Investigation into the use of the CUSUM technique in identifying changes in mean air pollution levels following introduction of a traffic management scheme

There is an increasing need for statistical techniques to identify and quantify the effects of traffic management schemes on ambient pollution levels. Cumulative sum (CUSUM) charts have been used extensively in industrial process control to detect deviations in production parameters from pre-determined values. This study investigates the use of the CUSUM procedure to identify change in ambient air pollution levels following the introduction of a traffic management scheme at a specific location in Central London.The CUSUM methods of Lucas first compute the standardised deviations of time series observations from the desired process mean. These are accumulated over time to compute the CUSUM at each time point.Data for the analysis were taken from a kerbside monitoring site on Marylebone Road, a six lane trunk route in Central London. In August 2001 the lane adjacent to the monitoring site was designated as a permanent bus lane. The CUSUM analysis clearly identifies a sustained decrease in carbon monoxide concentrations beginning in 2002. However, seasonality and other factors precluded precise characterisation of the timing of the change. When the analysis was repeated using a reference mean that extrapolated the pre-intervention trend in carbon monoxide concentrations, the CUSUM chart no longer identified a sustained decrease.CUSUM appears to offer a simple and rapid method for identifying sustained changes in pollution levels, but the range of confounding influences on carbon monoxide concentrations, most notably underlying trends, seasonality and independent interventions, complicate its interpretation. Its application in assessing the presence or timing of a stepped change in pollution or similar environmental time series data is recommended in its basic form only where the predicted change is large by comparison with other independent influences. The authors believe that further development of the technique beyond this initial study is worthwhile in order to improve the technique's sensitivity.     

Simplified Methods for Statistical Interpretation of Monitoring Data

Statistical techniques are useful for interpretation of monitoring data for pollutants, process variables, etc. Simplified nomographical methods are presented for relating numbers of samples to confidence intervals for their mean values, and for determining the proportion of the population exceeding a specified concentration and confidence intervals for the proportion. A chart also is given for design of a sampling program for quality control. Illustrative frequency distribution data are given for hourly-averaged methane concentrations in air over three-week periods. They show trimodal lognormal distributions. These charts are applicable to lognormal distributions, as well as to normal distributions. They are convenient for many common problems.     

Characterizing perfluorooctanoate in ambient air near the fence line of a manufacturing facility: comparing modeled and monitored values

In order to improve our understanding of the nature, measurement and prediction of salts of perfluorooctanoic acid (PFOA) in air, two studies were performed along the fence line of a fluoropolymer manufacturing facility. First, a six-event, 24-hr monitoring series was performed around the fence line using the OSHA versatile sampler (OVS) system. Perfluorooctanoate concentrations were determined as perfluorooctanoic acid (PFOA) via liquid chromatography and mass spectrometry. Those data indicated that the majority of the PFOA was present as a particulate. No vapor-phase PFOA was detected above a detection limit of approximately 0.07 microg/m3. A follow-up study using a high-volume cascade impactor verified the range of concentrations observed in the OVS data. Both studies aligned with the major transport direction and range of concentrations predicted by an air dispersion model, demonstrating that model predictions agreed with monitoring results. Results from both monitoring methods and predictions from air dispersion modeling showed the primary direction of transport for PFOA was in the prevailing wind direction. The PFOA concentration measured at the site fence over the 10-week sampling period ranged from 0.12 to 0.9 microg/m3. Modeled predictions for the same time period ranged from 0.12 to 3.84 microg/m3. Less than 6% of the particles were larger than 4 microm in size, while almost 60% of the particles were below 0.3 microm. These studies are believed to be the first published ambient air data for PFOA in the environment surrounding a manufacturing facility.     

Applications of open-path Fourier transform infrared for identification of volatile organic compound pollution sources and characterization of source emission behaviors

An open-path Fourier transform infrared spectroscopy (OP-FTIR) system was set up for 3-day continuous line-averaged volatile organic compound (VOC) monitoring in a paint manufacturing plant. Seven VOCs (toluene, m-xylene, p-xylene, styrene, methanol, acetone, and 2-butanone) were identified in the ambient environment. Daytime-only batch operation mode was well explained by the time-series concentration plots. Major sources of methanol, m-xylene, acetone, and 2-butanone were identified in the southeast direction where paint solvent manufacturing processes are located. However, an attempt to uncover sources of styrene was not successful because the method detection limit (MDL) of the OP-FTIR system was not sensitive enough to produce conclusive data. In the second scenario, the OP-FTIR system was set up in an industrial complex to distinguish the origins of several VOCs. Eight major VOCs were identified in the ambient environment. The pollutant detected wind-rose percentage plots that clearly showed that ethylene, propylene, 2-butanone, and toluene mainly originated from the tank storage area, whereas the source of n-butane was mainly from the butadiene manufacturing processes of the refinery plant, and ammonia was identified as an accompanying reduction product in the gasoline desulfuration process. Advantages of OP-FTIR include its ability to simultaneously and continuously analyze many compounds, and its long path length monitoring has also shown advantages in obtaining more comprehensive data than the traditional multiple, single-point monitoring methods.     

An analysis of continuous black carbon concentrations in proximity to an airport and major roadways

Black carbon (BC), a constituent of particulate matter, is emitted from multiple combustion sources, complicating determination of contributions from individual sources or source categories from monitoring data. In close proximity to an airport, this may include aircraft emissions, other emissions on the airport grounds, and nearby major roadways, and it would be valuable to determine the factors most strongly related to measured BC concentrations. In this study, continuous BC concentrations were measured at five monitoring sites in proximity to a small regional airport in Warwick, Rhode Island from July 2005 to August 2006. Regression was used to model the relative contributions of aircraft and related sources, using real-time flight activity (departures and arrivals) and meteorological data, including mixing height, wind speed and direction. The latter two were included as a nonparametric smooth spatial term using thin-plate splines applied to wind velocity vectors and fit in a linear mixed model framework. Standard errors were computed using a moving-block bootstrap to account for temporal autocorrelation. Results suggest significant positive associations between hourly departures and arrivals at the airport and BC concentrations within the community, with departures having a more substantial impact. Generalized Additive Models for wind speed and direction were consistent with significant contributions from the airport, major highway, and multiple local roads. Additionally, inverse mixing height, temperature, precipitation, and at one location relative humidity, were associated with BC concentrations. Median contribution estimates indicate that aircraft departures and arrivals (and other sources coincident in space and time) contribute to approximately 24–28% of the BC concentrations at the monitoring sites in the community. Our analysis demonstrated that a regression-based approach with detailed meteorological and source characterization can provide insights about source contributions, which could be used to devise control strategies or to provide monitor-based comparisons with source-specific atmospheric dispersion models.     

The Odor Impact Model

Abstract

This paper presents the simulation and field evaluation results of two approaches to localize pollutant emission sources with open-path Fourier transform infrared (OPFTIR) spectroscopy. The first approach combined the plume’s peak location information reconstructed from the Smooth Basis Function Minimization (SBFM) algorithm and the wind direction data to calculate source projection lines. In the second approach, the plume’s peak location was determined with the Monte Carlo methodology by randomly sampling within the beam segment having the largest path-integrated concentration. We first conducted a series of simulation studies to investigate the sensitivity of using different basis functions in the SBFM algorithm. It was found that fitting with the beta and Weibull basis functions generally gave better estimates of the peak locations than with the normal basis function when the plumes were mainly within the OP-FTIR’s monitoring line. However, for plumes that were symmetric to the peak position or spread over the OP-FTIR, fitting with the normal basis function gave better performance. In the field experiment, two tracer gases were released simultaneously from two locations and the OP-FTIR collected data downwind from the sources with a maximum beam path length of 97 m. For the first approach, the release locations were within the 0.25- to 0.5-probability area only after the uncertainty of the peak locations was included in the calculation process. The second approach was easy to implement and still performed as satisfactorily as the first approach. The distances from the sources to the best-fit lines (i.e., the regression lines) of the estimated locations were smaller than 10 m.     

Developing and evaluating techniques for localizing pollutant emission sources with open-path Fourier transform infrared measurements and wind data

This paper presents the simulation and field evaluation results of two approaches to localize pollutant emission sources with open-path Fourier transform infrared (OP-FTIR) spectroscopy. The first approach combined the plume's peak location information reconstructed from the Smooth Basis Function Minimization (SBFM) algorithm and the wind direction data to calculate source projection lines. In the second approach, the plume's peak location was determined with the Monte Carlo methodology by randomly sampling within the beam segment having the largest path-integrated concentration. We first conducted a series of simulation studies to investigate the sensitivity of using different basis functions in the SBFM algorithm. It was found that fitting with the beta and Weibull basis functions generally gave better estimates of the peak locations than with the normal basis function when the plumes were mainly within the OP-FTIR's monitoring line. However, for plumes that were symmetric to the peak position or spread over the OP-FTIR, fitting with the normal basis function gave better performance. In the field experiment, two tracer gases were released simultaneously from two locations and the OP-FTIR collected data downwind from the sources with a maximum beam path length of 97 m. For the first approach, the release locations were within the 0.25- to 0.5-probability area only after the uncertainty of the peak locations was included in the calculation process. The second approach was easy to implement and still performed as satisfactorily as the first approach. The distances from the sources to the best-fit lines (i.e., the regression lines) of the estimated locations were smaller than 10 m.     

Updated methods for assessing the impacts of nearby gas drilling and production on neighborhood air quality and human health

An explosive growth in natural gas production within the last decade has fueled concern over the public health impacts of air pollutant emissions from oil and gas sites in the Barnett and Eagle Ford shale regions of Texas. Commonly acknowledged sources of uncertainty are the lack of sustained monitoring of ambient concentrations of pollutants associated with gas mining, poor quantification of their emissions, and inability to correlate health symptoms with specific emission events. These uncertainties are best addressed not by conventional monitoring and modeling technology, but by increasingly available advanced techniques for real-time mobile monitoring, microscale modeling and source attribution, and real-time broadcasting of air quality and human health data over the World Wide Web. The combination of contemporary scientific and social media approaches can be used to develop a strategy to detect and quantify emission events from oil and gas facilities, alert nearby residents of these events, and collect associated human health data, all in real time or near-real time. The various technical elements of this strategy are demonstrated based on the results of past, current, and planned future monitoring studies in the Barnett and Eagle Ford shale regions.

Implications: Resources should not be invested in expanding the conventional air quality monitoring network in the vicinity of oil and gas exploration and production sites. Rather, more contemporary monitoring and data analysis techniques should take the place of older methods to better protect the health of nearby residents and maintain the integrity of the surrounding environment.     

Evaluation of weather conditions for methyl bromide ambient air monitoring in Monterey, Santa Cruz, and Kern Counties in the year 2000

Weather condition is one of the most important factors affecting spatial and temporal distributions of air pollutants, especially for short-term air dispersion. Abnormal weather conditions might lead to higher or lower ambient air concentrations than they would be under normal weather conditions. Therefore, testing for normality of weather conditions during the air monitoring period is an essential step for evaluating ambient air monitoring results. In this paper, a distance method was used to select a most representative weather station from the available candidates. An array of meteorological elements were identified that affect air dispersion and transportation. A statistical method was used to determine whether the weather conditions during the air monitoring period were significantly different from that of previous years. Using methyl bromide ambient air monitoring as a case study, this paper documents the methods, procedures, and results of weather analysis for Monterey, Santa Cruz, and Kern Counties during ambient air monitoring periods for methyl bromide in the year 2000. With a few exceptions, the meteorological elements and atmospheric stability factors, such as wind speeds, wind directions, and stability classes, during the monitoring period were in the normal range. Although there were higher frequencies of stable atmospheric conditions in Monterey/Santa Cruz Counties than in Kern County, weather conditions during the monitoring period were not significantly different from normal local weather conditions of previous years. Consequently, the subchronic air concentrations observed during the ambient air monitoring periods for methyl bromide in the year 2000 was taken under typical weather conditions of those areas at that time of the year.     

Site Redundancy in Urban Ozone Monitoring

This article describes two statistical methods that enable air pollution control agencies to assess the effectiveness of the spatial distribution of current stationary ozone monitoring networks by providing measures of site redundancy. These methods analyze site redundancy by determining the degree to which ozone measurements at one site can be successfully predicted from data collected at other monitoring sites. The first method, the similarity (SIM) measure, calculates redundancy based on the percentage of common operational days during which two monitoring stations report similar daily maximum ozone concentrations. The second method, a modeling technique, relates site redundancy in ozone measurement to an R-squared statistic from an autoregressive model. The model uses meteorological components recorded at a central location and ozone concentrations reported by the network’s other monitoring stations. Both techniques can assist in effective allocation of limited monitoring resources and offer a statistical approach to ambient air monitoring network design. The techniques are applied to data collected at six ozone monitoring stations in Harris County, Texas, during an eight-year period in the 1980s. The methods identified two sites in the six-site network that exhibit a high degree of redundancy.     

The Steubenville comprehensive air monitoring program (SCAMP): overview and statistical considerations

Average concentrations of particulate matter with an aerodynamic diameter less than or equal to 2.5 microm (PM2.5) in Steubenville, OH, have decreased by more than 10 microg/m3 since the landmark Harvard Six Cities Study associated the city's elevated PM2.5 concentrations with adverse health effects in the 1980s. Given the promulgation of a new National Ambient Air Quality Standard (NAAQS) for PM2.5 in 1997, a current assessment of PM2.5 in the Steubenville region is warranted. The Steubenville Comprehensive Air Monitoring Program (SCAMP) was conducted from 2000 through 2002 to provide such an assessment. The program included both an outdoor ambient air monitoring component and an indoor and personal air sampling component. This paper, which is the first in a series of four that will present results from the outdoor portion of SCAMP, provides an overview of the outdoor ambient air monitoring program and addresses statistical issues, most notably autocorrelation, that have been overlooked by many PM2.5 data analyses. The average PM2.5 concentration measured in Steubenville during SCAMP (18.4 microg/m3) was 3.4 microg/m3 above the annual PM2.5 NAAQS. On average, sulfate and organic material accounted for approximately 31% and 25%, respectively, of the total PM2.5 mass. Local sources contributed an estimated 4.6 microg/m3 to Steubenville's mean PM2.5 concentration. PM2.5 and each of its major ionic components were significantly correlated in space across all pairs of monitoring sites in the region, suggesting the influence of meteorology and long-range transport on regional PM2.5 concentrations. Statistically significant autocorrelation was observed among time series of PM2.5 and component data collected at daily and 1-in-4-day frequencies during SCAMP. Results of spatial analyses that accounted for autocorrelation were generally consistent with findings from previous studies that did not consider autocorrelation; however, these analyses also indicated that failure to account for autocorrelation can lead to incorrect conclusions about statistical significance.     

Statistical and diagnostic evaluation of a new-generation urban dispersion modelling system against an extensive dataset in the Helsinki area

We have developed a modelling system for predicting the traffic volumes, emissions from stationary and vehicular sources, and atmospheric dispersion of pollution in an urban area. This paper describes a comparison of the NO_x and NO₂ concentrations predicted using this modelling system with the results of an urban air quality monitoring network. We performed a statistical analysis to determine the agreement between predicted and measured hourly time series of concentrations at four permanently located and three mobile monitoring stations in the Helsinki Metropolitan Area in 1996–1997 (at a total of ten urban and suburban measurement locations). At the stations considered, the so-called index of agreement values of the predicted and measured time series of the NO₂ concentrations vary between 0.65 and 0.82, while the fractional bias values range from −0.29 to +0.26. In comparison with corresponding results presented in the literature, the agreement between the measured and predicted datasets is good, as indicated by these statistical parameters. The seasonal variations of the NO₂ concentrations were analysed in terms of the relevant meteorological parameters. We also analysed the difference between model predictions and measured data diagnostically, in terms of meteorological parameters, including wind speed and direction (the latter separately for two wind speed classes), atmospheric stability and ambient temperature, at two monitoring stations in central Helsinki. The modelling system tends to overpredict the measured NO₂ concentrations both at the highest (u⩾6 m s⁻¹) and at the lowest wind speeds (u<2 m s⁻¹). For higher wind speeds, the modelling system overpredicts the measured NO₂ concentrations in certain wind direction intervals; specific ranges were found for both monitoring stations considered. The modelling system tends to underpredict the measured concentrations in convective atmospheric conditions, and overpredict in stable conditions. The possible physico-chemical reasons for these differences are discussed.     

Pennsylvania’s Compuiterized Air Monitoring System

The Pennsylvania Air Pollution Commission has developed a regulatory program based upon the control of local air pollution problems and reduction of pollutant levels in air basins. The geographical boundaries of 10 air basins have been established. The Commission’s air basin regulations will provide for the reduction of over-all pollutant levels and for emergency procedures in the event of adverse meteorological conditions. The paper discusses the format and objectives of the program.

In order to effectively enforce the air basin regulations and maintain the necessary surveillance of the state’s air quality, a "computerized real time on-line integrated air monitoring-data handling system" has been designed. The system will incorporate a network to constantly monitor the air in each air basin.The primary objectives of the system are: 1. Constant surveillance of air pollution in the air basins.

2. Provide information on air pollution potential alerts.

3. Aid in further development of air quality criteria and regulations.

The air monitoring network is estimated to include approximately 25 remote stations. Each remote will contain air pollution and meteorological sampling equipment and hardware to telemeter to a central station. The data will be transmitted over leased telephone lines. The central station in Harrisburg will contain the necessary hardware to receive and process data, calculate and display results and permit supervisory control of the network. Output options will include immediate display of edited data, command and alarm information, and presentation of statistical results.

Although the air monitoring system is one of the principle ingredients of the program, the air basin concept encompasses other component systems designed to knit together the entire air pollution control program in Pennsylvania.     

Potential pathways and exposure to explain the human body burden of organochlorine compounds: a multivariate statistical analysis of human monitoring in Würzburg, Germany

By means of multivariate statistical analysis of human monitoring in the vicinity of Würzburg it is examined whether the organochlorine levels in blood data of 309 persons can in part be explained by pollution from outdoor and/or indoor air. Although diet is the main source for intake this study concentrated on the remaining sources. Using 12 personal and geographical variables, which describe some aspects or components of indoor or outdoor exposure, the quantities of the different pathways for the intake are calculated and shown. Besides the impact of the indoor air, which can be pointed out for intake at the workplace for some PCB-congeners, also the hypothesis of the existence of an intake from outdoor air for some organochlorine compounds cannot be rejected. Here, the origin is the spatial variation of the pollution, which can be directly traced back to micro-climatic processes, local relief conditions, and the functional set-up of the town.     

Tomographic reconstruction of tracer gas concentration profiles in a room with the use of a single OP-FTIR and two iterative algorithms: ART and PWLS

Computed tomographic (CT) reconstructions of air contaminant concentration fields were conducted in a room-sized chamber employing a single open-path Fourier transform infrared (OP-FTIR) instrument and a combination of 52 flat mirrors and 4 retroreflectors. A total of 56 beam path data were repeatedly collected for around 1 hr while maintaining a stable concentration gradient. The plane of the room was divided into 195 pixels (13 x 15) for reconstruction. The algebraic reconstruction technique (ART) failed to reconstruct the original concentration gradient patterns for most cases. These poor results were caused by the "highly underdetermined condition" in which the number of unknown values (156 pixels) exceeds that of known data (56 path integral concentrations) in the experimental setting. A new CT algorithm, called the penalized weighted least-squares (PWLS), was applied to remedy this condition. The peak locations were correctly positioned in the PWLS-CT reconstructions. A notable feature of the PWLS-CT reconstructions was a significant reduction of highly irregular noise peaks found in the ART-CT reconstructions. However, the peak heights were slightly reduced in the PWLS-CT reconstructions due to the nature of the PWLS algorithm. PWLS could converge on the original concentration gradient even when a fairly high error was embedded into some experimentally measured path integral concentrations. It was also found in the simulation tests that the PWLS algorithm was very robust with respect to random errors in the path integral concentrations. This beam geometry and the use of a single OP-FTIR scanning system, in combination with the PWLS algorithm, is a system applicable to both environmental and industrial settings.     

A 3D aerosol and visibility information system for urban areas using remote sensing and GIS

Currently, the depiction of urban air quality at boundary layer scale uses modelled climatic and land cover data. However, such models are difficult to verify, and only low to moderate accuracy may be achieved due to the complexity of the input data required and the reliance on assumptions about dispersion patterns. The provision of comprehensive air quality data to urban residents in city districts, at a level of detail commensurate with other Location-Based Services (LBS) which are time- and place-sensitive, has therefore not been possible. A method for urban air quality monitoring over cities at boundary layer scale, other than by the use of air quality models is presented here. The system presented uses empirical Aerosol Optical Thickness (AOT) data in near-real time, combining AOT data from AERONET with aerosol vertical profiles computed from twice-daily MODIS satellite images at 500 m resolution, to give three dimensional (3D) air quality data over the urban landscape. There has been no previous attempt to project the horizontal spatial distribution of aerosols from satellite image pixels into a vertical dimension to give a spatially comprehensive three dimensional record of air quality. The paper describes the sources and accuracy of the AOT data input to the system as well as its storage and retrieval on a Geographic Information System (GIS) platform, to provide air quality and visibility information according to user query at any 3D geographical location, including individual buildings or building floor.     

Source extraction information from air quality data monitored in an Argentinean steel mill

A statistical analysis of a series of ambient air concentrations of suspended particulate matter (SPM) and NO2 is presented. Measurements were taken at four sites that belong to an Argentinean steel mill and in another site located in its vicinity. The air pollutants were measured during a three-week exploratory sampling. The monitoring sites were selected on the basis of relevant characteristics of the emission sources and the corresponding climatological statistics of the last decade. Suspended particulate matter with aerodynamic diameter of less than 10 microm (PM10) and NO2 were continuously measured at only one site, while 1-hr samples of NO2 and 24-hr samples of total SPM and SO2 were collected at the other sites. The registered concentrations show that SPM was the pollutant of major concern. A first estimate about the nature of the contribution of the different sources of particles and NO2 present in the area was obtained through the statistical analysis of measured concentration data coupled with prevalent meteorological variables.     

Redundant Measurements of Urban Air Monitoring Networks in Air Quality Reporting

Abstract

This paper describes a statistical method to assess site redundancy of urban air monitoring networks in reporting daily Pollutant Standards Index (PSI), average concentrations, and the number of exceedances. Such a statistical method has identified significant redundancy in monitoring sites for one-year measurements of two air monitoring networks in Taiwan. There are five redundant sites out of 15 monitoring sites in the Taipei area and eight redundant sites out of 18 monitoring sites in the Kaohsiung area. By using the statistical method presented in this paper to downsize the monitoring networks, we can determine not only the number of redundant sites but also the priority of site removals. The derived sub-networks can maintain consistency in reporting air quality without significant changes in the spatial variations of air measurements for an existing air monitoring network.     

Detecting and characterising small changes in urban nitrogen dioxide concentrations

Detecting and quantifying abnormal changes in the concentration of urban air pollutants can be difficult due to the influence of meteorology and atmospheric chemistry. This study presents methods to detect and characterise small changes in the concentration of nitrogen dioxide (NO₂) that deviate from expected behaviour. Generalized additive models (GAMs) are used to describe daily mean NO₂ concentrations at roadside monitoring sites to determine how concentrations deviate from measured concentrations. Structural change methods are applied to time series describing this difference to identify change-points, where concentrations of NO₂ deviate significantly from expected behaviour. Methods are also used which quantify the timing and uncertainty associated in the timing of these change-points. For most time series data considered in London, concentrations of NO₂ underwent relatively abrupt changes rather than smoothly varying increases; changes which remain largely undetected in raw time series data. Most change-points occurred in late 2002 or early 2003, and the factors, which may have contributed to them, are discussed. These methods can also identify technical problems with monitoring equipment and are applicable to other instances where detecting sudden atmospheric composition change is important.