We hypothesized that the scale of the predicted Nankai Earthquake will increase gradually year by year as the time passes from the last occurrence in December 1946, and then we estimated strong motions chronologically from 2009 to 2060 for western Japan based on the up-to-date strong motion prediction method, i.e., the so-called statistical Green's function method with hierarchical source heterogeneities. Next, we input the derived chronological strong motions into a set of nonlinear response analysis models proposed by Nagato and Kawase and estimated building damage ratios for different structure types and numbers of floors. When we compare the evolutional damage ratios of buildings in the subject regions with those for the reference case (corresponding to the scenario that will happen in 2036), we found that for every 10 years of delay in occurrence of the mega-thrust Nankai event we must expect 40% of increase in building damage, mainly in the epicentral areas.
