The long-period strain seismograms from the Sumatra-Andaman earthquake of December 26, 2004 and the Nias earthquake of March 28, 2005 are analyzed by comparing the observed seismograms with synthetic seismograms calculated by normal mode theory. The multiple CMT source (Tsai et al., 2005) and Harvard CMT solution explain the seismograms recorded by both extensometers and volumetric strainmeters from the 2004 and 2005 earthquakes, respectively. The long-period strain seismograms observed in Japan are consistent with about 9 min source duration and 1200 km fault length proposed by Tsai et al. (2005) and Velasco et al. (2006).
The great Sumatra-Andaman earthquake of December 26, 2004 is studied by using extensometer (strain-meter) as very-broard-band seismograph. We analyze the strain seismograms obtained by 100 m extensometers installed at Matsushiro Seismological Observatory. Japan Meteorological Agency. We compare the observed strain seismograms with thouse calculated by normal mode theory and obtain fairly good agreement between the observed and calculated seismograms for a five-sourcse model with a moment magnitude of 9.3 and a source duration of about 600 sec. We find a peculiar strain change right after the passage of Rayleigh wave in the NS component of the strain seismograms. This strain change cannot be explained by the summation of normal modes and double-couple type sources.