High temperatures lead to heat-related human stress and an increased mortality risk. To quantify heat discomfort and the relevant dangers, heat stress indices combine different meteorological variables such as temperature, relative humidity, radiation and wind speed. In this paper, a set of widely-used heat stress indices is analyzed and compared to the heat index currently used to issue official heat warnings in Switzerland, considering 28 Swiss weather stations for the years 1981-2017. We investigate how well warnings based on the heat index match warning days and warning periods that are calculated from alternative heat stress indices. The latter might allow for more flexibility in terms of specific warning demands and impact-based warnings. It is shown that the percentage of alternative warnings that match the official warnings varies among indices. Considering the heat index as reference, the simplified wet bulb globe temperature performs well and has some further advantages such as no lower bound and allowing for the calculation of climatological values. Yet, other indices (e.g., with higher dependencies on humidity) can have some added value, too. Thus, regardless of the performance in terms of matches, the optimal index to use strongly depends on the purpose of the warning.
Keywords: heat stress index; heat warning systems; impact-based warnings; temperature extremes.