| Summary: | The degassing of magmatic volatiles during eruptions beneath ice sheets and glaciers, as recorded by the dissolved volatile content quenched in volcanic rocks, could provide powerful new constraints on former ice thicknesses in volcanic areas. As volcanic rocks are readily dateable using radiometric methods, subglacial volcanoes may therefore provide crucial information on Quaternary palaeo-environmental fluctuations. The use of a degassing-based reconstruction technique would be particularly valuable when studying deposits that were erupted entirely subglacially and therefore lack other diagnostic indicators of ice thickness such as subglacial-subaerial transitions. In order for magma degassing to potentially record palaeo-ice thicknesses a number of factors need to be considered[1,2], which include whether non-equilibrium degassing may have occurred, whether samples have undergone post-eruption hydration, are strongly compositionally heterogeneous, or have moved post-quenching, whether the quenching pressure reflected loading by rock, ice or meltwater, and whether pressure may have deviated significantly from glaciostatic due to meltwater drainage. Degassing during individual eruptions may be considerably more complex than anticipated[2], making interpretation of results challenging. Examples from both rhyolitic and basaltic eruptions in Iceland and elsewhere will be used to illustrate these important factors. The analytical techniques used to measure volatile concentrations need to improve on the common practise of using infra-red spectroscopy alone to determine H2O contents in one part of a sample. Multiple analyses are required to quantify the degree of heterogeneity within samples and techniques such as manometry, ion microprobe or electron microprobe are required to analyse other species (CO2, S, F, Cl). CO2 is particularly important as only trace amounts, beneath the detection limits of commonly-used analytical techniques (30 ppm), strongly affect the solubility-pressure relationships of water in ...
|
|---|