Air Pollution Problem in Chiang Mai Since 1995
A few years ago I started studying the air pollution data in Thailand. I wrote blogs about my finding, but that was just a starting point. I got in touch with Thai PCD for more data and was collaborating with UN-ESCAP Thailand in building air pollution models. Some of our report can be found here.
Air pollution data is rich. For Chiang Mai, Thai PCD start collecting air pollution data since 1995, the air pollutants was PM10, O3, CO, SO2. PM2.5 data came much latter at around 2021.
In the picture above, I converted the raw air pollution concentration to AQI index using US AQI standard so that I can compare different air pollutants side by side.
I learnt two things from this plots. First, the air pollution problem has been there since 1995. Let’s focus on PM10. The data shows that PM10 level has seasonal behavior and is in the moderate to unhealthy for sensitive groups range from the beginning.
Second, because PM2.5 and PM10 value are highly correlated (correlation coefficient = 0.96), we can use PM10 pattern to infer the value of PM2.5. As shown in the correlation plot below. Since PM10 pattern has not change since 1995, we can then infer that PM2.5 has had the same pattern since then. Which mean the PM2.5 was in the unhealthy range since 1995!
In fact, I build an ML model to fill back PM2.5 the historical data using the PM10 and got the same conclusion.
It is now widely accepted that the air pollution problem in Chiang Mai is from agricultural burning. The next logical question is was the burning the cause of air pollution since back then? The answer is yes.
The picture below shows average PM2.5, PM10 and total number of hotspots during winter season(December — March) for different year. I use the averages from these months focus on the air pollution seasons. Although we don’t have the hotspot data since 1995 (MODIS data from NASA start in 2002), we can already see the high correlation among these values. In the year with high burning activities, the air pollution level, both PM2.5 and PM10, are high. In the year with low burning activities such as 2010 and 2016, the air pollution problem improved.
Thank you for reading. Until next time!
The code can be found in my GitHub page. I would like to thank Mr. Matthew Perkins from UNESCAP for the collaboration and Thai PCD for the data.
My air pollution blogs:
https://www.unescap.org/blog/deciphering-black-box-air-pollution-data-thailand
https://medium.com/towards-data-science/identifying-the-sources-of-winter-air-pollution-in-bangkok-part-i-d4392ea608dc
