The behaviour of the multi-layer perceptron and the support vector regression learning methods in the prediction of NO and NO2 concentrations in Szeged, Hungary

The main aim of this paper is to predict NO and NO₂ concentrations 4 days in advance by comparing two artificial intelligence learning methods, namely, multi-layer perceptron and support vector machines, on two kinds of spatial embedding of the temporal time series. Hourly values of NO and NO₂ concentrations, as well as meteorological variables were recorded in a cross-road monitoring station with heavy traffic in Szeged, in order to build a model for predicting NO and NO₂ concentrations several hours in advance. The prediction of NO and NO₂ concentrations was performed partly on the basis of their past values, and partly on the basis of temperature, humidity and wind speed data. Since NO can be predicted more accurately, its values were considered primarily when forecasting NO₂. Time series prediction can be interpreted in a way that is suitable for artificial intelligence learning. Two effective learning methods, namely, multi-layer perceptron and support vector regression are used to provide efficient non-linear models for NO and NO₂ time series predictions. Multi-layer perceptron is widely used to predict these time series, but support vector regression has not yet been applied for predicting NO and NO₂ concentrations. Three commonly used linear algorithms were considered as references: 1-day persistence, average of several day persistence and linear regression. Based on the good results of the average of several day persistence, a prediction scheme was introduced, which forms weighted averages instead of simple ones. The optimization of these weights was performed with linear regression in linear case and with the learning methods mentioned in non-linear case. Concerning the NO predictions, the non-linear learning methods give significantly better predictions than the reference linear methods. In the case of NO₂, the improvement of the prediction is considerable, however, it is less notable than for NO.     

Application of a parallel photochemical air quality model to the Campania region (southern Italy)

During most of the year, the concentrations of both primary and secondary air pollutants over the Campania region (southern Italy) do not comply with the Italian air quality standards. To gain insight into the chemical and meteorological processes that lead to high air pollutant concentrations over this area, the parallel package PNAM (Parallel Naples Airshed Model) has been developed, for the numerical simulation of photosmog episodes on urban and regional scale domains. PNAM has been applied to a photosmog episode which occurred on 26 July 1995. On this day, due to the stagnant conditions and the intensive solar radiation, a high ozone concentration was reported for the Naples basin. The performance of PNAM has been assessed by comparing measured air quality data with simulated data for O₃, NO, NO₂ and CO. PNAM was able to reproduce temporal and spatial characteristics of measured air quality data, although some discrepancies were evident, probably mainly due to the emission inventory, which was based only on total annual emissions.     

Forecasting of traffic origin NO and NO2 concentrations by Support Vector Machines and neural networks using Principal Component Analysis

The main aim of this paper is to predict NO and NO₂ concentrations four days in advance comparing two artificial intelligence learning methods, namely, Multi-Layer Perceptron and Support Vector Machines on two kinds of spatial embedding of the temporal time series. Hourly values of NO and NO₂ concentrations, as well as meteorological variables were recorded in a cross-road monitoring station with heavy traffic in Szeged in order to build a model for predicting NO and NO₂ concentrations several hours in advance. The prediction of NO and NO₂ concentrations was performed partly on the basis of their past values, and partly on the basis of temperature, humidity and wind speed data. Since NO can be predicted more accurately, its values were considered primarily when forecasting NO₂. Time series prediction can be interpreted in a way that is suitable for artificial intelligence learning. Two effective learning methods, namely, Multi-Layer Perceptron and Support Vector Regression are used to provide efficient non-linear models for NO and NO₂ times series predictions. Multi-Layer Perceptron is widely used to predict these time series, but Support Vector Regression has not yet been applied for predicting NO and NO₂ concentrations. Grid search is applied to select the best parameters for the learners. To get rid of the curse of dimensionality of the spatial embedding of the time series Principal Component Analysis is taken to reduce the dimension of the embedded data. Three commonly used linear algorithms were considered as references: one-day persistence, average of several-day persistence and linear regression. Based on the good results of the average of several-day persistence, a prediction scheme was introduced, which forms weighted averages instead of simple ones. The optimization of these weights was performed with linear regression in linear case and with the learning methods mentioned in non-linear case. Concerning the NO predictions, the non-linear learning methods give significantly better predictions than the reference linear methods. In the case of NO₂ the improvement of the prediction is considerable; however, it is less notable than for NO.     

Prediction of tropospheric ozone concentrations: Application of a methodology based on the Darwin’s Theory of Evolution

This study aims to predict the next day hourly average tropospheric ozone (O₃) concentrations using genetic programming (GP). Due to the complexity of this problem, GP is an adequate methodology as it can optimize, simultaneously, the structure of the model and its parameters. It is an artificial intelligence methodology that uses the same principles of the Darwinian Theory of Evolution. GP enables the automatic generation of mathematical expressions that are modified following an iterative process applying genetic operations.The inputs of the models were the hourly average concentrations of carbon monoxide (CO), nitrogen oxide (NO), nitrogen dioxide (NO₂) and O₃, and some meteorological variables (temperature – T; solar radiation – SR; relative humidity – RH; and wind speed – WS) measured 24 h before. GP was also applied to the principal components (PC) obtained from these variables. The analysed period was from May to July 2004 divided in training and test periods.GP was able to select the most relevant variables for prediction of O₃ concentrations. The original variables, T, RH and O₃ measured 24 h before were considered significant inputs for prediction. The selected PC had also important contributions of the same variables and of NO₂. GP models using the original variables presented better performance in training period and worse performance in test period when compared with the models obtained using PC. The results achieved using the GP methodology demonstrated that it can be very useful to solve several environmental complex problems.     

Two-dimensional tomographic retrieval of MIPAS/ENVISAT measurements of ozone and related species

Observations from the Michelson Interferometer for Passive Atmospheric Sounding (MIPAS) were analysed with the two-dimensional GMTR retrieval system in order to obtain fields of ozone and several molecular species related to ozone chemistry: HNO₃, N₂O, NO₂, N₂O₅, ClONO₂, COF₂, CFC-11 and CFC-12. MIPAS measures mid-infrared emission of the atmosphere both during the day and at night time with global coverage. Observing the atmosphere with limb viewing geometries, the instrument is able to resolve finer vertical structures than with nadir instruments, thus enabling the investigation of ozone height-dependent processes. With the currently planned mission extended up to 2014, MIPAS can provide both short-term resolution and long-term trends needed for studying ozone. The adopted GMTR algorithm permits us to resolve the horizontal inhomogeneities of the atmosphere that are modelled using a two-dimensional discretization of the atmosphere. It is therefore especially suitable for analysing portions of the atmosphere where strong gradients such as at the ozone hole may be poorly reproduced by common horizontal homogeneous one-dimensional retrievals. The adopted strategy is well suited for a refined analysis and a correct monitoring of the ozone recovery, as required by the Montreal Protocol and successive amendments.     

Contour diagram fuzzy model for maximum surface ozone prediction

A contour diagram approach is presented for the identification of surface ozone concentration feature based on a set of rules by considering the meteorological variables such as the solar radiation, wind speed, temperature, humidity and rainfall. A fuzzy rule system approach is used because of the imprecise, insufficient, ambiguous and uncertain data available. The contour diagrams help to identify qualitative ozone concentration variability rules which are more general than conventional statistical or time series analysis. In the methodology, ozone concentration contours are based on a fixed variable as ozone precursor, namely, NO_x and as the third variable one of the meteorological factors. Such contour diagrams for ozone concentration variation are prepared for six months. It is possible to identify the maximum ozone concentration episodes from these diagrams and then to set up the valid rules in the form of IF-THEN logical statements. These rules are obtained from available daily ozone, NO_x and meteorological data as a first approximate reasoning step. In this manner, without mathematical formulations, expert maximum ozone concentration systems are identified. The application of the contour diagram approach is performed for daily ozone concentration measurements on European side of Istanbul city. It is concluded that through approximate reasoning with fuzzy rules, the maximum ozone concentration episodes can be identified and predicted without any mathematical expression.     

Analysis of spatial and seasonal distributions of air pollutants by incorporating urban morphological characteristics

Due to the worldwide trend of industrialization and urbanization, air pollutants were emitted heavily on a global scale particularly in developing countries, which produces adverse effects on human health by causing health problems such as respiratory and lung diseases. Many regression models based on land use types and urban fabrics have been built to analyze the spatiotemporal distribution of air pollutants, however, few of them examined the relationship between urban morphological characteristics and the distribution of air pollutants in a megacity. This study investigates such relationships for six types of air pollutants (PM2.5, PM10, SO₂, NO₂, O₃, and CO) and a composite AQI (Air Quality Index) based on hourly data at 35 monitoring stations in Beijing in 2016, with morphological characteristics (Morphological building index), meteorological factors (Land Surface Temperature, LST), land use (vegetation, road length, gas station and industry point data), and population distribution data. We also analyzed the results with spatiotemporal regression and SSH (Spatial Stratified Heterogeneity) models respectively. According to the spatiotemporal regression model, the morphological building index (MBI) shows a strong correlation with the dispersion of PM2.5 (R² = 0.81) and AQI (R² = 0.80) in the warm season and this finding was reinforced through the Leave-one-out-cross-validation (LOOCV) analysis. From the SSH analysis, the road length in a large proximal region impacts air pollutants the most, especially for O₃; and population density significantly affects PM 2.5, AQI, SO₂, and NO₂ in the cold season. From an integrated interpretation, distance to nearest industry impacts the spatial distribution of NO₂ in cold season, while it impacts that of PM2.5 and AQI in both warm and cold seasons. The research finds that these two models supplement each other well and together help to give us a better understanding of how air quality is affected in the urban landscape.     

Spatiotemporal variations of tropospheric column nitrogen dioxide over Jing-Jin-Ji during the past decade

Of all anthropogenic pollutants, nitrogen dioxide (NO₂) has the most negative effect on atmospheric chemistry. In this study, measurements of tropospheric column NO₂ obtained from the ozone monitoring instrument (OMI) are used to investigate temporal and spatial dynamics of NO₂. Temporal and spatial distributions of tropospheric NO₂ concentrations obtained from OMI over the Beijing-Tianjin-Hebei (Jing-Jin-Ji) region from 2007 to 2016 are presented, and annual changes and trends in the seasonal cycle are shown. Annual amounts of NO₂ are found to firstly increase then decrease, where after reaching a maximum in 2012 they begin a progressive yearly decline. NO₂ shows significant cyclical seasonal characteristics over Jing-Jin-Ji, with maximum values in winter and minimum in summer. In addition, the spatial distribution is unbalanced, and Beijing-Tianjin-Tangshan and Shijiazhuang-Xingtai-Handan are found to be highly polluted areas. The many complex factors affecting variations in NO₂ are analysed in this article, and the impact of meteorological factors and human activities are emphasized. It is considered that temperature and precipitation are natural factors influencing NO₂ concentration but there is a stronger negative relationship between tropospheric column NO₂ and temperature. Optimization of the energy structure is thus considered to be important and a reduction in energy consumption is required to control the concentration of pollutants. Coal combustion is a major anthropogenic factor in increasing NO₂ concentrations, and there is a strong correlation between higher amounts of NO₂ and coal consumption in the Jing-Jin-Ji region.     

Development of a land use regression model for daily NO2 and NOx concentrations in the Brisbane metropolitan area,Australia

Land use regression models are an established method for estimating spatial variability in gaseous pollutant levels across urban areas. Existing LUR models have been developed to predict annual average concentrations of airborne pollutants. None of those models have been developed to predict daily average concentrations, which are useful in health studies focused on the acute impacts of air pollution. In this study, we developed LUR models to predict daily NO₂ and NO_x concentrations during 2009–2012 in the Brisbane Metropolitan Area (BMA), Australia's third-largest city. The final models explained 64% and 70% of spatial variability in NO₂ and NO_x, respectively, with leave-one-out-cross-validation R² of 3–49% and 2–51%. Distance to major road and industrial area were the common predictor variables for both NO₂ and NO_x, suggesting an important role for road traffic and industrial emissions. The novel modeling approach adopted here can be applied in other urban locations in epidemiological studies.     

Impacts of urban form on air quality in metropolitan areas in the United States

Among 221 metropolitan areas (MAs) in the United States (US), this study explored the impact of urban form, either urban compactness or urban sprawl, on two types of air quality in 2014: NO_x emissions from road traffic and annual average NO2 concentrations. Urban form was quantified using Smart Growth America (SGA) sprawl indexes with density, land use mixing, centeredness, and street connectivity. NO_x emissions from road traffic were derived from the National Emissions Inventory (NEI). Through modeling NO2 concentrations using land use regression (LUR), with satellite-based estimates and kriging, this study measured NO₂ concentrations within MAs in the US The study results showed that higher levels of urban form scores (i.e., higher compactness) and land use mixing were associated with lower per-person NO_x emissions from road traffic. In addition, higher levels of centeredness were associated with lower NO₂ concentrations, but the effect was moderate. On the other hand, regional rainfall and solar insolation had more significant associations with NO₂ concentrations than metropolitan urban form. Meanwhile, localized emissions sources had significant associations with local-level NO₂ concentrations. This study provides additional evidence on the relationship between urban form and air quality in the US MAs. The study suggests that high compactness-oriented development and the reduction of localized emission sources may be effective in reducing NO_x emissions from road traffic and local NO₂ concentrations, respectively. However, future studies need to explore the impact of urban form at both the MA and local levels on NO₂ concentrations and develop a more accurate national NO₂ concentration prediction model.     

The use of AVHRR data to determine the concentration of visible and invisible tropospheric pollutants originating from a 1997 forest fire in Southeast Asia

A massive forest fire in Indonesia in 1997 affected the whole Asian region by producing a large smoke plume, with Malaysia bearing the brunt due to the wind direction and weather conditions and because of its proximity to the source. The five primary fire produced pollutants were carbon monoxide (CO), sulphur dioxide (SO₂), nitrogen dioxide (NO₂), ozone (O₃) and particulate matter less than 10 µm (PM10). The first four of these are, of course, invisible to conventional satellite-flown multispectral scanners operating in the visible and near infrared regions of the electromagnetic spectrum. The fifth, PM10, is present in the haze and therefore makes an observable contribution to the signal received by the Advanced Very High Resolution Radiometer (AVHRR). The haze in AVHRR channels 1 and 2 data for the fires of September 1997 has been used to study the concentration of PM10 directly. It has also been used to study the concentration indirectly--as a tracer or surrogate--for the four remaining materials, the gases CO, SO₂, NO₂ and O₃. Data from ground observations have been used to calibrate the results and the distributions of the fire pollutants over Peninsular Malaysia have been plotted.     

Assessment and prediction of tropospheric ozone concentration levels using artificial neural networks

This work deals specifically with the use of a neural network for ozone modelling in the lower atmosphere. The development of a neural network model is presented to predict the tropospheric (surface or ground) ozone concentrations as a function of meteorological conditions and various air quality parameters. The development of the model was based on the realization that the prediction of ozone from a theoretical basis (i.e. detailed atmospheric diffusion model) is difficult. In contrast, neural networks are useful for modelling because of their ability to be trained using historical data and because of their capability for modelling highly non-linear relationships. The network was trained using summer meteorological and air quality data when the ozone concentrations are the highest. The data were collected from an urban atmosphere. The site was selected to represent a typical residential area with high traffic influences. Three neural network models were developed. The main emphasis of the first model has been placed on studying the factors that control the ozone concentrations during a 24-hour period (daylight and night hours were included). The second model was developed to study the factors that regulate the ozone concentrations during daylight hours at which higher concentrations of ozone were recorded. The third model was developed to predict daily maximum ozone levels. The predictions of the models were found to be consistent with observations. A partitioning method of the connection weights of the network was used to study the relative percent contribution of each of the input variables. The contribution of meteorology on the ozone concentration variation was found to fall within the range 33.15–40.64%. It was also found that nitrogen oxide, sulfur dioxide, relative humidity, non-methane hydrocarbon and nitrogen dioxide have the most effect on the predicted ozone concentrations. In addition, temperature played an important role while solar radiation had a lower effect than expected. The results of this study indicate that the artificial neural network (ANN) is a promising method for air pollution modelling.     

基于多元线性回归的雾霾预测方法研究

提出了一种在线样本更新的多元线性回归分析的雾霾预测方法。首先搜集了北京市天气状况,包括平均气温、湿度、风级等气象数据以及PM2.5、CO、NO₂、SO₂等大气成分浓度数据,然后通过散点图对这些因素进行主要影响因素分析,筛选出对雾霾影响比较明显的因素作为雾霾预测的依据。通过在线样本更新的多元线性回归建立了PM2.5含量预测模型,并将气象要素作为雾霾的判断标准。最后给出实际例子,利用多元线性回归对北京未来一天、三天及一周的PM2.5含量进行较为精确的预测。     

Forecasting of the daily meteorological pollution using wavelets and support vector machine

The paper presents the method of daily air pollution forecasting by using support vector machine (SVM) and wavelet decomposition. Based on the observed data of NO₂, CO, SO₂ and dust, for the past years and actual meteorological parameters, like wind, temperature, humidity and pressure, we propose the forecasting approach, applying the neural network of SVM type, working in the regression mode. To obtain the acceptable accuracy of forecast we decompose the measured time series data into wavelet representation and predict the wavelet coefficients. On the basis of these predicted values the final forecasting is prepared. The paper presents the results of numerical experiments on the basis of the measurements made by the meteorological stations, situated in the northern region of Poland.     

Space–time modelling of air quality for environmental-risk maps: A case study in South Portugal

Since the 1960s, there has been a strong industrial development in the Sines area, on the southern Atlantic coast of Portugal, including the construction of an important industrial harbour and of, mainly, petrochemical and energy-related industries. These industries are, nowadays, responsible for substantial emissions of SO₂, NO_x, particles, VOCs and part of the ozone polluting the atmosphere. The major industries are spatially concentrated in a restricted area, very close to populated areas and natural resources such as those protected by the European Natura 2000 network. Air quality parameters are measured at the emissions’ sources and at a few monitoring stations. Although air quality parameters are measured on an hourly basis, the lack of representativeness in space of these non-homogeneous phenomena makes even their representativeness in time questionable. Hence, in this study, the regional spatial dispersion of contaminants is also evaluated, using diffusive-sampler (Radiello Passive Sampler) campaigns during given periods. Diffusive samplers cover the entire space extensively, but just for a limited period of time.In the first step of this study, a space–time model of pollutants was built, based on a stochastic simulation—direct sequential simulation—with local spatial trend. The spatial dispersion of the contaminants for a given period of time—corresponding to the exposure time of the diffusive samplers—was computed by ordinary kriging. Direct sequential simulation was applied to produce equiprobable spatial maps for each day of that period, using the kriged map as a spatial trend and the daily measurements of pollutants from the monitoring stations as hard data.In the second step, the following environmental risk and costs maps were computed from the set of simulated realizations of pollutants: (i) maps of the contribution of each emission to the pollutant concentration at any spatial location; (ii) costs of badly located monitoring stations.     

Enhanced SO2 concentrations observed over northern India: role of long-range transport

The combustion of fossil fuels (coal and petroleum products) constitutes a source of continuous release of anthropogenic SO₂ into the atmosphere. Furthermore, natural sources such as volcanoes can inject large amounts of SO₂ directly into the troposphere and sometimes even into the stratosphere. These event-based volcanic eruptions provide solitary opportunities to study the transport and transformation of atmospheric constituents. In this study, we present an episode of high SO₂ concentration over northern India as a result of long-range transport from Africa using multiple satellite observations. Monthly averaged column SO₂ values over the Indo-Gangetic Plain (IGP) were observed in the range of 0.6–0.9 Dobson units (DU) during November 2008 using observations from the Ozone Monitoring Instrument (OMI). These concentrations were conspicuously higher than the background concentrations (<0.3 DU) observed during 2005–2010 over this region. The columnar SO₂ loadings were highest on 6 November over most of the IGP region and even exceeded 6 DU, a factor of 10–20 higher than background levels in some places. These enhanced SO₂ levels were not reciprocated in satellite-derived NO₂ or CO columns, indicating transport from a non-anthropogenic SO₂ source. As most of the local aerosols over the IGP region occur below 3 km, a well-separated layer at 4–5 km was observed from the Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations (CALIPSO) satellite. Wind fields and back-trajectory analysis revealed a strong flow originating from the Dalaffilla volcanic eruption in Ethiopia during 4–6 November 2008. Although volcanic SO₂ plumes have been extensively studied over many parts of Asia, Europe, and the USA, analysis of such events for the IGP region is being reported for the first time in this study.     

Short-term ozone forecasting by artificial neural networks

In this work we report preliminary results of a study aiming to develop an intelligent tool for performing ozone forecasting in the polluted atmosphere of México City. This tool is based in the paradigm of neural networks. Two neural models are used in this work, namely, the Bidirectional Associative Memory (BAM) and the Holographic Associative Memory (HAM). We analyse and preprocess daily patterns of meteorological variables and concentrations of pollutants as measured by five monitoring stations in México City. These patterns are used to train both neural networks and then we use them to predict ozone at one point in the city. Preliminary results are reported and some conclusions are drawn.     

Estimated global nitrogen deposition using NO2 column density

Global nitrogen deposition has increased over the past 100 years. Monitoring and simulation studies of nitrogen deposition have evaluated nitrogen deposition at both the global and regional scale. With the development of remote-sensing instruments, tropospheric NO₂ column density retrieved from Global Ozone Monitoring Experiment (GOME) and Scanning Imaging Absorption Spectrometer for Atmospheric Chartography (SCIAMACHY) sensors now provides us with a new opportunity to understand changes in reactive nitrogen in the atmosphere. The concentration of NO₂ in the atmosphere has a significant effect on atmospheric nitrogen deposition. According to the general nitrogen deposition calculation method, we use the principal component regression method to evaluate global nitrogen deposition based on global NO₂ column density and meteorological data. From the accuracy of the simulation, about 70% of the land area of the Earth passed a significance test of regression. In addition, NO₂ column density has a significant influence on regression results over 44% of global land. The simulated results show that global average nitrogen deposition was 0.34 g m^?2 yr^?1 from 1996 to 2009 and is increasing at about 1% per year. Our simulated results show that China, Europe, and the USA are the three hotspots of nitrogen deposition according to previous research findings. In this study, Southern Asia was found to be another hotspot of nitrogen deposition (about 1.58 g m^?2 yr^?1 and maintaining a high growth rate). As nitrogen deposition increases, the number of regions threatened by high nitrogen deposits is also increasing. With N emissions continuing to increase in the future, areas whose ecosystem is affected by high level nitrogen deposition will increase.     

Analysis of the dispersion characteristics of the Navajo generating station plume using a Lagrangian Monte-Carlo model

The Monte-Carlo Lagrangian particle model MC-LAGPAR developed by Aero Vironment Inc., been used to analyze the dispersion characteristics of the Navajo Generating Station Power Plant plume over the Grand Canyon area. The study was performed on a 26 × 26 grid covering an area of 67,600 km². The simulations were performed for the period January 18–22, 1990, a high SO₂ concentration episode during which the plume from the power plant reached Hopi Point, a receptor site on the south rim of the Grand Canyon. The meteorological data used as input to the MC-LAGPAR model were generated by MM4, the Penn State/NCAR mesocale model. The dispersion characteristics of the NGS plume are presented using plan views (XY) of the particles in the computational domain, for different hours of the day. The four-hour average SO₂ concentrations predicted by the model, assuming no SO₂ oxidation and no deposition, are in good agreement with the measurements at Hopi Point. The time at which the peak SO₂ concentration occurs is predicted accurately. This suggests that the power plant was the major contributor to the levels of SO₂ measured during the episodic hours. However, simulations assuming a one-percent-per-hour SO₂ oxidation give sulfate levels that are different from the measurements both in magnitude and in the time at which maximum sulfate level occurs. This suggests other sources than NGS were the major contributor to sulfate at Hopi Point during the January episode.     

Performance of AERMOD at different time scales

As high-density monitoring networks observing pollutant concentrations are costly to establish and maintain, researchers often employ various models to estimate concentrations of air pollutants. The AMS/EPA Regulatory Model (AERMOD) is a fairly recent and promising model for estimating concentrations of air pollutants, but the effectiveness of this model at different time scales remains to be verified. This paper evaluates the performance of AERMOD in estimating sulfur dioxide (SO₂) concentrations in Dallas and Ellis counties in Texas. Results suggest that SO₂ concentrations simulated by AERMOD at the 8 h, daily, monthly, and annual intervals match their respective observed concentrations much better compared with the simulated 1 and 3 h SO₂ concentrations. In addition, AERMOD performs better in simulating SO₂ concentrations when combined point and mobile emission sources are used as model inputs rather than using point or mobile emission sources alone. Results also suggest that, at the monthly scale, AERMOD performs much better in simulating the high end of the spectrum of SO₂ concentrations in the study area compared to results at the 1, 3, 8 h, and daily scales. These results not only help us better understand the performance of AERMOD but also provide useful information to researchers who are interested in applying AERMOD in various applications, such as the utilization of AERMOD in chronic exposure assessment in epidemiological studies where long-term (i.e., monthly and/or annual) air pollution concentration estimations are often used.