Jökull - 01.01.2010, Page 91

Geothermal noise at Ölkelduháls, SW Iceland
350 to 450 m a.s.l. This variation in elevation occurs
on a length scale comparable to 1 km. Some micro-
seismicity is found in the area. Currently, most of the
events align roughly with the distribution of geother-
mal activity from Hengill in the NW under Ölkeldu-
háls to Grændalur in the SE. The microearthquakes
are concentrated at depths between 3 and 6 km and
may be connected with a recent magma intrusion into
the area (Feigl et al., 2000) and the heat mining of
geothermal fluids from its top. See Árnason et al.
(2010) for a more comprehensive account of the ge-
ology and geophysics of the area.
DATA
We used ten instruments from the Icelandic portable
seismic instrument pool, Loki. The instruments are
Lennartz 5 second three-component seismometers
registered with a Reftek 130 flash-card recorder. Sam-
pling rate was 100 samples per second. The in-
struments were placed on bedrock (typically hyalo-
clastite) and dug down where possible. A pail was
placed over the sensor and a plastic mat over that,
weighed down and stabilized with loose rock. First,
the 10 instruments were deployed along the profile la-
beled LAX in Figure 1 to the ENE of Ölkelduháls.
They were operated in this configuration for about two
days. Nine of the instruments were then pulled out
and redeployed along the profile labeled LBX in Fig-
ure 1. The westernmost instrument closest to Ölkeldu-
háls was left in the field in order to have a common tie
point for the two deployments. The instruments were
once again operated for about two days before they
were all pulled out of the field. The coordinates of
the instruments are listed in Table 1. For both parts
of the profile, all 10 stations recorded simultaneously
for about 40 hours. Weather remained good for most
of that time. However, wind picked up for some time
during the middle of both observational periods. This
caused a very clear increase of noise, also in the fre-
quency range of interest, in particular at those sites
where the sensor had not been dug down.
A part of both profiles lies adjacent to an electric
power line and associated road or track. Some vehi-
cle traffic is along the tracks, although not at night.
No noise was detected that could be directly associ-
ated with the power line. Wind noise is not clearly
elevated at sites adjacent to the power line.
Table 1. Coordinates of seismographs used in this
study. See also Figure 1. – Hnit jarðskjálftamæla-
stöðva við Ölkelduháls. Sjá einnig 1. mynd.
Station Name Latitude Longitude Elevation
LA9 64.07299 -21.18157 389
LA8 64.07075 -21.18679 379
LA7 64.06781 -21.18813 353
LA6 64.06721 -21.19632 401
LA5 64.06457 -21.20129 382
LA4 64.06235 -21.20701 389
LA3 64.06032 -21.21219 408
LA2 64.06068 -21.21847 458
LA1 64.05943 -21.22428 422
LA0 64.05975 -21.23039 391
LB0 64.05975 -21.23039 391
LB1 64.05872 -21.23651 365
LB2 64.05775 -21.23811 352
LB3 64.05659 -21.24072 346
LB4 64.05589 -21.24367 365
LB5 64.05478 -21.24553 378
LB6 64.05303 -21.24764 381
LB7 64.05229 -21.25038 380
LB8 64.05108 -21.25425 375
LB9 64.04867 -21.25411 392
SPECTRAL AMPLITUDES
We compute amplitude spectra from the data using an
FFT routine over 6 minute intervals. The 10 power
spectra thus computed within each hour are then av-
eraged and the resulting amplitude spectra plotted in
Figures 2 and 3, hour by hour. The spectra are com-
puted from all three components of motion (by sum-
ming their power spectra). The spectral density unit is
arbitrary, but uniform (all data are from identical in-
struments that are flat in their velocity response from
0.2 Hz. The spectra are plotted up to 10 Hz as that fre-
quency range contains that part of the spectrum that
can clearly be associated with the geothermal activ-
ity. In some cases the spectra are clearly disturbed by
wind noise or seismic events and have been removed
JÖKULL No. 60 91