Final Report Summary - RHEA (RHeophysics and Energy of mAgmas) To understand dynamic earth processes it is necessary to first understand how magmas flow. Volcanology, geodynamics and planetary sciences all need to know at which stresses magma will behave in a ductile or brittle manner and, even more importantly, the range of the transitional brittle-ductile field.RHEA aimed to separate the stable from the metastable flowing field of crystal bearing melts and estimate the onset of brittle behaviour. This question has been addressed by investigating experimentally and numerically the energy distribution within magmas. For the first time we directly compared numerical simulations to samples experimentally deformed at high pressure and temperature and obtained much better comprehension of the processes involved during magma deformation.This project involved collaboration between excellent universities in Europe and USA: the ETH Zürich renowned for its experimental expertise and the UC Berkeley for the originality of its numerical methods. Combining the individual strengths of these two groups allowed us to develop one of the first numerical rheometers for magmatic suspensions based on real measurements and to formulate new constitutive laws for larger scale models.During the return period, we experimentally tested our results on real volcanic systems in order to better constrain how the overall change in rheology may explain transitions in eruptive styles.Concerning the experimental section, well controlled synthetic magmas were produced with various crystal fractions. The apparent viscosity was measured in a high temperature high pressure Paterson press. From all the experiments the apparent viscosity has been extracted as well as the brittle onset. This is the first consistent study linking brittle onset in magmas to the crystal fraction. X-ray tomography was also employed to visualize different crack shapes for various melt relaxation times. As a complement to the high temperature - high pressure experiments, additional tests on analogue materials using cone and plate experimental apparatus were performed (Rheoscope 1- Haake) using various analogue fluids such as silicon oil and glycerol. Particles such as hollow spheres, glass beads and plastic particles were used to mimic the whole range of behaviour of magmas. In addition to regular measurements and strength tests, oscillatory measurements have been performed to investigate the visco-elastic properties of suspensions for various crystal fractions. From these experiments, we have determined the onset of non-Newtonian behaviour for particle-bearing fluids and shown that the behavior is similar to that observed at high pressure. Combining the two experimental approaches, it seems that non-Newtonian behaviour and brittle onset in crystal bearing magmas are both due to a geometrical factor that is based on the connectivity of particles. Concerning the numerical component of the study, two codes have been developed. The first is a finite element code aiming to reproduce the micro-hydrodynamics of suspensions. It has the capability to estimate the dimensionless drag-force, the apparent viscosity and stress localization of an elemental fluid volume. Additionally it can calculate the maximum and random packing fraction of a suspension for various particle shape and size distributions. From this code we first confirmed and theoretically explained the brittle trend observed in the experiments. We also established some first link between the dimensionless drag force and the apparent viscosity of suspensions. The second code is a Smoothed Hydrodynamic Particle (SPH) implemented within a MATLAB structure and solving the basic Navier-Stokes equation set. Ultimately we aim to develop a code that is able to solve non-Newtonian rheologies but for now it only computes simplified Newtonian flows. SPH are naturally three dimensional and consequently already investigate volcanic mass flow deposits and their predicted footprint. The aim is eventually to be able to on benchmark volcanic systems (see below). Although currently only focussed on gravity mass flow deposits, this code can potentially investigate the flow dynamics from the magmatic chamber to emplacement and will very soon be used to understand the flow within the volcanic conduit. In addition, numerical codes have also been developed for focussing on Depth Averaged Methods and the on the energy loss during lava flows emplacement.
