Mexico City has had a rapid growth in population and urban area in recent years. In previous studies it has been observed that the urban heat island unlike the traditional one also occurs during the day with TU-R up to 10 °C. This heat concentration can be considered as thermal pollution. The objective of this study was to determine whether the physiological equivalent temperature (PET) is able to assess this type of pollution because above 23 °C is said that people begin to show heat stress. To do that, micrometerological data generated by the RedMet which has 18 stations distributed in the city were used. PET analysis was performed using the proposed algorithm by Matzarakis et al. (1999). The maximum PET differences between the analyzed urban and rural stations were up to 8 °C in February and July, while in May and June was 7 °C. However, according to PET distributions, people could present thermal stress throughout the year in almost the entire territory of the city, and thus an intense thermal pollution. Probably, the categorization of the grade of stress of PET must be modified for the Mexico City inhabitants, as apparently this categorization does not take into account the natural acclimatization of people.

This study aims to analyze the response of the thermal comfort index, Physiological Equivalent Temperature - PET (°C), for acclimatized residents in the tropical city of Salvador – Brazil, by associating PET (°C) intervals with voted thermal sensation. PET defines neutral condition based on equalized human heat balance. However, as people are adapted to different climatic conditions, classes of thermal perception and sensation, including the comfort zone will differ according to the local climate. That means that calibration is necessary for different climates and this paper shows the results for Salvador, a city located at 13o S latitude. The methodology used involved carrying out meteorological measurements of air temperature and humidity, solar radiation, wind velocity and simultaneously interviewing acclimatized pedestrians about their thermal sensation. The population sample included those residents living in Salvador for more than a year and age between 20 and 59 years old. The research was carried out in two open urban spaces (two public squares) with different characteristics and environmental qualities. In total 1,435 interviews were conducted. Two techniques of exploratory data analysis were used to calibrate the PET index (°C): Box-Plot and Decision Tree - DT. The variable used was a seven scale categories: +3 (Very Hot); +2 (Hot); +1 (Warm); 0 (Thermal Comfort); -1 (Cool); -2 (Cold) and -3 (Very Cold). For the final evaluation, ‘very cold’ and ‘cold’ was cancelled because, as expected, no one reported them, keeping the classes in summer and winter alike. Through statistical analysis, using DT and considering 50% of the responses, only the ‘Very Hot’ classification showed a value above it (51.5%). As the 50% criterion was insufficient to distinguish the seven categories of thermal perception, we adopted the criterion of 30% in DT. With that criterion the results became: Very Hot (category 1) PET index more than 34.1 °C; Hot (category 2) PET index is between 29.4 to 34.1 °C; Warm (category 3) PET index is between 26.8 to 29.4 ºC; Comfort (category 4) only the upper limit for the PET index was found, up till PET less than 26.8 ºC, and the lower limit of comfort has not been established. With BOX-Plot technique ranges of PET for each category overlapped, invalidating their determination and highlighting DT option as the most suitable technique for determining the limits for each category of thermal sensation.

An universal thermal index is necessary for objective comparisons in different climatic zones for the research of human-biometeorology (e.g. for environmental evaluations, climate assessment for tourists, as well as assessments of climate change). Universal Thermal Comfort Index (UTCI) has been developed to try satisfying these requirements. However, UTCI was an operative statistic function based on investigations in Europe and Russia. Hence, UTCI shows limitations in its applicability in the other climatic zones, such as tropic and dry climates. Physiologically Equivalent Temperature (PET) is another thermal index which is at the beginning developed to evaluate the outdoor thermal conditions in temperate climates. It has been proved to effectively evaluate the impact of the air temperature, mean radiant temperature and wind speed on thermal comfort, but variations in air humidity and clothing insulation show weak influence on PET. Thus, this study aims to develop a thermal index for universal applications in all climate zones based on a modified PET (mPET).

mPET has two adjustments on PET: (1) physiological thermoregulation is improved to a simple multi-segment body model including a blood pool element and a bio-heat transfer principle; and (2) a multi-layer clothing model with clothing insulation and vapour resistance is implemented. Due to those two adjustments, PET has been improved to effectively evaluate of the impact of vapour pressure and clothing insulation on thermal conditions and mPET can be applied in all climatic zones.

A case study was shown to apply UTCI, PET and mPET to evaluate bioclimate using data investigated in Freiburg during 1981 to 2013. The wind speed measured at 10 m according to the guideline of WMO is adapted to 1.1 m using a wind reduction equation for the purpose on assessment of bioclimate. In summary, mPET evaluates less underestimations and overestimations for thermal conditions than PET and UTCI.

Keywords: Physiologically Equivalent Temperature, Modified Physiologically Equivalent Temperature, Thermal index, Universal Thermal Comfort Index