Jökull - 01.01.2014, Blaðsíða 127

Monitoring of volcanic eruptions using satellite images
Grímsvötn 2004
An explosive eruption began in Grímsvötn, Vatna-
jökull, on November 1st 2004, following an earth-
quake swarm under the caldera, associated land rise,
and then jökulhlaup in glacier river Skeiðará on Oc-
tober 30th (IMO, 2004). The eruption only lasted
five days. During that time the eruption plume was
visible in real time on NOAA AVHRR and MODIS
images, though affected by clouds. MODIS image
from November 2nd indicated ash fall on the glacier,
though it was not very clear due to the eruption plume.
It was possible to measure the height of the plume
on IMO’s weather radar (IMO, 2004), up to 13 km
(IMO, 2004; Guðmundsson and Sigurðsson, 2005).
One of the better images was acquired after the erup-
tion, on November 7th (Figure 2), when the tephra on
the glacier became apparent, in two cones heading N
and NNE, providing volcanologists with an overview
over the tephra fall for estimating its extent and select-
ing sampling sites (Oddsson et al., 2012).
Fimmvörðuháls 2010
An effusive eruption began on Fimmvörðuháls, S-
Iceland, on March 20th 2010. Having had clear
precursors (Sigmundsson et al., 2010) satellite im-
ages were being monitored when the eruption started,
MODIS and NOAA AHVRR in particular. A strong
thermal anomaly was observed in the early hours of
March 21st. The Icelandic Coast Guard (ICG) had
flown over the site on March 19th, looking for any
signs of increased geothermal or volcanic activity. Al-
though no such signs were observed, it was an im-
portant test for the coming weeks, when the new air-
craft, fitted with thermal cameras and radars, played
an important role in monitoring the eruption in Fim-
mvörðuháls and Eyjafjallajökull. Until then, the air-
craft had been used for sea-ice monitoring and re-
search, but the potential for other applications was
clear (Jónsdóttir and Friðriksson, 2009). The lava was
mapped on a regular basis using the ICG side scan
radar (Edwards et al., 2012) and relevant information
made available to the scientific community and the
public.
A team at NASA scheduled EO-1 ALI and HY-
PERION images to be acquired at the time, and they
were used to estimate effusion rate (Davies et al.,
2013). This was part of an effort to have thermal
anomalies on the MODIS images trigger EO-1 acqui-
sition (Davies et al., 2013).
Eyjafjallajökull 2010
An explosive eruption began in Eyjafjallajökull on
April 14th 2010, causing flooding and lahars. The
eruption lasted 39 days (Guðmundsson et al., 2012).
The eruption plume affected people in Iceland and
Europe, though for different reasons. Various satel-
lite images were merged with map data and published
in real time to show the distribution of the eruption
plume in relation to farms, roads and towns in Iceland
as well as in Europe (Jónsdóttir, 2010). This infor-
mation was sent to the relevant authorities and rescue
teams operating at the site, and published on the IES
web site (Figure 3).
Grímsvötn 2011
An explosive eruption began in Grímsvötn on May
21st, 2011, and lasted 8 days (IMO, 2011). The erup-
tion was observed by weather satellites like the previ-
ous eruptions. A NOAA AVHRR image of the begin-
ning of the eruption suggested that the height of the
plume could be estimated, in this case between 10.3
and 14.3 km, probably closer to the higher number
(Figure 2). The height could be calculated with more
accuracy if the top of the eruption plume could be po-
sitioned better, for example with a photograph from
the ground or webcameras.
Bárðarbunga/Nornahraun 2014
An effusive eruption began within the Holuhraun lava
field on August 31st, following an earthquake swarm
from Bárðarbunga (Sigmundsson et al., 2014; Gettel-
man et al., 2015) and a small, short lived eruption at
the same site on August 29th. By the end of 2014, the
new lava extended over 83 km2. All available satel-
lite images were used to monitor the eruption. Images
from weather satellites, NOAA AVHRR and MODIS
(Terra/Aqua) indicated the eruption site within 1 km
in near real time, and illustrated active regions within
the lava field every few hours during the next few
months, when clouds were not obstructing the view
of the site. LANDSAT-8 images were obtained ev-
ery week, and when conditions were favourable, these
JÖKULL No. 64, 2014 127