Doctoral Candidate 5

Objectives

Alkali-rich magmas occurring commonly at ocean islands show a large compositional range from highly primitive basaltic magmas to evolved rhyolitic or trachytic magmas, indicating a complex structuration of the magma plumbing system. The conditions (depth, temperature, volatile concentrations) and the differentiation in the reservoirs are poorly understood as well as the residence time of magmas at different levels of the crust and the connexion between magma reservoirs. The project will focus on two volcanic systems showing a diversity of mildly alkaline magmas (Pico, Azores) and highly alkaline magmas (Fogo, Cape Verde). The main aims are: 1) complementing classic and ML-based pyroxene- and amphibole-based thermobarometers to constrain the conditions of magma crystallization; 2) constraining volatiles (H₂O, CO₂, S, Cl) concentrations in silicate melt from the analysis of glass inclusions in minerals by complementing traditional and data-driven approaches; 3) performing diffusion chronometry on crystals with compositional zoning to evaluate residence time in magma reservoirs; 4) performing 3D-morphology of minerals, gas bubbles and inclusion by X-ray computed tomography. Machine learning algorithms will be also developed to automatize the segmentation of inclusions and host phases based on data from X-ray attenuation and to characterize compositional zoning of crystals-based scanning electron microscopy.

Epected Results

• By complementing state-of-the-art traditional techniques with ML investigations we expect to improve our knowledge on melt volatile contents and crystal cargo (pyroxenes and amphiboles) modelling to better constrain architecture of the volcanic plumbing system in two case studies (i.e., Fogo and Pico); 
• We expect to provide time scales estimates for the ascent rates of magmas with strong implications for the estimation of volcanic hazards and volcanic risk mitigation.