Air pollution poses a significant threat to both human health and the environment. Pollutants such as nitrogen dioxide (NO2), sulfur dioxide (SO2), and carbon monoxide (CO) are linked to a range of respiratory diseases and premature deaths, resulting in heightened healthcare costs and decreased productivity. Moreover, air pollution contributes to climate change, reduces agricultural productivity, and threatens biodiversity. Addressing this urgent issue is crucial for safeguarding public health and ensuring a sustainable future for our planet. Studies have explored the relationship between air pollutants and various socio-economic processes, as well as the dynamics of pollutant concentrations over extended temporal scales, there remains a significant gap in understanding these dynamics over shorter (annual) timeframes. This study aims to assess the spatio-temporal dynamics of NO2, SO2, and carbon monoxide (CO) in the North Island of New Zealand. We conducted our analysis using the cloud computing platform Google Earth Engine, the JavaScript APIs, and the extensive catalogue of remotely sensed earth observation data from the Sentinel-5P/TROPOMI satellite sensor. Additionally, a hybrid artificial neural network (ANN) model, integrating a long short-term memory (LSTM) model and a graph convolutional network (GCN) known as ConvLSTM was employed to forecast future pollutant levels for the year 2028. The results indicated a general decline in NO2 and CO concentrations, contrasted with a notable increase in SO2 levels across the study area. A decreased concentration of all three air pollutants was observed between 2020 and 2021 during the COVID-19 lockdown period. The corresponding Root Mean Square Errors (RMSE) were 0.779, 0.616, and 0.787 for NO2, SO2, and CO, respectively. This study offers valuable insights into air quality trends in the North Island from 2019 to 2023, providing critical information for policymakers and health practitioners. These insights are essential for formulating effective air quality management strategies to help reduce healthcare costs, improve public health outcomes, and guide policy interventions to enhance environmental sustainability.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

ANN-Based Pollution Forecasting Through Short-Term Spatio-Temporal Analysis: A North Island, New Zealand Case Study

  • Sara Zandi,
  • Aldridge Nyasha Mazhindu,
  • Akbar Ghobakhlou

摘要

Air pollution poses a significant threat to both human health and the environment. Pollutants such as nitrogen dioxide (NO2), sulfur dioxide (SO2), and carbon monoxide (CO) are linked to a range of respiratory diseases and premature deaths, resulting in heightened healthcare costs and decreased productivity. Moreover, air pollution contributes to climate change, reduces agricultural productivity, and threatens biodiversity. Addressing this urgent issue is crucial for safeguarding public health and ensuring a sustainable future for our planet. Studies have explored the relationship between air pollutants and various socio-economic processes, as well as the dynamics of pollutant concentrations over extended temporal scales, there remains a significant gap in understanding these dynamics over shorter (annual) timeframes. This study aims to assess the spatio-temporal dynamics of NO2, SO2, and carbon monoxide (CO) in the North Island of New Zealand. We conducted our analysis using the cloud computing platform Google Earth Engine, the JavaScript APIs, and the extensive catalogue of remotely sensed earth observation data from the Sentinel-5P/TROPOMI satellite sensor. Additionally, a hybrid artificial neural network (ANN) model, integrating a long short-term memory (LSTM) model and a graph convolutional network (GCN) known as ConvLSTM was employed to forecast future pollutant levels for the year 2028. The results indicated a general decline in NO2 and CO concentrations, contrasted with a notable increase in SO2 levels across the study area. A decreased concentration of all three air pollutants was observed between 2020 and 2021 during the COVID-19 lockdown period. The corresponding Root Mean Square Errors (RMSE) were 0.779, 0.616, and 0.787 for NO2, SO2, and CO, respectively. This study offers valuable insights into air quality trends in the North Island from 2019 to 2023, providing critical information for policymakers and health practitioners. These insights are essential for formulating effective air quality management strategies to help reduce healthcare costs, improve public health outcomes, and guide policy interventions to enhance environmental sustainability.