Development of a deep neural network for predicting 6 h average PM_2.5 concentrations up to 2 subsequent days using various training data

Abstract

Abstract. Despite recent progress of numerical air quality models, accurate prediction of fine particulate matter (PM2.5) is still challenging because of uncertainties in physical and chemical parameterizations, meteorological data, and emission inventory databases. Recent advances in artificial neural networks can be used to overcome limitations in numerical air quality models. In this study, a deep neural network (DNN) model was developed for a 3 d forecasting of 6 h average PM2.5 concentrations: the day of prediction (D+0), 1 d after prediction (D+1), and 2 d after prediction (D+2). The DNN model was evaluated against the currently operational Community Multiscale Air Quality (CMAQ) modeling system in South Korea. Our study demonstrated that the DNN model outperformed the CMAQ modeling results. The DNN model provided better forecasting skills by reducing the root-mean-squared error (RMSE) by 4.1, 2.2, and 3.0 µg m−3 for the 3 consecutive days, respectively, compared with the CMAQ. Also, the false-alarm rate (FAR) decreased by 16.9 %p (D+0), 7.5 %p (D+1), and 7.6 %p (D+2), indicating that the DNN model substantially mitigated the overprediction of the CMAQ in high PM2.5 concentrations. These results showed that the DNN model outperformed the CMAQ model when it was simultaneously trained by using the observation and forecasting data from the numerical air quality models. Notably, the forecasting data provided more benefits to the DNN modeling results as the forecasting days increased. Our results suggest that our data-driven machine learning approach can be a useful tool for air quality forecasting when it is implemented with air quality models together by reducing model-oriented systematic biases.

PM2.5 Estimation using Machine Learning Models and Satellite Data: A Literature Review

“…The feed-forward neural network [69] and Recurrent Neural Network (RNN) [41] are some of the fundamental algorithms to simulate the temporal variation of PM2.5 concentration by describing the stratigraphic characteristics of the predicted area. Observation data from monitoring stations in the forecast area and surrounding areas are utilized to develop Convolutional Neural Network (CNN) and Graph Neural Network (GNN) models that directly capture transportation characteristics [41], [70].…”

“…Theoretically, the convolutional LSTM (ConvLSTM) network structure makes it an ideal algorithm for combining transportation and formation features; however, these features cannot be accurately predicted after 12 hours [71]. The ensemble technique of Deep Neural Network (DNN) [69], RNN, CNN algorithmic models for real-time estimation of PM 2.5 is considered capable of reducing the average bias and improving the accuracy index of models that are substantially limited by the uncertainties in the input data of anthropogenic emissions and meteorological fields, as well as the inherent limitations of each model [65].…”

Forecast of Fine Particles in Chengdu under Autumn–Winter Synoptic Conditions

“…For instance, various studies have examined the direct impact of meteorological factor forecasts on the prediction of atmospheric pollutants [ 20 ], the utilization of micro-pulse lidar observations [ 21 ], the incorporation of Global Positioning System Zenith Total Delay data [ 22 ], and the implementation of an urban canopy parameterization scheme [ 23 ] to enhance the accuracy of fine particle forecasting using the CMAQ model. Researchers have also used machine learning [ 24 ], four-dimensional variational assimilation [ 25 ], and Kalman filter [ 26 ] methods to optimize fine particle forecasts by the CMAQ model. Also, the CMAQ model has been used to analyze the impact of changes in meteorological conditions on the regional transmission of fine particles [ 27 , 28 ], thereby facilitating policy-making related to the control of regional fine particulate emissions [ 29 ].…”

An optimized semi-empirical physical approach for satellite-based PM2.5 retrieval: embedding machine learning to simulate complex physical parameters

“…To date, machine learning (ML) has developed rapidly. It can detect complex nonlinear relationships of multiple data and model their interaction (Yuan et al, 2020;Lee et al,2022), which provides an idea for improving the accuracy of physical parameter acquisition, thereby estimating high-precision PM2.5 through semi-physical empirical models.…”