No abstract

Since late 1994, open-vent eruptive activity and degassing at Popocatépetl volcano, Mexico, have released large masses of CO 2 and SO 2 . Tephra and lava produced by these eruptions show evidence for mixing of mafi c and silicic magmas shortly before eruption. We present the fi rst measurements of dissolved CO 2 in the mafi c magma end member based on analyses of olivine-hosted melt inclusions that were trapped at pressures as high as ~400 MPa (~15 km depth) beneath the volcano. We combine our data with thermodynamic models to show that degassing of mafi c magma at ~150-350 MPa pressure can explain the CO 2 /SO 2 mass ratios (1-8) of volcanic gases released from the volcano during 1995-1997. Our results demonstrate that mafi c magma recharge was responsible for the high measured fl uxes of CO 2 and SO 2 from 1995 to 1997. The total SO 2 emission of 9 Mt during this period requires intrusion and degassing of a minimum of 0.8 km 3 of mafi c magma. Only ~0.3% of this new mafi c magma has been erupted in the form of mixed (hybrid) lava and tephra. Our results suggest that the ongoing eruption of Popocatépetl is essentially an intrusive event. More generally, we suggest that intrusion and deep degassing may explain the high gas fl uxes at some other open-vent volcanoes rather than convection of magma in the uppermost parts of subvolcanic conduits.

Submarine basaltic glasses from five widely separated sites on the Ontong Java Plateau (OJP) were analysed for major and volatile elements (H2O, CO2, S, Cl). At four of the sites (1183, 1185, 1186, 1187) the glass is from pillow basalt rims, whereas at Site 1184 the glass occurs as non-vesicular glass shards in volcaniclastic rocks. Glassy pillow rims from Site 1187 and the upper group of flows at Site 1185 have 8.3–9.3 wt% MgO compared with values of 7.2–8.0 wt% MgO for glasses from Sites 1183, 1184 1186, and the lower group of flows at Site 1185. Low-MgO glasses have slightly higher H2O contents (average 0.22 wt% H2O) than high-MgO glasses (average 0.19 wt%), with the exception of Site 1184, where low-MgO glasses have lower H2O (average 0.16 wt%). Average S concentrations are 910 ± 60 ppm for the high-MgO glasses v. 1030 ± 60 ppm for the low-MgO glasses. When compared with mid-ocean ridge basalt (MORB), the OJP glasses have lower S at comparable FeOT. This suggests that OJP basaltic magmas were not saturated with immiscible sulphide liquid during crystallization, but small decreases in S/K2O and S/TiO2 with decreasing MgO require some sulphide fractionation. Measurements of the wavelength of the S Kα peak in the glasses indicate low oxygen fugacities comparable to MORB values. Chlorine contents of the glasses are very high compared with MORB, and Cl/K ratios for all glasses are relatively high (>0.7). This ratio is sensitive to assimilation of hydrothermally altered material, so the high values indicate assimilation during shallow-level crystallization of OJP magmas. Ratios of H2O to Ce, which have similar incompatibility to each other, are higher than most depleted and enriched MORB. However, these high H2O/Ce values are probably also caused by the same assimilation process that results in high Cl. The water content of the high MgO-magmas before contamination is estimated to be approximately 0.07 wt% H2O, corresponding to H2O/Ce of 135 for OJP basalts, a value at the low end of the range for Pacific MORB. There is no evidence for high H2O contents that would have significantly increased extents of mantle melting beneath the OJP, and the estimated H2O content of the OJP mantle source region (170 ± 30 ppm H2O) is similar to that of the depleted MORB source (140 ± 40 ppm H2O). Instead, large extents of melting beneath the OJP must have been caused by a relatively high mantle potential temperature, consistent with upwelling of a hot mantle plume.

No abstract

No abstract