Global estimates of PM 2.5 from 2001-2006. (Source: van Donkelaar and Martin, 2010).

Global estimates of PM 2.5 from 2001-2006. (Source: van Donkelaar and Martin, 2010).


Canadian scientists Aaron van Donkelaar and Randall Martin at Dalhousie University, Halifax, Nova Scotia, Canada, have created a map of global PM 2.5 (air particulates with diameters 2.5 microns or less, which is less than 1/3 the width of a strand of human hair and more damaging to the lungs than larger particles) concentrations from 2001 and 2006 using satellite information from the MODIS and MISR sensors. They employed a method of combining total-column aerosol measurements obtained from the above-mentioned NASA satellites along with information regarding the vertical distribution of aerosols from a computer model. This map – which shows PM 2.5 concentrations from 2001 to 2006 – offers the most comprehensive look at global PM 2.5 concentrations around the world, according to NASA.

Why use satellite data instead of ground-level measurements? There are several considerations, which the authors point out in their article in Environmental Health Perspectives:

– Data for PM 2.5 concentrations do not often exist in developing countries. China, for example, which one can clearly see saturated in red in van Donkelaar and Martin’s map, does not measure PM 2.5, although post-Olympics, Beijing announced that they would start doing so and have since installed PM 2.5 air monitoring equipment.

– PM 2.5 measurements from satellite data could actually improve estimates of population exposure to the pollutant, argue the authors.

Of course, the use of satellite-derived data has its caveats: previous studies using satellite data to accurately gauge measurements of particles near the surface have had their fair share of challenges, due to the fact that the sensors measure particles in the air column and have trouble differentiating particles close to the ground from those higher in the atmosphere.  Additionally, as with all types of satellite image acquisition, data are generally limited by weather (clouds, snow, ice), albedo (reflected light), and bright land surfaces (desert, urban structures), which can all affect measurements.

The researchers also base a lot of their accuracy on ground-truthing the satellite results with ground-level measurements from the U.S. and Canada, where PM 2.5 is routinely measured. While this is certainly an improvement over previous chemical transport models mentioned in their article, the authors admit that data in developing countries in particular aren’t available to further refine and test the accuracy of their results. Remote sensing with Chinese characteristics, anyone?

A more detailed look at PM 2.5 concentrations over Eastern Asia:

From A. van Donkelaar and R. Martin (2010). Contours denote population density and surface elevation.

From A. van Donkelaar and R. Martin (2010). Contours denote population density and surface elevation.

The dark red areas over Northeastern China certainly has had the China blogosphere buzzing apocalyptically about the health impacts of such high levels of PM 2.5 concentrations. Despite daily readings of PM 2.5 and Ozone from a non-Chinese government monitoring station (Follow them on Twitter: @BeijingAir) that already provide information regarding these particulates to citizens with web access and a VPN in China, it sometimes takes a visual representation of the same or similar data, such as van Donkelaar and Martin’s global map, to realize that those of us who live in China are in the red on this one.

I’d like to dig into van Donkelaar and Martin’s methods to see if similar techniques might be able to be used to obtain measures of regional air quality in Eastern China. It would certainly help to localize pollution sources to some extent and provide more detail than is given with this map, whose spatial resolution is about 10 km by 10 km (to give some context, Beijing municipality is 16,000 km2).