Jökull - 01.01.2005, Page 63

Holocene tephra stratigraphy and eruption frequency of Katla, S-Iceland
Table 3: Mean SAR (soil accumulation rate) values calculated between tephra layers dated by 14C-method.
Layers UN, MN and LN are SILK layers, HS and H4 are Hekla layers (Larsen et al., 2001), RF4 is a basaltic
Katla layer and the A13 layer is of probable Torfajökull origin. – Reiknaður upphleðsluhraði jarðvegs (SAR)
milli lykilgjóskulaga sem hafa verið tímasett með geislakoli. Tímabilin milli þeirra eru mislöng. Gert er ráð
fyrir að upphleðsluhraðinn sé jafn á hverju tímabili og önnur gjóskulög fá aldur samkvæmt því. Lykillög merkt
UN, MN og LN eru súr Kötlulög, HS og H4 eru Heklulög, RF4 er basískt Kötlulag og A13 er e.t.v. ættað frá
Torfajökli.
Dated tephra layers Section Period length (years) Mean SAR (cm/year)
Eldgjá-Settlment Layer Atley 64 0.091
Settlement Layer-UN Atley 1720 0.047
UN-MN Atley 380 0.039
MN-LN Atley 210 0.034
LN-HS Atley 415 0.041
HS-H4 Atley 400 0.081
H4-RF4 Atley-Hillock 3100 0.027
RF4-A13 Rjúpnafell 1020 0.071
cm/year. In the HA outcrop, Atley, the calculated
SAR is unusually low, 0.027 cm/year. This most
likely reflects special conditions for soil accumula-
tion at the site rather than an absolute change in the
regional SAR, especially since the average SAR cal-
culated for the Rjúpnafell site, representing the lowest
part of the composite section, is 0.071 cm/year (Table
3).
The age of an individual tephra layer as calcu-
lated from the SAR should be taken as an approxi-
mate value. First of all, the 14C age measurements
involve considerable uncertainties, up to 130 years (at
the 95% confidence level) when combined with the
uncertainties inherent in the calibration procedure of
Stuiver et al., (1998). Moreover, the age is obtained
from charcoal or other organic material (i.e. peat or
tree trunks) found in or immediately below the tephra
layers. Therefore, the age of the organic material can
be slightly higher than the age of the tephra deposit it-
self (e.g. Thordarson et al., 1998; Sveinbjarnardóttir,
2003).
Another source of error in these calculations is the
assumption that the SAR is constant within individ-
ual periods. However, it is well known that SARs are
a function of many variables, including climate, to-
pography, drainage and vegetation (e.g. Thorarinsson,
1961). This type of error is likely to be enhanced in
the longer time intervals such as the one between the
marker layers H4 and RF4 (Table 3). However, sec-
tion locations are not a factor in the age calculations
of the tephra layers because each section begins and
ends with a 14C dated tephra layer and the SAR for
all but one time interval (H4-RF4) is calculated be-
tween dated layers of the same outcrops. Taking all
these uncertainties into account, the accuracy of the
ages reported here is estimated as ±250 years.
In order to check the accuracy of the SAR calcu-
lations, an estimated age of three Katla SILK layers
using a different approach than in this study (Larsen
et al., 2001 and Table 1b) was compared to the cal-
culated age derived from the SAR. The difference be-
tween the estimated and calculated ages ranges from
90 to 195 years, which is lower than the estimated 250
year uncertainty. Having constructed an age model for
the soil section, time-series of the tephra stratigraphy
can be established. The time-series is used to calcu-
late the Holocene frequency of explosive basalt erup-
tions within the Katla volcanic system and to address
the compositional variability of its basalts, during the
last 8400 years.
JÖKULL No. 55 63