Modeling the long-term frequency distribution of regional ozone concentrations

Efficient methods are developed for modeling emissions – air quality relationships that govern ozone and NO₂ concentrations over very long periods of time. A baseline model evaluation study is conducted to assess the accuracy and speed with which the relationship between pollutant emissions and the frequency distribution of O₃ concentrations throughout the year can be computed along with annual average NO₂ values using a deterministic photochemical airshed model driven by automated objective analysis of measured meteorological parameters. Methods developed are illustrated by application to the air quality situation that exists in Southern California. Model performance statistics for O₃ are similar to the results obtained in previous short-term episodic model evaluation studies that were based on hand-crafted meteorological inputs that are supplemented by expensive field measurement campaigns. Model predictions at one of the highest NO₂ concentration sites in the US indicate that measured violation of the US annual average NO₂ air quality standard at that site occurs because other species such as HNO₃ and PAN are measured as if they were NO₂ by the chemiluminescent NO_x monitors in current use.     

Estimates of AOT ozone indices from time-integrated ozone data and hourly air temperature measurements in southwest Sweden

Surface ozone concentration and surface air temperature was measured hourly at three coastal sites, four low elevation inland sites and two high elevation inland sites in southwestern Sweden. Diurnal ozone concentration range (DOR) and diurnal temperature range (DTR) were strongly correlated, both spatially and temporally, most likely because both depended on atmospheric stability. Accumulated ozone exposure above a threshold concentration of x nmol mol⁻¹ (AOTx) was estimated from time-integrated ozone concentration (as from diffusive sampling) and measures of ozone concentration variability. Two methods both estimated 24-h AOTx with high accuracy (modelling efficiencies >90% for x ≤ 40 nmol mol⁻¹). Daytime (08:00–20:00) AOTx could not be equally well estimated. Estimates were better for lower AOT thresholds. Diffusive ozone concentration sampling, combined with hourly temperature monitoring, could be a valuable complement to ozone concentration monitoring with continuous instruments.     

A multi-variate statistical model integrating passive sampler and meteorology data to predict the frequency distributions of hourly ambient ozone (O3) concentrations

A multi-variate, non-linear statistical model is described to simulate passive O3 sampler data to mimic the hourly frequency distributions of continuous measurements using climatologic O3 indicators and passive sampler measurements. The main meteorological parameters identified by the model were, air temperature, relative humidity, solar radiation and wind speed, although other parameters were also considered. Together, air temperature, relative humidity and passive sampler data by themselves could explain 62.5-67.5% (R(2)) of the corresponding variability of the continuously measured O3 data. The final correlation coefficients (r) between the predicted hourly O3 concentrations from the passive sampler data and the true, continuous measurements were 0.819-0.854, with an accuracy of 92-94% for the predictive capability. With the addition of soil moisture data, the model can lead to the first order approximation of atmospheric O3 flux and plant stomatal uptake. Additionally, if such data are coupled to multi-point plant response measurements, meaningful cause-effect relationships can be derived in the future.     

Prediction of hourly O3 concentrations using support vector regression algorithms

In this paper we present an application of the Support Vector Regression algorithm (SVMr) to the prediction of hourly ozone values in Madrid urban area. In order to improve the training capacity of SVMrs, we have used a recently proposed approach, based on reductions of the SVMr hyper-parameters search space. Using the modified SVMr, we study different influences which may modify the ozone prediction, such as previous ozone measurements in a given station, measurements in neighbors stations, and the influence of meteorologic variables. We use statistical tests to verify the significance of incorporating different variables into the SVMr. A comparison with the results obtained using a neural network (multi-layer perceptron) is also carried out. This study has been carried out in 5 different stations of the air pollution monitoring network of Madrid, so the conclusions raised are backed by real data. The final result of the work is a robust and powerful software for tropospheric ozone prediction in Madrid. Also, the prediction tool based on SVMr is flexible enough to incorporate any other prediction variable, such as city models, or traffic patters, which may improve the prediction obtained with the SVMr.     

Variation of ozone concentrations in the ambient air of Baghdad

Measurements of ozone concentration in the ambient air of the city of Baghdad were carried out for the period October 1983 to October 1984. The O₃, probably of local origin, showed a typical diurnal and seasonal variation. Maximum daily 1-h O₃ concentrations higher than the international ambient air quality standards were observed regularly during the summer months. High O₃ concentrations during the night were also observed. Scatter diagrams were used to relate the O₃ concentrations with temperature, solar radiation and humidity.     

The influence of the photolysis rates on modelled ozone concentrations

The sensitivity of ozone formation in a photochemical model to the photolysis rates has been studied using two photolytic schemes (old EMEP and UKPTM). The results from the comparison of these rates illustrate how the different parameterisation may influence the ozone development. The implementation of the UKPTM photolysis rates into the EMEP chemical scheme resulted in lower ozone levels than those predicted using the old EMEP photolytic scheme.     

Dynamic evaluation of a regional air quality model: Assessing the emissions-induced weekly ozone cycle

Air quality models are used to predict changes in pollutant concentrations resulting from envisioned emission control policies. Recognizing the need to assess the credibility of air quality models in a policy-relevant context, we perform a dynamic evaluation of the Community Multiscale Air Quality (CMAQ) modeling system for the “weekend ozone effect” to determine if observed changes in ozone due to weekday-to-weekend (WDWE) reductions in precursor emissions can be accurately simulated. The weekend ozone effect offers a unique opportunity for dynamic evaluation, as it is a widely documented phenomenon that has persisted since the 1970s. In many urban areas of the Unites States, higher ozone has been observed on weekends than weekdays, despite dramatically reduced emissions of ozone precursors (nitrogen oxides [NO_x] and volatile organic compounds [VOCs]) on weekends. More recent measurements, however, suggest shifts in the spatial extent or reductions in WDWE ozone differences. Using 18 years (1988–2005) of observed and modeled ozone and temperature data across the northeastern United States, we re-examine the long-term trends in the weekend effect and confounding factors that may be complicating the interpretation of this trend and explore whether CMAQ can replicate the temporal features of the observed weekend effect. The amplitudes of the weekly ozone cycle have decreased during the 18-year period in our study domain, but the year-to-year variability in weekend minus weekday (WEWD) ozone amplitudes is quite large. Inter-annual variability in meteorology appears to influence WEWD differences in ozone, as well as WEWD differences in VOC and NO_x emissions. Because of the large inter-annual variability, modeling strategies using a single episode lasting a few days or a few episodes in a given year may not capture the WEWD signal that exists over longer time periods. The CMAQ model showed skill in predicting the absolute values of ozone concentrations during the daytime. However, early morning NO_x concentrations were underestimated and ozone levels were overestimated. Also, the modeled response of ozone to WEWD differences in emissions was somewhat less than that observed. This study reveals that model performance may be improved by (1) properly estimating mobile source NO_x emissions and their temporal distributions, especially for diesel vehicles; (2) reducing the grid cell size in the lowest layer of CMAQ; and, (3) using time-dependent and more realistic boundary conditions for the CMAQ simulations.     

Neural network forecasting of air pollutants hourly concentrations using optimised temporal averages of meteorological variables and pollutant concentrations

The new method for the forecasting hourly concentrations of air pollutants is presented in the paper. The method was developed for a site in urban residential area in city of Zagreb, Croatia, for four air pollutants (NO₂, O₃, CO and PM₁₀). Meteorological variables and concentrations of the respective pollutant were taken as predictors. A novel approach, based on families of univariate regression models, was employed in selecting the averaging intervals for input variables. For each variable and each averaging period between 1 and 97 h, a separate model was built. By inspecting values of the coefficient of correlation between measured and modelled concentrations, optimal averaging periods for each variable were selected. A new dataset for building the forecasting model was then calculated as temporal moving averages (running means) of former variables. A multi-layer perceptron type of neural networks is used as the forecasting model. Index of agreement, calculated for the entire dataset including the data for model building, ranged from 0.91 to 0.97 for the respective pollutants. As suggested by the analysis of the relative importance of the input variables, different agreements for different pollutants are likely due to different sources and production mechanisms of investigated pollutants. A comparison of the new method with more traditional method, which takes hourly averages of the forecast hour as input variables, showed similar or better performance. The model was developed for the purpose of public-health-oriented air quality forecasting, aiming to use a numerical weather forecast model for the prediction of the part of input data yet unknown at the forecasting time. It is to expect that longer term averages used as inputs in the proposed method will contribute to smaller input errors and the greater accuracy of the model.     

Annual and weekly patterns of ozone and particulate matter in Jeddah,Saudi Arabia

Air pollution has been an increasing concern within the Kingdom of Saudi Arabia and other Middle Eastern countries. In this work the authors present an analysis of daily ozone (O₃), nitrogen oxide (NO_x), and particulate matter (<10 μm aerodynamic diameter; PM₁₀) concentrations for two years (2010 and 2011) at sites in and around the coastal city of Jeddah, as well as a remote background site for comparison. Monthly and weekly variations, along with their implications and consequences, were also examined. O₃ within Jeddah was remarkably low, and exhibited the so-called weekend effect—elevated O₃ levels on the weekends, despite reduced emissions of O₃ precursors on those days. Weekend O₃ increases averaged between 12% and 14% in the city, suggesting that NO_x/volatile organic compound (VOC) ratios within cities such as Jeddah may be exceptionally high. Sites upwind or far removed from Jeddah did not display this weekend effect. Based on these results, emission control strategies in and around Jeddah must carefully address NO_x/VOC ratios so as to reduce O₃ at downwind locations without increasing it within urban locations themselves. PM₁₀ concentrations within Jeddah were elevated compared with North American cites of similar climatology, though comparable to other large cities within the Middle East.

Implications:Daily concentrations of O₃, PM₁₀, and NO_x in and around the city of Jeddah, Saudi Arabia, are analyzed and compared with those of other reference cities. Extremely low O₃ levels, along with a significant urban weekend effect (higher weekend O₃, despite reduced NO_x concentrations), is apparent, along with high levels of PM₁₀ within the city. Urban O₃ in Jeddah was found to be lower than that of other comparable cities, but the strong weekend effect suggests that care must be taken to reduce downwind O₃ levels without increasing them within the city itself. Further research into the emissions and chemistry contributing to the reduced O₃ levels within the city is warranted.     

Integrated assessment modeling of atmospheric pollutants in the Southern Appalachian Mountains. Part I: hourly and seasonal ozone

Recently, a comprehensive air quality modeling system was developed as part of the Southern Appalachians Mountains Initiative (SAMI) with the ability to simulate meteorology, emissions, ozone, size- and composition-resolved particulate matter, and pollutant deposition fluxes. As part of SAMI, the RAMS/EMS-95/URM-1ATM modeling system was used to evaluate potential emission control strategies to reduce atmospheric pollutant levels at Class I areas located in the Southern Appalachians Mountains. This article discusses the details of the ozone model performance and the methodology that was used to scale discrete episodic pollutant levels to seasonal and annual averages. The daily mean normalized bias and error for 1-hr and 8-hr ozone were within U.S. Environment Protection Agency guidance criteria for urban-scale modeling. The model typically showed a systematic overestimation for low ozone levels and an underestimation for high levels. Because SAMI was primarily interested in simulating the growing season ozone levels in Class I areas, daily and seasonal cumulative ozone exposure, as characterized by the W126 index, were also evaluated. The daily ozone W126 performance was not as good as the hourly ozone performance; however, the seasonal ozone W126 scaled up from daily values was within 17% of the observations at two typical Class I areas of the SAMI region. The overall ozone performance of the model was deemed acceptable for the purposes of SAMI's assessment.     

Stability of two prediction schemes for hourly nitrogen oxide concentrations by a regression model

In a preceding paper (Inoue et al, 1986b, Atmospheric Environment20, 2325–2337), we proposed two prediction schemes for hourly nitrogen oxide (NO) concentration using the regression model with autocorrelated error terms, and applied these schemes to the prediction of NO-concentration at m h later. Data analyses based on the data sets observed at a measurement site for a year showed that these schemes are practical for the predictions 1 h later.In this follow up paper, we apply these schemes to the prediction for other years and/or other measurement sites, and investigate the stability of the assumed regression model and the adopted prediction schemes not only numerically but also theoretically. From these analyses, our prediction schemes are shown to be stable for the prediction of NO-concentration in other years.     

Trends in near-surface ozone concentrations in Switzerland: the 1990s

A time series analysis of ozone monitoring data from several locations in Switzerland from 1991 to 1999 is presented. Different methods are used to address changes in the ozone level during these years and to account for the influence of changing meteorological conditions. The results show a slight decrease of the peaks but a highly significant increase of the mean value of around 0.5–0.9 ppb yr⁻¹. The frequency distribution has changed in the sense that very low values have become less frequent and that there is a strong increase in frequency of occurrence of half-hourly mean values between about 45 and 55 ppb. A selection procedure reveals slight tendencies towards different trends of afternoon ozone peaks in summer depending on weather and pollution situations. Ozone peaks tend to decrease on fair weather days at rural sites (but increase at urban sites) and show a small increase on cloudy and windy days. A non-linear regression model is used to estimate trends of summertime afternoon ozone peaks in the presence of meteorological variability. In the model, the long-term signal is additively split into a linear part and a part which is modulated by global radiation. The coefficients for both terms are statistically significant at many sites, with an increasing linear trend at the sites north of the Alps of around 1 ppb yr⁻¹ and a decrease of ozone peaks under fair weather conditions relative to cloudy conditions. When additionally considering the effect of precursor concentrations in the regression models, both trends are weakened, which means that they can partly be explained by changes in local to regional emissions. However, at the sites north of the Alps remains a tendency towards a positive linear “base trend” of around 0.4 ppb yr⁻¹. This could possibly be due to increasing background ozone concentrations.     

Radial diffusive sampler for the determination of 8-h ambient ozone concentrations

The 8-h ozone radial diffusive sampler was evaluated according to the CEN protocol for the validation of diffusive samplers. All the parameters regarding the sampler characteristics were found to be consistent with the requirements of this protocol apart from the blank value, which must be evaluated and subtracted at each sampling. The nominal uptake rate was determined in laboratory conditions. However, the uptake rate depends on the mass uptake, temperature, humidity and on the combination of temperature and humidity. Based on laboratory experiments, an empirical model has been established which improved the agreement between the radial sampler and the reference method. This improvement was observed under several different meteorological and emission conditions of sampling. By using the model equation of uptake rate, the data quality objective of 30% for the expanded uncertainty included in the O(3) European Directive, is easily attained. Therefore, the sampler represents an appropriate indicative method.     

Spatiotemporal modelling of ozone distribution in the State of California

This paper is concerned with the spatiotemporal mapping of monthly 8-h average ozone (O₃) concentrations over California during a 15-years period. The basic methodology of our analysis is based on the spatiotemporal random field (S/TRF) theory. We use a S/TRF decomposition model with a dominant seasonal O₃ component that may change significantly from site to site. O₃ seasonal patterns are estimated and separated from stochastic fluctuations. By means of Bayesian Maximum Entropy (BME) analysis, physically meaningful and sufficiently detailed space–time maps of the seasonal O₃ patterns are generated across space and time. During the summer and winter months the seasonal O₃ concentration maps exhibit clear and progressively changing geographical patterns over time, suggesting the existence of relationships in accordance with the typical physiographic and climatologic features of California. BME mapping accuracy can be superior to that of other techniques commonly used by EPA; its framework can rigorously assimilate useful data sources that were previously unaccounted for; the generated maps offer valuable assessments of the spatiotemporal O₃ patterns that can be helpful in the identification of physical mechanisms and their interrelations, the design of human exposure and population health models, and in risk assessment. As they focus on the seasonal patterns, the maps are not contingent on short-time and locally prevalent weather conditions, which are of no interest in a global and non-forecasting framework. Moreover, the maps offer valuable insight about the space–time O₃ concentration patterns and are, thus, helpful for disentangling the influence of explanatory factors or even for identifying some influential ones that could have been otherwise overlooked.     

Understanding the effectiveness of precursor reductions in lowering 8-hr ozone concentrations

Analyses of ambient measured ozone data were used in conjunction with the application of photochemical modeling to determine the technical feasibility of attaining the federal 8-hr ozone standard in central California. Various combinations of volatile organic compound (VOC) and oxides of nitrogen (NOx) emission reductions were effective in lowering modeled peak 1-hr ozone concentrations. However, VOC emissions reductions were found to have only a modest impact on modeled peak 8-hr ozone concentrations. NOx emission reductions generally lowered 8-hr ozone concentrations, but their effectiveness was partially or, in some cases, wholly offset by the increase in the number of NO cycles and, hence, in the ozone produced per NO. As a result, substantial NOx emission reductions--70 to 90%--were required to reduce peak 8-hr ozone concentrations to the level of the standard throughout the modeling domain. These modeling results provide a possible physical explanation for recent analyses that have reported more prominent trends in peak 1-hr ozone levels than in peak 8-hr ozone concentrations or in occurrences of mid-level (60-90 parts per billion by volume) ozone concentrations. The findings also have serious implications for the feasibility of attaining the 8-hr ozone standard in central California. Further efforts are needed to clarify the applicability of the modeling results to the full set of days with ozone levels exceeding the 8-hr ozone standard, as well as their applicability to other geographical areas.     

Preliminary background ozone concentrations in the mountain and coastal areas of Bulgaria

Urban and non-urban rural ozone (O3) concentrations are high in Bulgaria and often exceed the European AOT40 ecosystem as well as the AOT60 human health standards. This paper presents preliminary estimates to establish background, non-urban O3 concentrations for the southern region of Bulgaria. Ozone concentrations from three distinctly different sites are presented: a mountain site influenced by mountain-valley wind flow; a coastal site influenced by sea-breeze wind flow; and a 1700-m mountain peak site without 'local' wind flow characteristics. The latter offers the best estimate of 46-50 ppb for a background O3 level. The highest non-urban hourly value, 118 ppb, was measured at the mountain-valley site.     

Factors influencing the ground level distribution of ozone in Europe

Ozone is a widely distributed pollutant in the atmospheric boundary layer over north west Europe. Three main sources have been identified: the stratosphere, the free troposphere and boundary layer photochemical production. The pattern of ground level ozone concentrations resulting from these three sources cannot be accurately specified. Ozone shows significant variations in space and time but because of the high cost of continuous monitoring equipment, spatial variations on a national and international basis have not been studied in detail. Variations in ozone concentrations at individual monitoring sites have been given a great deal of attention and experience gained from United Kingdom monitoring sites is described in some detail. The averaging time statistical model of Larsen is employed to relate the exposure levels measured over different averaging periods. Diurnal variations have a major influence on exposure levels at sites nominally exposed to the same regional ozone distribution. The physical and chemical mechanisms which give rise to diurnal variations are detailed so that sites can be screened for different diurnal behaviour characteristics.