Evaluation of land-use regression models used to predict air quality concentrations in an urban area

Cohort studies designed to estimate human health effects of exposures to urban pollutants require accurate determination of ambient concentrations in order to minimize exposure misclassification errors. However, it is often difficult to collect concentration information at each study subject location. In the absence of complete subject-specific measurements, land-use regression (LUR) models have frequently been used for estimating individual levels of exposures to ambient air pollution. The LUR models, however, have several limitations mainly dealing with extensive monitoring data needs and challenges involved in their broader applicability to other locations. In contrast, air quality models can provide high-resolution source–concentration linkages for multiple pollutants, but require detailed emissions and meteorological information. In this study, first we predicted air quality concentrations of PM_2.5, NO_x, and benzene in New Haven, CT using hybrid modeling techniques based on CMAQ and AERMOD model results. Next, we used these values as pseudo-observations to develop and evaluate the different LUR models built using alternative numbers of (training) sites (ranging from 25 to 285 locations out of the total 318 receptors). We then evaluated the fitted LUR models using various approaches, including: 1) internal “Leave-One-Out-Cross-Validation” (LOOCV) procedure within the “training” sites selected; and 2) “Hold-Out” evaluation procedure, where we set aside 33–293 tests sites as independent datasets for external model evaluation. LUR models appeared to perform well in the training datasets. However, when these LUR models were tested against independent hold out (test) datasets, their performance diminished considerably. Our results confirm the challenges facing the LUR community in attempting to fit empirical response surfaces to spatially- and temporally-varying pollution levels using LUR techniques that are site dependent. These results also illustrate the potential benefits of enhancing basic LUR models by utilizing air quality modeling tools or concepts in order to improve their reliability or transferability.     

Note on a Method to Estimate Missing Air Pollution Data

An iterative regression procedure is presented to estimate missing air pollution measurements when the data are measured at two or more sampling stations in the same vicinity. The procedure utilizes the measurements taken at other stations, on neighboring days, and of other pollutants.

The procedure is applied to a set of Philadelphia pollution data with from five to seventeen per cent of the observations missing. The method is tested by comparing the known observed pollutant values, with their estimates given by the procedure. Correlations between the observations and their estimates are uniformly high, ranging from 0.87 to 0.91. These correlations compare favorably with those estimates given by a simple linear interpolation. The magnitude of the correlations suggests that estimates given by this iterative regression procedure may be used where missing observations pccur without fear of undesirable effects on subsequent work. Therefore, this procedure may be a valuable tool in handling the problem of missing observations in air pollution data.     

Splitting error estimation for micro-physical–multiphase chemical systems in meso-scale air quality models

Operator splitting applied to cloud micro-physical and multiphase chemical process causes the so-called operator splitting errors in addition to other numerical errors when used in numerical models. Operator splitting is mainly used due to limited computer resources or for historical reasons. Unfortunately, it is impossible so far to theoretically estimate either the order of magnitude or the tendency of the splitting errors in complex non-linear systems such as mutually interacting cloud chemical and micro-physical processes.The present study systematically investigates the splitting error mentioned above, by numerical means to define valid ranges of the applicability of the method of operator splitting to those systems. Results of the current study show that de-coupling intervals larger than 100 s cause an underestimation of the total liquid water content as well as the particle radius of the order of 10% for simulation periods of 1000 s. The maximum overprediction of the total content of dissolved material in the particles is of the order of about 20% for de-coupling intervals of 10–15 min. The error in the sulphate production contributes about 50% to the discrepancy in total aerosol content. Since the de-coupling intervals between dynamical, micro-physical, and chemical processes in most recent air quality models are considerably longer than 15 min, the consequences of the application of operator splitting requires further investigation with respect to predicted aerosol formation, cloud water content, deposition rates, photo-chemistry, cloud optical properties, etc.     

Numerical simulation of photochemical smog formation in Athens, Greece—A case study

High emission levels and the unfavourable topography are the main reasons for the alarming photochemical air pollution levels in Athens. An analysis of available air quality data proves that air pollution levels in Athens are largely affected by local wind circulation systems. The most frequent of these systems is dominated by the phenomenon of the sea breeze. Severe air pollution episodes occur, however, primarily under synoptic situations leading to stagnant conditions in the atmosphere over Athens. Photosmog formation in the Athens Basin is studied with the photochemical dispersion model MARS. The implicit solution algorithm incorporated in MARS is characterized by a variable time increment and a variable order. This solver allows avoiding unnecessary operator splitting by a coupled treatment of vertical diffusion and chemical kinetics. In this paper, MARS is used to analyse the situation on 25 May 1990, a day for which very high air pollution levels were reported in Athens. The simulation results elucidate the characteristics of a photosmog episode under stagnant conditions in Athens. In general, the model results reproduce satisfactorily the observed air pollution patterns.     

The Use of Artificial Activable Trace Elements to Monitor Pollutant Source Strengths and Dispersal Patterns

A tracer technique using certain of the rare earth elements which are easily activated by neutrons has been developed for the analysis of air pollution problems. Studies employing these tracers were made to determine whether the available meteorological dispersion models can be used effectively to describe pollution emissions from selected industries in the vicinity of Albany, Oregon. The Gaussian plume model was found to be satisfactory for the moderately intense turbulence fields which characterize Stability Types B, C, and D, including cases in which the pollution was trapped by an inversion layer aloft. For sources near ground level, however, it was necessary to make allowance for urban influences on plume dispersion. A box model best described the observed dispersal pattern when the upward penetration of the very intense turbulence of Stability Type A was limited by an inversion layer aloft. These meteorological models were applied using a “blind” experimental procedure to predict the emission rates of the effluent from multiple sources of air pollution in the Albany area. It was found that these techniques can be used to predict the rate of emission within a factor of two for multiple sources consisting of three stacks.     

Confidence limits for air quality model evaluations,as estimated by bootstrap and jackknife resampling methods

Air quality models are used to make decisions regarding the construction of industrial plants, the types of fuel that will be burnt and the types of pollution control devices that will be used. It is important to know the uncertainties that are associated with these model predictions. Standard analytical methods found in elementary statistics textbooks for estimating uncertainties are generally not applicable since the distributions of performance measures related to air quality concentrations are not easily transformed to a Gaussian shape. This paper suggests several possible resampling procedures that can be used to calculate uncertainties or confidence limits on air quality model performance. In these resampling methods, many new data sets are drawn from the original data set using an empirical set of rules. A few alternate forms of the socalled bootstrap and jackknife resampling procedures are tested using a concocted data set with a Gaussian parent distributions, with the result that the jackknife is the most efficient procedure to apply, although its confidence bounds are slightly overestimated. The resampling procedures are then applied to predictions by seven air quality models for the Carpinteria coastal dispersion experiment. Confidence intervals on the fractional mean bias and the normalized mean square error are calculated for each model and for differences between models. It is concluded that these uncertainties are sometimes so large for data sets consisting of about 20 elements that it cannot be stated with 95% confidence that the performance measure for the ‘best’ model is significantly different from that for another model.     

Screening Models for Estimating Toxic Air Pollution Near a Hazardous Waste Landfill

This paper is directed to environmental scientists concerned with assessing toxic air pollution downwind of hazardous waste landfills to determine whether potential health threats or exceedances of air quality standards exist. The purpose of this paper is to evaluate the performance of four air quality screening models.

The emission rate of vinyl chloride from the BKK co-disposal landfill in West Covina, California is estimated. Ambient vinyl chloride concentrations are estimated using a ground level point source model, two virtual point source models, arid the simple box model with meteorological and landfill input data representative of periods when ambient monitoring was conducted. The two virtual point source models are most precise and accurate in estimating 24-hour vinyl chloride concentrations. However, the results could include compensating errors in the emission rate and dispersion calculations because the emission rate estimate could not be independently evaluated.     

A synoptic climatological approach to forecast concentrations of sulfur dioxide and nitrogen oxides in Hong Kong

The objective of this study is to develop an automated synoptic climatological procedure to forecast high air pollution concentrations in the most polluted synoptic categories. The procedure is able to identify air masses historically associated with high air pollution concentrations. The arrival of air mass can be predicted 24 or 48 h in advance with the use of the weather forecast data. The development and statistical basis of the procedure are discussed, and an analysis of the procedure's ability to forecast weather conditions associated with high air pollution concentrations is presented. In addition, the dataset of 24 weather variables from 1993 to 1995 is used to validate the procedure. The procedure predicts that 70.3 and 83.3% of total high and severe SO2 concentration days fall into the identified most polluted categories, and the corresponding figures for NOx are 47.8 and 73.7%. The agreement between observed and predicted values is generally good. The prediction models can explain about 58 and 45% of total variance for NOx and SO2 with RMSEs of 42.5 and 16.5 microg m(-3), respectively. They are smaller than 1 SD of the observations.     

Fine particle (PM2.5) measurement methodology, quality assurance procedures, and pilot results of the EXPOLIS study

EXPOLIS is a European multicenter (Athens, Basel, Grenoble, Helsinki, Milan, and Prague) air pollution exposure study. It is the first international, population-based, large-scale study, where personal exposures to PM2.5 aerosol particles (together with volatile organic compounds and carbon monoxide) are being monitored. EXPOLIS is performed in six different centers across Europe, the sampled aerosol concentrations vary greatly, and the microenvironmental samples are not collected with the same equipment as the personal samples. Therefore careful equipment selection, methods development and testing, and thorough quality assurance and quality control (QA & QC) procedures are essential for producing reliable and comparable PM2.5 data. This paper introduces the equipment, the laboratory test results, the pilot results, the standard operating procedures, and the QA & QC procedures of EXPOLIS. Test results show good comparability and repeatability between personal and microenvironmental monitors for PM2.5 at different concentration levels measured across Europe in EXPOLIS centers.     

Deep learning architecture for air quality predictions

With the rapid development of urbanization and industrialization, many developing countries are suffering from heavy air pollution. Governments and citizens have expressed increasing concern regarding air pollution because it affects human health and sustainable development worldwide. Current air quality prediction methods mainly use shallow models; however, these methods produce unsatisfactory results, which inspired us to investigate methods of predicting air quality based on deep architecture models. In this paper, a novel spatiotemporal deep learning (STDL)-based air quality prediction method that inherently considers spatial and temporal correlations is proposed. A stacked autoencoder (SAE) model is used to extract inherent air quality features, and it is trained in a greedy layer-wise manner. Compared with traditional time series prediction models, our model can predict the air quality of all stations simultaneously and shows the temporal stability in all seasons. Moreover, a comparison with the spatiotemporal artificial neural network (STANN), auto regression moving average (ARMA), and support vector regression (SVR) models demonstrates that the proposed method of performing air quality predictions has a superior performance.     

Evaluation of Episode Control Schemes Through Air Quality Data Analysis

Prediction performance of various air pollution episode models are first compared with that of a persistence model which is based on the assumption that present concentrations persist to a future time. The comparisons are made by computing a correlation coefficient for different lead times between the observed and predicted values, and an auto-correlation function of the air quality data to which the episode model is applied. The persistence of high levels of air pollution is next examined, using existing air quality data, by constructing frequency distributions of air pollution episode duration for various concentration thresholds. Based on the results of persistence analysis, the flaws of currently used episode management schemes are discussed and some alternative episode management schemes are presented. Methodologies and parameters to evaluate the anticipated performances of episode management schemes are developed and some examples are worked out. In conclusion, it is suggested that a combination of episode persistence analysis and air pollution meteorological forecasting could lead to a workable air pollution episode management scheme.     

Author’s Reply

A technique is developed to compute precision requirements for component parts of an emissions inventory to ensure (at a given confidence level) an overall acceptable precision in the estimate of total emissions. Since the emissions inventory is a basic requirement of air quality control implementation plans and provides a valuable management tool for planning air pollution control activities, it isi appropriate to state in quantitative terms the confidence that can be associated with each inventory. The approach reported here uses weighted sensitivity analysis methods to distribute both percentage and physical errors in source class emissions according to their contribution to the total emissions, and utilizes Chebyshev’s inequality to establish confidence levels for total emissions. The analysis has been extended to cover the case where one or more of the error components in a given inventory source class can be fixed by the analyst. The utility of the technique is manifold and several practical applications are reported. In particular, it serves to establish percentage error requirements for source categories to satisfy given error bounds for the overall emissions inventory at a given level of statistical confidence. The weighted sensitivity analysis technique possesses a high degree of generality, being applicable to compute component error requirements for any kind of data inventory which exhibits a hierarchical (tree-like) structure, as exemplified by NEDS Emissions Summary Reports. This work should be of interest to air pollution control planners at all levels of government and to anyone responsible for the air pollution portion of environmental impact statements.     

An investigation of distributed lag models in the context of air pollution and mortality time series analysis

In particulate air pollution mortality time series studies, the particulate air pollution exposure measure used is typically the current day's or the previous day's air pollution concentration or a multi-day moving average air pollution concentration. Distributed lag models (DLMs) that allow for differential air pollution effects that are spread over multiple days are seen as an improvement over using a single- or multi-day moving average air pollution exposure measure. However, at the current time, the statistical properties of DLMs as a measure of air pollution exposure have not been investigated. In this paper, a simulation study is used to investigate the performance of DLMs as a measure of air pollution exposure in comparison with single- and multi-day moving average air pollution exposure measures under various forms for the true effect of air pollution on mortality. The simulation study shows that DLMs offer a more robust measure of the effect of air pollution on mortality and avoid the potential for a large negative bias compared with single- or multi-day moving average air pollution exposure measures. This is important information. In many U.S. cities, particulate air pollution concentrations are observed only once every six days, meaning it is often only possible to use single-day particulate air pollution exposure measures. The results from this paper will help quantify the magnitude of the negative bias that can result from using single-day exposure measures. The implications of this work for future air pollution mortality time series studies are discussed. The data used in this paper are concurrent daily time series of mortality, weather, and particulate air pollution from Cook County, IL, for the period 1987-1994.     

The use of wind fields in a land use regression model to predict air pollution concentrations for health exposure studies

A methodology is developed to include wind flow effects in land use regression (LUR) models for predicting nitrogen dioxide (NO₂) concentrations for health exposure studies. NO₂ is widely used in health studies as an indicator of traffic-generated air pollution in urban areas. Incorporation of high-resolution interpolated observed wind direction from a network of 38 weather stations in a LUR model improved NO₂ concentration estimates in densely populated, high traffic and industrial/business areas in Toronto-Hamilton urban airshed (THUA) of Ontario, Canada. These small-area variations in air pollution concentrations that are probably more important for health exposure studies may not be detected by sparse continuous air pollution monitoring network or conventional interpolation methods. Observed wind fields were also compared with wind fields generated by Global Environmental Multiscale-High resolution Model Application Project (GEM-HiMAP) to explore the feasibility of using regional weather forecasting model simulated wind fields in LUR models when observed data are either sparse or not available. While GEM-HiMAP predicted wind fields well at large scales, it was unable to resolve wind flow patterns at smaller scales. These results suggest caution and careful evaluation of regional weather forecasting model simulated wind fields before incorporating into human exposure models for health studies. This study has demonstrated that wind fields may be integrated into the land use regression framework. Such integration has a discernable influence on both the overall model prediction and perhaps more importantly for health effects assessment on the relative spatial distribution of traffic pollution throughout the THUA. Methodology developed in this study may be applied in other large urban areas across the world.     

Air pollution detection by satellites: The transport and deposition of air pollutants over oceans

Research is continuing towards the possible detection of air pollution by remote sensing techniques, and satellite imagery has been examined to find evidence of cross-Atlantic transport of air pollution. Pollution masses from industrial areas are often carried out over the Atlantic Ocean by tropospheric winds. However, the pollution mass is generally steered by convergent flows and fronts of extra-tropical cyclones, and wet deposition and scavenging of air pollutants within clouds occur primarily over the cold ocean, especially during the occlusion stage of a cyclone. As a result, the oceanic area from Cape Hatteras to 1500 km ENE of Newfoundland (the SW sector of the Icelandic low area) is often a ‘dumping ground’ (sink region) for air pollution from N America.However, a dust cloud generated by a volcanic eruption and a smoke plume from large-forest fires in western N America have been observed near the W coast of Europe. Saharan dust carried to N America by trade winds have been identified on satellite imagery. The massive smoke generation by large forest fires in Siberia is also identified in the present study. The results of research on forest fire smoke are currently being used by scientists studying the atmospheric effects of a large-scale nuclear war. It is suggested that the area between the S of Japan and the SW section of the Aleutian low is another principal sink of air pollutants and dust originating from NE Asia.     

SAGA: a decision support system for air pollution management around a coal-fired power plant

Air quality models are currently feasible approaches to prevent air pollution episodes. From one of the first source-oriented modelling approaches for air pollution forecasting (Souto et al., 1994, 1996, 1998), a new decision support system for air quality management, SAGA, was developed to provide support to As Pontes Power Plant (APPP) staff. SAGA can provide air pollution forecasts and manage meteorological and air quality measurements. Power plant decisions are supported by the results of a non-hydrostatic meteorological model (ARPS, Xue et al., 2001) to produce Meteorological Forecasts (MFs), and to be coupled to different Lagrangian dispersion models.     

Box models versus Eulerian models in air pollution modeling

Box models are widely used in air pollution modeling. They allow the use of simple computational tools instead of the simulation of 3D Eulerian grid models, given by a large set of partial differential equations. We investigate here the theoretical justification of such box models. The key point is the comparison with the underlying Eulerian model describing the dispersion of pollutants in the atmosphere. We restrict the study to a vertical monodimensional case for more clarity. The main result is that the nonlinearity of the chemical kinetics, which is a characteristic feature of chemistry, induces the loss of accuracy.     

A Pilot Indoor-Outdoor Study of Organic Particulate Matter and Particulate Mutagenicity

Abstract

Many large metropolitan areas experience elevated concentrations of ground-level ozone pollution during the summertime “smog season”. Local environmental or health agencies often need to make daily air pollution forecasts for public advisories and for input into decisions regarding abatement measures and air quality management. Such forecasts are usually based on statistical relationships between weather conditions and ambient air pollution concentrations. Multivariate linear regression models have been widely used for this purpose, and well-specified regressions can provide reasonable results. However, pollution-weather relationships are typically complex and nonlinear—especially for ozone—properties that might be better captured by neural networks. This study investigates the potential for using neural networks to forecast ozone pollution, as compared to traditional regression models. Multiple regression models and neural networks are examined for a range of cities under different climate and ozone regimes, enabling a comparative study of the two approaches. Model comparison statistics indicate that neural network techniques are somewhat (but not dramatically) better than regression models for daily ozone prediction, and that all types of models are sensitive to different weather-ozone regimes and the role of persistence in aiding predictions.     

The effect of improved nowcasting of precipitation on air quality modeling

The predictive potential of air quality models and thus their value in emergency management and public health support are critically dependent on the quality of their meteorological inputs. The atmospheric flow is the primary cause of the dispersion of airborne substances. The scavenging of pollutants by cloud particles and precipitation is an important sink of atmospheric pollution and subsequently determines the spatial distribution of the deposition of pollutants. The long-standing problem of the spin-up of clouds and precipitation in numerical weather prediction models limits the accuracy of the prediction of short-range dispersion and deposition from local sources. The resulting errors in the atmospheric concentration of pollutants also affect the initial conditions for the calculation of the long-range transport of these pollutants. Customary the spin-up problem is avoided by only using NWP (Numerical Weather Prediction) forecasts with a lead time greater than the spin-up time of the model. Due to the increase of uncertainty with forecast range this reduces the quality of the associated forecasts of the atmospheric flow.In this article recent improvements through diabatic initialization in the spin-up of large-scale precipitation in the Hirlam NWP model are discussed. In a synthetic example using a puff dispersion model the effect is demonstrated of these improvements on the deposition and dispersion of pollutants with a high scavenging coefficient, such as sulphur, and a low scavenging coefficient, such as cesium-137. The analysis presented in this article leads to the conclusion that, at least for situations where large-scale precipitation dominates, the improved model has a limited spin-up so that its full forecast range can be used. The implication for dispersion modeling is that the improved model is particularly useful for short-range forecasts and the calculation of local deposition. The sensitivity of the hydrological processes to proper initialization implies that the spin-up problem may reoccur with changes in the model and increased model resolution. Spin-up should be an ongoing concern for atmospheric modelers.     

Internationally harmonized approach to biomonitoring trace element atmospheric deposition

The International Atomic Energy Agency (IAEA) has been systematically supporting work on biomonitoring air pollution using plants since 1997. Such studies are presently being supported by the IAEA in 14 countries within a co-ordinated research project. The main emphasis of this project is on (1) identification of suitable biomonitors of atmospheric pollution for local and/or regional application, and (2) their validation for general environmental monitoring, whenever possible. Although the participants are using different plants as biomonitors in their research in geographically and climatically diverse parts of the world, they are harmonising sampling approaches and analytical procedures. In this paper, an overview of these activities is given, along with the details, where possible. In all of these activities, proficiency testing and analytical quality assurance are important issues, which merit special attention. Within the scope of an intercomparison exercise, two lichen materials were distributed among the participating laboratories and a proficiency test was organised. The results obtained proved satisfactory performance for most participating laboratories.