Deep Air Quality Forecasts: Suspended Particulate Matter Modeling With Convolutional Neural and Long Short-Term Memory Networks

Abstract

Public health risks arising from airborne pollutants, e.g., Total Suspended Particulate (TSP) matter, can significantly elevate ongoing and future healthcare costs. The chaotic behaviour of air pollutants posing major difficulties in tracking their three-dimensional movements over diverse temporal domains is a significant challenge in designing practical air quality systems. This research paper builds a deep learning hybrid CLSTM model where convolutional neural network (CNN) is amalgamed with the long short-term memory (LSTM) network to forecast hourly TSP. The CNN model entails a data processer including feature extractors that draw upon statistically significant antecedent lagged predictor variables, whereas the LSTM model encapsulates a new feature mapping scheme to predict the next hourly TSP value. The hybrid CLSTM model is comprehensively benchmarked and is seen to outperform an ensemble of five machine learning models. The efficacy of the CLSTM model is elucidated in model testing phase at study sites in Queensland, Australia. Using performance metrics, visual analysis of TSP simulations relative to observations, and detailed error analysis, this study ascertains the CLSTM model's practical utility for air pollutant forecasting systems in health risk mitigation. This study captures a feasible opportunity to emulate air quality at relatively high temporal resolutions in global regions where air pollution is a considerable threat to public health.