Satellite Altimeters Measure Tsunami

satellites obtained profi les of sea surface height on transects across the Indian Ocean between two and nine hours after the December 26 Sumatra earthquake. The data are received hours to days af-ter “real time,” too late to be used in detection and warning of tsunamis. We compared the sea level anomaly profi les of December 26 measured along the satellite tracks (Figure 1D-G) with the measurements on previous passes of the same satellites 10 days, 35 days, and 17 days earlier. This allowed us to remove the majority of permanent and slowly varying features of sea level, revealing transient signals. The altimeters also provide wind speed and wave height data, and these allowed us to interpret a sea-level anomaly at 16°S in the Jason-1 profi le (Figure 1D) as being due to a severe storm. The remaining sea-level anomaly signal appears to be associ-ated with the tsunami. The signal of the leading edge two hours after the earth-quake is particularly prominent, with an amplitude of 60 cm and two narrow peaks where the NOAA tsunami model forecast shows two overlapping peaks coalescing into one broad (250 km) crest. Increased sea-surface roughness at spatial scales from 150 to 15 km wave-lengths also appears inside the portion of the ocean excited by the tsunami.The fi rst model simulation results of the Indian Ocean tsunami (Figure 1A-C) were obtained from the “MOST” (Meth-od of Splitting Tsunamis) model (Titov and Synolakis, 1998) and were posted by V.V. Titov on the Internet Tsunami Bul-letin Board less than 12 hours after the earthquake. MOST is part of the tsunami forecasting and warning system under development for the Pacifi c Ocean (Titov et al., 2005) that will provide fast real-time estimates of tsunami amplitudes using preset models, real-time seismic data, and, most importantly, deep-ocean tsunami amplitude data from a network of deep-ocean pressure sensors. Other researchers also ran models and posted results. Results of MOST and other mod-el runs have been widely used worldwide by the media for early planning of relief efforts and for post-tsunami fi eld sur-veys. Unlike the Pacifi c, the Indian Ocean does not yet have a network of deep-ocean pressure sensors, and so coastal tide gauges provide the only direct mea-surement of Indian Ocean tsunami am-plitudes. The satellite altimeter data we present here are the only measurements of the amplitude of the December 26 tsunami in the deep, open ocean. At the time of the fi rst MOST model simulation, earthquake source mecha-nism models described a rupture con-fi ned to only the southernmost portion of the broad region mapped out by the aftershock pattern. However, it seemed clear that the tsunami should have been generated by displacements distributed along the entire aftershock zone. The initial conditions for the MOST model were set assuming this more spatially distributed source, with initial amplitude guesses based on preliminary estimates of the earthquake magnitude and one coastal tide-gauge measurement from Cocos Island. Because of the lack of