Neural network modelling and prediction of hourly NOx and NO2 concentrations in urban air in London

Multilayer perceptron (MLP) neural networks were trained to model hourly NO_x and NO₂ pollutant concentrations in Central London from basic hourly meteorological data. Results have shown that the models perform well when compared to previous attempts to model the same pollutants using regression based models. This work also illustrates that MLP neural networks are capable of resolving complex patterns of source emissions without any explicit external guidance.     

A comparison of nonlinear regression and neural network models for ground-level ozone forecasting

A hybrid nonlinear regression (NLR) model and a neural network (NN) model, each designed to forecast next-day maximum 1-hr average ground-level O3 concentrations in Louisville, KY, were compared for two O3 seasons--1998 and 1999. The model predictions were compared for the forecast mode, using forecasted meteorological data as input, and for the hindcast mode, using observed meteorological data as input. The two models performed nearly the same in the forecast mode. For the two seasons combined, the mean absolute forecast error was 12.5 ppb for the NLR model and 12.3 ppb for the NN model. The detection rate of 120 ppb threshold exceedances was 42% for each model in the forecast mode. In the hindcast mode, the NLR model performed marginally better than the NN model. The mean absolute hindcast error was 11.1 ppb for the NLR model and 12.9 ppb for the NN model. The hindcast detection rate was 92% for the NLR model and 75% for the NN model.     

Forecasting peak daily ozone levels: part 2--A regression with time series errors model having a principal component trigger to forecast 1999 and 2002 ozone levels

A modified time series approach, a Box-Jenkins regression with time series errors (RTSE) model plus a principal component (PC) trigger, has been developed to forecast peak daily 1-hr ozone (O3) in real time at the University of Wisconsin-Milwaukee North (UWM-N) during 1999 and 2002. The PC trigger acts as a predictor variable in the RTSE model. It tries to answer the question: will the next day be a possible high O3 day? To answer this question, three PC trigger rules were developed: (1) Hi-Low Checklist, (2) Discriminant Function Approach I, and (3) Discriminant Function Approach II. Also, a pure RTSE model without including the PC trigger and a persistence approach were tested for comparison. The RTSE model with DFA I successfully forecast the only two 1-hr federal exceedances (124 ppb), one in 1999 and one in 2002. In terms of the O3 100-ppb exceedances, 60-80% of the incorrect forecasts occurred with incorrect PC decisions. A few others may have been caused by unexpected O3-weather relations. When the three approaches used UWM-N data to forecast a 100-ppb exceedance somewhere in the Milwaukee, WI, metropolitan area, their performance was dramatically improved: the false alarm rate was reduced from 0.89 to 0.44, and the probability of detection was increased from 0.71 to 0.88.     

Predicting frequency distributions for ozone,NO2 and TSP from restricted data sets

The results of Ott and Mage (1981) for CO indicate that random sampling yields accurate estimates of the mean and standard deviation. This paper examines the estimates for a cumulative frequency distribution using four different types of data set: (1) continuous monitoring throughout the year, (2) continuous monitoring one week out of every four, (3) random sampling at any time and (4) random sampling only on weekdays between 9.00 a.m. and 5.00 p.m. Data sets for O₃, NO₂ and TSP are used. The results show that random sampling should only be used when there is no restriction on the time of sampling and that option (2) yields very good results indicating a cost-effective way of recording results.     

Benzene,toluene, ozone,NO2 and SO2 measurements in an urban street canyon in Thessaloniki,Greece

Benzene, toluene, sulphur dioxide, ozone and nitrogen dioxide were measured at a mean level of 13.5 m above ground in a narrow, four-lane street canyon (height 30 m, width 20 m) in Thessaloniki, Greece during the period January–July 1997 by means of a commercial differential optical absorption spectrometer (OPSIS DOAS). Primary pollutant levels were found to be 2.5–4.4 times higher during the cold part of the year than during the warm part of the year, the winter/summer ratio increasing with the reaction rate constant with OH for each of the measured species. Ozone, on the other hand, exhibited a winter/summer ratio of 0.36. NO₂ originates from both primary and secondary sources; its winter/summer concentration ratio of 1.4 lies, therefore, between those of primary pollutants and ozone. Pollution levels were influenced considerably by wind speed, while for the street canyon under study wind direction did not influence pollutant levels considerably. While primary pollution was found to decrease with increasing wind speed, ozone increased. Benzene mean levels during the study period were around 6 ppb and hence much higher than the EU annual limit value of 5 μg m⁻³ (1.44 ppb at STP). Toluene mean levels were around 14 ppb and hence also several times above the WHO recommendation of 2 ppb for 24 h. The apportionment of traffic emissions in four time zones used in most inventories in urban airshed models was tested using benzene and toluene measurements at low (<1 m s⁻¹) wind speeds. The agreement between model emissions and calculated emissions apportionment into the four time zones was good, except for Zone D (23:00–1:59), where model inventory emissions were somewhat too low.     

Trend analysis of urban NO2 concentrations and the importance of direct NO2 emissions versus ozone/NOx equilibrium

The annual air quality standard of NO₂ is often exceeded in urban areas near heavy traffic locations. Despite significant decrease of NO_x emissions in 1986–2005 in the industrial and harbour area near Rotterdam, NO₂ concentrations at the urban background remain at the same level since the end of the nineties. Trend analysis of monitoring data revealed that the ozone/NO_x equilibrium is a more important factor than increasing direct NO₂ emissions by traffic. The latter has recently been identified as an additional NO₂ source due to the introduction of oxy-catalytic converters in diesel vehicles and the growing number of diesel vehicles. However, in Rotterdam over the period 1986–2005 direct NO₂ emissions by road traffic only increased 3–4%. Due to the importance of the ozone/NO_x equilibrium, it is concluded that local NO_x emissions in Rotterdam need substantial reduction to achieve lower NO₂ urban background levels. This is a relatively costly abatement strategy and, therefore, a “hotspot” approach aiming at reducing NO_x emissions by local traffic measures is more effective to meet European air quality standards.     

Odour impact assessment by means of dynamic olfactometry,dispersion modelling and social participation

This work discusses how it is possible to assess odour impact in presence of multiple similar sources by illustrating a case study. The study was conducted on an area of northern Italy comprising three small municipalities where four rendering plants are located near to each other. Based on the emission data resulting from olfactometric surveys conducted in different periods of the year the overall odour emission rate emitted by each plant were evaluated, showing that the major contributor to the odour impact on the territory was plant 2. These data were linked with meteorological and orographical data in order to evaluate odour dispersion with a model (Calpuff). The results of the odour dispersion modelling confirmed the outcomes of the olfactometric survey and they were further validated through a “questioning” survey, conducted with the aim of involving the population by means of questionnaires for reporting the perceived odour episodes, which showed a good correspondence (86.5%) between odour perceptions and simulated odour immissions.     

Seasonal evolutions of ozone (O3) and its nitrogen precursors (NO, NO2) in urban Sfax (Tunisia)

Seasonal evolution of ozone (O3) and its nitrogen precursors (NO, NO2) in downtown Sfax (Tunisia) was monitored. Nitrogen oxides are shown to be closely related to local vehicle sources. Seasonal ozone levels, however, are shown to be dependent on regional meteorological conditions. High ozone levels are due to the effect of anticyclones and stratosphere intrusions (cut-off lows). Low levels are associated with cyclonic conditions of small vertical range of motion. Other than these particular conditions, ozone levels are shown to be relatively higher in fall and winter seasons, characterised by a very steady atmosphere. Overall, the examined meteorological conditions, the ozone concentrations observed in downtown Sfax are characterised by clear day/night cycles, which can be explained by the significant ventilation of the region.     

Field observations of regional and urban impacts on NO2, ozone,UVB, and nitrate radical production rates in the Phoenix air basin

In the May and June of 1998, field measurements were taken at a site near the Usery Pass Recreation Area, ∼27 miles from the downtown Phoenix area, overlooking Phoenix and Mesa, Arizona. This site was selected to examine the impacts of the Phoenix urban plume on the Usery Pass Recreation Area and surrounding regions. Data were obtained for ultraviolet-B (UVB) radiation, nitrogen dioxide (NO₂), peroxyacetyl nitrate (PAN), ozone (O₃), and carbon monoxide (CO). Nocturnal plumes of NO₂ (in tens of ppb), observed near midnight, were correlated with CO and anti-correlated with O₃. This behavior was consistent with the titration of locally generated NO by boundary layer O₃ to form the nighttime NO₂ plumes that were subsequently transported into the Usery Pass Recreation area. Nitrate radical (NO₃) production rates were calculated to be very high on the edges of these nocturnal plumes. Examination of O₃ and PAN data also indicates that Phoenix is being affected by long-range transport of pollutants from the Los Angeles to San Diego areas. A regional smoke episode was observed in May, accompanied by a decrease in UVB of factor of two and a decrease in O₃ and an increase in methyl chloride. Low level back trajectories and chemical evidence confirm that the smoke event originated in northern Mexico and that the reduced O₃ levels observed at Usery Pass could be partially due to reduced photolysis rates caused by carbonaceous soot aerosols transported in the smoke plume. The results are discussed with regard to potential effects of local pollution transport from the Phoenix air basin as well as an assessment of the contributions from long-range transport of pollutants to the background levels in the Phoenix-Usery Pass area.     

Regression modelling of hourly NOx and NO2 concentrations in urban air in London

Based on hourly measurements of NO_x NO₂ and O₃ and meteorological data, an ordinary least squares (OLS) model and a first-order autocorrelation (AR) model were developed to analyse the regression and prediction of NO_x and NO₂ concentrations in London. Primary emissions and wind speed are the most important factors influencing NO_x concentrations; in addition to these two, reaction of NO with O₃ is also a major factor influencing NO₂ concentrations. The AR model resulted in high correlation coefficients (R > 0.95) for the NO_x and NO₂ regression based on a whole year's data, and is capable of predicting NO₂ (R = 0.83) and NO_x (R = 0.65) concentrations when the explanatory variables were available. The analysis of the structure of regression models by Principal Component Analysis (PCA) indicates that the regression models are stable. The results of the OLS model indicate that there was an exceptional NO₂ source, other than primary emission and reaction of NO with O₃, in the air pollution episode in London in December 1991.     

Decomposition of 2-nitrophenol in aqueous solution by ozone and UV/ozone processes.

The decomposition of 2-nitrophenol in aqueous solutions by ozone and UV/ozone processes was found to be technically feasible under adequate experimental conditions. Formation of nitrate ions was observed following the decomposition of 2-nitrophenol by ozone and UV/ ozone processes. Increasing ozone dosage and UV light intensity accelerated the decomposition rate of 2-nitrophenol in an aqueous solution. The species distribution of 2-nitrophenol under various solution conditions plays a significant role in determining decomposition behavior. In most experiments conducted in this study, the decomposition of 2-nitrophenol by ozone and UV/ozone processes was favored to occur in alkaline conditions. The addition of 2-butanol accelerated the rate of gaseous ozone transfer to an aqueous phase by reducing the surface tension of aqueous solution and therefore enhancing the decomposition rate of 2-nitrophenol by ozone and UV/ozone processes.     

In situ experiments for element species-specific environmental reactivity of tin and mercury compounds using isotopic tracers and multiple linear regression

The fate of mercury (Hg) and tin (Sn) compounds in ecosystems is strongly determined by their alkylation/dealkylation pathways. However, the experimental determination of those transformations is still not straightforward and methodologies need to be refined. The purpose of this work is the development of a comprehensive and adaptable tool for an accurate experimental assessment of specific formation/degradation yields and half-lives of elemental species in different aquatic environments. The methodology combines field incubations of coastal waters and surface sediments with the addition of species-specific isotopically enriched tracers and a mathematical approach based on the deconvolution of isotopic patterns. The method has been applied to the study of the environmental reactivity of Hg and Sn compounds in coastal water and surface sediment samples collected in two different coastal ecosystems of the South French Atlantic Coast (Arcachon Bay and Adour Estuary). Both the level of isotopically enriched species and the spiking solution composition were found to alter dibutyltin and monomethylmercury degradation yields, while no significant changes were measurable for tributyltin and Hg(II). For butyltin species, the presence of light was found to be the main source of degradation and removal of these contaminants from surface coastal environments. In contrast, photomediated processes do not significantly influence either the methylation of mercury or the demethylation of methylmercury. The proposed method constitutes an advancement from the previous element-specific isotopic tracers’ approaches, which allows for instance to discriminate the extent of net and oxidative Hg demethylation and to identify which debutylation step is controlling the environmental persistence of butyltin compounds.     

Modeling and monitoring of CO,NO and NO2 in a modern townhouse

Mass-balance models for indoor concentrations of CO, NO and NO₂ were applied to an energy-efficient townhouse. Model parameters included source emission rates, infiltrating airflows, and, for NO₂, loss rate coefficients. Emission factors for CO, NO and NO₂ were estimated for each of the gas-fired appliances in the house. Airflows were estimated using sulfur hexafluoride (SF₆) decay techniques. Loss rates for NO₂ were calculated as the difference between NO₂ removal rates and estimated air exchange rates; CO and NO concentrations decayed at a rate not significantly different than that for SF₆. Comparing model predictions with measured concentrations yielded differences averaging 17% for CO and NO, and 28% for NO₂.     

Assessing vegetation exposure to ozone: properties of the AOT40 index and modifications by deposition modelling

A discussion is presented on the application of micrometeorological deposition modelling principles to improve the characterisation of vegetation exposure to ozone and thus the use of critical levels as the basis of targeted emission control. The AOT40 (accumulated exposure over a threshold of 40 ppb or nl l(-1)) ozone exposure index is shown to impose a differential weighting that results in a high sensitivity, by a factor of two to 10 depending on the pollution climate, with respect to concentration. This makes it necessary to correct for systematic effects, such as the concentration profile below the measurement height, in order to justify a comparison with the biological data obtained from well-mixed exposure chambers. Available studies indicate a 50-70% lower AOT40 at the vegetation height. The resistance method for estimating the profile is extended to allow for stomatal effects that potentially bias the plant response predicted with an exposure index. This integrated profile-uptake correction refines the current approach and serves as a transitional step towards a real flux-based approach. For the latter, a new deposition parameterisation is tested against field observations.     

Short-term prediction of means of artificial neural urban NO2 pollution by networks

A neural network model for the short-term prediction of concentrations of urban pollutants was developed and applied to the Turin (Northern Italy) air quality network. In particular, the study was focused on NO₂ concentrations measured at five stations; t + 3 and t + 24 hour NO₂ concentration forecasting based on hourly meteorological and concentration data gave good agreement with observed concentrations. This is particularly true for the mean concentration values and concentration distribution. The time of occurrence of peak values was correctly forecast but the amounts were generally underestimated. To reduce this underestimation, an empirical step function was applied in the t + 24 case. This allowed an accurate estimate to be obtained of the few cases in which 50% of the air quality monitoring stations exceeded the attention level (200 μg m^-3) during the following day for at least one hour.     

Measurements and modelling of the ozone deposition velocity to concrete tiles,including the effect of diffusion

A range of models were fitted to the experimental time-dependent curves for the deposition velocity of O₃ to concrete floor tile samples. The models included modified Langmuir isotherms assuming adsorption of O₃ on the material surfaces and models assuming direct reaction on and diffusion of O₃ into the material from the air. The best fit was obtained with a simple two-parameter model assuming direct reaction of O₃ with adsorbed surface water and direct diffusion of O₃ into the material from the air. However, models assuming an additional second-order reaction of O₃ with an adsorbed surface species with a given start mass gave improved curve fit in the first 500 min. Applying the best model to experimental data obtained over the whole air humidity range resulted in markedly lower equilibrium deposition velocities than those measured after 48 h. The modelling gave a deposition velocity minimum in the 50–70% relative air humidity range in agreement with observations. The deposition velocity minimum could be explained with a reduced constant for the reaction of O₃ with water or OH⁻ ions on the surface.     

Quantification of ozone levels in indoor environments generated by ionization and ozonolysis air purifiers

Indoor air purifiers are advertised as safe household products for health-conscious individuals, especially for those suffering from allergies and asthma. However, certain air purifiers produce ozone (O3) during operation, either intentionally or as a byproduct of air ionization. This is a serious concern, because O3 is a criteria air pollutant regulated by health-related federal and state standards. Several types of air purifiers were tested for their ability to produce ozone in various indoor environments at 40-50% relative humidity, including office rooms, bathrooms, bedrooms, and cars. O3 levels generated by personal wearable air purifiers were also tested. In many cases, O3 concentrations were well in excess of public and/or industrial safety levels established by U.S. Environmental Protection Agency, California Air Resources Board, and Occupational Safety and Health Administration. Simple kinetic equations were obtained that can predict the steady-state level of O3 in a room from the O3 emission rate of the air purifier and the first-order decay rate of O3 in the room. The additivity of O3 levels generated by independent O3 generators was experimentally demonstrated.     

Proposed formulation of surface water quality and modelling using gene expression,machine learning,and regression techniques

Environmental Science and Pollution Research - The rising water pollution from anthropogenic factors motivates further research in developing water quality predicting models. The available models...     

Single and interactive effects of low levels of O3, SO2 and NO2 on the growth and yield of spring rape

During three consecutive seasons (1987-1989), the effects of low-levels of O3, SO2 and NO2 singly and in all possible combinations (NO2 in 1988 and 1989 only) on growth and yield of potted plants of spring rape (Brassica napus L. var. napus, 'callypso') were investigated by means of factorial fumigation experiments in open-top chambers. Plants were exposed from the early vegetative stage of development until seed harvest, to charcoal-filtered air (CF; control) and CF which was supplemented for 8-h per day (8.00-16.00) with O3, for 16-h per day with NO2 (16.00-8.00) and continuously with SO2. Including the controls, the 24-h daily mean concentrations [microg m(-3)] ranged between 6-44 (O3), 9-88 (SO2) and 10-43 (NO2). The corresponding daily mean concentrations during the time of fumigation were 10-121 and 11-60 microg m(-3) for O3 and NO2, respectively. Single effects of O3 on growth and yield parameters were mostly negative and the magnitude of this effect was dependent on the season. O3 reduced plant dry weight by 11.3-18.6% and yield of seeds by 11.4-26.9%. While medium levels of SO2 stimulated the weight of pods up to 33%, higher concentrations (88 microg m(-3)) caused a decline of yield of 12.3%. From the significant interactive effects which were observed, it could be established that SO2 and NO2 alone mostly acted positively, but that their interaction with each other and especially with O3 was antagonistic, as some of the detrimental effects of O3 were mitigated by these pollutants. An important antagonistic effect between SO2 and O3 or NO2 was observed on yield. While 56 microg m(-3) SO2 increased yield by 9.9% compared to the control treatment, it aggravated the yield loss caused by O3 from -16.18% to -21.4%, and it reduced the yield stimulation caused by NO2 from +11.8% to +4.2%. Leaf area was the only parameter which was negatively affected by all pollutants, their joint action being synergistic.