Future probabilistic hot summer years for overheating risk assessments

As the 2003 Paris heatwave showed, elevated temperatures in buildings can cause thousands of deaths. This makes the assessment of overheating risk a critical exercise. Unfortunately current methods of creating example weather time series for the assessment of overheating are based on a single weather variable, and hence on only one driver of discomfort or mortality. In this study, two alternative approaches for the development of current and future weather files are presented: one (pHSY-1) is based on Weighted Cooling Degree Hours (WCDH), the other (pHSY-2) is based on Physiologically Equivalent Temperature (PET). pHSY-1 and pHSY-2 files were produced for fourteen locations. These were then compared with the existing probabilistic future Design Summer Year (pDSY) and the probabilistic future Test Reference Year. It was found that both pHSY-1 and pHSY-2 are more robust than the pDSY. It is suggested that pHSY-1 could be used for assessing the severity and occurrence of overheating, while pHSY-2 could be used for evaluating thermal discomfort or heat stress. The results also highlight an important limitation in using different metrics to compare overheating years. If the weather year is created by a ranking of a single environmental variable, to ensure consistent results assessment of the building should be with a similar single metric (e.g. hours >28 °C or WCDH), if however the weather year is based upon several environmental variables then a composite metric (e.g. PET or Fanger's PMV) should be used. This has important implications for the suitability of weather files for thermal comfort analysis.