Comparison of wet-bulb globe temperature (WBGT) and mean temperature for assessment of heat-related mortality

evidence from 47 Japanese prefectures

Abstract

Objectives: We analyzed the association between wet bulb globe temperature (WBGT) and mortality, and compared WBGT with mean temperature for evaluating mortality risk in 47 Japanese prefectures using data from 1972-2012.

Methods: We first calculated the prefecture-specific effect of WBGT on mortality using a time series regression model combined with a distributed lag non-linear model, and then used the same modelling strategy to examine the mean temperature. Second, we compared the minimum mortality WBGT (MMW) and minimum mortality (mean) temperature (MMT) for all the prefectures.

Results: We found that the curves of WBGT- and mean temperature-mortality were similar and an inverse J-shape was identified among most prefectures. In northern areas, the heat effect was comparable to the cold effect, and in southern areas, the cold effects were more pronounced. In general, MMW and MMT increased from north to south, except for some prefectures. The MMW and MMT percentile values ranged between the 80th and 90th percentile for most prefectures. We compared MMW and MMT for all the prefectures and found that Kochi is an outlier. We conducted an identical analysis for Kochi except for setting shorter lags (seven days from the original 21 days), because the heat effect usually has a short lag. After this modification, Kochi became a non-outlier. This implies that Kochi appeared to be an outlier because the model had an extraneous long lag.

Conclusions: We found mean temperature and WBGT were highly correlated in the evaluation of mortality in most prefectures. Therefore, in cases where WBGT data are not available for heat warnings, mean temperature is a good substitute.