Abstract
A paradigm shift to the use of indoor rather than outdoor temperature to estimate the exposure risk of low and high temperatures is vital for better prediction of temperature health effects and timely health warnings, and will also assist in understanding the influence of temperature on energy consumption and comfort. This study aimed to quantify the percentage of hours during the year that indoor temperature (living room) was in the extended comfort band (18-28 degrees C) of a subtropical climate, and identify the diurnal pattern of indoor temperatures in different seasons. Data used was collected in a previous study on the association between indoor and outdoor temperature. A k-shape cluster analysis resulted in two clusters of indoor temperature patterns for both weekdays and weekends. A bimodal pattern was identified during the cool season and a flat top pattern for the warm season, with many variations at weekends. These patterns can be attributed to the influence of cooling and heating processes depending on the season as well as occupancy, occupants' interference, and building materials. During the intermediate season, a sinusoidal pattern was observed for both weekdays and weekends because occupants likely relied on outdoor temperature conditions which were similar to those expected indoors without heating or cooling devices. The percentage of hours in which the indoor temperature of the houses ranged within the extended comfort band was 72-97% throughout the year, but for the coldest and hottest months it was 50-75%. These findings show that Brisbane residents are at possible risk of exposure to cold and hot temperatures due to the poor thermal performance of houses, and confirm that there is no standard indoor temperature pattern for all houses. (C) 2019 Elsevier B.V. All rights reserved.