Jökull - 01.01.2010, Side 53

Upptyppingar seismic swarms
LOCATION RESULTS
To illustrate the effects of processing techniques, net-
work geometry, network size, and phase-picking pre-
cision on hypocentral locations, we highlight five in-
dividual earthquake clusters that occurred from 20–
24 July 2007 and that represent 154 of the 288 events
in our study period. Courtesy of the IMO, we have
access to SIL-system manually refined time picks for
each event in the dataset as well as double-difference
locations and uncertainty estimates. Catalogue loca-
tions produced by the SIL system’s automated loca-
tion software and reviewed manually are available to
the public on the worldwide web and are also pre-
sented here.
All locations described in this manuscript are de-
rived using the IMO’s linear-gradient Vatnajökull ve-
locity model. The Vatnajökull model, shown in Fig-
ure 4, is based on extrapolation from the ICEMELT
refraction profile (Darbyshire et al., 1998, 2000). Al-
though most SIL on-line catalogue locations are de-
rived using the standard SIL model (Stefánsson et
al., 1993; Hjaltadóttir et al., 2009), locations in the
Upptyppingar area default to the more locally suitable
Vatnajökull model.
I. Catalogue locations
We begin with an examination of CMM output com-
pared with on-line SIL catalogue locations. Firstly, it
is important to note that SIL catalogue locations are
computed with respect to the elevation of the nearest
one or two seismic stations to an event. For Upptypp-
ingar, this value is approximately 700 m; however, the
offset could be substantially greater in regions with
higher station elevations. Furthermore, the SIL cata-
logue locations are derived using data from up to 18
SIL stations. The CMM locations are derived using
data from the ASN (i.e., 21 Cambridge University in-
struments plus six nearby SIL stations).
In Figure 5, we display the positions of SIL cata-
logue locations relative to the output of CMM as di-
rectly reported by each method. A systematic offset in
depth is apparent; however, since CMM locations are
reported as depth below sea level and SIL catalogue
locations are reported as depth beneath the average el-
evation of the nearest receivers, the SIL locations ap-
pear 700 m deeper relative to the CMM locations than
they ought to. A correction of 700 m, however, does
not account for the entire offset in depth (Table 1, first
column). We also note that the CMM locations, which
are derived from ASN data, are noticeably tighter than
the SIL catalogue locations, particularly in northing
and depth.
3 4 5 6 7 8
0
5
10
15
20
25
30
35
Velocity [km/s]
D
e
p
th
[k
m
]
Figure 4. One-dimensional P-wave velocity model
used to locate hypocentres. The linear-gradient,
’Vatnajökull’ model is extrapolated from the
ICEMELT seismic refraction profile (Darbyshire
et al., 1998, 2000), and assumes a constant VP /VS
ratio of 1.78. Depth is defined as below the average
surface elevation of the two receivers nearest to the
events, or approximately 700 m in this case (K.
Vogfjörd and R. Slunga, personal communications).
– Einvítt P-hraðalíkan, Vatnajökulslíkanið, sem notað
er við staðsetningar. Líkanið byggir á niðurstöðum
úr ICEMELT verkefninu (Darbyshire og fleiri 1998,
2000) og gerir ráð fyrir VP /VS hlutfallinu 1,78. Dýpi
miðast við tvær næstu stöðvar, sem eru í um 700
metra hæð á þessu svæði samkvæmt munnl. upp-
lýsingum frá Kristínu Vogfjörð og Ragnari Slunga.
JÖKULL No. 60 53