A refined and easily accessible database for SIL data is under construction. Since 1991, 100.000 earthquakes have been recorded by the SIL system. The data were automatically evaluated and manually corrected. Facilities have been developed to store all the data on-line on hard disks. Because of the enormous amount of data they have to be compressed very much, to make this practically possible. Seismogram data in digital form is stored using packed binary format, where only the number of bits that is required to store sample to sample variation is stored. Parameter data of events are stored in relational database tables, which contain all parameters of the events, both observed parameters like onset time of phases and amplitudes as well as calculated parameters such as hypocenter location, magnitude and fault plane solution.
Other data is stored in relational database tables. Station parameters such as coordinates, instrument characteristics and time corrections are stored in separate tables. This information is incorporated into headers when data are extracted from the database [51].
In order to insure against loss of data, procedures and facilities have been developed to back up all data onto magnetic tapes. All new data and modifications are written to tape each day and all data are written to tape approximately every two weeks. Periodically, a set of tapes is moved for storage to a different site. As magnetic tapes only last a few years, and the long-term stabilty of optical storage media is not well established, this is possibly the most effective way to permanently preserve the data, and it has the advantage that the data is always readily accessible.
An interface to the database through the World Wide Web is in operation. Currently anyone with access to the Internet can search through a list of over 65.000 earthquakes that have been manually checked since January 1, 1995.
Data from the SIL database have been provided to the various projects of PRENLAB as requested. It has not yet been possible to provide data in standardized form covering the whole of the SIL period, i.e. since 1991. This is, however, very significant for pattern recognition research. It is aimed at that this will be possible shortly and for this purpose reevaluation of old automatic data is being speeded up as much as possible.
Work has been carried out for a new, reevaluated and refined catalogue of earthquakes in Iceland since 1926. The catalogue from 1926-1963 has been reevaluated and put on digital form. The refinement of the more recent catalogues is in progress [].
Work has been carried out for refined estimation of magnitudes and locations of historical earthquakes as well as felt events, not instrumentally detected since 1926 [36].
A long-term overview (since 1979) of 7 volumetric strainmeters in the SISZ has been worked out. Methods have been developed for correcting the strainmeter record for weather influences [5]. As a result of this work change in strainrate 5-6 months before the start of the 1996 eruption in Vatnajökull was discovered, which has been proposed to be caused by magma intrusion there, i.e. more than 150 km from the strainmeter stations [].
Basic evaluation was carried out on the seismic activity related to the volcanic eruption in Vatnajökull, especially as concerns hypocenters and mechanism of the earthquakes. Much effort was put in saving data on this remarkable eruption from the seismic networks, both earthquake data as well as data on volcanic tremor. Vatnajökull is directly above the center of the Iceland mantle plume and changes of the plume activity greatly affect the seismicity along all of the plate boundary in Iceland, and is thus of a great significance for the PRENLAB objectives []. Although the SIL system is a seismic data acquisition system, that is primarily designed for automatic acquisition and evaluation of data from local microearthquakes, it can also be used for collecting teleseismic and regional data for deep structure studies. It broadens the scientific use of the network and has made it easier to obtain funds for extending the network to a large part of the plate boundary in Iceland. The SIL station software has now been modified allowing for selection of waveform data at 20 and 4 samples per second in addition to saving data at 100 samples per second, which is routinely done. This makes it economically possible to save long time intervals of seismological data from the SIL stations. An automatic procedure has been developed to select and store teleseismic data in the SIL system, based on USGS/NEIC information on teleseismic events in the whole world, which are measurable in Iceland. From USGS/NEIC we receive E-mail messages with a single-line information on earthquakes they have determined, the so-called ,,E`` type messages. A selection program reads the messages and selects events that fulfill certain criteria of magnitude and epicentral distance. The program uses the IASPEI91 model to compute the first arrival time at each station. The teleseismic body wave data are fetched with a sampling rate of 20 samples (in some cases 100 samples) per second and the surface wave data with sampling rate of 4 samples per second from the 1-3 days long ringbuffer of the SIL site stations. Since mid-year 1996 waveform data from 230 teleseismic events have been stored by this automatic procedure [].
A real-time filter has been introduced into the on-line process of the SIL system, to be tuned for detecting harmonic tremor and signals which now are not identified automatically. The continuous seismic signal at the SIL site stations is bandpass filtered at 0.5-1 Hz, 1-2 Hz and 2-4 Hz and the 1 minute mean amplitude is scanned and sent to the SIL center. Visual presentation of this data gives a useful indication of the multiplicity of activity in real-time. This data has been calibrated and procedures developed to estimate magnitudes of local events larger than magnitude 2, independent of the waveform processing [].
The extension of the SIL system into the highlands of Iceland has lead to many problems in the automatic detection and analysis which are gradually being solved. The SIL system was developed for use in the seismic zones. Automatic monitoring in the highlands revealed in many ways new problems because the crustal structure is not as well known and the earthquake sources are often more complicated. Much work has been carried out to lower the detection threshold for earthquakes in Central Iceland to be able to acquire more information from microearthquakes in this area. The new real-time filter mentioned above is significant for this purpose as well as further tuning of all detection parameters [41].
Work has started on a method to use cross-correlation of waveforms to accurately and automatically determine onsets and classify earthquakes in cooperation with UUPP.DGEO. There has been close cooperation with Uppsala in various other fields, such as stress tensor inversion procedures and mapping of faults.
Work has been carried out to overview the state of knowledge and modelling in the areas of catastrophic earthquakes in Iceland, the SISZ, the Reykjanes Peninsula and the north coast of Iceland [,,].