The Impact of Extreme Heat Events on Mortality in Hong Kong: A 10-year Time-series Study (2006-2015)
Other conference paper

Full Text

Other information
AbstractClimate change and aggravated urban heat island effect increased the frequency, intensity, and duration of extreme heat events in cities. Numerous epidemiological studies suggested that extreme heat events are the predominant contributor to heat-related mortality (Kravchenko et al., 2013), whereas standards in defining extreme heat events to evaluate associated health risks in the local context are generally lacking. In Hong Kong, the extremely high building and population density exacerbate the vulnerability to extreme heat events, particularly to vulnerable population groups like elderly people, children and people with chronic diseases. Studies suggested that there is excess mortality during daytime extreme heat events (D'Ippoliti et al., 2010) (Chan et al., 2012) while our previous study also highlighted the noticeable mortality risk during consecutive hot nights (Ho et al., 2017). However, as most of the typical definitions of extreme heat events use daily maximum temperature as an indicator, the assessment of extreme heat events may not be able to accurately describe a population’s typical routine of heat exposure. This paper aims to provide a useful reference for an accurate and applicable definition of extreme heat events in Hong Kong. Based on the city-wide ecological information on mortality and hourly air temperature from 2006 to 2015, the impacts of different types of extreme heat events were assessed in terms of their duration and timing. In particular, the difference and interactive effect of extreme heat events between daytime and nighttime were also examined.

A time-series daily dataset of mortality, temperature, and air quality of Hong Kong from 2006 to 2015 was collected. Daily counts of all-cause deaths with records of the death date, cause, and place were obtained from the Census and Statistics Department. Based on the hourly temperature record acquired from the Hong Kong Observatory, two indicators, namely Very Hot Days (VHDs; daily maximum temperature ≥ 33°C) and Hot Night (HN; daily minimum temperature ≥ 28°C), were calculated to define extreme hot weather in Hong Kong. Three durations of extreme heat events were observed for daytime and nighttime independently: 1-2, 3, >3 consecutive VHDs and HNs respectively. With an extreme heat event during daytime and nighttime defined as the occurrence of at least three consecutive VHDs and HNs respectively, the timing of a new extreme heat event was grouped into 1) 1-3 days after the previous heat event; 2) >3 days after the previous extreme heat event. Furthermore, two VHDs with no HN (2D0N), one HN (2D1N), or two HNs (2D2N), as well as three VHDs with no HN (3D0N), one HN (3D1N), two HNs (3D2N), and three HNs (3D3N) in between were identified to capture the interactive effect between daytime and nighttime extreme heat events. The daily average concentration of seven air pollutants (CO, NO2, NOx, SO2, PM2.5, PM10, O3) was acquired from the Environmental Protection Department. The relative risk (RR) of all-cause mortality between the extreme heat events and non-extreme heat events was estimated by adopting Poisson generalized additive regression models, adjusting for the long-term trend, regular seasonal trend, day of the week, and daily concentration of air pollutants. Autocorrelations between covariates were removed by checking the residuals. Subgroup analysis was also conducted to explore the modifying effect of individual age and sex, which could be used to identify vulnerable populations.
Acceptance Date23/02/2018
All Author(s) ListWang Dan, Lau Kevin. K.L., Ren Chao, Shi Yuan
Name of Conference10th International Conference on Urban Climate/14th Symposium on the Urban Environment
Start Date of Conference06/08/2018
End Date of Conference10/08/2018
Place of ConferenceNew York, NY, US
Country/Region of ConferenceUnited States of America
LanguagesEnglish-United States

Last updated on 2019-26-11 at 09:35