A systematic literature review of deep learning neural network for time series air quality forecasting

Environmental Science and Pollution Research - Rapid progress of industrial development, urbanization and traffic has caused air quality reduction that negatively affects human health and...     

Deep learning architecture for air quality predictions

With the rapid development of urbanization and industrialization, many developing countries are suffering from heavy air pollution. Governments and citizens have expressed increasing concern regarding air pollution because it affects human health and sustainable development worldwide. Current air quality prediction methods mainly use shallow models; however, these methods produce unsatisfactory results, which inspired us to investigate methods of predicting air quality based on deep architecture models. In this paper, a novel spatiotemporal deep learning (STDL)-based air quality prediction method that inherently considers spatial and temporal correlations is proposed. A stacked autoencoder (SAE) model is used to extract inherent air quality features, and it is trained in a greedy layer-wise manner. Compared with traditional time series prediction models, our model can predict the air quality of all stations simultaneously and shows the temporal stability in all seasons. Moreover, a comparison with the spatiotemporal artificial neural network (STANN), auto regression moving average (ARMA), and support vector regression (SVR) models demonstrates that the proposed method of performing air quality predictions has a superior performance.     

Forecasting of particulate matter time series using wavelet analysis and wavelet-ARMA/ARIMA model in Taiyuan,China

Particulate matter with aerodynamic diameter below 10 μm (PM₁₀) forecasting is difficult because of the uncertainties in describing the emission and meteorological fields. This paper proposed a wavelet-ARMA/ARIMA model to forecast the short-term series of the PM₁₀ concentrations. It was evaluated by experiments using a 10-year data set of daily PM₁₀ concentrations from 4 stations located in Taiyuan, China. The results indicated the following: (1) PM₁₀ concentrations of Taiyuan had a decreasing trend during 2005 to 2012 but increased in 2013. PM₁₀ concentrations had an obvious seasonal fluctuation related to coal-fired heating in winter and early spring. (2) Spatial differences among the four stations showed that the PM₁₀ concentrations in industrial and heavily trafficked areas were higher than those in residential and suburb areas. (3) Wavelet analysis revealed that the trend variation and the changes of the PM₁₀ concentration of Taiyuan were complicated. (4) The proposed wavelet-ARIMA model could be efficiently and successfully applied to the PM₁₀ forecasting field. Compared with the traditional ARMA/ARIMA methods, this wavelet-ARMA/ARIMA method could effectively reduce the forecasting error, improve the prediction accuracy, and realize multiple-time-scale prediction.

Implications: Wavelet analysis can filter noisy signals and identify the variation trend and the fluctuation of the PM₁₀ time-series data. Wavelet decomposition and reconstruction reduce the nonstationarity of the PM₁₀ time-series data, and thus improve the accuracy of the prediction. This paper proposed a wavelet-ARMA/ARIMA model to forecast the PM₁₀ time series. Compared with the traditional ARMA/ARIMA method, this wavelet-ARMA/ARIMA method could effectively reduce the forecasting error, improve the prediction accuracy, and realize multiple-time-scale prediction. The proposed model could be efficiently and successfully applied to the PM₁₀ forecasting field.     

Single-source impact analysis using three-dimensional air quality models

Isolating the effects of an individual emissions source on secondary air pollutants such as ozone and some components of particulate matter must incorporate complex nonlinear processes, be sensitive to small emissions perturbations, and account for impacts that may occur hundreds of kilometers away. The ability to evaluate these impacts is becoming increasingly important for efficient air quality management. Here, as part of a recent compliance enforcement action for a violation of the Clean Air Act and as an evaluation of ozone response to single-source emissions plumes, two three-dimensional regional photochemical air quality models are used to assess the impact on ozone from approximately 2000 to 3000 excess t/month of nitrogen oxides emitted from a single power plant in Ohio. Periods in May, July, and August are evaluated. Two sensitivity methods are applied: the "brute-force" (B-F) method and the decoupled direct method (DDM). Using DDM, maximum 1-hr averaged ozone concentrations are found to increase by up to 1.8, 1.3, and 2.2 ppbv during May, July, and August episodes, respectively, and concentration increases greater than 0.5 ppbv occur in Ohio, Pennsylvania, Maryland, New York, West Virginia, Virginia, and North and South Carolina. B-F results for the August episode show a maximum 1-hr averaged ozone concentration increase of 2.3 ppbv. Significant localized decreases are also simulated, with a maximum of 3.6 ppbv in Ohio during the August episode and decreases of 0.50 ppbv and greater in Ohio, Pennsylvania, Maryland, West Virginia, and Virginia. Maximum increases are compared with maximum decreases for the August period using second-order DDM and are found, in aggregate, to be greater in magnitude by 42%. When evaluated during hours when ozone concentrations exceed 0.060 ppm, the maximum increases in ozone are higher than decreases by 82%. The spatial extent of ozone increase in both cases is about triple that of reduction.     

A methodology to urban air quality assessment during large time periods of winter using computational fluid dynamic models

The representativeness of point measurements in urban areas is limited due to the strong heterogeneity of the atmospheric flows in cities. To get information on air quality in the gaps between measurement points, and have a 3D field of pollutant concentration, Computational Fluid Dynamic (CFD) models can be used. However, unsteady simulations during time periods of the order of months, often required for regulatory purposes, are not possible for computational reasons. The main objective of this study is to develop a methodology to evaluate the air quality in a real urban area during large time periods by means of steady CFD simulations. One steady simulation for each inlet wind direction was performed and factors like the number of cars inside each street, the length of streets and the wind speed and direction were taken into account to compute the pollutant concentration. This approach is only valid in winter time when the pollutant concentrations are less affected by atmospheric chemistry. A model based on the steady-state Reynolds-Averaged Navier–Stokes equations (RANS) and standard k-? turbulence model was used to simulate a set of 16 different inlet wind directions over a real urban area (downtown Pamplona, Spain). The temporal series of NO_x and PM₁₀ and the spatial differences in pollutant concentration of NO₂ and BTEX obtained were in agreement with experimental data. Inside urban canopy, an important influence of urban boundary layer dynamics on the pollutant concentration patterns was observed. Large concentration differences between different zones of the same square were found. This showed that concentration levels measured by an automatic monitoring station depend on its location in the street or square, and a modelling methodology like this is useful to complement the experimental information. On the other hand, this methodology can also be applied to evaluate abatement strategies by redistributing traffic emissions.     

GIS-based air quality modelling: spatial prediction of PM10 for Selangor State,Malaysia using machine learning algorithms

Environmental Science and Pollution Research - Rapid urbanization has caused severe deterioration of air quality globally, leading to increased hospitalization and premature deaths. Therefore,...     

Modelling of ecological status of Polish lakes using deep learning techniques

Environmental Science and Pollution Research - Since 2000, after the Water Framework Directive came into force, aquatic ecosystems’ bioassessment has acquired immense practical importance for...     

Economic damages of ozone air pollution to crops using combined air quality and GIS modelling

This study aims at presenting a combined air quality and GIS modelling methodological approach in order to estimate crop damages from photochemical air pollution, depict their spatial resolution and assess the order of magnitude regarding the corresponding economic damages. The analysis is conducted within the Greater Thessaloniki Area, Greece, a Mediterranean territory which is characterised by high levels of photochemical air pollution and considerable agricultural activity. Ozone concentration fields for 2002 and for specific emission reduction scenarios for the year 2010 were estimated with the Ozone Fine Structure model in the area under consideration. Total economic damage to crops turns out to be significant and estimated to be approximately 43 M€ for the reference year. Production of cotton presents the highest economic loss, which is over 16 M€, followed by table tomato (9 M€), rice (4.2 M€), wheat (4 M€) and oilseed rape (2.8 M€) cultivations. Losses are not spread uniformly among farmers and the major losses occur in areas with valuable ozone-sensitive crops. The results are very useful for highlighting the magnitude of the total economic impacts of photochemical air pollution to the area’s agricultural sector and can potentially be used for comparison with studies worldwide. Furthermore, spatial analysis of the economic damage could be of importance for governmental authorities and decision makers since it provides an indicative insight, especially if the economic instruments such as financial incentives or state subsidies to farmers are considered.     

Urban air quality

Since 1950 the world population has more than doubled, and the global number of cars has increased by a factor of 10. In the same period the fraction of people living in urban areas has increased by a factor of 4. In year 2000 this will amount to nearly half of the world population. About 20 urban regions will each have populations above 10 million people.Seen over longer periods, pollution in major cities tends to increase during the built up phase, they pass through a maximum and are then again reduced, as abatement strategies are developed. In the industrialised western world urban air pollution is in some respects in the last stage with effectively reduced levels of sulphur dioxide and soot. In recent decades however, the increasing traffic has switched the attention to nitrogen oxides, organic compounds and small particles. In some cities photochemical air pollution is an important urban problem, but in the northern part of Europe it is a large-scale phenomenon, with ozone levels in urban streets being normally lower than in rural areas. Cities in Eastern Europe have been (and in many cases still are) heavily polluted. After the recent political upheaval, followed by a temporary recession and a subsequent introduction of new technologies, the situation appears to improve. However, the rising number of private cars is an emerging problem. In most developing countries the rapid urbanisation has so far resulted in uncontrolled growth and deteriorating environment. Air pollution levels are here still rising on many fronts.Apart from being sources of local air pollution, urban activities are significant contributors to transboundary pollution and to the rising global concentrations of greenhouse gasses. Attempts to solve urban problems by introducing cleaner, more energy-efficient technologies will generally have a beneficial impact on these large-scale problems. Attempts based on city planning with a spreading of the activities, on the other hand, may generate more traffic and may thus have the opposite effect.     

Characterization of urban air quality using GIS as a management system

Keeping the air quality acceptable has become an important task for decision makers as well as for non-governmental organizations. Particulate and gaseous emissions of pollutant from industries and auto-exhausts are responsible for rising discomfort, increasing airway diseases, decreasing productivity and the deterioration of artistic and cultural patrimony in urban centers. A model to determine the air quality in urban areas using a geographical information system will be presented here. This system permits the integration, handling, analysis and simulation of spatial and temporal data of the ambient concentration of the main pollutant. It allows the users to characterize and recognize areas with a potential increase or improvement in its air pollution situation. It is also possible to compute past or present conditions by changing basic input information as traffic flow, or stack emission rates. Additionally the model may be used to test the compliance of local standard air quality, to study the environmental impact of new industries or to determine the changes in the conditions when the vehicle circulation is increased.     

利用地衣移植法对乌鲁木齐市大气质量的评价

利用地衣移植法把中国树花移植到乌鲁木齐市7个监测点,用电感耦合等离子光谱直读仪检测中国树花体内8种重金属含量.结果显示,随着移植时间的延长,不同监测点的中国树花体内重金属含量明显增长.Nemerow综合污染指数显示,红雁池(23.79)＞钢铁厂(11.87)＞开发区(8.07)＞儿童公园(6.95)＞人民公园(5.93...     

Analysis of air quality data near roadways using a dispersion model

We used a dispersion model to analyze measurements made during a field study conducted by the U.S. EPA in July–August 2006, to estimate the impact of traffic emissions on air quality at distances of tens of meters from an eight-lane highway located in Raleigh, NC. The air quality measurements consisted of long path optical measurements of NO at distances of 7 and 17 m from the edge of the highway. Sonic anemometers were used to measure wind speed and turbulent velocities at 6 and 20 m from the highway. Traffic flow rates were monitored using traffic surveillance cameras. The dispersion model [Venkatram, A., 2004. On estimating emissions through horizontal fluxes. Atmospheric Environment 38, 2439–2446] explained over 60% of the variance of the observed path averaged NO concentrations, and over 90% of the observed concentrations were within a factor of two of the model estimates.Sensitivity tests conducted with the model indicated that the traffic flow rate made the largest contribution to the variance of the observed NO concentrations. The meteorological variable that had the largest impact on the near road NO concentrations was the standard deviation of the vertical velocity fluctuations, σ_w. Wind speed had a relatively minor effect on concentrations. Furthermore, as long as the wind direction was within ±45° from the normal to the road, wind direction had little impact on near road concentrations. The measurements did not allow us to draw conclusions on the impact of traffic-induced turbulence on dispersion. The analysis of air quality and meteorological observations resulted in plausible estimates of on-road emission factors for NO.     

The differential impact of air pollutants on acute urticaria and chronic urticaria: a time series analysis

Environmental Science and Pollution Research - Several studies have revealed a relationship between short-term exposure to air pollution and the exacerbation of certain skin conditions. This study...     

An investigation of distributed lag models in the context of air pollution and mortality time series analysis

In particulate air pollution mortality time series studies, the particulate air pollution exposure measure used is typically the current day's or the previous day's air pollution concentration or a multi-day moving average air pollution concentration. Distributed lag models (DLMs) that allow for differential air pollution effects that are spread over multiple days are seen as an improvement over using a single- or multi-day moving average air pollution exposure measure. However, at the current time, the statistical properties of DLMs as a measure of air pollution exposure have not been investigated. In this paper, a simulation study is used to investigate the performance of DLMs as a measure of air pollution exposure in comparison with single- and multi-day moving average air pollution exposure measures under various forms for the true effect of air pollution on mortality. The simulation study shows that DLMs offer a more robust measure of the effect of air pollution on mortality and avoid the potential for a large negative bias compared with single- or multi-day moving average air pollution exposure measures. This is important information. In many U.S. cities, particulate air pollution concentrations are observed only once every six days, meaning it is often only possible to use single-day particulate air pollution exposure measures. The results from this paper will help quantify the magnitude of the negative bias that can result from using single-day exposure measures. The implications of this work for future air pollution mortality time series studies are discussed. The data used in this paper are concurrent daily time series of mortality, weather, and particulate air pollution from Cook County, IL, for the period 1987-1994.     

A model study for ambient air quality analysis using routine meteorological observations

In the present study, more realistic and easily adaptable input parameters have been used with a view to investigating the long-range air quality analysis for the dispersion of air pollutants emitted from an area source with a multiple box model. The model formulation has been discussed at length for the ground level sources when convective conditions prevail. The routine meteorological observations have been used for the computation of sensible surface heat flux, friction velocity and mixing depth. A radiation model provides the estimates of the sensible surface heat flux. Based on the similarity theory, an iterative procedure has been adopted for the estimation of friction velocity which provides a coupling of radiation computation and the surface layer of the planetary boundary layer through surface heat flux expression. The important parameters—wind speed and eddy diffusivity profiles—have been derived and have been used to obtain the concentration patterns as hourly averages. The procedure could be easily adopted where observed meteorological parameters may be used for studying the dispersal of pollutants from the ground level sources.