Suitability of different comfort indices for the prediction of thermal conditions in tree-covered outdoor spaces in arid citie

Abstract

Thermal comfort of people in outdoor environments is one of the most influential factors in the habitability of spaces. The comfort level is determined not only by climate variables but also by the adaptation of people to the environment. This work presents a comparative study among five thermal comfort indices and its contrast with subjective answers in order to identify which of them can be better used to predict thermal comfort in forested outdoor spaces of the Mendoza Metropolitan Area, a city located in an arid zone whose configuration corresponds with the "oasis city" type. The microclimatic measurements and the subjective surveys were carried out during winter of 2010 and summer of 2011. Five indices used internationally were selected: the Temperature-Humidity Index (THI), the Vinje Index (PE), the thermal sensation index (TS), the Predicted Mean Values (PMV), and the COMFA model. The results show that all the mathematical models evaluated present percentages of predictive correctness lower than 25%. Despite this low indicator, the use of inductive methods is recommended to know if a certain space is under comfort or discomfort conditions, whereas deductive methods should be used to intervene on urban design strategies. In addition, it is necessary to develop local models to evaluate the thermal comfort perceived by the inhabitants of the studied city more adequately. These tools are very useful in the design and evaluation of the thermal behaviour of open spaces according not only to climate criteria but also to the subjective characteristics of users.