Jökull - 01.12.2006, Qupperneq 67

Pleistocene rhyolitic volcanism at Torfajökull, Iceland
Table 1. Ar-Ar ages and uncertainties, along with the epoch, glacial/interglacial period, and OI stage indicated
by the age. – Yfirlit yfir argon-argon aldursákvarðanir, mæliskekkjur og myndunarskeið.
Location Sample 40Ar/39Ar Uncertainty Epoch Glacial (G) or OI Stage
ID age Interglacial (I) Period
Rauðfossafjöll TJ97-29 67 ka ±9 ka Late-Pleistocene Weichselian (G) 4 (or 5)
Illihnúkur TJ97-18 72 ka ±7 ka Late-Pleistocene Weichselian (G) 4 (or 5)
Un-named ridge TJ97-9 83 ka ±6 ka Late-Pleistocene Eemian (I) 5
Gvendarhyrna TJ97-22 278 ka ±18 ka Mid-Pleistocene Drenthe G 8
Hábarmur TJ97-14 384 ka ±20 ka Mid-Pleistocene Holsteinian (I) 11
and/or compositional conditions during their forma-
tion, which suggest that they are in equilibrium with
the surrounding glass and therefore probably formed
not long before eruption. Based on feldspar crystal-
lization rates and residence times reported in the lit-
erature (e.g. Anderson, 1984; Hawkesworth et al.,
2004), the difference between feldspar crystallization
and eruption is likely to be less than 5 ka, which
is less than the uncertainty on the Ar-Ar ages (7–
9 ka). The balance of evidence therefore suggests that
the eruptions of Illihnúkur and SW Rauðfossafjöll
must have occurred at or soon after 67–72 ka, which
(even allowing for the uncertainties on the ages) defi-
nitely confirms their eruption during the last (Weich-
selian) glacial period. Furthermore, the near-identical
Ar-Ar ages for these two widely-separated edifices of
the ring fracture rhyolites also lends support to their
eruption during one large (c. 16 km3) eruptive event,
which corroborates the body of evidence described
above (McGarvie, 1984). Whilst the uncertainties on
the Ar-Ar ages also permit the interpretation that the
magma was erupted over a period lasting thousands of
years, this does not negate the main conclusion of the
Ar-Ar dating, that a large magma batch (c. 16 km3)
was erupted during the last (Weichselian) glacial pe-
riod at Torfajökull.
Ice-volcano interactions
An important application of the Ar-Ar dating of tuyas
is that if a clearly understood relationship between
tuya morphology and ice thickness exists, then it
should be possible to provide minimum estimates of
ice thicknesses (and therefore ice sheet elevations) at
the time of tuya eruptions (Smellie, 2000). Using rhy-
olite tuyas for such purposes is a hitherto-unexploited
proxy of past ice sheet thickness. Such information
could be of considerable value to researchers mod-
elling the rate of growth and decay of land-based ice
sheets (see for example Hubbard et al., 2006).
Tuffen (2001) developed a model for rhyolite tuya
formation related to ice thicknesses based on obser-
vations of tuyas at Torfajökull, which was refined and
developed by Stevenson (2005) using detailed investi-
gations of early-erupted material during rhyolite tuya-
forming eruptions at Kerlingarfjöll in central Iceland.
In this model (Figure 4), the lava caps of some well-
developed tuyas consist of two tiers, an upper and a
lower tier. The lower tier which comprises thicker and
steeply-ramped lavas, is interpreted to reflect confine-
ment of the flowing lava by ice walls within a steep-
sided ice chimney within the glacier (lava A in Fig-
ure 4a). In contrast, the upper tier comprises much
thinner (∼10 m thick) sub-horizontal lava flows oc-
casionally with pumiceous carapaces, which is in-
terpreted as the unconfined flow of subaerial lavas
across the ice surface in the lowermost parts of ice
cauldrons (shallow-angled basins/depressions) in the
glacier surface (lava B in Figure 4b). Ice cauldrons de-
velop above zones of basal ice melting, such as above
active geothermal areas and above subglacial erup-
tion sites in Iceland (e.g. Guðmundsson, 2005). In re-
cent years, cauldrons formed in glaciers 400–750 m
thick have been typically 1–2 km across and up to
200 m in depth (e.g. Guðmundsson, 2005). Of particu-
JÖKULL No. 56 65