Jökull - 01.01.2012, Síða 48

Dugmore and Newton
Vegetated plateau and terraced areas are likely to
contain the best indication of fallout thickness, slopes
are likely to have experienced either erosion of the
primary fallout thickness, or thickening due to tephra
mobilisation down-slope. Sediment traps such as
well-vegetated basins are likely to have the most com-
plete tephra records and if their catchments are sta-
ble, they are unlikely to contain multiple layers of re-
worked tephra (e.g. Kirkbride and Dugmore, 2001a).
Lowland, ecologically-favoured areas are likely to re-
cover more rapidly from the impacts of tephra fall
than upland, ecologically marginal areas, and as a re-
sult lowland tephra sequences are likely to have less
disturbance. Crucially, if a tephra layer is found in
multiple profiles in contrasting geomorphological set-
tings, then it is not likely to be the product of lo-
calised tephra re-mobilisation and is likely to define
an isochron.
Tephra deposits will experience varying degrees
of reworking and redistribution while they are ex-
posed to the surface environment. Original obser-
vations of the season-by-season changes to the fall-
out from the 2010 AD eruption of Eyjafjallajökull
show that the small-scale variability of tephra layer
thickness (a good indication of the cumulative amount
of post-depositional change) is a reflection of land-
scape stability and the completeness and depth of
vegetation cover. The mobilisation of a tephra de-
posit - and its potential movement across the land-
scape - will be minimised if the full thickness of the
tephra layer is rapidly stabilised by a spatially con-
tinuous vegetation cover. Redistribution will result in
areas stripped of primary tephra deposits. This pro-
cess has been observed happening to Ey2010 in the
un-vegetated forelands of the southern margin of Ey-
jafjallajökull ice cap, and also to the fallout of the c.
1357 eruption of Katla on vegetated surfaces at Fell í
Mýrdalur (Figure 1). The stripping of unconsolidated
tephra from exposed, unvegetated surfaces affected by
winds, water and frost action is to be expected. The
reasons for the near-complete removal of tephra from
grass-covered, slopes of aggrading soil are less obvi-
ous. Fell í Mýrdalur lay beneath the principal axis of
fallout in 1357 AD, was comparatively close to the
eruption site and received coarse (sand-grade) fallout
(Einarsson et al., 1980; Figure 1). If, as is likely, hill
slopes of around 30◦ near to the farm were covered
by a well-grazed sward, there would have been little
opportunity for a decimetre-scale deposit of coarse-
grained tephra to stabilise, especially as this is one
of the wettest parts of Iceland. In contrast, the accu-
mulation of a continuous ’rain’ of silt-grade aeolian
sediment did take place, as did the discrete episodes
of silt-grade, mm-thickness fallout from both Hekla
1300 and 1341 eruptions; the crucial difference being
that the fine-grained silts could work into the vegeta-
tion mat and thus be incorporated into the stratigra-
phy, whereas the coarse grained tephra from c. 1357
AD evidently did not.
THE USE OF POORLY PROVENANCED
TEPHRA STRATIGRAPHIES
In contrast to the generally well-known tephrochrono-
logy of the last 1200 years, pre-Landnám tephra strati-
graphies may contain many unidentified layers. The
basic framework is secure and built around one of
Thórarinsson’s great legacies, knowledge of the great
silicic layers from Hekla; Hekla-S, Hekla 3, Hekla 4
and Hekla 5 (Thórarinsson, 1944; Larsen and Thórar-
insson, 1977), now supplemented with a thorough un-
derstanding of the volcanic history of Katla (Larsen
et al., 2000, 2001; Óladóttir et al., 2005), Gríms-
vötn, Bárdarbunga and Kverkfjöll volcanic systems
(Óladóttir et al., 2011b). These studies have iden-
tified over 550 Holocene tephra layers, established
their chemical characteristics and revealed the erup-
tion frequency of key volcanic systems, but despite
these monumental achievements the spatial distribu-
tions of most Holocene layers is yet to be established.
As a result, local details can be usefully added using a
’barcode’ approach that replicates recognisable strati-
graphic sequences over short distances (Figure 2b).
The key to the ’barcode’ approach is that it uses
the thicknesses and stratigraphic order of layers to
make very short range correlations typically over dis-
tances of 10–100m. This approach is unlikely to work
over longer (km scale) distances because of the effects
of different tephra plume orientations. Even if a short
stratigraphic sequence of tephra layers are all from the
46 JÖKULL No. 62, 2012