Using seismographs and GPS displacement measurements, we have estimated the seafloor deformation history of the December 2004 Sumatra-Andaman earthquake and the March 2005 Nias earthquake by separating their deformation period into intervals of 800-s, 1-h, and 6-months. We have then calculated their corresponding gravity changes (induced by the seafloor deformation), which are 11.3, 12.5, and 14.9 microgalileo, respectively. We show that the seismographs and GPSderived values are consistent with the known postseismic to coseismic moment ratio of 30% and the Gravity Recovery and Climate Experiment (GRACE) satellites measurements of 15 microgalileo for the same period of 6 months. However, the vertical component of the accumulated seafloor deformation during the tsunami formation period (~30 min) could only generate a potential energy of 1.2 × 1015 Joules and account for only one third of the actual tsunami height. The evidence is overwhelmingly contrary to the long-held theory that the vertical deformation of seafloor is the primary source of tsunamis.
Furthermore, we have carefully examined the pioneering wave-maker experiment that initially conceived the ubiquitous tsunami genesis theory. Surprisingly, we found that the experimental ratio of the horizontal slip distance to the water depth – the non-dimensional parameter that allows comparing the experiment with reality on an apple-to-apple basis – was 200 times of realistic earthquake parameters. The experimental conclusion is problematic in conceiving the tsunami theory. By including the horizontal momentum energy transferred by the faulting continental slope in a three-dimensional tsunami model, we have re-examined the December 2004 tsunami using both seismographs and GPS measurements. Our results show that the new theory is more consistent with altimetry and tide data than the conventional theory of using the vertical force alone, suggesting that the tsunami formation mechanism is not as simple as previously thought.