Diverse mantle components with invariant oxygen isotopes in the 2021 Fagradalsfjall eruption, Iceland

Iceland sits at the intersection of a mantle plume and the Mid-Atlantic Ridge, with the Reykjanes Peninsula (RP) forming the onshore extension of the Reykjanes Ridge and hosting significant population and infrastructure. After 781 years of dormancy, a low-intensity effusive basaltic eruption began at Fagradalsfjall on March 19, 2021, likely marking the onset of a new eruptive period. Geophysical data indicate a brittle–ductile transition at 6–7 km and a Moho at ~15–20 km depth, suggesting magma storage in the lower crust/upper mantle. Understanding the mantle source(s) feeding the eruption, particularly their oxygen isotope composition, is critical for assessing magmatic and hydrothermal processes on the RP. A long-standing debate concerns the nature of plume components and whether Iceland’s large δ18O diversity in basalts (≈ +2 to +6‰) reflects mantle heterogeneity, crustal assimilation of hydrothermally altered low-δ18O crust, or both. This study investigates major and trace elements and oxygen isotopes (δ18O, Δ17O), along with hydrogen isotopes (δD) and H2O contents, in a time-resolved suite of Fagradalsfjall samples to test for δ18O variability, link geochemical changes to mantle components, and evaluate crustal assimilation versus mantle-source controls.

Prior work identifies geochemically enriched and depleted components beneath Iceland delivered by a deep plume, characterized by trace-element ratios (e.g., Zr/Y, Nb/Zr, Nb/Y) and radiogenic isotopes. Ambient upper mantle (MORB) typically has δ18O ≈ 5.5‰, but many Icelandic basalts and their crystals/melt inclusions show broader δ18O ranges, often attributed to low-temperature crustal processes or mantle heterogeneity. Studies at Theistareykir, Askja, and Holuhraun document relatively low δ18O values, commonly interpreted as assimilation of hydrothermally altered crust or heterogeneous mantle. Mixing models along Reykjanes Ridge and Iceland invoke multiple endmembers: depleted components (D1, D2), enriched plume components including EMORB-like and OIB-like (e.g., Eldgjá, Snæfellsnes), and a Depleted Icelandic Plume (DIP). Recent MORB δ18O reference values based on laser fluorination of glass refine the mantle baseline to ~5.5‰, improving comparisons with Icelandic data.

Sampling and temporal coverage: 30 lava and tephra samples for bulk-rock major and trace elements; 47 glassy samples for δ18O; 3 samples for Δ17O; 10 analyses for δD and H2O; additional 17 H2O measurements (several below detection). Samples span 160 days from March to August 2021, covering initial to late eruption stages and multiple vents. Petrography and EPMA: Groundmass glasses and minerals analyzed by EPMA (JEOL JXA-8230, 15 kV, 10 nA, 10 µm beam); A99 standard used for drift monitoring. Whole-rock geochemistry: Major and trace elements measured at Activation Laboratories (package 4LITHORES). Samples fused with Li metaborate/tetraborate; majors by ICP-OES; traces by ICP-MS (Perkin Elmer Sciex ELAN 6000/6100/9000). Detection limits: majors 0.01 wt.% (MnO, TiO2 at 0.001 wt.%), trace elements 0.01–30 ppm. Precision: better than 5% (majors) and 10% (traces). Major elements normalized volatile-free. Oxygen isotopes (δ18O): Laser fluorination of 1–1.5 mg glass at University of Oregon (UO). BrF5 reagent; excess F2 removed by mercury diffusion pump; O2 converted to CO2 for dual-inlet MAT253 analysis. Calibration with San Carlos olivine (δ18O = 5.25‰), UWG2 garnet (5.80‰), and in-house UOG garnet (6.52‰). Sessions included 4–6 standards; day-to-day corrections 0.0–0.2‰; typical 1σ on standards ≈ ±0.06‰. Triple oxygen isotopes (Δ17O): O2 measured in dual-inlet MAT253 with dedicated triple-O line; purification via GC and 5 Å zeolite traps; UWG standards used to correct to VSMOW; SCO standard used for reference. Hydrogen isotopes and H2O: TC/EA-MAT253 at UO on 50–200 µm groundmass glass. Samples (8–18 mg) reduced at 1450 °C in glassy C reactor; H2 measured relative to reference gas. Standards: USGS57/58 micas and MORB glass UOB. H2O by peak integration if >~0.02–0.04 wt.%; δD precision ±8‰ (<0.1 wt.% H2O) to ±5‰ (higher H2O). Degassing modeling: Batch and Rayleigh models from starting compositions similar to MORB glass standard (UOB, 0.40–0.53 wt.% H2O) to assess δD-H2O trends. Data quality: One duplicate whole-rock showed virtually identical results; low LOI values; uncertainties on plotted major/trace elements better than 5/10%, respectively.

- Trace elements and majors: SiO2 = 48.1–49.8 wt.%; MgO = 8.1–9.7 wt.% (increased early, then plateaued). K2O/TiO2 increased during first two months then stabilized. Nb/Zr increased continuously over 160 days (≈0.10 to 0.20). Zr/Y ranged 2.5–4.6; Th/Yb 0.2–0.7; early products more heterogeneous, later products leveled. All samples enriched relative to Nb/Zr > 0.08. - Oxygen isotopes: δ18O in glasses = 5.0–5.7‰; mean 5.4 ± 0.3‰ (2 SD, N = 47); monthly means (Mar–Aug) 5.3–5.5‰ and indistinguishable within ±0.2‰ natural variability. Values overlap MORB (~5.5‰) and show no coupling with trace element variability. Δ17O (N=3) = −0.031 to −0.048 ± 0.012‰, within mantle range. - Hydrogen isotopes and H2O: δD = −69.2 to −109.7‰; mean −81.8 ± 26‰ (2 SD, N = 10). Except for the most degassed sample (−109.7‰), δD overlaps canonical MORB (−80 ± 10‰). H2O in Icelandic glasses = 0.00–0.23 wt.% (subaerial, degassed), lower than submarine MORB UOB standard (0.38–0.53 wt.%). δD-H2O trends consistent with near-surface degassing (Batch and Rayleigh). - Mantle source inferences: Incompatible trace element ratios (Zr/Y, Nb/Zr, Nb/Y, Th/Yb) indicate mixing of multiple mantle components (depleted D1/D2, EMORB-like, and OIB-like Eldgjá/Snæfellsnes), with apparent increase of enriched component over time. Despite this, δ18O remains invariant and mantle-like, indicating absence of 18O-depleted mantle or significant assimilation of low-δ18O crust beneath RP. - Eruption and storage: Geophysics indicates lower crustal storage near Moho (~15–20 km). Mineral zoning and evolving trace elements suggest a replenished, zoned reservoir with short residence times preserving mantle δ18O.

The study aimed to test whether the mantle feeding the 2021 Fagradalsfjall eruption exhibits oxygen isotope heterogeneity or reflects crustal assimilation. The invariant, mantle-like δ18O (~5.4‰) throughout 160 days of eruption, despite large variations in incompatible trace element ratios, shows that oxygen isotopes are decoupled from trace-element-defined mantle components. Combined δD and low H2O contents indicate subaerial degassing without evidence for seawater or meteoric hydration, and argue against significant assimilation of hydrothermally altered, low-δ18O crust. Alternative explanations for trace element evolution—fractional crystallization or variable degrees of melting of a single source—are inconsistent with the major element trends and simultaneous variability of similarly incompatible element ratios. The most consistent model is mixing of closely spaced mantle components (depleted peridotite, EMORB-like, and enriched OIB-like domains), with replenishments into a lower crustal reservoir near the Moho and short pre-eruptive residence that preserved mantle δ18O. Regionally, the results imply relatively uniform primary mantle δ18O beneath Reykjanes and suggest that low δ18O basalts elsewhere in Iceland (e.g., Holuhraun) likely reflect assimilation of low-δ18O crust or delamination of such crustal materials into the source, rather than pervasive low-δ18O mantle.

Fagradalsfjall basalts record mixing of diverse mantle components (depleted, EMORB-like, and enriched OIB-like) feeding a lower crustal reservoir, yet maintain invariant, mantle-like oxygen isotopes (mean δ18O = 5.4 ± 0.3‰). Δ17O and δD also overlap mantle ranges, and H2O–δD systematics indicate near-surface degassing. These data collectively show no evidence for 18O-depleted mantle or significant interaction with low-δ18O crust beneath the Reykjanes Peninsula. The Fagradalsfjall δ18O value thus provides a useful mantle reference for this part of the Icelandic plume system. Future work should expand spatial coverage across Iceland to evaluate the lateral uniformity of mantle δ18O, further integrate radiogenic isotopes with triple oxygen and hydrogen isotopes, and quantify the geometry and scales of mantle heterogeneity and melt transport that govern trace element evolution.

- Δ17O dataset is limited (N = 3), constraining precision on triple oxygen isotope comparisons. - δD and H2O datasets are affected by subaerial degassing and include several anhydrous samples below detection, limiting volatile-system interpretations; δD is sensitive to degassing path. - The study focuses on subaerial, partly degassed glasses; submarine quenched materials were not available for direct comparison at the eruption site. - Incompatible trace element interpretations rely on literature mixing frameworks; Th/Yb literature data for RP are relatively sparse. - Localized crustal assimilation cannot be completely excluded for individual outliers, though inferred to be minor. - Findings pertain to the Reykjanes Peninsula and may not directly generalize to all Icelandic volcanic zones without further sampling.