Evolution of the crustal reservoir feeding La Palma 2021 eruption. Insights from phase equilibrium experiments and petrologically derived time scales

Crystallization experiments were performed on three representative samples of the 2021 La Palma eruption at variable temperatures (920–1150 °C), pressures (100–400 MPa), and H₂O-CO₂ ratios in order to shed light on the pre-eruptive reservoir architecture and evolution. Experimental data reveal that La Palma crustal reservoir was at 300 MPa (10 km), at a temperature of 1065 °C with a melt water content of 2–3 wt%, lying in the stability field of amphibole. Mineral compositional zoning along with experimental constraints and whole rock data, show that a cold magma body (850–950 °C), likely a remnant of previous eruptive episodes at La Palma, was rejuvenated by hotter magmas that increased the temperature of the bottom portion of the reservoir up to 1135 °C. Time scales derived from olivine diffusion profiles show that such a reactivation started 10–15 years prior to eruption, and was marked by at least four different injections from a deep mantle reservoir at ≥25 km. This sequence is corroborated by geophysical signals and changes of surficial fluid geochemistry monitored during that period. Olivine zoning further indicates that the last mafic recharge prior to eruption onset occurred in mid-October 2018, and was followed by a post-injection cooling phase which continued up to the date of the eruption, during which the top portion of the rejuvenated body re-entered the stability field of amphibole. This cooling period preceding the eruption could in part explain the absence of pre-eruptive seismic signals at ∼10 km, as revealed by the revision of the precursory seismic catalogue since 2017. Once initiated, the eruption drained the 300 MPa body, which in turn activated the deep-seated mantle reservoir, lying at >500–600 MPa, which supplied fresh, hotter and volatile-rich magma, that was emitted during the second half of the eruptive episode. Amphibole breakdown documented in first emitted magmas is related to decompression and not to overheating of the resident magmas. The fact that the rocks emitted during the first half of the eruption do not bear textural or compositional evidence for a mafic recharge occurring a short time prior the eruption suggest that the eruption triggering is linked to the internal evolution of the reservoir (volatile build-up) or to external factors.