Flow and sedimentation of pyroclastic density currents: from large scale to boundary layer processes.

Our Earth is a living Planet in which rocks are exhumed at the surface, and subjected to erosion, transport and deposition. Pyroclastic density currents (PDCs) concentrate all these steps in a single phenomenon. During explosive volcanic eruptions, rock fragments known as pyroclasts are ejected from the inside of the Earth to the surface. They can then be transported as a mixture of gas and particles, the PDC, possibly entraining air and eroding their bed until they lack kinetic energy and dissipate as sediment. PDCs are thus a fundamental transport mode associated with explosive eruptions, and belong to the class of particulate density currents. Those are mixtures of particles and ambient fluid, which behave, as a whole, as a fluid, and where the agent of excess density driving the momentum is the particles. Particles forming PDCs thus combine the roles of driving the momentum and be the resulting deposit when momentum comes to lack. With the improvements of technology, models and experiments become more and more sophisticated. Numerical simulations benefit from greater calculation capacities, whereas analogue models and experiments can be recorded at high speed and resolution in 4D. These tools are fundamental to the understanding of the otherwise inaccessible internal dynamics of PDCs. Yet the question remains, what to model? How to know which effects are acting, if a model is accurate, and what should be the input parameters? The answers must be entirely based on observations of the natural phenomenon, and the field sedimentologist has the role to answer these questions. This study investigates deposited PDCs, with the goal of understanding their former flow dynamics. Whereas others have made the link between the eruption and the parti- cle transport, the present target is to understand the particle transport from the study of the sediment. The field examples are predominantly based on the deposits of the Au- gust 17. 2006 eruption of Tungurahua volcano (Ecuador). This small volume eruption deposited sediments that present a great variety of sedimentary characteristics including dune bedforms in an well-preserved state. A combination of methods are used: direct field observation benefit from laboratory instruments and sedimentary peel techniques. Ground Penetrating Radar and Terrestrial Laser Scanner record the 3D architecture of bedforms. Unfortunately, the latter results, acquired in collaboration with the Ecole et Observatoire des Sciences de la Terre -EOST, Strasbourg, France- (and financed by THESIS and LMU for this PhD), cannot be detailed in this work since they are present in the PhD dissertation from Jean Remi Dujardin (2014, EOST). The structure of the dissertation is a a voyage through the scales, from the eruption scale to the particle scale. The mapping of the deposits highlights the large-scale behavior driven by the whole flow. The formation of dilute PDCs is linked to flow stripping pro- cesses at cliffs and curves of valleys confining the main flows.The lower density of dilute PDCs permitted them to overflow valley-walles on outer overbanks of curves and rapidly deposit. At the meter scale, pristine dune bedforms shape the surface of these deposits, and show an evolution in their dimension, form as well as internal patterns. Beyond the usual interpretation as antidunes, those bedforms are found to contain a great richness of information. Antidunes are indeed almost systematically evoked for the formation of PDC bedforms. However, this type of structures represents anecdotic phenomena related to transcritical, stationary gravity waves developing at density interfaces. Several alterna- tives to the antidune interpretation are presented throughout the dissertation. Finally, the scale of the basal boundary interaction between the flow and substrate is scoped. Wind tunnel measurements of the saltation threshold permit to derive quantitative flow param- eters from the recognition of the transport mechanism and grain size of involved particles. The basal flow processes and sediment state are approached from the study of soft sed- iment deformation. Sedimentary peels (lacquer peels of undisturbed deposits) enlighten levels of details that permit to recognize features previously only suspected from analogue experiments. The variation of deposit grain size between laminae with sub-mm thickness can be linked with the wind tunnel measurements and tell about levels of turbulence in dilute PDCs. The answer is blowing in the wind. All these results represent but a prime in volcanic sedimentology. Whereas under- standing eruption mechanisms helps to forecast these events in time, the study of the deposit permit to constrain their dispersion in space. As particulate density currents, many of the mechanisms evidenced from PDCs are also valid for economically relevant sedimentary targets such as glacial hyperpycnal flows or turbidity currents