Assessment of the global effects of air pollution requires a comprehensive set of estimated exposures for all populations. The primary source of information for estimating exposures has been measurements from ground monitoring networks but, although coverage is increasing, there remain regions in which monitoring is limited.
Ground monitoring data therefore needs to be supplemented with information from other sources, such as satellite retrievals of aerosol optical depth, chemical transport models. In addition to these estimates of air quality, other information can be used in estimating air pollution, including population estimates and topography (e.g. elevation). The Data Integration Model for Air Quality (DIMAQ) was developed by the WHO Data Integration Task Force to integrate data from multiple sources in order to provide estimates of exposures to PM2.5 at high spatial resolution globally.
Sources of data include ground measurements from 9,690 monitoring locations around the world from the WHO cities database together with satellite remote sensing, population estimates, topography, and information on local monitoring networks and measures of specific contributors of air pollution from chemical transport models.
It is estimated that over 90% of the world's population reside in areas exceeding the World Health Organization's Air Quality Guidelines. Estimated exposures from the model are combined with risk estimates to produce a global assessment of exposures to PM2.5 and to estimate the associated burden of disease attributable to air pollution.EndFragment
Air pollution is a major risk factor for global health, with both ambient and household air pollution contributing substantial components of the overall global disease burden. One of the key drivers of adverse health effects is fine particulate matter ambient (outdoor) pollution (PM2.5) to which an estimated 4.2 million deaths can be attributed annually. Together with household air pollution, it is estimated that globally 7 million deaths can be attributable to air pollution annually.
Button
Click here for interactive map
FOR FURTHER DETAILS PLEASE CONTACT
DUNCAN SANDES
UNIVERSITY OF EXETER PRESS OFFICE
Telephone: 01392 722391
Mob: 07789 874838
Email: d.sandes@exeter.ac.uk
https://www.exeter.ac.uk/news/contact/
OR
PROFESSOR GAVIN SHADDICK
G.SHADDICK@EXETER.AC.UK
ESTIMATES OF GLOBAL FINE PARTICULATE MATTER AIR POLLUTION (PM2.5) FOR 2016
GLOBAL AIR QUALITY
HOME
WHO TASKFORCE
LINKS
AMBIENT/ DIMAQ
HOUSEHOLD / GHHEM
DATA INTEGRATION MODEL FOR AIR QUALITY (DIMAQ)
The Data Integration Model for Air Quality (DIMAQ) was developed by the WHO Data Integration Task Force to integrate data from multiple sources in order to provide estimates of exposures to PM2.5 at high spatial resolution globally.The primary source of information for estimating exposures has been measurements from ground monitoring networks but, although coverage is increasing, there remain regions in which monitoring is limited. Ground monitoring data therefore needs to be supplemented with information from other sources, such as satellite retrievals of aerosol optical depth and chemical transport models.
A hierarchical modelling approach for integrating data from multiple sources is proposed allowing spatially-varying relationships between ground measurements and other factors that estimate air quality. Set within a Bayesian framework, the resulting Data Integration Model for Air Quality (DIMAQ) is used to estimate exposures, together with associated measures of uncertainty, on a high-resolution grid covering the entire world.
Modelled exposure to PM2.5 provides more comprehensive exposure information for countries than measured data that provide data only for a selection of towns and cities. A comprehensive set of estimated exposures such as this will be required when estimating the health impacts of ambient air pollution for a country.
The DIMAQ model calibrates data from these sources with ground measurements. The relationships between the different sources of data may be complex and will vary between regions due to differences in the composition of PM2.5 and other factors. Within DIMAQ, these relationships are expressed in terms of a series of calibration equations which are allowed to change over space. DIMAQ has a hierarchical structure that uses individual country’s data, where available, to construct these equations.
The WHO has been working with the University of Exeter on recent developments to DIMAQ which, where there is sufficient data, now allows the calibration equations to vary both within and between countries. Where there is insufficient data within a country, to produce accurate equations, it is supplemented with information from the wider region.
When calibration equations have been established and tested, the model is used to estimate exposures, together with associated measures of the uncertainty, across the globe. These high-resolution estimates of PM2.5 can be used to produce air quality profiles for individual countries, regions and globally. Further to this, by defining areas as either urban or rural (based on land-use, derived from satellite images, and population estimates) at a resolution to match the estimates of air pollution from DIMAQ, the profiles can presented separately for urban and rural areas.
In order to produce the global estimates of PM2.5 DIMAQ was implemented on the University of Exeter’s high-performance computing systems, ISCA.
EndFragment
A full description of the model development and evaluation is available at the Journal of the Royal Statistical Society. Click on the image to the right to see further details.
GLOBAL HIERARCHICAL HOUSEHOLD ENERGY MODEL (GHHEM)
Information on the types of technologies and fuels used by households for cooking is regularly collected on nationally‐representative household surveys or censuses. WHO regularly collects and compiles such household energy data in the WHO Household energy database In order to estimate the health burden of household air pollution, accurate and consistent estimates of primary polluting fuel usage, by country, are required.
Although coverage of the surveys is increasing, there remain areas in which detailed information may not be available and, in those areas, the proportion of the population who are exposure to pollution fuels will need to be estimated.
Together with the University of Exeter, the WHO has developed a global hierarchical household energy model (GHHEM) for producing estimates of overall polluting (and clean) fuel usage together with estimates of a number of sub-classes including solid fuels, wood and kerosene.
Set within a Bayesian hierarchical modelling framework, trends in the proportions (of each type of fuel use) are estimated for each country, based primarily on information from surveys within that country. Where data is not available within a country, or is insufficient to produce accurate estimates, the model structure derives information from regional trends and, in such cases, estimates are complimented by increased uncertainty.
Bayesian hierarchical models provide an extremely useful and flexible framework, in which to model complex relationships and dependencies in data that has an inherently hierarchical structure. Here, for example solid fuel use is modelled as a proportion of all fuel use, with Kerosene modelled as a proportion of one minus solid fuel use. The proportions of the individual classes of fuel types are then modelled as proportions of solid fuel use. This both ensures consistency, between the sub-classes and the total and, where there are consistent patterns over time and space (country), enables information to be borrowed between the different categories of fuel usage.
The GHHEM was implemented at the University of Exeter using using Markov chain Monte Carlo (MCMC) and, as with any Bayesian analysis, results in not just (point) estimates of the proportions (of the usage of each fuel type, by country and by year), but a set of full posterior distributions for those estimates. Summaries of these distributions can be taken to provide both point estimates (e.g. means) and measures of uncertainty (e.g. 95% credible and 95% prediction intervals).
The GHHEM is applied to the WHO household energy database to produce a comprehensive set of estimates, together with associated measures of uncertainty, of the use of polluting fuels from cooking, by country, for each year for which survey data was available (1990-2016), although it may be noted that the model can also produce predictions into the future.
Estimates from the model are used to create a global assessment of exposures to household PM2.5 which are combined with risk estimates to produce the burden of disease attributed to household (indoor) air pollution.
WHO INTERACTIVE MAP
UNIVERSITY OF EXETER PRESS OFFICE
LINKS
BREATHELIFE
PROFESSOR GAVIN SHADDICK
WHO AIR QUALITY
WHO ESTIMATES AND DATA DOWNLOAD
UNIVERSITY OF EXETER
WHO DATA INTEGATION TASKFORCE
The Data Integration Model for Air Quality (DIMAQ) was developed by the WHO Data Integration Task Force to integrate data from multiple sources in order to provide estimates of exposures to PM2.5 at high spatial resolution globally.
DIMAQ was developed by a multi-disciplinary group of experts established as part of the recommendations from the first meeting of the WHO Global Platform for Air Quality, Geneva, January 2014.
The resulting Data Integration Task Force consists of
Gavin Shaddick, Department of Mathematics , University of Exeter, Exeter, U.K.
Michael Brauer, School of Population and Public Health, The University of British Columbia, Vancouver, British Colombia, Canada
Aaron van Donkelaar, Department of Physics and Atmospheric Science, Dalhousie University, Halifax, Nova Scotia, Canada
Rick Burnett, Health Canada, Ottawa, Ontario, Canada
Howard Chang, Department of Biostatistics, Rollins School of Public Health, Emory University, Atlanta, Georgia, USA
Aaron Cohen, Health Effects Institute, Boston, Massachusetts, USA
Yang Liu, Department of Environmental Health, Rollins School of Public Health, Emory University, Atlanta, Georgia, USA
Randall Martin, Department of Physics and Atmospheric Science, Dalhousie University, Halifax, Nova Scotia, Canada
Jason West, Department of Environmental Sciences and Engineering, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA
Lance Waller, Department of Biostatistics, Rollins School of Public Health, Emory University, Atlanta, Georgia, USA
Jim Zidek, Department of Statistics, University of British Columbia, Vancouver, British Columbia, Canada
Initial developments of the DIMAQ model and its implementation were also performed by:
Amelia Green, Department of Mathematical Sciences, University of Bath, Bath, U.K.
Matthew Lloyd Thomas, Department of Mathematical Sciences, University of Bath, Bath, U.K.
The Taskforce was supported by members of the WHO:
Sophie Gumy, World Health Organisation, Geneva, Switzerland
Pierpaulo Mudu, World Health Organisation, Geneva, Switzerland
Annette Pruss-Ustun, World Health Organisation, Geneva, Switzerland
Giulia Ruggeri, World Health Organisation, Geneva, Switzerland
Recent developments to DIMAQ and the implementation for the global estimates of air pollution for 2016 have included a wider group of researchers including:
Roman Aguirra-Perez, Department of Mathematics , University of Exeter, Exeter, U.K.
James Salter, Department of Mathematics , University of Exeter, Exeter, U.K.
Daniel Simpson, Department of Statistics, University of Toronto, Canada
The development group of the Global Hierarchical Household Energy Model included:
Theo Economou, Department of Mathematics , University of Exeter, Exeter, U.K.
Gavin Shaddick, Department of Mathematics , University of Exeter, Exeter, U.K.
Oliver Stoner, Department of Mathematics , University of Exeter, Exeter, U.K.
Sophie Gumy, World Health Organisation, Geneva, Switzerland
Jessica Lewis Mudu, World Health Organisation, Geneva, Switzerland
Heather Adair-Rohani, World Health Organisation, Geneva, Switzerland
The views expressed in these webpages and associated publications are those of the members of the Taskforce and the wider research group and they do not necessarily represent the views, decisions or policies of the institutions with which they are affiliated.