Task 1.1: Data collection

Much more work was carried out in data collection and data evaluation than anticipated when the workprogramme was prepared.

This was partly due to general extensions of the applied monitoring systems, but partly due to very high seismic, volcanic and deformation activity at various places in southern Iceland.

The high seismic and deformation activity that started in 1994 in the Hengill-Ölfus area in SW-Iceland, culminated in June and November 1998, with magnitude 5 earthquakes (Figure 1).

  

Figure 2: Yellow and red circles show well located earthquakes (horizontal error less than 1 km and vertical error less than 2 km) from November 13-15, 1998, in Ölfus, SW-Iceland. The largest earthquake had magnitude 5.0. Red and white spheres present fault plane solutions in a conventional manner. Active faults found by accurate relative location of groups of microearthquakes are shown as blue line segments. White circles show locations of a few hundred earthquakes of the preceding earthquake swarm in Hengill in June 1998, maximum magnitude 5.1. The course of events was such that seismic activity had mostly been concentrated in the Hengill area since 1994, but ripped a fault to the south on June 4, 1998, followed by high seismic activity in an E-W elongated zone in Ölfus on November 13-15, 1998.

Enormously significant data were collected in this area which contain earthquake premonitory activity and short-term precursors to earthquakes. The data collection and evaluations carried out in relation to this activity have been of basic significance in understanding crustal processes leading to earthquakes, and for creating algorithms for short-term warnings, for modelling motions involved in earthquakes and for understanding large-scale stress modifications that were caused by the two earthquakes. Most of the participants of the project have been making use of these data (Rögnvaldsson et al. 1998c). Other significant earth activity which have created significant data are the eruptions in Grímsvötn in the Vatnajökull glacier in December 1998, and the eruption in volcano Hekla in February 2000 (Stefánsson et al. 2000a; Ágústsson et al. 2000). Since July 1999 there has been a volcanic crisis in volcano Katla and in nearby volcanoes (Stefánsson et al. 2000b; Geirsson et al. 2000), based on observed activity, and on the threat which these volcanoes are for the inhabitants of the area and travellers on the nearby roads.

The two large earthquakes in the South Iceland seismic zone (SISZ) in June 2000, shortly before the end of this project and the data collected from these are of enormous significance for earthquake prediction research in general (Figure 4). Earthquakes in the SISZ, which can reach magnitude 7 have been a threat for the inhabitants in this area. Understanding large earthquakes in this area has been a basic objective for the PRENLAB projects (Stefánsson et al. 2000c). The seismic data and GPS data of the hazards shortly described here above have all been used in research actions as well as of other actions. Also significant warnings and information about these events have been provided, so significant experience has also been gained in predictions and early warnings to the public and to authorities. The extension of the SIL acquisition and evaluation system, the SIL system, has continued during the year. The number of operating SIL stations was increased from 33 to 41.

Quite often during high earthquake activity the incoming data of small earthquakes is so high in the SIL system that the communication system and the computers have problems to cope with the data stream, and jams were created, which sometimes could delay the data, so the system evaluation was delayed. This could even lead to loss of data. As it is very significant to gather earthquake data down to the smallest earthquakes that provide information about crustal conditions, it was necessary to design and implement more effective procedures for doing this. For this purpose a new compression algorithm was developed for the system, i.e. the bit compression. This algorithm compresses the data very effectively at the site stations and the compressed data go directly into the evaluation procedures at the SIL center, much faster than the earlier procedures. A new format for saving the digital earthquake waveform data will be described shortly in following:

The output of the seismometer digitizer is a series of integer values. The sample-to-sample variation is usually much less than the maximum values, which for most of SIL stations are between +/-3276800. In the AH format which was used by the SIL software, each value is stored in 32 bits.

A reduction in size of the data files of approximately a factor of 5 is achieved by storing the sample-to-sample variation in packed, variable size integers.

The access to data is thus much faster than to data that is compressed using general purpose compression programs such as gzip or compress and the files are typically 2-3 times smaller.

This new bit-compress format (bc) (Kjartansson 1996) was incorporated into the data acquisition in the SIL system during the autumn of 1998. The software on each station writes in ascii files in format that is called the SIL format. The program bc-tool can convert these files to the bc-format and back. All information in the headers are preserved.

The bc-files are then transferred to the SIL center (currently using uucp). All files from each day are kept together, with a directory for each station. An index file that contains a list of all waveforms for each day is maintained.

The index files are stored on binary form, and are sorted by the programs that read them. A major performance bottleneck in previous version resulted from sorting index files on ascii form, each time that waveform data arrived.

The new software is able to keep up with much larger levels of earthquake activity than previous software. Because the routines that read and uncompress the data are very fast and files are small, performance of all programs that use the data has been improved.

There are now 7 continuous GPS stations in S-Iceland. These stations collected valuable data during the June 2000 earthquake sequence. Although the installation cost of these stations is paid by Icelandic authorities, PRENLAB-2 has contributed significantly to the build-up and development of these measurements. The continuous monitoring of these stations has provided data which are significant for the objectives of the PRENLAB-2 project for data collection. Four of the continuous GPS stations were installed in the Hengill-Ölfus area, two south of Mýrdalsjökull and one south of Eyjafjallajökull volcanic area, as shown in Figure 3.

The project of building the continuous GPS measurements is a collaboration between IMOR.DG, NVI, and UICE.SI, with significant support for development work from PRENLAB-2. The funding for purchasing the equipment comes from the Icelandic government and the Reykjavík Municipal District Heating Service. We use Trimble 4700 CORS and Trimble 4000 SSI dual frequency receivers, and Trimble Choke Ring antennas, to ensure the best data quality.

The data are automatically downloaded once every 24 hours to IMOR.DG via phone lines. Data from the IGS stations Reykjavík (REYK) and Höfn (HOFN) are included in our analysis. The data are automatically processed at IMOR.DG using the Bernese v4.2 software, and Center for Orbit Determination in Europe (CODE) predicted orbits. The displacements relative to REYK are calculated and the results posted automatically on the IMOR.DG website. The URL is http://www.vedur.is/ja/gps.html. A description of the network, data processing and results from the first year of observations are described by Árnadóttir et al. (2000).

  

Figure 3: Location of the continuous GPS stations in S-Iceland (red squares). The continuous GPS station in Reykjavík (REYK) is shown with a blue square. Thin black lines denote mapped faults (Einarsson and Eiríksson 1982; Einarsson and Sæmundsson 1987; Erlendsson and Einarsson 1996). The yellow areas are volcanic fissure swarms, and the calderas are shown with black lines with tick marks.

Stations for continuous monitoring of conductivity at depth are presently operated at three sites in Iceland. A station was installed at Skrokkalda in the central highland of Iceland in July 1999, after being operated for a year at a site at the eastern end of the SISZ. This station is linked to the SIL seismic system and real-time observations are made at the SIL center at IMOR.DG. A second station is operated at Húsafell in Borgarfjörður, W-Iceland, since spring 2000, and a station started operation in Tjarnarland in Eyjafjörður, N-Iceland at the beginning 2000. This work has been carried out through cooperation between Axel Björnsson at the University of Akureyri, and IMOR.DG. The two last continuous MT stations operate off-line.