 1 a 3 kbar) donde la temperatura de los magmas puede descender hasta 1156 1181 °C, y las variaciones entre volcanes se explica por variaciones temporales en la temperatura potencial y tasa de fusión parcial de la pluma mantélica. O Higgins y Robinson Crusoe también muestran una fase de volcanismo rejuvenecido formada por coladas de lava basanítica eruptadas tras un periodo de inactividad máximo de ~ 0.25 Ma en O Higgins, y ~ 1.73 Ma en Robinson Crusoe. Su mayor enriquecimiento geoquímico y signatura isotópica relativamente similar al escudo confirman que también se origina a partir de una pluma mantélica, pero posiblemente con sutiles variaciones en la proporción de sus constituyentes (peridotita y piroxenita), temperatura y grado de fusión parcial (ambas menores a la etapa de escudo). Estos magmas ascienden de manera directa (> 1300 °C), capturando xenocristales, con cristalización polibárica y poca diferenciación, ya que solo algunas son almacenadas por breves periodos en pequeños reservorios someros (a ~ 1256 1295 °C).

Since 2000, a relatively new electric thruster concept has been in research, development, and production at Thales Electron Devices in Germany. This High Efficiency Multistage Plasma Thruster, or HEMPT, has promising lifetime capabilities due to its plasma confinement system. However, the permanent magnet system that offers this and other benefits also creates a hollow plume, where ions are accelerated at angles rather than up the thruster centerline, causing a dip in ion current along the centerline. A laboratory model, built at JPL, was run at Cal Poly to characterize this plume shape and implement a shield to restore a conical shape to the plume. A similar solution was used on a different type of thruster, a cylindrical hall thruster, at Princeton with excellent results. A shield was designed to shunt the magnetic field outside the thruster, where the Princeton experiments have identified a radial magnetic field as the cause for this hollow plume. The thruster was run with and without the shield, taking measurements of the ion current in the plume using a linear probe drive. The shield fixed the plume shape, increasing centerline current by 48%, however it also had detrimental effects on thruster performance, causing a decrease in thrust, specific impulse, and cut the total efficiency in half. The shield design was reexamined and a new design has been suggested for future testing of the HEMPT to restore performance while still fixing the plume shape.

Low wind speed conditions should be studied because these conditions can present risk, particularly for areas immediately surrounding the release point, where high concentrations can occur and not dissipate. The following research attempts to clarify the processes governing both the general and low wind speed cases by determining the accuracy and uncertainty of standard prediction methods for contaminant plume transport in low wind speed plume modeling. Multiple techniques were utilized to incorporate field measured data, previously gathered for a different purpose, to generate parameter distributions and ground-truth data that could be used in stochastic models for chemical plume prediction. These data were taken during a multi-day experiment performed on Frenchman Flats, a flat, dry lakebed, at the Nevada Test Site (NTS) in February of 2007 and include weather data and chemical concentrations throughout the chemical release time. I organized these data into continuous time series for each sampling location, which were represented as vectors for the statistical and mathematical analysis. I then animated these vectors with respect to time and performed a stochastic analysis which I compared to these observed vectors. Predicted vectors of chemical concentrations, based on the statistical parameter distributions generated from the observed vectors were developed and a statistical analysis was performed on the results of the stochastic process to determine how well the model predicted the plume. It was found that stochastically modeling, with SCIPuff, of contaminant plume releases in low wind speed conditions is not accurate. This was expected because below 2 m/s, plumes no longer have a Gaussian distribution and are difficult to predict because of fluctuating winds. In fact, the model only accurately predicts the period before the plume arrives at the sensor when no plume is present. It is possible, and even probable, that stochastic modeling of contaminant plumes will provide a means to compute the bounds of a release, when coupled with a model that is accurate for low wind speed conditions and includes all the complexities of the wind field. An unexpected finding is the fact that the vertical dimension of wind movement cannot be ignored in low wind speed conditions. When planning future experiments, special attention should be paid to obtaining a good representation of the 3-D wind profile.

Electrical activity at volcanoes has been recently recognized as a potential new remote sensing technique for plume-forming eruptions. Volcanic electrical activity takes place in the conduit and plume and therefore has the benefit of being a direct indicator of surface activity. This is unlike seismic signals, which indicate magma/gas movement underground, and infrasound signals, which indicate a surface explosion but not necessarily the formation of an ash plume. There are two distinct types of volcanic electrical discharges: volcanic lightning and continual radio frequency (CRF) impulses. This dissertation explores the relationships between these two electrical signals and other commonly monitored volcanic parameters. For volcanic electrical activity to be widely adopted into monitoring platforms it is important to understand how electrical discharges at volcanoes are related to other monitored signals. I present a case study of the electrical activity at Sakurajima Volcano, Japan. The lightning mapping array (LMA) is used to record both lightning and CRF. I relate CRF to ash properties and show that CRF corresponds to eruptions containing more juvenile magma that has undergone milling as it is transported out of the conduit. Seismic, infrasound, and video data are used in conjunction with multivariable statistical methods on a suite of electrical parameters to show that high levels of volcanic electrical activity are related to eruptions with large infrasound signals (> 107 J), high initial velocities (> 55 m/s), and relatively tall plume heights (> 1 km). Finally, an examination of globally detected lightning at Bogoslof Volcano, AK shows the potential for volcanic lightning in plume tracking (0-100 km), even after the end of the explosive phase of the eruption.