- © 2014 by the Seismological Society of America
Online Material: Table of macroseismic intensity symbols and color codes.
The description of the seismicity of the European region is today fragmented into an increasing number of earthquake archives, databases, and catalogs related to individual countries or even to part of them. Therefore, the compilation of a comprehensive, European earthquake history requires dealing with a puzzle of partially overlapping, only partially public catalogs, the background of which is compiled according to varied schemes. One of the consequences is that earthquakes in the frontier areas are often interpreted in a conflicting way by the catalogs of the bordering countries.
In the framework of the European Commission (EC), 2006–2010 Network of Research Infrastructures for European Seismology (NERIES) Project, the task of Networking Activity 4 (NA4) was defined precisely to conceive and develop solutions to bridge over these differences. NA4 promoted the cooperation among existing national online archives, and contributed establishing new regional online archives compiled according to common standards. As a result, a first release of the distributed European archive of historical earthquake data, for the time‐window 1000–1899 and for the large earthquakes, was published in 2010.
Special attention was devoted to retrieve the earthquake background information, that is, the results of historical earthquake investigation—referenced to as studies in the following—in terms of a paper, a report, a book chapter, a map, etc. As the most useful studies are those supplying a set of macroseismic data points (MDPs)—that is a list of localities (name and coordinates) with a macroseismic intensity assessment and the related macroseismic scale—a dedicated effort was addressed to make such data available.
The Archive of Historical Earthquake Data (AHEAD) distributed archive was improved and updated in the frame of the 2010–2012 EC Project Seismic Hazard Harmonization in Europe (SHARE), Task 3.1 European earthquake database, with the contribution of a number of European institutions. For the time window 1000–1899, it was AHEAD (AHEAD Working Group) that supported the compilation of the SHARE European Earthquake Catalog (SHEEC; Stucchi et al., 2013). This paper describes the AHEAD portal (http://www.emidius.eu/AHEAD/; last accessed March 2014), and how it was conceived to network the local components of the distributed archive. Although local historical macroseismic databases usually supply one set of information for each earthquake, at a European scale an earthquake still might be described by several studies, available from different archives. The AHEAD portal inventories and gives access to multiple sets of information concerning each earthquake, and allows users to get comprehensive information about individual earthquakes, providing the answers to the following questions:
which sets of earthquake parameters (time, location, magnitude, magnitude type, maximum intensity, etc.) are available for each earthquake?
what is the background information, or supporting material, upon which each set of earthquake parameter determination is based?
ACCESSING THE DATA
The AHEAD portal considers earthquakes in the Euro‐Mediterranean area in the time window 1000–1899. To date, users may access 4722 earthquakes and the related background information coming from 337 historical earthquake studies. All these data can be queried by earthquake and by study, through a user‐friendly web interface. The distinguishing feature of AHEAD is to grant access not only to one study, but also to all the available (published) studies dealing with each individual earthquake, allowing the researchers to take into account the different point of view and interpretations. To introduce this peculiar approach, the 20 February 1743, Ionian Sea earthquake is presented as an example (see Earthquake Information section). The 1743 earthquake is a transfrontier event that affected a large area split among three bordering countries, at that time ruled by the Republic of Venice (part of Albania and the Ionian Islands, today Greece), the Ottoman Empire (part of Albania and mainland Greece), and the Kingdom of Two Sicilies (southern Italy). The studies inventoried by AHEAD make it possible to show that the earthquake was investigated according to alternative historical research strategies, in their turn leading to alternative earthquake scenarios.
Data about an earthquake can be retrieved entering the portal section called “query by earthquake.” Users can select earthquakes either from the list on the left or from the map on the right, which can be zoomed and panned. The search for an earthquake is facilitated by the grouping of the earthquakes into four predefined classes of magnitude and seven predefined time windows (Fig. 1). The location and size of each earthquake are taken from SHEEC 1000–1899 catalog (Stucchi et al., 2013). Users may customize the map background by adding a flat‐shaded or a hill‐shaded digital elevation model, change the aspect and size of symbols, or temporarily add reference points.
Once the earthquake is selected, the user may investigate the available information which is organized into three sections:
“Catalogues,” supplying the parameters extracted from the main catalogs;
“Studies,” showing the list of the main earthquake investigations;
“Seismicity 1000–1899,” showing the seismicity (1000–1899) of the area surrounding the selected earthquake epicenter.
Each section shows the textual content on the right and a map on the left (Fig. 2).
The section “Catalogues” shows the earthquake parameters as assessed by the most recent, published catalogs (Fig. 2a). Each catalog is identified by the short citation, and the main earthquake parameters such as the epicentral coordinates, the hypocentral depth, the magnitude and its type, the epicentral intensity, and the reference earthquake study are listed. Among all listed catalogs, relevance is given to SHEEC 1000–1899 because of its close relation to AHEAD. Each epicenter can be individually or simultaneously plotted on the map, in order to easily show the possible differences in the solutions proposed by the different catalogs. For the 1743 earthquake (Fig. 2a), a significant difference in the epicentral location results from the listed catalogs, whereas the assessed magnitudes are similar.
The background information on each earthquake is presented in the section “Studies”, in which the list of studies dealing with the selected earthquake is presented. For each study, the short citation, the date and area of the earthquake as reported in the study, and, when available, the number of MDPs and maximum observed intensity are presented. If the copyright status allows a free publication over the Internet, the complete text can be downloaded as a PDF file. A link to the website that originally provided the study is presented, if available. The MDP set from each study is shown and users can switch from one MDP set to the other. To allow a quick and easy comparison among different MDP sets, the AHEAD portal plots MDP distributions using a symbol set expressly devised for allowing an intuitive vision of the effect distribution. The list of color codes is available in the electronic supplement to this article. The case of the 1743 earthquake shows that two of the available MDP sets are different (Fig. 2c,d) as it concerns both the spatial distribution and the evaluation of intensities.
The section “Seismicity 1000–1899” shows a map with the seismicity (1000–1899) of the area surrounding the selected earthquake, as from SHEEC 1000–1899 (Fig. 2b); the seismicity is also presented in a diagram with magnitude versus time. The user can alter the considered circular area proposed by default or draw a polygon directly on the map.
As mentioned above, AHEAD information today comes from 337 studies which are listed in the “query by data source” portal section (Fig. 3). For each data source, the list gives access to its full reference and to the complete text as a PDF file. Online data sources (e.g., databases and catalogs) are linked when available. In particular, eight regional online archives, run by European research institutions dealing with historical earthquakes, contribute two thirds of AHEAD data: ASMI (ASMI Working Group), ECOS‐09 (Fäh et al., 2011), SisFrance (BRGM‐EDF‐IRSN/SisFrance, 2010), Macroseismic Data of Southern Balkan area (University of Thessaloniki, 2003), Base de Datos Macrosísmica (Instituto Geografíco Nacional, 2010), UK Historical Earthquake Database (British Geological Survey, 2010), Hellenic Macroseismic Database (Kouskouna and Sakkas, 2013) and Base de Dades Macrosísmica (Institut Geològic de Catalunya, 2010).
Besides regional online archives, the AHEAD portal inventoried other sources of data such as publicly available catalogs, studies, and reports. One third of the historical earthquake data available through AHEAD come from the latter sources, which had to be parsed manually and one‐by‐one in order to extract the relevant information.
From the IT point of view, implementing an archive of historical earthquake data involves an approach completely different from that used for instrumental data. The key distinctive points are:
historical earthquake data do not always come in a digital format, but in terms of printed texts or maps. In order to build a digital archive, historical studies have to be carefully analyzed in order to (1) check the reliability of the study, (2) extract the relevant data, and (3) put them in a format compatible with the database structure;
earthquake parameters such as date, epicentral coordinates, and magnitude may vary considerably from one study to another, particularly for the earliest events. Sometimes the very same earthquake is present in a catalog, but absent in another, sometimes the differences are clearly explained, sometimes the explanation is completely lacking;
macroseismic data are generally less in volume than instrumental data, making a big difference in the technologies involved when storing, manipulating, and searching through data, thus allowing a simpler IT infrastructure. Exchanging macroseismic data does not need the creation of complex data interfaces between data centers.
Consequently, the IT framework of AHEAD was focused on the human interaction and the design of more comfortable and intuitive graphical user interfaces, in order to support both the internal data curators’ activities and external users who browse the archive. The final IT framework results from a long and fruitful interaction between IT staff and data compilers, and it is structured using a relational database.
The database schema design is focused on the creation of a dynamic and efficient structure, carefully calibrated between the need of avoiding data redundancy and data management simplicity. The selected Relational Database Management System is MySQL, the reliable and well‐documented software developed by Oracle. This standard tool is fitted for dynamic database structures that will evolve over time. In addition, MySQL is released under an open‐source license, and this guarantee against future unresolved software bugs, and any dead‐end development path.
The database is made of ten tables (Fig. 4), in which the content is logically subdivided by type of data and in which each stored item is identified by means of a unique identifier or key. By grouping data in this way, data entry and data analysis can be performed easily, because of the small number of tables involved to query the database.
The main index table of the archive is called “Earthquakes” and contains the lists of earthquakes, each with a unique identifier and a minimum set of parameters (date, epicentral area, and magnitude). The data supporting each earthquake are stored in the DataSources table; the information extracted from the studies is split into as many records as the number of mentioned earthquakes that each study describes. As an earthquake can be investigated by multiple studies, more than one record representing each study can be related to an earthquake (one‐to‐many relationship). The information retrieved from a study includes the study unique identifier, the origin time, the name of the area of maximum effects, the epicentral coordinates, the magnitude, and related uncertainties whenever available. All bibliographical metadata describing the DataSources are stored in the References; it details the author, title, journal issue date, copyright restrictions, and a link to the online resources. PDFs and image covers are stored in the server file system. Users of the portal can notify a new study using the corresponding facility on the portal homepage; these new items are stored separately from the main archive in the table “NewMaterial.”
If a study contains an MDP set, its unique identifier, the maximum observed intensity, and the number of MDPs are stored in the table “MDPsets.” MDP sets extracted from historical studies are stored into the “MDPdata” table; such table contains the MDP set identifier and the list of MDPs (place names, coordinates, and assigned macroseismic intensity). A unique place identifier is assigned to each place mentioned in the MDP sets and a table with a Gazetteer is then compiled using the information retrieved. Information about the administrative subdivision of the European countries is stored in the table “AdminDivisions.”
The portal deals with some features or sets of data that require user authentication; the table “Users” contains the list of registered users and their access rights. Registered users can exchange private comments about an earthquake using an internal messaging system; these comments are stored in the table “Comments.”
The adoption of a client‐server data management approach makes it possible to access the data from several client front‐end tools at the same time, and perform simultaneous tasks, such as adding earthquake parameters, adding new MDP sets, updating the bibliographical information, or performing data‐coherency checks. The connection between the data curator’s front‐end tool and the data server is established using an Open Database Connectivity (ODBC) connection that transparently enables ODBC‐aware software to remotely manipulate the data. The adopted ODBC‐aware client is Microsoft Access, a widely used software with a friendly user interface. As data curators are not usually skilled IT professionals, only a basic knowledge of the rules of a relational database is required; it involves the concept of data redundancy, unique identifiers, the use of keys, the relations between tables, and some structured query language (SQL) keywords. All activities carried out by the data curators that involve a change in the database content are monitored by the database administrator (usually called DBA), in order to avoid faulty tables, and to keep the data consistent. From a technical point of view, the control activity is carried out sending SQL commands to the MySQL server using a remote text console that performs a series of data integrity checks. An incremental backup of the database tables is performed daily. It should be stressed that the role of the DBA is not strictly bounded to data management, but should involve an active support of the curators’ specific activities, such as optimizing their interaction with the stored data.
Website Technical Solutions
The solution adopted for visualizing maps is based on a custom‐developed mapping technique, designed to fully exploit the computational power of the client browser. Devices able to access the Internet are nowadays powerful enough to handle demanding tasks by themselves, and the standardization of web‐browsers functions offers a reliable chance to obtain acceptable results.
The web‐mapping solution described above was developed in house and was originally conceived for publishing the Database Macrosismico Italiano 2004 (DBMI04; Stucchi et al., 2007), the Database Macrosismico Italiano 2011 (DBMI11; Locati et al., 2011) and the Catalogo Parametrico dei Terremoti Italiani 2011 (CPTI11; Rovida et al., 2011). By adopting the same web‐mapping solution, AHEAD developed a tool called Macroseismic Intensity Data Publisher (MIDOP; Locati and Cassera, 2010), which is meant to facilitate the establishment of new regional macroseismic archives, and at the same time the promotion of common compilation procedures and data exchange standards. The tool is released under an open source license and it is freely available (http://www.emidius.eu/MIDOP/; last accessed March 2014) together with an exhaustive documentation.
DISCUSSION AND PERSPECTIVES
The AHEAD portal was tailored to allow the users to better exploit European, historical earthquake data spanning over a millennium; in particular it was conceived to bridge the gap between data producers (e.g., authors of historical earthquake studies) and data users (e.g., catalog compilers). The AHEAD portal helps users to easily retrieve and compare the largest amount of available, historical earthquake data and to better understand the origins of different interpretations. This goal is achieved by (1) tracing back, preserving, and granting access to the studies and the parameters for each earthquake and (2) creating relations among these studies.
Special attention was devoted to retrieve MDPs, because they provide both a better image of an earthquake than epicenters, and the basis for the application of modern, repeatable procedures for determining earthquake parameters (e.g., Bakun and Wentworth, 1997; Gasperini et al., 1999, 2010). With respect to this issue, AHEAD today offers at least one study providing MDPs for 2710 earthquakes out of 4722, that is the 57% of the total. The availability of MDPs increases for the most energetic events, that is 74% of the most damaging earthquakes (Mw≥5.6).
The picture provided by the geographical distribution of the earthquakes with or without MDPs (Fig. 5) highlights the different level of knowledge of the European historical earthquakes. Figure 5 tells us that the largest amount of MDPs comes from most of the western European countries, which could count on past, dedicated investments. The situation is not that good in the Eastern countries, in spite of a pioneer project initiative in the Balkans (Shebalin, 1974).
The user interface is designed to emphasize the distinctive features of the AHEAD portal and to be user‐friendly and straightforward. The underlying IT framework is tailored on the peculiarities of the historical earthquake data and fulfills both the data curators’ and the users’ needs. The portal website and its database schema are conceived to be sustainable in the long term. As research progresses, the portal is ready to evolve accordingly, incorporating new data, supporting an extension in both time and space.
After creating and establishing AHEAD with the support of two European initiatives (2006–2013), two issues are currently under consideration: the way to make the AHEAD portal durable and how to continue feeding the distributed archive, with main reference to the time window after 1900. Both issues can be addressed by the increased awareness and cooperative action of the European and national research institutions, now that a consistent description of the Euro‐Mediterranean long‐term earthquake history does exist.
The authors wish to acknowledge the contribution of Daniele Viganò and Andrea Cassera. The Archive of Historical Earthquake Data (AHEAD) portal was partially supported by two European Commission initiatives, the FP6 project Network of Research Infrastructures for European Seismology (NERIES) (Contract Number 026130) and the FP7 project SHARE (Seismic Hazard Harmonization in Europe) (Grant Agreement number 226769). We thank an anonymous reviewer and the associate editor, whose constructive feedback improved the manuscript.