Back-extrapolation of estimates of exposure from current land-use regression models

Land use regression has been used in epidemiologic studies to estimate long-term exposure to air pollution within cities. The models are often developed toward the end of the study using recent air pollution data. Given that there may be spatially-dependent temporal trends in urban air pollution and that there is interest for epidemiologists in assessing period-specific exposures, especially early-life exposure, methods are required to extrapolate these models back in time. We present herein three new methods to back-extrapolate land use regression models. During three two-week periods in 2005–2006, we monitored nitrogen dioxide (NO₂) at about 130 locations in Montreal, Quebec, and then developed a land-use regression (LUR) model. Our three extrapolation methods entailed multiplying the predicted concentrations of NO₂ by the ratio of past estimates of concentrations from fixed-site monitors, such that they reflected the change in the spatial structure of NO₂ from measurements at fixed-site monitors. The specific methods depended on the availability of land use and traffic-related data, and we back-extrapolated the LUR model to 10 and 20 years into the past. We then applied these estimates to residential information from subjects enrolled in a case–control study of postmenopausal breast cancer that was conducted in 1996.Observed and predicted concentrations of NO₂ in Montreal decreased and were correlated in time. The estimated concentrations using the three extrapolation methods had similar distributions, except that one method yielded slightly lower values. The spatial distributions varied slightly between methods. In the analysis of the breast cancer study, the odds ratios were insensitive to the method but varied with time: for a 5 ppb increase in NO₂ using the 2006 LUR the odds ratio (OR) was about 1.4 and the ORs in predicted past concentrations of NO₂ varied (OR～1.2 for 1985 and OR～1.3–1.5 for 1996). Thus, the ORs per unit exposure increased with time as the range and variance of the spatial distributions decreased, and this is due partly to the regression coefficient being approximately inversely proportional to the variance of exposure. Changing spatial variability complicates interpretation and this may have important implications for the management of risk. Further studies are needed to estimate the accuracy of the different methods.     

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.     

Comparison of artificial neural network and regression models in the prediction of urban stormwater quality.

Urban stormwater quality is influenced by many interrelated processes. However, the site-specific nature of these complex processes makes stormwater quality difficult to predict using physically based process models. This has resulted in the need for more empirical techniques. In this study, artificial neural networks (ANN) were used to model urban stormwater quality. A total of 5 different constituents were analyzed-chemical oxygen demand, lead, suspended solids, total Kjeldahl nitrogen, and total phosphorus. Input variables were selected using stepwise linear regression models, calibrated on logarithmically transformed data. Artificial neural networks models were then developed and compared with the regression models. The results from the analyses indicate that multiple linear regression models were more applicable for predicting urban stormwater quality than ANN models.     

Mass balances of sulphur and nitrogen oxides over Great Britain

From aircraft measurements in air entering and leaving Great Britain conclusions about the cycles of nitrogen and sulphur oxides on a regional scale are drawn. Using a tracer it is possible to follow the dispersion of plumes over distances of hundreds of km. For sulphur oxides, in the absence of precipitation, overall removal is shown not to be very efficient with 70 % of the emitted SO₂ still airborne over the coast. Estimates are made of the deposition velocity of SO₂ and its rate of oxidation to sulphate. For nitrogen oxides the removal mechanisms in dry conditions appear to be of similar magnitude, though individual loss mechanisms are less certain. The importance of background levels in air entering the country is assessed. The roles of other airborne materials, such as O₃ and ammonium (of which observations were made) are related to the S and NO_x cycles in the atmosphere. It is argued that they and their degree of mixing with polluted air may have a strong influence on removal in rain.     

Influence of air pollution on the foliar nutrition of conifers in Great Britain

The nutritional status of needles from Sitka spruce, Norway spruce and Scots pine in a total of 108 stands was assessed. There was little evidence of nutritional deficiency, although potassium levels were frequently quite low. Analysis of some heavy metals (lead and copper) failed to reveal any likely toxicity problems. Sulphur, nitrogen and iron levels in/on the foliage were all related to various measures of sulphur and nitrogen pollution, determined using improved deposition models that take into account cloud deposition and the seeder-feeder mechanism. The analysis strongly suggested that direct air pollution has a greater effect on sulphur, nitrogen and iron foliar analyses than indirect pollution (wet deposition). The relationships were identified for levels of pollution that were generally lower than those seen in traditional gradient studies.     

The mixing height and mass balance of SO2 in the atmosphere above Great Britain

Measurements of the SO₂ concentration at heights ranging from 150 to 1200 m were made at Cardington, Bedfordshire, using lightweight samplers carried on the cable of the tethered balloon. A value of 1200 m for the mean mixing height for SO₂ was deduced from the measurements. This value indicates that dry deposition limits the mean lifetime of SO₂ to about 2 days. The results of the National Survey of Air Pollution were used to derive an area-mean concentration of 36 μg m⁻³ for SO₂ over Great Britain (excluding the North of Scotland, West Wales and the South West peninsula) and this result is used in a discussion of the mass balance of atmospheric sulphur over this area.     

An empirical map of critical loads of acidity for soils in Great Britain

The method used to produce a critical load map of acidity for soils in Great Britain is described. Critical loads were assigned to the dominant soil in each 1 km grid square of the UK national grid. Mineral soils were assigned a critical load based on mineralogy and chemistry, using approaches appropriate to UK conditions. Critical loads for peat soils are based primarily on a maximum acceptable reduction of peat pH, and results from laboratory equilibration studies. The map shows that soils with small critical loads (<0.5 kmol_c ha⁻¹ year⁻¹) i.e. highly sensitive to acidic deposition, dominate in the north and west of Britain; the south and east are dominated by soils with large critical loads, with small areas of more sensitive soils associated with sandy soil-forming materials. A modified critical load map illustrates the potential impact of agricultural liming on soil critical loads.     

Validity and uncertainty in the calculation of an emission inventory for ammonia arising from agriculture in Great Britain

Emissions of ammonia have received increasing attention recently, following concern about the environmental consequences, especially in The Netherlands where levels are high due to intensive livestock farming. Direct local effects and more widespread consequences for a range of ecosystems have been attributed to ammonia emissions. As the most prevalent alkaline gas in the atmosphere, ammonia interacts with acidic species, changing their characteristics, chemical and physical behaviour, and enhancing their potential for acidification of soils. Ammonia also forms an important component of the nitrogen cycle and of nitrogen deposition. In the UK, as in many other European countries, there has been a considerable increase in the emission of ammonia within the last 30 years, estimated at about 50%. This results mainly from increases in agricultural production based on the steadily rising number of livestock and increasing fertilizer consumption. This paper discusses the various sources to emissions of ammonia from agricultural sources in the United Kingdom, and some of the uncertainties involved in constructing a national emissions inventory.     

The sensitivity of surface waters of Great Britain to acidification predicted from catchment characteristics

Using a combination of soil, land use and geological information, a map of Great Britain has been derived which indicates the sensitivity of surface waters to acidification. For the geological information, a slightly modified version of an available map was used which indicated the sensitivity of groundwaters to acidification. For soils, 1-km databases of soil information for England and Wales and for Scotland were employed to map the soil sensitivity as determined by buffering capacity. The derived soils map was modified to take account of agricultural liming in arable and managed grassland areas using the ITE Land Classification. The final map of surface water sensitivity was obtained by using a geographic information system overlay procedure which enabled each combination of soil and geology sensitivity to be uniquely defined. The final sensitivity classification was based upon expert knowledge and the experience of a similar sensitivity mapping exercise for Wales.     

Comparison of mercury chemistry models

Five mercury (Hg) chemistry models are compared using the same data set for model initialisation. All five models include gas-phase oxidation of Hg(0) to Hg(II) (except for one model), fast reduction–oxidation aqueous reactions between Hg(0) and Hg(II), and adsorption of Hg(II) species to soot particles within droplets. However, the models differ in their detailed treatments of these processes. Consequently, the 48-h simulations reveal similarities but also significant discrepancies among the models. For the simulation that included all Hg species (i.e., Hg(0), Hg(II) and Hg(p)) as well as soot in the initial conditions, the maximum simulated Hg(II) aqueous concentrations ranged from 55 to 148 ng l⁻¹ whereas the minimum concentrations ranged from 20 to 110 ng l⁻¹. These results suggest that further experimental work is critically needed to reduce the current uncertainties in the formulation of Hg chemistry models.     

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.     

Metals in tissues of European otters (Lutra lutra) from Denmark, Great Britain and Ireland.

Samples of livers of European otters from Denmark, Great Britain and Ireland were analyzed for manganese, chromium, zinc, copper, nickel and cobalt. Concentrations were generally significantly higher in samples from Great Britain. It was considered that this reflected the catchment geology from where the specimens originated.     

Validation of regression models for nitrate concentrations in the upper groundwater in sandy soils

For Dutch sandy regions, linear regression models have been developed that predict nitrate concentrations in the upper groundwater on the basis of residual nitrate contents in the soil in autumn. The objective of our study was to validate these regression models for one particular sandy region dominated by dairy farming. No data from this area were used for calibrating the regression models. The model was validated by additional probability sampling. This sample was used to estimate errors in 1) the predicted areal fractions where the EU standard of 50 mg l⁻¹ is exceeded for farms with low N surpluses (ALT) and farms with higher N surpluses (REF); 2) predicted cumulative frequency distributions of nitrate concentration for both groups of farms.Both the errors in the predicted areal fractions as well as the errors in the predicted cumulative frequency distributions indicate that the regression models are invalid for the sandy soils of this study area.     

Trends in large-scale VOC concentrations in the Southern Netherlands between 1991 and 1997

Results of C₆–C₁₂ hydrocarbon measurements at three sites in the southern part of The Netherlands, a polluted region in Western Europe, are presented. The measurements were carried out over the period March 1991–February 1997. The concentrations at the sites, with 100–150 km distance between them, are quite similar and they are predominantly determined by large-scale transport. The concentrations in this part of the country are substantially higher than those observed at a coastal site in the north of The Netherlands, but much lower compared to the concentrations in cities and near streets. A distinct difference between the trends of aromatics and aliphatics was observed. The concentrations of the aromatic components display trends that are systematically 4–5% yr^-1 lower than the trends of the aliphatics, which is possibly related to the increased use of catalysts in cars and, partly, to an enhanced atmospheric chemical activity. For the chlorinated species the trends are highly significant. The trend of 1,1,1-trichloroethane is in the order of 8–12% yr^-1 downward while for tetrachloromethane an annual downward trend of 4–6% is found. These downward trends suggest that measures have been taken to fulfil the requirements of the Montreal Protocol to ban the production of these species in a few years time from now.     

Comparison of models used for the determination of odor thresholds

Several methods of data analysis used for the evaluation of odor detection thresholds have been examined through application to two samples of n-butanol. Panels of seven-ten people, working with a six level, IITRI, ternary forced choice olfactometer, were presented with initial concentrations of 99.5 and 52.1 ppm n-butanol during three trials. The ranking-plotting and ASTM E-679 methods were applied to the evaluation of discrimination-recognition thresholds of the odorous samples. It was found that single evaluations of detection or discrimination-recognition thresholds by either method were always ± 50%of the mean of six trials.The effects of successful guessing on the magnitudes of detection thresholds were examined in terms of a model based on the principle of maximum likelihood estimation of one, two and three trials of panel response. The magnitude of the discrimination threshold obtained by this method always fell between the detection and discrimination-recognition thresholds evaluated by the currently used models. The mean discrimination threshold of n-butanol for six trials was found to be 0.65 ± 0.25 ppm. It appears that the magnitude obtained from one trial with seven panel members would be sufficiently reliable for regulatory purposes when only one field sample is available, since any subsequent trials did not produce threshold values better than ± 40 % of the mean of six tests involving seven and ten panel members exposed to two different initial concentrations.     

Comparing universal kriging and land-use regression for predicting concentrations of gaseous oxides of nitrogen (NOx) for the Multi-Ethnic Study of Atherosclerosis and Air Pollution (MESA Air)

BACKGROUND: Epidemiological studies that assess the health effects of long-term exposure to ambient air pollution are used to inform public policy. These studies rely on exposure models that use data collected from pollution monitoring sites to predict exposures at subject locations. Land use regression (LUR) and universal kriging (UK) have been suggested as potential prediction methods. We evaluate these approaches on a dataset including measurements from three seasons in Los Angeles, CA. METHODS: The measurements of gaseous oxides of nitrogen (NOx) used in this study are from a "snapshot" sampling campaign that is part of the Multi-Ethnic Study of Atherosclerosis and Air Pollution (MESA Air). The measurements in Los Angeles were collected during three two-week periods in the summer, autumn, and winter, each with about 150 sites. The design included clusters of monitors on either side of busy roads to capture near-field gradients of traffic-related pollution. LUR and UK prediction models were created using geographic information system (GIS)-based covariates. Selection of covariates was based on 10-fold cross-validated (CV) R(2) and root mean square error (RMSE). Since UK requires specialized software, a computationally simpler two-step procedure was also employed to approximate fitting the UK model using readily available regression and GIS software. RESULTS: UK models consistently performed as well as or better than the analogous LUR models. The best CV R(2) values for season-specific UK models predicting log(NOx) were 0.75, 0.72, and 0.74 (CV RMSE 0.20, 0.17, and 0.15) for summer, autumn, and winter, respectively. The best CV R(2) values for season-specific LUR models predicting log(NOx) were 0.74, 0.60, and 0.67 (CV RMSE 0.20, 0.20, and 0.17). The two-stage approximation to UK also performed better than LUR and nearly as well as the full UK model with CV R(2) values 0.75, 0.70, and 0.70 (CV RMSE 0.20, 0.17, and 0.17) for summer, autumn, and winter, respectively. CONCLUSION: High quality LUR and UK prediction models for NOx in Los Angeles were developed for the three seasons based on data collected for MESA Air. In our study, UK consistently outperformed LUR. Similarly, the 2-step approach was more effective than the LUR models, with performance equal to or slightly worse than UK.