Data-Driven Temporal-Spatial Model for the Prediction of AQI in Nanjing

AbstractAir quality data prediction in urban area is of great significance to control air pollution and protect the public health. The prediction of the air quality in the monitoring station is well studied in existing researches. However, air-quality-monitor stations are insufficient in most cities and the air quality varies from one place to another dramatically due to complex factors. A novel model is established in this paper to estimate and predict the Air Quality Index (AQI) of the areas without monitoring stations in Nanjing. The proposed model predicts AQI in a non-monitoring area both in temporal dimension and in spatial dimension respectively. The temporal dimension model is presented at first based on the enhanced k-Nearest Neighbor (KNN) algorithm to predict the AQI values among monitoring stations, the acceptability of the results achieves 92% for one-hour prediction. Meanwhile, in order to forecast the evolution of air quality in the spatial dimension, the method is utilized with the help of Back Propagation neural network (BP), which considers geographical distance. Furthermore, to improve the accuracy and adaptability of the spatial model, the similarity of topological structure is introduced. Especially, the temporal-spatial model is built and its adaptability is tested on a specific non-monitoring site, Jiulonghu Campus of Southeast University. The result demonstrates that the acceptability achieves 73.8% on average. The current paper provides strong evidence suggesting that the proposed non-parametric and data-driven approach for air quality forecasting provides promising results.

Download Full-text

Probabilistic Deep Learning to Quantify Uncertainty in Air Quality Forecasting

Sensors ◽

10.3390/s21238009 ◽

2021 ◽

Vol 21 (23) ◽

pp. 8009

Author(s):

Abdulmajid Murad ◽

Frank Alexander Kraemer ◽

Kerstin Bach ◽

Gavin Taylor

Keyword(s):

Deep Learning ◽

Air Quality ◽

Probabilistic Models ◽

Performance Metrics ◽

Current Data ◽

Data Driven ◽

Quality Data ◽

Empirical Performance ◽

Adversarial Training ◽

Scalable Methods

Data-driven forecasts of air quality have recently achieved more accurate short-term predictions. However, despite their success, most of the current data-driven solutions lack proper quantifications of model uncertainty that communicate how much to trust the forecasts. Recently, several practical tools to estimate uncertainty have been developed in probabilistic deep learning. However, there have not been empirical applications and extensive comparisons of these tools in the domain of air quality forecasts. Therefore, this work applies state-of-the-art techniques of uncertainty quantification in a real-world setting of air quality forecasts. Through extensive experiments, we describe training probabilistic models and evaluate their predictive uncertainties based on empirical performance, reliability of confidence estimate, and practical applicability. We also propose improving these models using “free” adversarial training and exploiting temporal and spatial correlation inherent in air quality data. Our experiments demonstrate that the proposed models perform better than previous works in quantifying uncertainty in data-driven air quality forecasts. Overall, Bayesian neural networks provide a more reliable uncertainty estimate but can be challenging to implement and scale. Other scalable methods, such as deep ensemble, Monte Carlo (MC) dropout, and stochastic weight averaging-Gaussian (SWAG), can perform well if applied correctly but with different tradeoffs and slight variations in performance metrics. Finally, our results show the practical impact of uncertainty estimation and demonstrate that, indeed, probabilistic models are more suitable for making informed decisions.

Download Full-text

A Cluster Analysis of PM2.5 Using CMAQ Model Results for Representativeness of Air Quality Monitoring Networks in Busan, Korea

10.5194/egusphere-egu2020-3189 ◽

2020 ◽

Author(s):

Woo-Sik Jung ◽

Woo-Gon Do

Keyword(s):

Air Pollution ◽

Cluster Analysis ◽

Air Quality ◽

Monitoring Network ◽

Hierarchical Cluster ◽

Quality Monitoring ◽

Quality Data ◽

Air Quality Monitoring ◽

Monitoring Networks ◽

Monitoring Stations

<p><strong>With increasing interest in air pollution, the installation of air quality monitoring networks for regular measurement is considered a very important task in many countries. However, operation of air quality monitoring networks requires much time and money. Therefore, the representativeness of the locations of air quality monitoring networks is an important issue that has been studied by many groups worldwide. Most such studies are based on statistical analysis or the use of geographic information systems (GIS) in existing air quality monitoring network data. These methods are useful for identifying the representativeness of existing measuring networks, but they cannot verify the need to add new monitoring stations. With the development of computer technology, numerical air quality models such as CMAQ have become increasingly important in analyzing and diagnosing air pollution. In this study, PM2.5 distributions in Busan were reproduced with 1-km grid spacing by the CMAQ model. The model results reflected actual PM2.5 changes relatively well. A cluster analysis, which is a statistical method that groups similar objects together, was then applied to the hourly PM2.5 concentration for all grids in the model domain. Similarities and differences between objects can be measured in several ways. K-means clustering uses a non-hierarchical cluster analysis method featuring an advantageously low calculation time for the fast processing of large amounts of data. K-means clustering was highly prevalent in existing studies that grouped air quality data according to the same characteristics. As a result of the cluster analysis, PM2.5 pollution in Busan was successfully divided into groups with the same concentration change characteristics. Finally, the redundancy of the monitoring stations and the need for additional sites were analyzed by comparing the clusters of PM2.5 with the locations of the air quality monitoring networks currently in operation.</strong></p><p><strong>This research was supported by Basic Science Research Program through the National Research Foundation of Korea(NRF) funded by the Ministry of Education(2017R1D1A3B03036152).</strong></p>

Download Full-text

Ambient air quality data reported at Sasol Secunda monitoring stations during COVID-19 lockdown – Mpumalanga, South Africa

Clean Air Journal ◽

10.17159/caj/2021/31/1.8954 ◽

2021 ◽

Vol 31 (1) ◽

Author(s):

Andrew Venter ◽

Sandra De Vos

Keyword(s):

South Africa ◽

Air Quality ◽

Ambient Air ◽

Human Behaviour ◽

Quality Data ◽

Ambient Air Quality ◽

Industrial Facilities ◽

Ambient Temperatures ◽

Pollutant Concentrations ◽

Monitoring Stations

Various local and international research has been published on the effects of COVID-19 lockdown on ambient air quality. In most cases, a reduction in ambient NOx and PM concentrations have been observed with varying changes in ambient SO2 levels. Secunda, located in the Highveld Priority Area in Mpumalanga, South Africa is known for its large industrial facilities utilising coal as primary feedstock. The towns of Secunda and eMbalenhle provide the majority of the workforce to Sasol and has therefore been the focus of this study. The ambient air quality in the Secunda region was assessed due to the changes in human behaviour during lockdown, familiarity with the Sasol facility and the strategic locations of ambient air quality stations.Results show a clear decrease in ambient CO, NO2 and PM concentrations, especially during the first two weeks of lockdown. Only subtle changes were observed for ambient H2S and SO2 pollutant concentrations at the ambient monitoring stations. An increasing trend in all ambient species was observed towards the end and post lockdown, in contrast to declining ambient temperatures with the onset of winter. This is also contrary to the reduction in emissions from the factory that conducted annual maintenance in the month following lockdown (phase shutdown). This article concludes that human behaviour has a material local ambient impact on CO, NO2 and PM pollutant species, while H2S concentration profiles are more directly related to the industrial complex’s levels of activity. Ambient SO2 trends did not show a similar correlation with the facility’s activities (as H2S), but a stronger correlation was observed with the diverse local and regional sources in close proximity to Secunda and eMbalenhle. The influence of better dispersion especially on a local scale, brought about by more effective emission heights, is considered material. Moreover, meteorological factors, on local air quality, has been shown to be a material contributor to observed ambient air quality levels in the study domain

Download Full-text

Evidence of air quality data misreporting in China: An impulse indicator saturation model comparison of local government-reported and U.S. embassy-reported PM2.5 concentrations (2015–2017)

PLoS ONE ◽

10.1371/journal.pone.0249063 ◽

2021 ◽

Vol 16 (4) ◽

pp. e0249063

Author(s):

Jesse S. Turiel ◽

Robert K. Kaufmann

Keyword(s):

Air Quality ◽

Local Government ◽

Model Comparison ◽

Quality Data ◽

Random Chance ◽

Pollution Levels ◽

Reported Data ◽

Impulse Indicator Saturation ◽

Air Quality Data ◽

Monitoring Stations

This paper analyzes hourly PM2.5 measurements from government-controlled and U.S. embassy-controlled monitoring stations in five Chinese cities between January 2015 and June 2017. We compare the two datasets with an impulse indicator saturation technique that identifies hours when the relation between Chinese and U.S. reported data diverges in a statistically significant fashion. These temporary divergences, or impulses, are 1) More frequent than expected by random chance; 2) More positive than expected by random chance; and 3) More likely to occur during hours when air pollution concentrations are high. In other words, relative to U.S.-controlled monitoring stations, government-controlled stations systematically under-report pollution levels when local air quality is poor. These results contrast with the findings of other recent studies, which argue that Chinese air quality data misreporting ended after a series of policy reforms beginning in 2012. Our findings provide evidence that local government misreporting did not end after 2012, but instead continued in a different manner. These results suggest that Chinese air quality data, while still useful, should not be taken entirely at face value.

Download Full-text

VARMA-EGARCH Model for Air-Quality Analyses and Application in Southern Taiwan

Atmosphere ◽

10.3390/atmos11101096 ◽

2020 ◽

Vol 11 (10) ◽

pp. 1096

Author(s):

Edward Ming-Yang Wu ◽

Shu-Lung Kuo

Keyword(s):

Time Series ◽

Air Quality ◽

Air Pollutants ◽

Quality Monitoring ◽

Quality Data ◽

Air Quality Monitoring ◽

Egarch Model ◽

Southern Taiwan ◽

Air Quality Monitoring Stations ◽

Monitoring Stations

This study adopted the Exponential Generalized Autoregressive Conditional Heteroscedasticity (EGARCH) model to analyze seven air pollutants (or the seven variables in this study) from ten air quality monitoring stations in the Kaohsiung–Pingtung Air Pollutant Control Area located in southern Taiwan. Before the verification analysis of the EGARCH model is conducted, the air quality data collected at the ten air quality monitoring stations in the Kaohsiung–Pingtung area are classified into three major factors using the factor analyses in multiple statistical analyses. The factors with the most significance are then selected as the targets for conducting investigations; they are termed “photochemical pollution factors”, or factors related to pollution caused by air pollutants, including particulate matter with particles below 10 microns (PM10), ozone (O3) and nitrogen dioxide (NO2). Then, we applied the Vector Autoregressive Moving Average-EGARCH (VARMA-EGARCH) model under the condition where the standardized residual existed in order to study the relationships among three air pollutants and how their concentration changed in the time series. By simulating the optimal model, namely VARMA (1,1)-EGARCH (1,1), we found that when O3 was the dependent variable, the concentration of O3 was not affected by the concentration of PM10 and NO2 in the same term. In terms of the impact response analysis on the predictive power of the three air pollutants in the time series, we found that the asymmetry effect of NO2 was the most significant, meaning that NO2 influenced the GARCH effect the least when the change of seasons caused the NO2 concentration to fluctuate; it also suggested that the concentration of NO2 produced in this area and the degree of change are lower than those of the other two air pollutants. This research is the first of its kind in the world to adopt a VARMA-EGARCH model to explore the interplay among various air pollutants and reactions triggered by it over time. The results of this study can be referenced by authorities for planning air quality total quantity control, applying and examining various air quality models, simulating the allowable increase in air quality limits, and evaluating the benefit of air quality improvement.

Download Full-text

Developing Web-Based 3D Health-Aware Routing for Pedestrians and Cyclists

E3S Web of Conferences ◽

10.1051/e3sconf/202017102009 ◽

2020 ◽

Vol 171 ◽

pp. 02009

Author(s):

Rosanny Sihombing ◽

Sabo Kwada Sini ◽

Matthias Fitzky

Keyword(s):

Air Pollution ◽

Air Quality ◽

Health Concern ◽

Quality Data ◽

Quality Level ◽

3D Navigation ◽

Index Numbers ◽

Web Based ◽

Air Quality Data ◽

Monitoring Stations

As the population of people migrating to cities keeps increasing, concerns have been raised about air quality in cities and how it impacts everyday life. Thus, it is important to demonstrate ways of avoiding polluted areas. The approach described in this paper is intended to draw attention to polluted areas and help pedestrians and cyclists to achieve the lowest possible level of air pollution when planning daily routes. We utilise real-time air quality data which is obtained from monitoring stations across the world. The data consist of the geolocation of monitoring stations as well as index numbers to scale the air quality level in every corresponding monitoring stations. When the air quality level is considered having a moderate health concern for people with respiratory disease, such as asthma, an alternative route that avoid air pollution will be calculated so that pedestrians and cyclists can be informed. The implementation can visualize air quality level in several areas in 3D map as well as informs health-aware route for pedestrian and cyclist. It automatically adjusts the observed air quality areas based on the availability of monitoring stations. The proposed approach results in a prototype of a health-aware 3D navigation system for pedestrian and cyclist.

Download Full-text

The representativeness of ground-based air quality monitoring stations: observation and recommendation from Indian cities

10.5194/egusphere-egu21-13197 ◽

2021 ◽

Author(s):

Arindam Roy ◽

Satoshi Takahama ◽

Athanasios Nenes ◽

Sumit Sharma ◽

Anju Goel

Keyword(s):

Land Use ◽

Air Quality ◽

Premature Mortality ◽

Quality Monitoring ◽

Quality Data ◽

Population Exposure ◽

Air Quality Monitoring ◽

The City ◽

Indian Cities ◽

Monitoring Stations

<div> <p>It is well established that the high level of particulate matter is a leading cause of premature mortality and disease worldwide and especially in South Asia (Global Burden of Disease Study, 2019). The ground-based air quality (AQ) monitoring stations are used&#160;to calculate economic loss, premature mortality and validate the conversed PM2.5&#160;concentration from satellite-based Aerosol Optical Depth (AOD) data. Over India, 793 manual monitoring air quality (AQ) monitoring stations and 307 automated AQ monitoring station are presently operating under the aegis of National Air Quality Monitoring&#160;Programme&#160;and Central Pollution Control Board respectively. However, studies addressing the spatial representativeness of the data generated from the AQ monitoring stations over India are very limited and therefore, it is unclear that whether the existing stations are sufficient to reflect the average ambient AQ over different Indian cities.&#160;</p> </div><div> <p>The present study intends to classify the existing AQ monitoring stations on the basis of spatial representativeness and derive a general conceptual framework for commissioning representative AQ monitoring sites for Indian cities. The methodology involves analysis of land use, populations and air quality data for the existing air quality stations in million plus Indian cities. A case study was conducted for Pune (18.5&#176; N, 73.8&#176; E), a western Indian metro city with 3.15 million population (Census, 2011). Using the night-time light data and high resolution PM2.5, population exposure hotspots over Pune city were identified. It was observed that not only at the midst of the municipal area, population exposure hotspots can be identified at the peripheral region of PMC/PNMC which certainly signify the role of rapid developmental activity and urban agglomeration over Pune city. The existing air quality monitoring sites are located majorly in the pollution hotspots in the city center region and therefore installing AQ monitoring stations (co-located&#160; with weather station) at the rapidly developing parts of the city is highly recommended. The present land use pattern and the location of existing monitoring sites suggests lack of urban background monitoring stations which indicates the gap of knowledge in monitoring the average air quality responsible of long-term health effect over Pune. The prevalence of AQ monitoring stations in the road junction points and near to metro construction works might overestimate the exposure estimate of the general population in the city.&#160;&#160;&#160;</p> </div>

Download Full-text

Modelling NO<sub>2</sub> concentrations at the street level in the GAINS integrated assessment model: projections under current legislation

Atmospheric Chemistry and Physics Discussions ◽

10.5194/acpd-13-22687-2013 ◽

2013 ◽

Vol 13 (8) ◽

pp. 22687-22732

Author(s):

G. Kiesewetter ◽

J. Borken-Kleefeld ◽

W. Schöpp ◽

C. Heyes ◽

P. Thunis ◽

...

Keyword(s):

Air Quality ◽

Integrated Assessment ◽

Integrated Assessment Model ◽

Assessment Model ◽

Quality Data ◽

Limit Values ◽

Street Level ◽

Regional Background ◽

The Eu ◽

Monitoring Stations

Abstract. NO2 concentrations at the street level are a major concern for urban air quality in Europe and have been regulated under the EU Thematic Strategy on Air Pollution. Despite the legal requirements, limit values are exceeded at many monitoring stations with little or no improvement during recent years. In order to assess the effects of future emission control regulations on roadside NO2 concentrations, a downscaling module has been implemented in the GAINS integrated assessment model. The module follows a hybrid approach based on atmospheric dispersion calculations and observations from the AirBase European air quality data base that are used to estimate site-specific parameters. Pollutant concentrations at every monitoring site with sufficient data coverage are disaggregated into contributions from regional background, urban increment, and local roadside increment. The future evolution of each contribution is assessed with a model of the appropriate scale – 28 × 28 km grid based on the EMEP Model for the regional background, 7 × 7 km urban increment based on the CHIMERE Chemistry Transport Model, and a chemical box model for the roadside increment. Thus, different emission scenarios and control options for long-range transport, regional and local emissions can be analysed. Observed concentrations and historical trends are well captured, in particular the differing NO2 and total NOx = NO + NO2 trends. Altogether, more than 1950 air quality monitoring stations in the EU are covered by the model, including more than 400 traffic stations and 70% of the critical stations. Together with its well-established bottom-up emission and dispersion calculation scheme, GAINS is thus able to bridge the scales from European-wide policies to impacts in street canyons. As an application of the model, we assess the evolution of attainment of NO2 limit values under current legislation until 2030. Strong improvements are expected with the introduction of the Euro 6 emission standard for light duty vehicles; however, for some major European cities, further measures may be required, in particular if aiming to achieve compliance at an earlier time.

Download Full-text

Multivariate data driven prediction of COVID-19 dynamics: Towards new results with temperature, humidity and air quality data

Environmental Research ◽

10.1016/j.envres.2021.112348 ◽

2021 ◽

pp. 112348

Author(s):

Dunfrey P. Aragão ◽

Emerson V. Oliveira ◽

Arthur A. Bezerra ◽

Davi H. dos Santos ◽

Andouglas G. da Silva Junior ◽

...

Keyword(s):

Air Quality ◽

Multivariate Data ◽

Data Driven ◽

Quality Data ◽

Air Quality Data

Download Full-text

Short-term environmental impact of ambient air quality trends in during the COVID-19 pandemic in India

Annals of Civil and Environmental Engineering ◽

10.29328/journal.acee.1001031 ◽

2021 ◽

Vol 5 (1) ◽

pp. 017-025

Author(s):

Karuppasamy Manikanda Bharath ◽

Natesan Usha ◽

Periyasamy Balamadeswaran ◽

S Srinivasalu

Keyword(s):

Air Quality ◽

Ambient Air ◽

Air Pollutant ◽

Quality Monitoring ◽

Quality Data ◽

Air Quality Monitoring ◽

The North ◽

Air Quality Monitoring Stations ◽

Air Quality Data ◽

Monitoring Stations

The lockdown, implemented in response to the COVID-19 epidemic, restricted the operation of various sectors in the country and its highlights a good environmental outcome. Thus, a comparison of air pollutants in India before and after the imposed lockdown indicated an overall improvement air quality across major Indian cities. This was established by utilizing the Central Pollution Control Board’s database of air quality monitoring station statistics, such as air quality patterns. During the COVID-19 epidemic, India’s pre-to-post nationwide lockdown was examined. The air quality data was collected from 30-12-2019 to 28-04-2020 and synthesized using 231 Automatic air quality monitoring stations in a major Indian metropolis. Specifically, air pollutant concentrations, temperature, and relative humidity variation during COVID-19 pandemic pre-to-post lockdown variation in India were monitored. As an outcome, several cities around the country have reported improved air quality. Generally, the air quality, on a categorical scale was found to be ‘Good’. However, a few cities from the North-eastern part of India were categorized as ‘Moderate/Satisfactory’. Overall, the particulate matters reduction was in around 60% and other gaseous pollutants was in 40% reduction was observed during the lockdown period. The results of this study include an analysis of air quality data derived from continuous air quality monitoring stations from the pre-lockdown to post-lockdown period. Air quality in India improved following the national lockdown, the interpretation of trends for PM 2.5, PM 10, SO2, NO2, and the Air Quality Index has been provided in studies for major cities across India, including Delhi, Gurugram, Noida, Mumbai, Kolkata, Bengaluru, Patna, and others.

Download Full-text