Neural Network Model for Predicting Peak Photochemical Pollutant Levels

ABSTRACT

In this paper, an attempt is made for the 24-hr prediction of photochemical pollutant levels using a neural network model. For this purpose, a model is developed that relates peak pollutant concentrations to meteorological and emission variables and indexes. The analysis is based on measurements of O₃ and NO₂ from the city of Athens. The meteorological variables are selected to cover atmospheric processes that determine the fate of the airborne pollutants while special care is taken to ensure the availability of the required input data from routine observations or forecasts. The comparison between model predictions and actual observations shows a good agreement. In addition, a series of sensitivity tests is performed in order to evaluate the sensitivity of the model to the uncertainty in meteorological variables. Model forecasts are generally rather insensitive to small perturbations in most of the input meteorological data, while they are relatively more sensitive in changes in wind speed and direction.     

Systematic development of an artificial neural network model for real-time prediction of ground-level ozone in Edmonton, Alberta, Canada

Ground-level ozone is a secondary pollutant that has recently gained notoriety for its detrimental effects on human and vegetation health. In this paper, a systematic approach is applied to develop artificial neural network (ANN) models for ground-level ozone (O3) prediction in Edmonton, Alberta, Canada, using ambient monitoring data for input. The intent of these models is to provide regulatory agencies with a tool for addressing data gaps in ambient monitoring information and predicting O3 events. The models are used to determine the meteorological conditions and precursors that most affect O3 concentrations. O3 time-series effects and the efficacy of the systematic approach are also assessed. The developed models showed good predictive success, with coefficient of multiple determination values ranging from 0.75 to 0.94 for forecasts up to 2 hr in advance. The inputs most important for O3 prediction were temperature and concentrations of nitric oxide, total hydrocarbons, sulfur dioxide, and nitrogen dioxide.     

Assimilating concentration observations for transport and dispersion modeling in a meandering wind field

Assimilating concentration data into an atmospheric transport and dispersion model can provide information to improve downwind concentration forecasts. The forecast model is typically a one-way coupled set of equations: the meteorological equations impact the concentration, but the concentration does not generally affect the meteorological field. Thus, indirect methods of using concentration data to influence the meteorological variables are required. The problem studied here involves a simple wind field forcing Gaussian dispersion. Two methods of assimilating concentration data to infer the wind direction are demonstrated. The first method is Lagrangian in nature and treats the puff as an entity using feature extraction coupled with nudging. The second method is an Eulerian field approach akin to traditional variational approaches, but minimizes the error by using a genetic algorithm (GA) to directly optimize the match between observations and predictions. Both methods show success at inferring the wind field. The GA-variational method, however, is more accurate but requires more computational time. Dynamic assimilation of a continuous release modeled by a Gaussian plume is also demonstrated using the genetic algorithm approach.     

Improving pollutant source characterization by better estimating wind direction with a genetic algorithm

In homeland security applications, it is often necessary to characterize the source location and strength of a potentially harmful contaminant. Correct source characterization requires accurate meteorological data such as wind direction. Unfortunately, available meteorological data is often inaccurate or unrepresentative, having insufficient spatial and temporal resolution for precise modeling of pollutant dispersion. To address this issue, a method is presented that simultaneously determines the surface wind direction and the pollutant source characteristics. This method compares monitored receptor data to pollutant dispersion model output and uses a genetic algorithm (GA) to find the combination of source location, source strength, and surface wind direction that best matches the dispersion model output to the receptor data. A GA optimizes variables using principles from genetics and evolution.The approach is validated with an identical twin experiment using synthetic receptor data and a Gaussian plume equation as the dispersion model. Given sufficient receptor data, the GA is able to reproduce the wind direction, source location, and source strength. Additional runs incorporating white noise into the receptor data to simulate real-world variability demonstrate that the GA is still capable of computing the correct solution, as long as the magnitude of the noise does not exceed that of the receptor data.     

Predicting hourly air pollutant levels using artificial neural networks coupled with uncertainty analysis by Monte Carlo simulations

Recent progress in developing artificial neural network (ANN) metamodels has paved the way for reliable use of these models in the prediction of air pollutant concentrations in urban atmosphere. However, improvement of prediction performance, proper selection of input parameters and model architecture, and quantification of model uncertainties remain key challenges to their practical use. This study has three main objectives: to select an ensemble of input parameters for ANN metamodels consisting of meteorological variables that are predictable by conventional weather forecast models and variables that properly describe the complex nature of pollutant source conditions in a major city, to optimize the ANN models to achieve the most accurate hourly prediction for a case study (city of Tehran), and to examine a methodology to analyze uncertainties based on ANN and Monte Carlo simulations (MCS). In the current study, the ANNs were constructed to predict criteria pollutants of nitrogen oxides (NO_x), nitrogen dioxide (NO₂), nitrogen monoxide (NO), ozone (O₃), carbon monoxide (CO), and particulate matter with aerodynamic diameter of less than 10 μm (PM₁₀) in Tehran based on the data collected at a monitoring station in the densely populated central area of the city. The best combination of input variables was comprehensively investigated taking into account the predictability of meteorological input variables and the study of model performance, correlation coefficients, and spectral analysis. Among numerous meteorological variables, wind speed, air temperature, relative humidity and wind direction were chosen as input variables for the ANN models. The complex nature of pollutant source conditions was reflected through the use of hour of the day and month of the year as input variables and the development of different models for each day of the week. After that, ANN models were constructed and validated, and a methodology of computing prediction intervals (PI) and probability of exceeding air quality thresholds was developed by combining ANNs and MCSs based on Latin Hypercube Sampling (LHS). The results showed that proper ANN models can be used as reliable metamodels for the prediction of hourly air pollutants in urban environments. High correlations were obtained with R ² of more than 0.82 between modeled and observed hourly pollutant levels for CO, NO_x, NO₂, NO, and PM₁₀. However, predicted O₃ levels were less accurate. The combined use of ANNs and MCSs seems very promising in analyzing air pollution prediction uncertainties. Replacing deterministic predictions with probabilistic PIs can enhance the reliability of ANN models and provide a means of quantifying prediction uncertainties.     

Coupling of the Weather Research and Forecasting Model with AERMOD for pollutant dispersion modeling. A case study for PM10 dispersion over Pune,India

The prediction of spatial variation of the concentration of a pollutant governed by various sources and sinks is a complex problem. Gaussian air pollutant dispersion models such as AERMOD of the United States Environmental Protection Agency (USEPA) can be used for this purpose. AERMOD requires steady and horizontally homogeneous hourly surface and upper air meteorological observations. However, observations with such frequency are not easily available for most locations in India. To overcome this limitation, the planetary boundary layer and surface layer parameters required by AERMOD were computed using the Weather Research and Forecasting (WRF) Model (version 2.1.1) developed by the National Center for Atmospheric Research (NCAR). We have developed a preprocessor for offline coupling of WRF with AERMOD. Using this system, the dispersion of respirable particulate matter (RSPM/PM10) over Pune, India has been simulated. Data from the emissions inventory development and field-monitoring campaign (13–17 April 2005) conducted under the Pune Air Quality Management Program of the Ministry of Environment and Forests (MoEF), India and USEPA, have been used to drive and validate AERMOD. Comparison between the simulated and observed temperature and wind fields shows that WRF is capable of generating reliable meteorological inputs for AERMOD. The comparison of observed and simulated concentrations of PM10 shows that the model generally underestimates the concentrations over the city. However, data from this single case study would not be sufficient to conclude on suitability of regionally averaged meteorological parameters for driving Gaussian models like AERMOD and additional simulations with different WRF parameterizations along with an improved pollutant source data will be required for enhancing the reliability of the WRF–AERMOD modeling system.     

Systematic Development of an Artificial Neural Network Model for Real-Time Prediction of Ground-Level Ozone in Edmonton,Alberta, Canada

Abstract

Ground-level ozone is a secondary pollutant that has recently gained notoriety for its detrimental effects on human and vegetation health. In this paper, a systematic approach is applied to develop artificial neural network (ANN) models for ground-level ozone (O₃) prediction in Edmonton, Alberta, Canada, using ambient monitoring data for input. The intent of these models is to provide regulatory agencies with a tool for addressing data gaps in ambient monitoring information and predicting O₃ events. The models are used to determine the meteorological conditions and precursors that most affect O₃ concentrations. O₃ time-series effects and the efficacy of the systematic approach are also assessed. The developed models showed good predictive success, with coefficient of multiple determination values ranging from 0.75 to 0.94 for forecasts up to 2 hr in advance. The inputs most important for O₃ prediction were temperature and concentrations of nitric oxide, total hydrocarbons, sulfur dioxide, and nitrogen dioxide.     

Assessing sensitivity of source term estimation

Source term estimation algorithms compute unknown atmospheric transport and dispersion modeling variables from concentration observations made by sensors in the field. Insufficient spatial and temporal resolution in the meteorological data as well as inherent uncertainty in the wind field data make source term estimation and the prediction of subsequent transport and dispersion extremely difficult. This work addresses the question: how many sensors are necessary in order to successfully estimate the source term and meteorological variables required for atmospheric transport and dispersion modeling?The source term estimation system presented here uses a robust optimization technique – a genetic algorithm (GA) – to find the combination of source location, source height, source strength, surface wind direction, surface wind speed, and time of release that produces a concentration field that best matches the sensor observations. The approach is validated using the Gaussian puff as the dispersion model in identical twin numerical experiments. The limits of the system are tested by incorporating additive and multiplicative noise into the synthetic data. The minimum requirements for data quantity and quality are determined by an extensive grid sensitivity analysis. Finally, a metric is developed for quantifying the minimum number of sensors necessary to accurately estimate the source term and to obtain the relevant wind information.     

Prediction of sulfur dioxide concentrations at a site near downtown Santiago,Chile

We have analyzed the possibility to predict hourly averages of sulfur dioxide concentrations in the atmosphere at a site not far from the downtown area in the city of Santiago, Chile. We have compared the forecasts produced assuming persistence, linear regressions and feed forward neural networks. The effect of meteorological conditions is included by using forecasted values of temperature, relative humidity and wind speed at the time of the intended prediction as inputs to the different models. The best predictions for hourly averages are obtained with a three-layer neural network that has hourly averages of sulfur dioxide concentrations every 6 h on the previous day plus the actual values of the meteorological variables as input. Training the network with 1995 data, error in 8 h in advance prediction for 1996 data is of the order of 30%.     

A Test Reference Year (TRY) approach to gas dispersion in the atmosphere

For environmental analysis such as the dispersion of pollutants in the atmosphere, it is essential to have meteorological data that are relevant for a long period. In this paper, we explore the possibility of using an environmental Test Reference Year (TRY), i.e., a set of real, contemporaneous and hourly meteorological variables, 'extracted' from a hourly series of at least 10 years, for modelling pollutant dispersion in the atmosphere. The classical approach, based on a statistical data set, implies the loss of important information such as the real correlation between the different meteorological variables, and this implies crude approximation in the simulation results. We compare the simulation results with the long hourly 10 years data set (which can be considered a 'brute force' approach, since it requires a huge amount of data and time processing, but it is here considered the most severe benchmark) and with the statistical data set commonly used. It is shown that the results obtained using the TRY have a good agreement with the ones obtained with the simulation of the 10 years and they are also much better than those obtained using the statistical data set.     

Tracer transport model at Japan meteorological agency and its application to the ETEX data

This paper describes the JMA tracer transport model and its sensitivity to model physics and initial conditions by using the ETEX data. Compared with observations, the model overestimates ground-level concentration in the early stage within about one day from the emission, possibly due to underestimation of vertical diffusion. In the early stage, the enhanced vertical diffusion effectively transports the tracer upward and decreases the ground-level concentration, while in the later stage, it enhances downward transports from the upper layer and increases the ground-level concentration. A conceptual model is given for understanding the vertical transport due to vertical diffusion. The horizontal diffusion introduced to the model has preferable impacts on forecasts especially in the early stage. Finally, we discuss the predictability of this model based on sensitivity to initial conditions of emission time and height.     

Artificial neural network for the identification of unknown air pollution sources

Artificial neural networks (ANN), whose performances to deal with pattern recognition problems is well known, are proposed to identify air pollution sources. The problem that is addressed is the apportionment of a small number of sources from a data set of ambient concentrations of a given pollutant. Three layers feed-forward ANN trained with a back-propagation algorithm are selected. A test case is built, based on a Gaussian dispersion model. A subset of hourly meteorological conditions and measured concentrations constitute the input patterns to the network that is wired to recover relevant emission parameters of unknown sources as outputs. The rest of the model data are corrupted adding noise to some meteorological parameters and we test the effectiveness of the method to recover the correct answer. The ANN is applied to a realistic case where 24 h SO₂ concentrations were previously measured. Some of the limitations of the ANN approach, together with its capabilities, are discussed in this paper.     

Assimilation of conventional and satellite wind observations in a mesoscale atmospheric model for studying atmospheric dispersion

A mesoscale atmospheric model PSU/NCAR MM5 is used to provide operational weather forecasts for a nuclear emergency response decision support system on the southeast coast of India. In this study the performance of the MM5 model with assimilation of conventional surface and upper-air observations along with satellite derived 2-d surface wind data from QuickSCAT sources is examined. Two numerical experiments with MM5 are conducted: one with static initialization using NCEP FNL data and second with dynamic initialization by assimilation of observations using four dimensional data assimilation (FDDA) analysis nudging for a pre-forecast period of 12 h. Dispersion simulations are conducted for a hypothetical source at Kalpakkam location with the HYSPLIT Lagrangian particle model using simulated wind field from the above experiments. The present paper brings out the differences in the atmospheric model predictions and the differences in dispersion model results from control and assimilation runs. An improvement is noted in the atmospheric fields from the assimilation experiment which has led to significant alteration in the trajectory positions, plume orientation and its distribution pattern. Sensitivity tests using different PBL and surface parameterizations indicated the simple first order closure schemes (Blackadar, MRF) coupled with the simple soil model have given better results for various atmospheric fields. The study illustrates the impact of the assimilation of the scatterometer wind and automated weather stations (AWS) observations on the meteorological model predictions and the dispersion results.     

Surface Air Pollutant Concentration Frequency Distributions: Implications for Urban Modeling

With the development of ambient air quality standards (AAQS), the need arises to describe the characteristics of regional surface air-pollutant concentration frequency distributions. In the evaluation of land use plans, numerous agencies will be concerned with evaluating the effectiveness of emission zoning and/or control actions. On a regional basis, one means of performing this assessment lies in determining the changes in the pollutant frequency distributions resulting from control actions.

This study presents new data concerning the surface air-pollutant concentration frequency distributions observed for area sources and continuous point sources, and compares these distributions with those of the pertinent meteorological variables describing the transport and diffusion of the pollutant. The observed surface air pollutant frequency distributions are compared to those corresponding to simple modeling concepts from either an urban area source or a continuous point source. For an urban source and a relatively inert pollutant like CO, we found that the observed frequency distribution for CO surface air concentration parallels the approximately log-normal frequency distribution of the reciprocal of the wind speed. We show that the constant relating these two well-correlated frequency distributions can be determined either experimentally or with a numerical simulation model of air pollution. The usefulness of numerical models in air pollution is discussed.     

Development of the ClearSky smoke dispersion forecast system for agricultural field burning in the Pacific Northwest

The post-harvest burning of agricultural fields is commonly used to dispose of crop residue and provide other desired services such as pest control. Despite careful regulation of burning, smoke plumes from field burning in the Pacific Northwest commonly degrade air quality, particularly for rural populations. In this paper, ClearSky, a numerical smoke dispersion forecast system for agricultural field burning that was developed to support smoke management in the Inland Pacific Northwest, is described. ClearSky began operation during the summer through fall burn season of 2002 and continues to the present. ClearSky utilizes Mesoscale Meteorological Model version 5 (MM5v3) forecasts from the University of Washington, data on agricultural fields, a web-based user interface for defining burn scenarios, the Lagrangian CALPUFF dispersion model and web-served animations of plume forecasts. The ClearSky system employs a unique hybrid source configuration, which treats the flaming portion of a field as a buoyant line source and the smoldering portion of the field as a buoyant area source. Limited field observations show that this hybrid approach yields reasonable plume rise estimates using source parameters derived from recent field burning emission field studies. The performance of this modeling system was evaluated for 2003 by comparing forecast meteorology against meteorological observations, and comparing model-predicted hourly averaged PM_2.5 concentrations against observations. Examples from this evaluation illustrate that while the ClearSky system can accurately predict PM_2.5 surface concentrations due to field burning, the overall model performance depends strongly on meteorological forecast error. Statistical evaluation of the meteorological forecast at seven surface stations indicates a strong relationship between topographical complexity near the station and absolute wind direction error with wind direction errors increasing from approximately 20° for sites in open areas to 70° or more for sites in very complex terrain. The analysis also showed some days with good forecast meteorology with absolute mean error in wind direction less than 30° when ClearSky correctly predicted PM_2.5 surface concentrations at receptors affected by field burns. On several other days with similar levels of wind direction error the model did not predict apparent plume impacts. In most of these cases, there were no reported burns in the vicinity of the monitor and, thus, it appeared that other, non-reported burns were responsible for the apparent plume impact at the monitoring site. These cases do not provide information on the performance of the model, but rather indicate that further work is needed to identify all burns and to improve burn reports in an accurate and timely manner. There were also a number of days with wind direction errors exceeding 70° when the forecast system did not correctly predict plume behavior.     

Near-road air pollutant concentrations of CO and PM2.5: A comparison of MOBILE6.2/CALINE4 and generalized additive models

The contribution of vehicular traffic to air pollutant concentrations is often difficult to establish. This paper utilizes both time-series and simulation models to estimate vehicle contributions to pollutant levels near roadways. The time-series model used generalized additive models (GAMs) and fitted pollutant observations to traffic counts and meteorological variables. A one year period (2004) was analyzed on a seasonal basis using hourly measurements of carbon monoxide (CO) and particulate matter less than 2.5 μm in diameter (PM_2.5) monitored near a major highway in Detroit, Michigan, along with hourly traffic counts and local meteorological data. Traffic counts showed statistically significant and approximately linear relationships with CO concentrations in fall, and piecewise linear relationships in spring, summer and winter. The same period was simulated using emission and dispersion models (Motor Vehicle Emissions Factor Model/MOBILE6.2; California Line Source Dispersion Model/CALINE4). CO emissions derived from the GAM were similar, on average, to those estimated by MOBILE6.2. The same analyses for PM_2.5 showed that GAM emission estimates were much higher (by 4–5 times) than the dispersion model results, and that the traffic-PM_2.5 relationship varied seasonally. This analysis suggests that the simulation model performed reasonably well for CO, but it significantly underestimated PM_2.5 concentrations, a likely result of underestimating PM_2.5 emission factors. Comparisons between statistical and simulation models can help identify model deficiencies and improve estimates of vehicle emissions and near-road air quality.     

An enhanced ozone forecasting model using air mass trajectory analysis

An enhanced ozone forecasting model using nonlinear regression and an air mass trajectory parameter has been developed and field tested. The model performed significantly better in predicting daily maximum 1-h ozone concentrations during a five-year model calibration period (1993–1997) than did a previously reported regression model. This was particularly true on the 28 “high ozone” days ([O₃]>120 ppb) during the period, for which the mean absolute error (MAE) improved from 21.7 to 12.1 ppb. On the 77 days meteorologically conducive to high ozone, the MAE improved from 12.2 to 9.1 ppb, and for all 580 calibration days the MAE improved from 9.5 to 8.35 ppb. The model was field-tested during the 1998 ozone season, and performed about as expected. Using actual meteorological data as input for the ozone predictions, the MAE for the season was 11.0 ppb. For the daily ozone forecasts, which used meteorological forecast data as input, the MAE was 13.4 ppb. The high ozone days were all anticipated by the ozone forecasters when the model was used for next day forecasts.     

Fuzzy system models combined with nonlinear regression for daily ground-level ozone predictions

This study focuses on applying a Takagi–Sugeno fuzzy system and a nonlinear regression (NLR) model for ozone predictions in six Kentucky metropolitan areas. The fuzzy “c-means” clustering technique coupled with an optimal output predefuzzification approach (least square method) was used to train the Takagi–Sugeno fuzzy system. The fuzzy system was tuned by specifying the number of rules and the fuzziness factor. The NLR models were based in part on a previously reported, trajectory-based hybrid NLR model that has been used for years for forecasting ground-level ozone in Louisville, KY. The NLR models were each composed of an interactive nonlinear term and several linear terms. Using a common meteorological parameter set as input variables, the NLR models and the Takagi–Sugeno fuzzy systems model exhibited equivalent forecasting performance on test data from 2004. For all 2004 ozone season forecasts for the six metropolitan areas, the mean absolute error was 8.1 ppb for the NLR model and 8.0 ppb for the Takagi–Sugeno fuzzy model. When a nonlinear term (which was part of the NLR model) was included in the fuzzy model, the combined NLR–fuzzy model had slightly better performance than the original NLR model. For all 2004 metropolitan area forecasts, the mean absolute error of the NLR–fuzzy model forecasts was 7.7 ppb. These small differences may be statistically significant, but for practical purposes the performance of the fuzzy models was equivalent to that of the NLR models.     

Screening and Scenarios of Traffic Emissions at Trier,Germany

SCOPE AND BACKGROUND: In the course of the European Council Directive on permissible air pollutant limit values, valid starting from 2005 there is an urgent call for action, particularly for fine dust (PM10). Current investigations (Junk & Helbig 2003, Reuter & Baumüller 2003) show that the limit values in certain places in congested areas are exceeded. Only if it is possible to locate these Hot Spots purposeful measures to reduce the ambient air pollution can be conducted. For an efficient identification of these Hot Spots numerical computer models or establishing special measurements networks are too expensive. Using the statistical model STREET 5.0 (KTT 2003) a cost-effective screening of the air pollution situation caused by the traffic can be done. METHODS: STREET is based on the 3-dimensional micro-scale non-hydrostatic flow- and dispersion model MISCAM (Eichhorn 1989). The results of over 100.000 different calculations with MISCAM are stored in a Database and used to calculate the emissions with STREET. In collaboration with the city council of Trier more than 150 streets were investigated, mapped, and calculated. A special urban climate measuring network supplies the necessary meteorological input data about the wind field and precipitation events in the valley of the Moselle. Information about road width and road orientation as well as building density was derived from aerial photographs. Traffic censuses and mobile air pollutants measurements supplied the remaining input data. We calculated the mean annual air pollutant concentrations for NO2, CO, SO2, O3, benzene as well as PM10. RESULTS: A comparison of the model results with the values obtained from the stations of the central emission measuring network of Rhineland-Palatinate (ZIMEN, annual report 2002) shows very good agreements. The model was not only used to calculate the annual air pollutant but also for urban planning and management. The absolute level of the air pollutant is mainly dependent on the amount of traffic in the street canyons. Therefore four different case-scenarios with varying quantity of traffic were calculated and interpreted for each street. The results of the calculation show that on the basis of the mean values for both NO2 and benzene, it is not to be expected that the limits PERSPECTIVES: Furthermore the model can be used to find the maximum tolerable numbers of cars for a street without exceeding the air pollutant thresholds.     

Analysis of the Trend and Seasonal Cycle of Carbon Monoxide Concentrations in an Urban Area (4 pp)

Background, Aim and Scope Air quality is an field of major concern in large cities. This problem has led administrations to introduce plans and regulations to reduce pollutant emissions. The analysis of variations in the concentration of pollutants is useful when evaluating the effectiveness of these plans. However, such an analysis cannot be undertaken using standard statistical techniques, due to the fact that concentrations of atmospheric pollutants often exhibit a lack of normality and are autocorrelated. On the other hand, if long-term trends of any pollutant’s emissions are to be detected, meteorological effects must be removed from the time series analysed, due to their strong masking effects. Materials and Methods The application of statistical methods to analyse temporal variations is illustrated using monthly carbon monoxide (CO) concentrations observed at an urban site. The sampling site is located at a street intersection in central Valencia (Spain) with a high traffic density. Valencia is the third largest city in Spain. It is a typical Mediterranean city in terms of its urban structure and climatology. The sampling site started operation in January 1994 and monitored CO ground level concentrations until February 2002. Its geographic coordinates are W0°22′52″ N39°28′05″ and its altitude is 11 m. Two nonparametric trend tests are applied. One of these is robust against serial correlation with regards to the false rejection rate, when observations have a strong persistence or when the sample size per month is small. A nonparametric analysis of the homogeneity of trends between seasons is also discussed. A multiple linear regression model is used with the transformed data, including the effect of meteorological variables. The method of generalized least squares is applied to estimate the model parameters to take into account the serial dependence of the residuals of this model. This study also assesses temporal changes using the Kolmogorov-Zurbenko (KZ) filter. The KZ filter has been shown to be an effective way to remove the influence of meteorological conditions on O₃ and PM to examine underlying trends. Results The nonparametric tests indicate a decreasing, significant trend in the sampled site. The application of the linear model yields a significant decrease every twelve months of 15.8% for the average monthly CO concentration. The 95% confidence interval for the trend ranges from 13.9% to 17.7%. The seasonal cycle also provides significant results. There are no differences in trends throughout the months. The percentage of CO variance explained by the linear model is 90.3%. The KZ filter separates out long, short-term and seasonal variations in the CO series. The estimated, significant, long-term trend every year results in 10.3% with this method. The 95% confidence interval ranges from 8.8% to 11.9%. This approach explains 89.9% of the CO temporal variations. Discussion The differences between the linear model and KZ filter trend estimations are due to the fact that the KZ filter performs the analysis on the smoothed data rather than the original data. In the KZ filter trend estimation, the effect of meteorological conditions has been removed. The CO short-term componentis attributable to weather and short-term fluctuations in emissions. There is a significant seasonal cycle. This component is a result of changes in the traffic, the yearly meteorological cycle and the interactions between these two factors. There are peaks during the autumn and winter months, which have more traffic density in the sampled site. There is a minimum during the month of August, reflecting the very low level of vehicle emissions which is a direct consequence of the holiday period. Conclusions The significant, decreasing trend implies to a certain extent that the urban environment in the area is improving. This trend results from changes in overall emissions, pollutant transport, climate, policy and economics. It is also due to the effect of introducing reformulated gasoline. The additives enable vehicles to burn fuel with a higher air/fuel ratio, thereby lowering the emission of CO. The KZ filter has been the most effective method to separate the CO series components and to obtain an estimate of the long-term trend due to changes in emissions, removing the effect of meteorological conditions. Recommendations and Perspectives Air quality managers and policy-makers must understand the link between climate and pollutants to select optimal pollutant reduction strategies and avoid exceeding emission directives. This paper analyses eight years of ambient CO data at a site with a high traffic density, and provides results that are useful for decision-making. The assessment of long-term changes in air pollutants to evaluate reduction strategies has to be done while taking into account meteorological variability