Evaluation of human thermal comfort using by meteorological approach

For the evaluation of the climate impacts on humans, we often use an unweighted average such as monthly mean. Thermal comfort for indoor condition is sometimes evaluated by the monthly mean air temperatures. However, these techniques are insufficient due to the following reasons: (1) human temperature regulation is described as a negative feedback control system and climatology is described as its disturbance factor, (2) the most suitable period for the determination of the impact on humans corresponds to the time lag between the experiences of climatic events and the observation of the effects on humans. The objective of this study is to improve the techniques for the assessment of the climatic impacts on human health or comfort by considering the abovementioned points with feedback-optimized models or no-feedback models.

In order to consider the feedback effect of human temperature regulation, we applied the AR model as the feedback model and the MA model as the no-feedback model for each daily temperature. To obtain the thermal comfort data, from May 10, 2006 to July 12, 2006, we conducted field experiments by means of questionnaire surveys and meteorological observation in cooperation with 919 undergraduate engineering students in Sapporo, Japan. We also collected data of thermal environmental investigate in San Ramon and San Francisco from ASHRAE RP-884 project to compare to the results in Sapporo.

Results are as follows. The most suitable order of AR model in Sapporo is 10th. In the same term, that of in San Francisco is also 10th. Lags of approximately 10 days is corresponds up to 12 days requiring period for human thermal acclimatization. On the other hand, both the AIC values of the MA model continue to decrease with increase in the order up to 48th. These result of Sapporo applied to thermal comfort data. The tendency between the “not discomfort temperature” and climatic values are classified into (1) one-year MA model and 30-day averaged model, and (2) MA model, except the one-year cycle, and AR model. The results suggested group (2) is better to evaluated thermal comfort conditions. Thermal comfort data in San Ramon and San Francisco was also analyzed in same way.