The 1257 Samalas eruption (Lombok, Indonesia): the single greatest stratospheric gas release of the Common Era

Abstract

Large explosive eruptions inject volcanic gases and fine ash to stratospheric altitudes, contributing to global cooling at the Earth's surface and occasionally to ozone depletion. The modelling of the climate response to these strong injections of volatiles commonly relies on ice-core records of volcanic sulphate aerosols. Here we use an independent geochemical approach which demonstrates that the great 1257 eruption of Samalas (Lombok, Indonesia) released enough sulphur and halogen gases into the stratosphere to produce the reported global cooling during the second half of the 13th century, as well as potential substantial ozone destruction. Major, trace and volatile element compositions of eruptive products recording the magmatic differentiation processes leading to the 1257 eruption indicate that Mt Samalas released 158 ± 12 Tg of sulphur dioxide, 227 ± 18 Tg of chlorine and a maximum of 1.3 ± 0.3 Tg of bromine. These emissions stand as the greatest volcanogenic gas injection of the Common Era. Our findings not only provide robust constraints for the modelling of the combined impact of sulphur and halogens on stratosphere chemistry of the largest eruption of the last millennium, but also develop a methodology to better quantify the degassing budgets of explosive eruptions of all magnitudes.

Figure 1. Map of the Lesser Sunda Islands and their active volcanoes.

SRTM DEM at 3 arcsecond (~90 m) resolution (http://srtm.csi.cgiar.org) of Bali, Lombok and Sumbawa. This map was generated using the Esri ArcMapI 10.1 software (http:www.esri.com).

Figure 2. Melt inclusions record the magma evolution leading to the 1257 eruption.

(a) K₂O vs SiO₂ variation diagram for the Rinjani-Samalas calc-alkaline suite including whole-rocks, melt inclusions in olivine of the 712 high alumina basalt (Fo is olivine forsterite content, i.e. 100 × Mg/(Mg + Fe)), in plagioclase of 2550 B.P. and 1257 pumice clasts (An is anorthite content, i.e. 100 × Ca/(Ca + Na)), in clino-(cpx), ortho-(opx) pyroxene, amphibole, and matrix glasses of the 1257 eruptive products. Compositions are normalised to 100 wt%, free of volatiles. The Rinjani-Samalas suite plots in between the whole-rock compositional fields of 1815 Tambora and 1963 Agung products, highlighting the enrichment in K₂O of magmas towards the East of the Lesser Sunda arc. (b) Rb-Th positive correlation indicates that the 1257 trachydacite derived from its parent basaltic magma through a dominant process of fractional crystallization. Plagioclase An_82–75-hosted melt inclusions are representative of the 1257 whole magma composition, whereas plagioclase An_50–46-hosted melt inclusions record its shallow depth in-situ crystallization. (c) Cu vs Th variation diagram showing strong Cu fractionation during magma differentiation recording the prevalent Cu-sulphide segregation. See Supplementary Figures S2 and S3 for more major and trace element variation diagrams.

Figure 3. Volatile repositories in the 1257 trachydacitic system.

Composition of the residual melt is recorded by melt inclusions trapped in cpx, opx and amphibole whereas that of the whole magma (melt + mineral phases + pre-eruptive vapour) is preserved by melt inclusions in plagioclase An_82–75. The existence of pre-eruptive vapour is illustrated by the occurrence of water-rich fluid inclusions in minerals (Supplementary Figure S5). CO₂ is very likely present in pre-eruptive vapour. The whole system is likely the result of the mixing of trachydacitic magma batches displaying distinct volatile contents (Supplementary Figure S4f).

Figure 4. Melt inclusions record the volatile evolution through magma differentiation and in-situ crystallization.

(a) H₂O vs Th and (b) H₂O vs FeO* contents in melt inclusions reflect water exsolution through magma differentiation. Reported average H₂O content of basalt are calculated considering melt inclusions unaffected by H⁺ diffusion through host mineral (Supplementary Figure S6b). (c) S vs Th variation diagram in melt inclusions. (d) Positive correlation of S and FeO* in the 1257 melt inclusions is consistent with sulphide saturation. The S concentration at sulphide saturation (SCSS, dashed curve) was calculated with model B of Fortin et al.. (e) Cl vs Th contents in melt inclusions and whole-rocks suggest that Cl has an incompatible behaviour during basalt differentiation. (f) Cl vs FeO* contents in melt inclusions, whole-rocks and matrix glasses reflect mixing between trachydacitic melts and Cl exsolution during in-situ crystallization of the 1257 trachydacite. The process of mixing is also demonstrated by the positive correlation of S and Cl (Supplementary Figure S4f). Symbols as in Fig. 2. Average initial volatile contents of melt inclusions representative of the basaltic magma and of the whole 1257 system as well as whole-rock compositions are reported with error bars (1σ).

Figure 5. The 1257 Samalas eruption produced the largest volatile emissions of the Common Era.

Plots of SO₂, Cl and Br emissions (in Tg, i.e. megatons) for climate-impacting plinian eruptions. Yields of eruptions before 1980 are from petrological studies. Fine lines represent the maximum estimates (associated with the contribution of the parent basaltic magma in the case of the 1257 eruption). References: Minoan 3,600 y B.P. (Greece); Peaktu 946 A.D. (DPRK/China); Huaynaputina 1600 (Peru); Tambora 1815 (Indonesia), emissions re-calculated at 36–45 Tg S (73–91 Tg SO₂) and 18–23 Tg Cl based on the syn-eruptive losses (400 ppm S; 200 ppm Cl) and new volume estimates of 41 ± 4 km³ DRE and 51 km³ DRE69); Cosigüina 1835 (Nicaragua); Krakatau 1883 (Indonesia); El Chichón 1982 (Mexico); Pinatubo 1991 (Philippines).