Tropospheric Ozone Formation Estimation in Urban City, Bangi, Using Artificial Neural Network (ANN)

Area of Science:

• Environmental Science
• Atmospheric Chemistry
• Artificial Intelligence

Background:

• Urban areas globally face rising ground-level ozone, a harmful air pollutant, due to economic and societal development.
• Accurate air pollutant concentration monitoring is crucial for public health and environmental agencies' precautionary measures.

Purpose of the Study:

• To apply artificial neural networks (ANNs) for estimating and forecasting ozone concentration in Bangi.
• To identify key factors influencing ozone levels and assess the efficacy of a single ANN model.

Main Methods:

• Utilized an artificial neural network (ANN) framework with input variables including temperature, humidity, nitrogen dioxide, time, and UV radiation.
• Employed a ten-hidden-layer ANN architecture to optimize and predict ozone concentration as the output.

Main Results:

• Meteorological conditions and emission patterns significantly influence ground-level ozone concentration.
• A single ANN model proved sufficient for reliable ozone concentration estimation under various circumstances.

Conclusions:

• Artificial neural networks provide reliable and satisfactory estimations for daily ozone concentration forecasting.
• ANNs are a viable tool for environmental agencies to monitor and predict air quality.