Estimating the spatiotemporal variation of NO2 concentration using an adaptive neuro-fuzzy inference system
Bijan Jeganeh, Michael G. Hewson, Samuel Clifford, Ahmad Tavassoli, Luke D. Knibbs, Lidia Morawska
Environmental Modelling & Software
Volume 100, February 2018, Pages 222–235
The study was able to combine various data to estimate average annual NO2 concentrations in south west Queensland including Brisbane in 2011 (right). Darker colours mean higher NO2 concentrations. The Australian maximum national standard for average annual NO2 concentration is 30 ppb (parts per billion)
What did the study find?
The study developed and used a cutting-edge method called ‘adaptive neuro-fuzzy inference system’ (ANFIS) to calculate nitrogen dioxide (NO2) pollution in urban and rural areas. ANFIS is an artificial intelligence-based method which can analyse complicated phenomena. The study combined data from satellites, traffic congestion, meteorology and geography to estimate the concentration of NO2 in South-east Queensland. The study is the first to use data on traffic flow and congestion to estimate NO2 and did so because traffic that is ‘stuck’ produces higher emissions than that which is free-flowing.
Why does it matter?
NO2 is a major component of air pollution and is a strong indicator of emissions from traffic. It has been shown that exposure to NO2 affects people’s health in negative ways. NO2 is normally measured by sensors on the ground. But the sparseness of these sensors means that we aren’t able to measure NO2 accurately. This novel technique will allow scientists and governments to more accurately determine air pollution levels in our cities and in turn estimate the adverse health effects on its residents. The more data we have about pollutants, the more we can do to protect the health of our populations.
CAR members involved in this study
Bijan Yeganeh (first author)