19 December 2014

Quality control of aerosol measurements filters out important readings

Posted by kwheeling

By Kim Smuga-Otto

Sky cover over South Korea for July 13, 2014 (circled is Seoul) from TERRA-MODIS  Credit: NASA

Sky cover over South Korea for July 13, 2014 (circled is Seoul) from TERRA-MODIS
Credit: NASA

The hardworking AERONET (AErosol RObotic NETwork) instrument in Baengyeong, South Korea was having a rough day. Every 15 minutes, the telescope-like device pointed its barrel at the sun to record its light and measure how much was blocked by airborne particles, or aerosols. July 13, 2012 was an overcast day and the light absorbed by the clouds dominated the measurements. But then, just after 1 p.m., the clouds parted, the instrument looked up, and data was collected.

Only no one saw it.

Heavy clouds absorb so much light that the AERONET unit can’t detect aerosols and these useless measurements are discarded. On days where there are less than three successful readings, the whole day is filtered out and hidden from most users.

That’s what happened on July 13, 2012. But, when NASA atmospheric scientist Thomas Eck went through the raw data from China and South Korea’s AERONET he “found it misidentified a lot of high pollution data as clouds”including the data from July 13.

Developing countries without robust ground-based emission detection systems rely on atmospheric measurements to detect and measure aerosols, which are a major component of air pollution. Underreporting of aerosol levels because clouds interfere with AERONET’s readings can hinder health monitoring and cause computer models that use this data to inaccurately predict pollution levels, according to a poster presented by Eck at the American Geophysical Union’s Fall Meeting in San Francisco.

There are over 400 AERONET units worldwide and for more than 20 years they have been providing public data on aerosol optical measurements.

When Eck looked at data gathered from the first five months of 2012 from a unit about 40 miles (60 kilometers) outside of Beijing, he found only one day of extreme pollution levels —the type where the smog is so thick, you can barely see. But when he went back to the original, unfiltered data, he found five such days.

These five days were cloudy and  only had one or two readings. Eck explained the data was probably collected during brief gaps in the clouds.

The cloudiness and high concentrations of aerosols might not be coincidental, Eck said. Pollution can build up in a low wind area and the rotating of a cloud mass around it can concentrate the pollution, Eck explained.

Kim Smuga-Otto is a science communication graduate student at UC Santa Cruz.