Down Previous Home Next CD-ROM

Data Selection and Processing

In order to be able to analyse and use effectively the data selected for dissemination on this CD-ROM it was necessary to process the strong-motion records and to determine a number of parameters relating to the earthquake and the recording site using consistent procedures. Another important task was to present, for dissemination, the records and the associated parameters in a uniform format.

Strong-motion records and their associated parameters have been acquired from networks, agencies and data centres operated by a multitude of private, academic or governmental establishments. The records have been disseminated in different formats and the associated parameters were, quite often, incomplete, inaccurate or wrong. On the other hand the dataset supplied for engineering applications has to fulfil the following fundamental conditions:
the strong-motion records should be accurate and reflect the "ground response".
the strong-motion records should be processed with consistent procedures.
the associated parameters should characterise the earthquake source, the source-to-site path and the site of the recording station.
the associated parameters should have a real physical significance and suitable for engineering applications.
the values of the associated parameters should be reliable and easy to determine.
it should be possible to determine the associated parameters consistently using primary sources, such us earthquake catalogues, bulletins, and systematic calculations.
strong-motion records and their associated parameters must be presented in an uniform and easy to use data-format.
the organisation of the data should allow a straightforward extraction of strong-motion records and associated parameters using any selection criteria.

The strong-motion records provided for publication on the CD-ROM have been selected according to the following criteria:
the uncorrected strong-motion record is of good quality.
the associated parameters are as complete as possible.
the ground-motion represented by the strong-motion record is of engineering importance.
the strong-motion record was produced at 'ground response' stations.
the final dataset should 'represent' all tectonic region of Europe and adjacent areas.
the contributing organisation has given its consent to release the record and associated parameters.

Finally 1,068 strong-motion records from a number of sources have been selected, processed and included to the CD-ROM databank. A complete list of all selected strong-motion records is presented in the following tables. The records are identified by the origin time of the earthquake and the location of the recording station. All selected strong-motion records are properly associated with the earthquake and recording station to which they have been attributed in the original sources:
Records grouped by earthquakes
Records grouped by stations

The owners of the strong-motion records and the source of data are indicated either in the databank and database and in all datafiles archived on the CD-ROM.

The geographical distribution of these records is shown in the following figure, where the surface-projection of the path of the direct wave, which has triggered the instrument, are drawn as solid red lines.

Distribution 1

The distribution of the data among the Flinn Engdahl regions and countries is given in the following table:
Flinn Engdahl Region Records
Central Italy 285
Northern Italy 145
Southern Italy 123
Turkey 92
Greece 60
France 42
Western Caucasus 41
Turkey-Georgia-Armenia border region 37
Southern Greece 33
Northern Algeria 27
Western Iran 24
Switzerland 23
Northwestern Balkan Peninsula 21
Sicily, Italy 17
Albania 11
Adriatic Sea 11
Romania 11
Austria 10
Germany 9
Northern and central Iran 9
Near south coast of France 7
Crete, Greece 6
Greece-Albania border region 5
Aegean Sea 4
Strait of Gibraltar 4
Dodecanese Islands, Greece 3
Ionian Sea 3
Azores Islands, Portugal 1
Northwestern Iran 1
Northwestern Uzbekistan 1
Turkey-Iran border region 1
United Kingdom 1
Country Records
Italy 542
Greece 109
Turkey 97
France 54
Armenia 41
Yugoslavia 39
Georgia 37
Iran 35
Switzerland 32
Algeria 27
Slovenia 12
Romania 11
Austria 10
Germany 10
Spain 4
Croatia 2
Bosnia-Herzegovina 1
Liechtenstein 1
Macedonia 1
Portugal 1
United Kingdom 1
Uzbekistan 1
The present dataset contains digitised and digital strong-motion records that we were able to acquire and which were recorded at permanent, ground level recording stations located in Europe, the Mediterranean area and Middle East. Records from temporary stations were also included, if they are of engineering importance. Most of the data have been recorded with instruments described as 'free-field', others have been recorded with instruments located in the basement or on the ground floor of buildings or structures, which should then be considered as "ground response" recordings.

Only uncorrected records have been acquired and selected for the CD-ROM dataset. Many accelerograms were already digitised from film or paper records. They all were adjusted to have zero origin time, scaled in time and acceleration and a straight baseline has been fitted to the records using a least square fit. They often have been supplied in the form of unequally spaced time histories. More recently, time histories have been digitally recorded at evenly spaced intervals.

All uncorrected records have been pre-processed to account for any irregularities in the time-histories, such as spurious large amplitude or very high frequency acceleration points, non-zero sloping or translated baselines, poor quality digitisation, zero or negative time steps or duplicate records. Uncorrected records have been re-formatted using a standard format for uncorrected data, plotted and visually inspected. The digitisation quality of accelerograms has been checked and where it has been judged to be sufficiently poor, records were excluded from further consideration. Other records displaying recoverable anomalies have been treated individually in order to remove errors at source wherever possible.

Following completion of pre-processing, uncorrected records have been subjected to a uniform standard correction procedure to reduce the low and high frequency noise levels. The records from the Umbria-Marche earthquake sequence of 1997 and 1998 were processed with the method proposed by Rinaldis et al., 1996. Unfortunately reliable values of the instrument transducers natural frequencies and damping levels were not available for a large proportion of the selected records. Therefore de-coupling of the instrument response was not performed for any of the selected records. However, the majority of the data have been recorded with SMA-1 instruments for which the transducers typically have natural frequencies around 25Hz and damping of about 65%. Significant distortion of the ground motions only occurs for frequencies above this level, which is generally outside the range of engineering interest. The acceleration and velocity time histories have been re-formatted using the standard format for corrected records.

Elastic response spectra have been calculated by the method of Nigam & Jennings (1969) and presented in a standard format for spectral data. Spectra have been determined from the corrected records only. The period range is between 0.04s and 4.0s. Spectra have been calculated at 5 damping levels of 0, 2, 5, 10 and 20%. Absolute acceleration, SA, relative-velocity, SV, relative-displacement, SD, and pseudo-velocity spectra, PSV, are presented.

All time histories and spectra have been transferred into a uniform file format, with one file for each component of the recording station. Unfortunately there is no standard data format or database system for ground motion data and associated parameters. There are nearly as many data formats for strong-motion data as there are data sources (e.g. GSE, SAC, SMC, V1, V2, V3 (Trifunac & Lee, 1973), SEEDS, numerous ASCII). It has been finally decided to develop a database system and to support a data-format, which is tailored to the specific requirement mentioned above and the data and associated parameters selected for dissemination on this CD-ROM.

Up Previous Home Next