Jökull - 01.01.2015, Page 65

The Eldgjá lava flow on Mýrdalssandur
18º48’W
63
º3
0’
N
63
º3
5’
N
18º36’W 18º24’W
1
2
3
4
5
6
9
10
11
12
1314
7 8
Álftaver
lava field
Hafursey
Kötlujökull
0 4km
Figure 4. Magnetic profile layout at Mýrdalssandur. The study area is shown on a Landsat 8 image from USGS,
2013. Profiles are marked 1–14. The Álftaver lava field is from Larsen (2000). – Segulsnið sem mæld voru á
Mýrdalssandi. Mælisvæðið er sýnt á Landsat 8 gervitunglamynd, 2013. Snið eru merkt 1–14.
quency) and anomaly amplitude, were identified. Spa-
tial frequency changes were identified visually across
the profiles which correlate with depth to magnetic
sources (Hinze et al., 2013). Places where sudden de-
creases of several meters in depth to magnetic sources
occur, to the east of the estimated lava edge indicate
the presence of possible ledges. The Peters half slope
method was used to estimate maximum depth to the
magnetic sources (Figure 5). The rule can simply be
stated as (Peters, 1949; Hinze et al., 2013).
h = d/c (1)
where d is the half slope distance, c a proportionality
constant that depends on the shape of the source body
but for most cases lies between 1.2 and 2.0, and h the
depth to magnetic source.
The method is useful to determine the approxi-
mate depth to a source but requires that the anomaly
comes from a single magnetic body. This is not al-
ways the case and the observed field may be a com-
bination of several sources. The results obtained are
therefore very approximate, the uncertainty due to the
proportionality constant alone is about 30%. Other
possible error sources include irregular shape and dip
of source bodies. Thus, the error in absolute values
can easily be about 50%. However, the relative com-
parison between areas should be reliable, e.g. a dou-
bling of the depth to source bodies should show up
clearly as rapid attenuation of the field. Importantly,
the method should work well for a layer of strongly
magnetized sources covered by non-magnetic sedi-
ments.
Each profile was plotted as total magnetic field
(nT) vs. distance (m). The profiles were separated into
numbered sections where anomaly spatial frequency
was similar. The Peters half slope method was used
on ca. six anomalies in every 1000 m of each pro-
file, the number depending on usable anomalies and
avoiding anomalies that were obviously affected by
composite sources. The height of the sensor (2 m)
has been subtracted from the depths and they are thus
given relative to the surface. Identification of the lava
flow on the surface was done in the field, on aerial
photographs and satellite images.
JÖKULL No. 65, 2015 65