EGU24-10056, updated on 08 Mar 2024
https://doi.org/10.5194/egusphere-egu24-10056
EGU General Assembly 2024
© Author(s) 2024. This work is distributed under
the Creative Commons Attribution 4.0 License.

 Air Emission Inventory and AI- based Air Quality Forecasting Models for Developing Countries: A Case Study of Ho Chi Minh City, Vietnam

Quoc Bang Ho1,2, Khue Hoang Ngoc Vu1,3, Tam Thoai Nguyen1, and Ricardo Simon Carbajo4
Quoc Bang Ho et al.
  • 1Vietnam National University Ho Chi Minh City (VNU-HCM), Institute for Environment & Resources (IER), Air Pollution and Climate Change Research Center (APAC), Ho Chi Minh city, Viet Nam (hqbang@vnuhcm.edu.vn)
  • 2Department of Academic Affairs, Vietnam National University, Ho Chi Minh City 700000, Vietnam
  • 3Global Change Research Institute, The Czech Academy of Sciences
  • 4Ireland’s National Centre for Applied Artificial Intelligence (CeADAR), University College Dublin, NexusUCD, Belfield Office Park, Dublin, Ireland

Outdoor air pollution damages the climate and causes many diseases, including cardiovascular diseases, respiratory infections, and lung damage. Understating of air pollution sources and accurate hourly forecasting of air pollution concentrations is thus of significant importance for public health, helping the citizens to plan the measures to alleviate the harmful effects of air pollution on health. This study conducts air emision inventory (EI), analyses and discusses the temporal characteristics of air pollutants at different locations in Ho Chi Minh City (HCMC), Vietnam - an economic center and a megacity in a developing country with a population of 8.99 million people and more than 8 million of private vehicles.

A combination of bottom-up and top-down approaches was employed to conduct air pollution EI, in which EMISENS model was utilized to generate the EI for road traffic sources. The results showed that the motorcycles were the main reasons of emission in HCMC, contributing 90% of CO, 68% of non-methane volatile organic compounds (NMVOC), 63% of CH4, 41% of SO2, 29% of NOx, and 18% of patriculate matter (PM2.5).

We developed several AI-based one-shot multi-step PM2.5 forecasting models, with both an hourly forecast granularity (1h to 24h) and a 24-hour rolling mean. These Machine Learning algorithms include Stochastic Gradient Descent Regressor, hybrid 1D CNN-LSTM, eXtreme Gradient Boosting Regressor, and Prophet. We collected the data from six monitoring stations installed by the HealthyAir project partners at different locations in HCMC, including traffic, residential and industrial areas in the city. In addition, we developed a suitable model training protocol using data from a short period to address the non-stationarity of PM2.5 time series. Our proposed PM2.5 forecasting models achieve state-of-the-art accuracy and will be deployed in our HealthyAir mobile app to warn HCMC citizens of air pollution issues in the city. 

How to cite: Ho, Q. B., Vu, K. H. N., Nguyen, T. T., and Carbajo, R. S.:  Air Emission Inventory and AI- based Air Quality Forecasting Models for Developing Countries: A Case Study of Ho Chi Minh City, Vietnam, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10056, https://doi.org/10.5194/egusphere-egu24-10056, 2024.