Abstract
We have carried out the benchmark tests of ground motion calculation using the numerical methods. In the case of the Kanto Plain in the previous reports, difference in later phases of ground motion can be seen mainly due to the grid interval modelling from participants. In the 2013 benchmark tests, the foreshock of the 2004 Kii Peninsula earthquake is targetted to focus on the long-period ground motions caused by faults in the Nankai Trough. In the blind calculation, differences were found in synthetic waveforms provided by 6 teams. For the cases which used the unified and simple models, results becomes similar. Finally we examined the effect of the Q values in surrounding damping area on reflected motion in waveforms of the long-period ground motion.