Tide gauge data collected from Sri Lanka (three stations) and Western Australia (eleven stations) during the Indian Ocean tsunamis, which occurred in December 2004, March 2005, July 2006, and September 2007, and incorporated five tsunamis, were examined to determine tsunami behaviour during these events. During the December 2004 tsunami, maximum wave heights of 3.87 m and 1.75 m were recorded at Colombo (Sri Lanka) and Bunbury (Western Australia), respectively. The results indicated that although the relative magnitudes of the tsunamis varied, the tsunami behaviour at each station was similar. This was due to the effect of the local and regional topography. At all tide gauges, the spectral energy corresponding to periods between 20 and 85 minutes increased during the tsunami. The sea-level data obtained from the west and south coasts of Sri Lanka (Colombo and Kirinda) indicated the importance of wave reflections from the Maldives Island chain, which produced the maximum wave two to three hours after the arrival of the first wave. In contrast, Trincomalee on the east coast did not show evidence of a reflected wave. Similarly, along the west coast of Australia, the highest waves occurred 15 hours after the arrival of the first wave. Here, based on travel times, we postulated that the waves were reflected from the Mascarene Ridge and/or the Island of Madagascar. Reflected waves were not present in the 2006 tsunami, where the primary waves propagated away from topographic features. One of the main influences of the tsunami was to set up oscillations at the local resonance frequency. Because Sri Lanka and Western Australia have relatively straight coastlines, these oscillations were related to the fundamental period of the shelf oscillation. For Colombo, this corresponded to 75-minute period, whereas in Geraldton and Busselton (Australia), the four-hour period was most prominent; at Jurien Bay and Fremantle, the resonance period was 2.7 hours.