The base of glaciers and ice sheets provide environments suitable for the production of methane. High pressure conditions beneath the impermeable ‘cap’ of overlying ice promote entrapment of methane reserves that can be released to the atmosphere during ice thinning and meltwater evacuation. However, contemporary glaciers and ice sheets are rarely accounted for as methane contributors through field measurements. Here, we present direct field-based evidence of methane production and release from beneath the Icelandic glacier Sólheimajökull, where geothermal activity creates sub-oxic conditions suited to methane production and preservation along the meltwater flow path. Methane production at the glacier bed (48 tonnes per day, or 39 mM CH4 m−2 day−1), and evasion to the atmosphere from the proglacial stream (41 tonnes per day, or 32 M CH4 m−2 day−1) indicates considerable production and release to the atmosphere during the summer melt season. Isotopic signatures (−60.2‰ to −7.6‰ for δ13Cch4 and −324.3‰ to +161.1‰ for Dch4), support a biogenic signature within waters emerging from the subglacial environment. Temperate glacial methane production and release may thus be a significant and hitherto unresolved contributor of a potent greenhouse gas to the atmosphere.

Understanding patterns of geothermal and volcanic activity at many of Iceland's most active volcanic systems is hampered by thick overlying ice, which prevents direct observation and complicates interpretation of geophysical signals. Katla is a prime example, being a large and restless volcanic system covered by the 740 m thick Mýrdalsjökull ice cap, whose eruptions have triggered some of the most powerful known meltwater floods in historical times. To shed new light on geothermal and subglacial hydrological processes at Katla, we have determined the sulphate isotopic composition of a series of glacial meltwater samples discharged from Sólheimajökull, a valley glacier of Mýrdalsjökull, between 2009 and 2012. Dual isotopic analysis of δ 34 S and δ 18 O in dissolved sulphate allows identification of source mixing processes and chemical evolution during subglacial meltwater transport. Strikingly, meltwater δ 18 OSO 4 signatures indicate redox conditions at the glacier bed, which are inverse to those normally encountered at Arctic and Alpine glaciers. Discharge of reduced, anoxic meltwater in summer, rather than winter, points towards seasonal release of geothermally derived volatile gases. We attribute this to headward expansion of the channelized subglacial drainage system during the summer melt season, accessing key areas of geothermal activity within the Katla caldera. Volatile release may be further enhanced by unloading of overburden pressure due to snowpack melting in the summer season. In winter, restriction of subglacial channels to lower elevations effectively seals geothermal fluids and dissolved gases beneath the ice cap, with only sporadic release permitted by periodic increases in subglacial water pressure. When the subglacial drainage configuration permits access to key geothermal areas, sulphate isotopic signatures thereby form sensitive indicators of geothermal activity occurring deep beneath the Mýrdalsjökull ice cap.

Researchers have analyzed various properties of drumlins within individual drumlin fields in order to provide evidence to help in identifying the processes involved in drumlin formation. One property which has been examined is the spatial distribution of drumlins within a field. Traditionally, in such endeavours the individual drumlins have been represented as points and their distribution examined using techniques of point-pattern analysis. We suggest that not only is such a representation inappropriate at this scale, it also introduces statistical bias which makes the results of such analyses questionable. Consequently, we propose an alternative approach which involves representing individual drumlins as areal phenomena and considering their pattern as a two-phase mosaic. The advantages of such an approach are discussed and it is illustrated by applying it to two different drumlin fields.

No abstract