The aim of this study was to investigate changes in the prevalence of adverse stress reactions in different areas of Hat Yai during the year immediately following a major natural disaster—the flood of 2000. Samples of around l00 residents of each of four areas exposed to different degrees of severity of the stressors caused by the flood were followed-up at 8 to l0 week intervals. When first interviewed on entry into the study, the respondents were invited to complete a structured questionnaire that sought information on their personal characteristics, data on the various flood-related stressors to which they were exposed, and their responses to these stressors. They also completed a Thai version of the 12-item General Health Questionnaire and the Impact of Events Scale at each of the five data collection times. There was a gradual decline in the scores on these instruments over the l2 month period but, on the anniversary of the flood, there was an increase in scores among residents confined to those areas of Hat Yai where the risk and/or consequences of severe flooding was highest and where rescue work was most difficult. This is considered to be the first epidemiological study of a major civilian disaster that has demonstrated an increase in prevalence of stress-related symptoms as an anniversary reaction within a community. Copyright © 2006 John Wiley & Sons, Ltd.