Vol.2 No.3 2009

Research paper : Evaluation of earthquake occurrence from active faults (T. Yoshioka)−179−Synthesiology - English edition Vol.2 No.3 (2009) inconsistencies. This model was employed for the active faults in California, U.S.A. as an earthquake model for a very long active fault running several hundred kilometers[4]. In this model, a continuous active fault is divided into several behavioral segments that have their own unique amount of slip and rupture intervals[5]. These segments rupture while maintaining their unique cycles and sometimes in conjunction with the neighboring segments.The greatest characteristic of the Cascade Earthquake Model is that the length of the segment can be fixed regardless of the range value of fault rupture, by considering the large-scale fault rupture as a co-movement of the behavioral segments. By doing so, the earthquake time-varying position at a certain point can be kept constant while maintaining the proportional relationships of the fault length, the amount of displacement by earthquake, and the earthquake size. Therefore, the cyclic nature of fault rupture at a certain point can be explained extremely easily. We decided to employ the Cascade Earthquake Model as the most realistic model at this point, to achieve predictions utilizing the geomorphological and geological research results.The schematic diagram of the Cascade Earthquake Model is shown in Fig. 3. Conventionally, active faults were given the name “X Fault” from their geographical distribution without any particular criterion. However, since the range that was given a name may not necessarily cause the next earthquake, it is impossible to estimate the earthquake scale from the length of that fault. Therefore, these active faults were categorized into behavioral segments (here four categories from A to D) as “units” that may cause earthquakes, with certain criteria such as past rupture history and distribution, regardless of the conventional fault names. By thinking that an earthquake occurs when these segments rupture individually or in conjunction with the neighboring segments, the size of the earthquake can be predicted from the length of the individual fault (behavioral segment) and its co-movement relationship with the neighboring segments.To apply this model to the small-scale and complex faults of Japan, the data of earthquake faults with records of rupture were listed, and we prepared a trial plan for the relationship of earthquake scale and criterion for behavioral segment category[6]. Using the Kinki area as an example, the probability of future earthquake occurrence for each behavioral segment was calculated and publicized[7]. Although this calculation was only a result of one research, it was meaningful as an assessment obtained by logical deduction based on a model.3 Publication of the Rupture Probability Map of Major Active Faults in JapanTo develop this research further and to actually make it useful in society to reduce the risk of earthquake damage, it was necessary to apply this model on a national scale. Therefore, the author and others started to compile the Rupture Probability Map of Major Active Faults in Japan[8].The criteria for categorizing the behavioral segments are summarized in Fig. 4. The behavioral segments were categorized according to this criterion based on the distribution map of active faults. When covering all active faults in Japan, it was inevitable that some active faults were surveyed heavily and had plenty of data while others had hardly any specific data. In our map, we placed emphasis on obtaining uniform values for the entire country. Therefore, even in cases where no data was available, some assessment was obtained using provisional and empirical values rather than labeling them “unknown.” By placing priority on national coverage, we aimed for the Product Realization in Fig. 3 Schematic diagram of the behavioral segment categories.Behavioral segmentABCDs : Distribution intervalDifferent rupture historyX FaultY FaultX FaultZ Fault2 km


page 6