A DRASTIC REVISION OF ACTIVE FAULT DATABASE IN JAPAN BASED ON THE REDEFINED RELATIONAL DATA MODEL

Outline of the Database

Active Fault Database of Japan (http://www.aist.go.jp/RIODB/activefault/cgi-bin/index_e.cgi) contains comprehensive information about active faults in Japan, sorted by the concept of "behavioral segments" [McCalpin, 1996]. Each fault is subdivided into behavioral segments based on surface trace geometry and rupture history determined through paleoseismic studies. Faults shown on the index map are linked to a database of behavioral segments, which contains information about geologic and paleoseismic parameters including slip rate, slip per event, recurrence interval, and calculated rupture probability in the future. Behavioral segments can be searched also by name or combination of fault parameters. All those data are compiled from journal articles, theses, and other documents.

Problems on the Data Input Method

The database was first formulated in 2002 by the Active Fault Research Center, Geological Survey of Japan, National Institute of Advanced Industrial Science and Technology (AIST). The first edition of the database was launched online in March 2005, as part of the Research Information Database (RIO-DB), which is managed by AIST. Through the relational database management system (RDBMS) of ORACLE 9i, many users could read easily many information like the kind of a catalog or a handbook using a web browser. However, searching information using composite key words was not possible. Furthermore, the order of the data could not be changed using composite attributes. These shortcomings were caused by the data input method. The data were inputted using the MS-Excel spread-sheet software. Attempts were made to solve the problem. The inputted data were reconstructed into a pseudo data model, defined for a serialized data on 2-dimentional spread-sheet schema. However, the pseudo data model could not solve the problem of data redundancy. Furthermore, standardization of the data specifications is also not possible. Because of the high level of data redundancy and the low level of standardization, flexible searching functions in the first edition of the database was not possible.

Redefined Data Model of the Database

In general, all data of relational database must be inputted by RDBMS on a well planned and designed data model. In this study, the data model was redefined to formulate a genuine relational database. The redefinition of the data model is based on the general contexture of published journal articles describing active faults. Therefore, the titles of the constitutive entities/tables are Survey, Locality, Horizon, Displacement, Event, Discussion, Figure and Reference. Sixty five tables are defined. Many attributes and their data types, domains and other specifications are also defined. Many more tables are constructed for appending legend code numbers of these attributes. Another attributes describing meta-data are also defined. All these tables are normalized and connected by relationships. Top-level data model i.e. Entity Relationship Diagram in IDEF1X is described in Figure 1. The completed data model, which contain detailed definitions, will be presented in a poster at the conference.

Revision of the Database

Based on the redefined data model, a data input user interface system on an MS-Access RDBMS was formulated. The new data were inputted stepwise by the input system of the MS-Access and finally merged into the main RDBMS of ORACLE 10g. In August 2006, the database containing large amount of data was revised base on the redefined data model. The flexible searching functions, which allow searching using composite key words, other advanced searching methods, and changing the order of data using composite attribute are successfully implemented using the revised database. Coincidentally, new field data collected from over ten thousands number of research localities, are added into the new Japanese edition database. The database also shows meta data, bibliographies, figures, chronostratigraphies and more detailed information. The new Japanese edition is presently translated into English. The new edition is developed to include analytical and GIS functions and more.

References Cited

McCalpin, J. P. 1996, Application of paleoseismic data to seismic hazard assessment and neotectonic research. McCalpin ed. "Paleoseismology", Academic Press, p.439-493.