Key points are not available for this paper at this time.
Abstract This study describes the impact of weather on human mortality at numerous locations around the United States. We evaluate forty-eight cities and determine the differential impact of weather on mortality on an intercity and interregional level. The mortality data are analyzed separately for different age, race, and cause of death categories. The possible impact of geographical and within-season acclimatization is also analyzed. Thresh-old temperatures, which represent the temperature beyond which mortality increases, are identified for all the categories and all cities in summer and winter. We correlate numerous weather variables with mortality for days with temperatures beyond the threshold. In summer, warm, humid, calm conditions (especially at night) relate to the highest mortality. The strongest relationships occur in regions where hot weather is uncommon and the weakest relationships in the hottest locales. Regional acclimatization appears to be especially important in summer. A strong intra-seasonal acclimatization factor is also apparent, and hot weather early in the season produces a more pronounced response than similar weather late in the season. Winter relationships are generally weaker, and cloudy, damp, snowy conditions are associated with the greatest mortality. In both seasons, the elderly appear to be disproportionately stressed when compared to other age groups. Racial responses appear similar on a national and seasonal level, but there are differential responses on a regional level. Non-whites appear more sensitive than whites in the South during both seasons. Key Words: climate impact assessmenthuman mortalityacclimatizationweather/mortality relationshipsweather stress Notes This research was supported by the Environmental Protection Agency under contract no. CR-814301-01-0. The authors thank Guanri Tan, Kathleen Valimont, and Jon Skindlov for their technical support and helpful suggestions. We also thank Lisa Davis and Darcy Kenton for their cartographic assistance.
Kalkstein et al. (Wed,) studied this question.