An increasing proportion of the world’s population lives in densely populated three dimensional
(3D) urban landscapes. Many of the world’s megacities identified as having the most severe air
pollution problems are characterised by these high-rise cities. Despite this, current
traffic-related air pollution (TRAP) exposure estimates are strictly two dimensional (2D).
Further, nearly all epidemiologic studies of TRAP health impacts estimate exposure based on 2D
residential location and do not consider population mobility or time activity patterns during
the course of the day. Hong Kong is an ideal test site for development of an exposure model that
incorporates population mobility and a 3D urban landscape. It is a high profile example of a
high density city, with elevated levels of air pollution, a well-developed network of vehicle
flow and pollution monitoring sites, dominant use of public transport and a supportive
government administration.
The study has three main objectives:
- to investigate the behaviour and distribution of vehicle
emissions in a 3D urban landscape using air quality sensor networks
- to develop, evaluate and
demonstrate a dynamic 3D air pollution exposure model for Hong Kong
- to create an
incremental exposure assessment methodology that can be applied in megacities across Asia
and
the developing world
The first stage in producing the Hong Kong D3D model will be to create a 2D land use regression
(LUR) model of the type used in many population exposure assessments worldwide. This will
provide a foundation for the vertical and dynamic layers subsequently developed and a comparator
for exposure misclassification assessment. Second, arrays of pollution sensors measuring PM2.5,
BC, CO, NO, NO2 and O3 will be deployed on the facing façades of selected buildings at four
heights above street level within a range of street canyons. Sensors will be paired inside and
outside the building to assess infiltration efficiencies. The resulting characterization of TRAP
behaviour will produce a canyon decay rate ‘typology’, dependent on canyon aspect ratio and
orientation. This typology will be applied across the city using an existing 3D cityscape model
of HK, thereby creating a 3D LUR model capable of producing exposure estimates that distinguish
residential height above street level. Finally, travel behaviour surveys will be used to
classify population mobility over space and time. Combining population mobility with the 3D LUR
model will produce a dynamic 3D exposure model capable of producing time-weighted total and
major micro-environmental TRAP exposure estimates for population subgroups stratified by
residential and destination locations, age and sex.
The final phase of the study will apply a staged approach to quantify the sources of bias
associated with epidemiologic effect estimates derived from each increasingly sophisticated
exposure model. 3D and animated visualizations of exposure estimates will provide a powerful
interpretation tool. Guidance will be produced explaining generalized development of D3D models
in other megacities.
HKD3D is funded by the US Health Effects Institute and is a collaboration between King’s College
London, the University of Hong Kong, the University of British Columbia, and Simon Fraser
University.
