Timing of Magmatic Processes and Unrest Associated with Mafic Historical Monogenetic Eruptions in Tenerife Island

The historical eruptive activity on Tenerife (Canary Islands, Spain) consists of relatively mafic magmas that produced mildly explosive monogenetic eruptions; this is the most likely style of eruption that might occur in the near future. Here we investigate the processes and time scales that lead to such eruptions with the aim of better interpreting volcanic unrest in the island and improving eruption forecasts. We combine geochemical and petrological information with a compilation of accounts of seismicity for the historical eruptions of Siete Fuentes, Fasnia, and Arafo. These occurred between December 1704 and February 1705 along the NE rift of Tenerife from three vents at different altitudes and more than 10 km apart from each other. The erupted volume increased with time from about 1 × 10 6 m 3 for Siete Fuentes, to 7 × 10 6 m 3 for Fasnia, and to 28 × 10 6 m 3 for Arafo. All the erupted magmas were basanitic; the bulk-rock compositions became more mafic with time owing to accumulation of olivine with Cr-spinel inclusions and clinopyroxene, rather than to the arrival of a truly more primitive melt. We have identified four olivine compositional populations [crystal cores at Fo = 79–80, Fo = 81–82, Fo = 83–84, and Fo = 85–87; Fo = 100 × Mg/(Mg + Fe)] that are either unzoned or normally or reversely zoned in major and minor elements. The variety of olivine core populations and zoning patterns reflects mixing and mingling between different magmas; their proportions changed with time from Siete Fuentes to Arafo. Clinopyroxene crystals also show a variety of zoning patterns and changes in Mg/Fe and Cr contents indicative of open-system magma interactions, but are more difficult to interpret because of fast disequilibrium growth conditions. Modelling of the zoning profiles of olivine shows that early mixing occurred between two relatively evolved magmas, probably at shallow depths 1 year prior to the eruption. Another mixing event between two more primitive magmas stored presumably at deeper levels occurred less than 2 months prior to the eruption. Finally, movement from depth of the pre-mixed primitive melts and interaction with the also pre-mixed, shallower and more evolved melts occurred only 2 weeks before the eruption. Although the first and last eruptions were fed by the same magma, each also records renewed magma movement from depth and mixing before eruption and thus they were probably fed by different dike systems. The shorter transport times of weeks are also seen in our compilation of the historical accounts of seismicity associated with these eruptions. The sequence of events and their timing should be useful for interpretation of monitoring data of unrest in Tenerife, and the short times between the last processes before eruption could be considered for the implementation of emergency plans during volcanic crises.