Interplate coupling of the central Nankai Trough and shortening of the central Japan estimated from GPS data

Interplate megathrust earthquakes have repeatedly occurred with a repeat time of 100-150 years along the Nankai subduction zone. The latest megathrust events there were the 1944 Tonankai (M7.9) and the 1946 Nankai (M8.0) earthquakes. It is very important to investigate tectonic loading processes toward the next large earthquake along the Nankai Trough. For this purpose, we estimate slip deficit distribution of the source region of the 1944 Tonankai earthquake from GPS data inversion analysis and also examine E-W shortening in Central Japan where tectonic contraction between the northeastern and the southwestern Japan arcs. We invert GPS displacement rate data estimated from daily coordinates from April 1996 to March 2000 to estimate interplate coupling in the Tonankai region by using the method of Yoshioka et al. (1993). We reconstruct the plate boundary configuration based on hypocenters determined by the Japan Meteorological Agency and results of seismic explorations in this region. We find that GPS data analysis with a single fixed point is inappropriate in the Tonankai region. With such analysis, the resultant slip deficit distribution becomes unrealistic, showing large slip deficit vectors deviating from the direction of the plate convergence. Such a result can be attributed to the E-W shortening in the inland region. Therefore, this shortening effect must be eliminated to estimate correct distribution of slip deficits. For the elimination of E-W shortening, we divide the GPS stations into the eastern and western groups and introduce one fixed point for each group. We calculate displacement rate of GPS stations with respect to the fixed point of each group, and invert those relative displacement rates to estimate the slip deficit on the plate boundary. We test various cases of partitioning of GPS stations into two groups, and find that a group division near the Yoro-Kuwana-Yokkaichi fault yields a reasonable result. Slip deficit is estimated to be the 4.0-4.3cm/yr and largest below the Sea of Kumano, southeastern of Kii Peninsula, where the plate boundary depth is 10-25km. Vertical displacement rates calculated from our slip deficit distribution are consistent with those obtained from leveling survey during 1983-2000, indicating the result represent a persistent feature for the interseismic period. Estimated slip deficit distribution is consistent with the result of a similar analysis in the Tokai region by Ohta et al. (2004). In addition, the large slip deficit region corresponds to the coseismic slip distribution of the 1944 Tonankai earthquake estimated by Sagiya and Thatcher (1999), but is inconsistent with the seismological analysis results by Ichinose et al. (2003) and Yamanaka (2004). This discrepancy suggests a possibility of significant postseismic slip under the Sea of Kumano, but these exists an open question whether a strongly locked patch can exhibit a velocity-strengthening behavior or not. As for the inland deformation, 5mm/yr E-W shortening is inferred between two fixed points. This inland contraction rate is consistent with the geological displacement rate of Yoro-Kuwana -Yokkaichi fault, indicating that this active fault is playing an important role in the long-term E-W shortening of the Japan mainland.