Magmatic timescales prior to the 1951, 1995 and 2014-15 eruptions at Fogo, Cape Verde: Insights from diffusion chronometry of clinopyroxenes and olivines

Ross Hassard¹, Eleanor Jennings¹, Hilary Downes¹ and Simon Day²

¹School of Natural Sciences, Birkbeck University of London, Malet Street, London WC1E 7HX, UK

²Formerly Institute for Risk and Disaster Reduction, University College London, Gower Street, London, WC1E 7HX, UK.

Historical records of recent volcanic eruptions on the intraplate oceanic island of Fogo (Cape Verdes) reveal significant impacts on the local population, including loss of livestock, destruction of buildings and farming areas and, occasionally, death of inhabitants. The limited volcano monitoring network on the island and the lack of diffusion chronometry studies of the erupted products means there are significant gaps in our understanding of the timescales of magmatic processes at this active volcano.

This research aims to improve eruption forecasting and hazard mitigation at Fogo by using diffusion chronometry data from clinopyroxene and olivine phenocrysts in lava flows from the eruptions of 1951, 1995 and 2014-15 to obtain pre-eruption magmatic timescales.

Key research objectives are (i) to determine diffusion timescales from normally-zoned clinopyroxene and olivine crystals to understand magma and/or lava flow residence times and ascent rates; (ii) to determine whether there a link between seismicity and eruptions at Fogo through the comparison of diffusion timescales with recorded seismic activity.

Findings include an assessment of diffusion chronometry methods which shows that published Fe-Mg diffusion coefficients underestimate timescales in Al-rich clinopyroxenes in strongly alkaline magmas, such as those at Fogo. Results for the 2014-15 eruption using published Mg self-diffusion coefficients, indicate that clinopyroxene rims formed ~5-6 months prior to eruption. Diffusion chronometry from 16 olivine phenocrysts reveals timescales of hours for residence in the lava flow for proximal samples, and up to ~1 month for distal samples, comparable with recorded seismicity prior to the eruption.