Goal
The goal of Task 1 is to compile a list of earthquakes to serve as input for the next tasks.
 
 

Procedure
Due to the puzzling situation of the current parametric earthquake catalogues (PEC), the safest way to accomplish Task 1 would be to start again from the seismological literature (earthquake studies and compilations). As this way would require an amount of time and expertise exceeding the constraints of the project, the task will be accomplished by retrieving, evaluating and recompiling the current PEC in a single, qualified, comprehensive datafile.
As this datafile will mainly serve to inventory earthquakes, the evaluation of current values of earthquake parameters, such as Lat, Lon, Io, M, H and so on, is not among the scopes of Task 1.

Definition
The BEECD Working File (WF) is a file where:

• several input PEC are compiled together in a single file where each entry clearly shows its origin, that is, the file where it comes from;

the parameters of each input PEC are converted into those adopted for the comprehensive format;
entries considered by the compiler as multiple determinations of the same event are flagged;
in case of multiple determinations, one "preference" entry is flagged.
• Time-space-windows and size threshold

The main BEECD time-space-window will be 1400-1899, West of 31° lon E; however, the WF will be compiled also for the time-windows before 1400 and after 1900. The WF will incorporate entries with Io or Imax >= 4/5, or M (at least one type) >= 3.5. For the time-windows before 1900 entries with Io and Ix and M = 0 will also be incorporated.
In the time-window after 1900 problems are complicated by the large number of entries and by the presence of two sets of data: historical/macroseismic and instrumental. Parameters derived from the two sets are currently mixed together in a single string, with some exceptions which report instrumental data only (Makropoulos and Burton, 1981; Legros and Bonnin, 1996; ISC, etc.). No PEC reports distinct locations from macroseismic and instrumental data; in many cases it is hard to assess the origin of the available parameters.
Despite the current issue, also adopted by GSHAP, that "instrumental" is always better than "macroseismic", in Europe we deal with a lot of low quality instrumental data before 1964, and even later. Moreover, tests carried out by Rebez and Slejko (1997) showed that some files of instrumental determinations cover only a portion of the earthquakes which actually occurred. Therefore, it was concluded that the time-window 1900-1963 (at least) should be dealt with, in the future, with the same criteria of the period before 1900.

• PEC with references are the most useful ones;
• published versions are in principle preferable to late, unpublished versions belonging to the same family.

Following the adopted criteria, PEC listed in Tab. 1.1 were selected as input PEC for BEECD WF. Three input PEC (Grigorova et al., 1978; Herak, 1995; Labak and Broucek, 1995) do not meet completely the established requirements, as they are not published; however, they were adopted because they carry references. A few more PEC which did not fulfil the assumptions made above (not published and not carrying references: Lenhardt, 1996; Verbeiren et al., 1994; the Bulgarian, 1993, datafile), were adopted as side "sources of additional information".
Tab. 1.2 summarises the number of entries of each PEC in varied time-windows, and the contribution of each PEC to the WF (three time-windows).
 
 

BEECD parameters
Parameters and format of input PEC are varied; there is no international standard, and this is not surprising, as they represent the final step of processing performed by each compiler according to his own needs and choice; however, this variety may cause problems when compiling a comprehensive PEC.
Actually, many parameters are found in most PEC, while some others are found in a few only. A comprehensive set of parameters would require a very extended format, but many columns would result scarcely populated. On the other hand, to adopt the criteria of selecting only the most widespread parameters would lead to very poor sets. This is the case, for instance, of the PEC by Van Gils and Leydecker (1991), where references are lacking and magnitudes of varied types are packed together in the same column.
Hints for BEECD format came from the official GSHAP format (Basham and Giardini, 1993) and from the selected PEC. The final set of parameters (Tab. 1.3)adopted for BEECD contains:

• a selection of the most widespread, current parameters in the selected PEC (C)
• some parameters on the purpose for BEECD (B)
• some parameters proposed by GSHAP (G).

• management parameters (1 to 6)
• parameters describing the supporting data set (7 to 20)
• earthquake parameters (21 to 44).

WF 1400-1899 compilation
The considerations and the figures which follow refer to the main time-window.
The re-compilation of input PEC in terms of BEECD parameters is in principle similar to all translations, which usually present problems and pitfalls, and require a lot of care; therefore this step is sometimes responsible for large inhomogeneities present in many comprehensive PEC.
Tab. 1.4 summarises the relationships among BEECD and input PEC parameters. It also gives an idea of how inhomogeneous input PEC are with respect to some parameters. For some parameters the translation was easy; for others, such as those evidenced with @, it required some decisions.
Tab. 1.5 summarises how parameters of type C were translated and how parameters of type B were determined. Parameters of type G were not compiled at this stage.
 
 

Multiple determinations assessment
More entries from varied PEC may refer to the same event; they will be called multiple determinations of the event. They can provide a set of coincident or conflicting parameters, as the use of different data sets or determination procedures may lead to gaps of years in time and tens of km in location.
A careful assessment of multiple determinations would require a complete study in many cases. Therefore, in the WF 1400-1899 they were assessed by expert judgement, comparing earthquake date, time and other parameters; obviously, the adoption of other input PEC or new expertise may change the assessment. Each multiple determination of the same event was given the code MD (parameter Dt) and grouped using the same value of parameter Fn (family number). In all, MD entries represent about the 20% of the total number of entries; they belong to families which consist of a number of members ranging from 2 (most cases) to 6 (one case).
A special case of multiple determinations is represented by duplications. Actually, the selected PEC are not completely independent one from another, and many PEC incorporate "clone" entries from other PEC covering the same area and/or the neighbouring countries. Such clones are usually converted into the format of the other PEC, sometimes undergoing slight changes. These entries have been defined "duplications", given the code DD instead of MD and the corresponding family number (Fn). The WF 1400-1899 contains 553 duplications (4%).
Tab. 1.6 presents also the distribution of single, multiple and duplicated determinations in the input PEC (absolute and percent figures).
The sorting of multiple determinations and the flagging of "preference" entries will be performed and described at the end of Task 2.
 
 

Plots
The global plot of the input PEC (1400-1963) by input PEC is presented in Fig. 1.1; single input PEC are presented in Fig. 1.2 a, b, c. Similarly, the plot of the WF 1400-1899 by input PEC is presented globally in Fig. 1.3 and per time-windows in Fig. 1.4; single input PEC 1400-1899 are presented in Fig. 1.5 a, Fig. 1.5 b, Fig. 1.5 c. The most energetic earthquakes (Io/Ix _ 7/8 or M _ 5.2) of the WF 1400-1899 are presented in Fig. 1.6.
Fig. 1.7 gives the distribution of multiple and duplicated determinations.
Finally, Fig. 1.8 shows the plot of the WF 1400-1899 by size (class of Io/Ix/M); separate time-windows are presented in Fig. 1.9.