Jökull - 01.01.2021, Blaðsíða 31

The 1918 Katla eruption
Isopach map
The isopach map, based on all available data (Figures
1–4) and our reconstructed locations and thicknesses
inside the caldera, is shown in Figures 5 and 6, with
the latter being a blowup covering Mýrdalsjökull. Fig-
ure 6 shows the isopachs and the estimated locations
of fallout and calculated thicknesses of tephra at these
sites. The map in Figure 5 shows that no single direc-
tion was dominant in the distribution of tephra from
this eruption. Three dispersal axes are most promi-
nent, directed towards north, northeast and southeast.
Three other axes are indicated by the map, towards
south, west and northwest. The 1 cm isopach extends
about 100 km to the north but less than 10 km to the
southwest. Details of how dispersal of tephra varied
with time during the eruption are given in Larsen et
al. (this issue).
Volume of airborne tephra in 1918
The bulk volume of the tephra carried by the plume
and forming the tephra layer has been estimated by:
(1) direct integration of the map for contours >0.5
cm, using Surfer (Golden Software) after generating
the map using the kriging option. (2) Plotting the
logarithm of thickness against the square root of the
area within each isopach and integrating the curve
as four exponential segments, following Fierstein and
Nathenson (1992) (Figure 7a). (3) By integrating be-
tween successive contours we obtain 10 exponential
segments. In all cases, the fallout outside the outer-
most, 0.5 cm isopach, was estimated by extrapolat-
ing the exponential curve obtained for the interval be-
tween 1 and 0.5 cm out to infinity.
The results on volume by the three methods
(Table 1) all lie in the range 0.9 to 1.0 km3. They
are not fully independent, as the same method is used
in all cases for the region outside the 0.5 cm contour.
Integration of the map also shows that about half of
the total volume of airborne tephra (0.45 – 0.50 km3)
fell on Mýrdalsjökull.
The uncertainty of the volume estimate of tephra
that fell on Mýrdalsjökull can be crudely estimated.
The tephra thickness near the vents on Figure 2a
has an estimated uncertainty of 40%, while thick-
nesses for the northern part (Sléttjökull) are better
constrained due to the limited effect of post 1918 ice
flow on layer thickness, as shown above. By using
uncertainties of 40% for the caldera and 20% for the
northern part of the glacier, the resulting combined
error for the glacier part is 0.15 km3. For the areas
outside the glacier, the large number of survey points
results in lower uncertainty, which we cautiously as-
sume to be no more than 20%, or 0.1 km3. By us-
ing the mean of the three values in Table 1 as the
best available estimate for the tephra layer we obtain
a rounded off volume of 0.95±0.25 km3.
Table 1. Estimates of the bulk volume of the Katla
1918 tephra layer. – Rúmmál gjóskulagsins frá 1918.
Thickness Map Exponential method Exponential method
integration1 4 segments2 10 segments2
(cm) km3 km3 km3
<0.5 (0.16) 0.16 0.16
>0.5 0.74 0.84 0.87
Total 0.90 1.00 0.93
1Integration made using Surfer (Golden Software) for
thickness >0.5 cm and exp. integration results for <0.5 cm.
2The fallout thickness as function of the square root of area
is shown on Figure 7a.
Fallout in the ocean to the south and southeast of
the volcano is estimated as about 10% of the total,
a value obtained from integration of the ocean part
of the map in Figure 5. We estimate the total mass
of the layer using a density of 1200 kg/m3, a rea-
sonable number for basaltic, fine grained, and to a
large degree phreatomagmatic tephra (Oddsson et al.,
2012). The result is a tephra layer deposit mass of
1.15±0.30×1012 kg (about 0.4±0.1 km3 DRE).
It should be noted that the volume/mass obtained
here is far from being the total amount of material pro-
duced in the eruption, as water-transported pyroclasts
are not included. This material may have been of com-
parable quantity (Tómasson, 1996; Larsen, 2000) as
that of the airborne tephra. Moreover, any material
deposited at the eruption site, forming a subglacial ed-
ifice, is not included.
JÖKULL No. 71, 2021 29