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.     

Seasonal and spatial patterns of metals at a restored copper mine site II. Copper in riparian soils and Bromus carinatus shoots

Soil and plants were sampled throughout winter and spring near a perennial stream traversing a restored mine site in a winter-rainy climate. Within 1m of an acidic reach of the stream, soil had pH 3-5 and 50-100 microg/g "bioavailable" copper (extractable with 0.01 M CaCl2). Soil 2-3 m from the stream had pH 5-8 and lower (less than 3 microg/g) bioavailable copper. "Oxide-bound" copper (extractable with 2N HCl) was 50-100 microg/g at most locations. Copper concentrations in the shoots of field-collected Bromus carinatus declined from 20 microg/g in winter to 2 microg/g in spring at all sampling sites. A similar temporal pattern was found in plants grown under controlled conditions. Thus B. carinatus has a developmental program for control of shoot copper concentration, causing a seasonally-varying pattern of copper phytoaccumulation over a large range of copper availability in the soil.     

Carbon monoxide emission rates from roasted whole bean and ground coffee

Carbon monoxide (CO) emitted from roasted coffee is a potential occupational respiratory exposure hazard to workers within the coffee industry. The current study objective was to estimate CO emission factors from commercially available roasted whole bean and ground coffee measured in loose form, not packaged, and to assess the utility of CO monitoring in nonventilated storage spaces such as within coffee roasting and packaging facilities, transport vessels, and cafés. Determinants affecting CO emissions from coffee were investigated, including form (whole bean vs. ground), roast level (light, medium, medium-dark, dark), and age (time since the package was opened). CO emission factors were estimated for roasted coffee samples from a variety of manufacturers purchased from local grocery stores and online. Emission tests were performed on 36 brands of coffee, some with more than one sample per brand and with various roast levels. Decaying source equations or smoothing functions were fitted to the CO concentration measurements. Maximum observed emission factors at the peak of the predicted concentration curve were adjusted by the time required to reach the maximum CO concentration and reported as emission factors (EF_buildup). Ground coffee had a significantly increased EF_buildup (P < 0.0001) compared with whole bean. Roast level did not significantly affect emissions for whole bean (P = 0.72) but did for ground (P < 0.001) coffee. For ground coffee, medium-dark and dark roasts had significantly higher emissions than medium and light roasts. Worst-case emission factors from commercially available whole bean and ground coffee measured in loose form, not packaged, showed that roasted coffee can rapidly emit CO. CO concentrations should be monitored in storage spaces in service and manufacturing facilities as well as transport vessels to ensure exposures do not exceed occupational exposure limits. Storage spaces may need to be ventilated to control CO concentrations to safe levels.

Implications: Emission rates of carbon monoxide (CO) from roasted coffee showed that unventilated or underventilated storage spaces should be monitored and ventilated, if necessary, to control CO concentrations to safe levels.     

Systematic comparison of ILWAS, MAGIC, and ETD watershed acidification models: 3. Mass balance budgets for acid neutralizing capacity

Three watershed acidification models-ILWAS, MAGIC, and ETD-were quantitatively compared to determine model structural differences by using a combination of input mapping and ANC mass balance budgets. Input mapping is a set of rules and algorithms to ensure that consistent input values were simultaneously derived for all three models. ANC budget analysis under current SO4(2-) deposition and a 70% reduction in SO4(2-) deposition allows examination of the relative importance of biogeochemical processes in affecting predictions of ANC or predicted changes in ANC. Model inputs were based on two dissimilar watersheds having characteristics typical of watersheds in the northeastern US. After mapping inputs, the three models predicted values of outflow ANC fluxes that were similar among the models for each watershed and deposition scenario. Within each watershed, the changes in outflow ANC fluxes between the scenarios were similar for the three models. Terrestrial weathering was the major source of ANC for all three models for both watersheds and deposition scenarios. The contributions of other processes to the ANC of the two watersheds were, under certain conditions, model-specific. Cation exchange was responsible for changes in ANC when deposition decreased for the three models. Other processes responsible for changes in ANC between scenarios were SO4(2-) sorption (for MAGIC) and in-lake weathering (for ETD). The processes responsible for the change in ANC from a change in deposition (cation exchange, SO4(2-) sorption, and in-lake weathering) were different from the processes contributing to the absolute ANC for a given deposition scenario (terrestrial weathering). The budget analysis complements an earlier Monte Carlo analysis that showed that the three models are structurally different and that predictions viewed on a relative scale are more similar than absolute scale predictions.     

The prognostic factors in the prenatal diagnosis of the echogenic fetal lung

The prenatal diagnosis of an echogenic fetal lung (EFL) is now often made in the early second trimester using high-resolution ultrasound. This ultrasound appearance is usually caused by a congenital cystic adenomatoid lung malformation (CCAM), an intrapulmonary lung sequestration or obstruction of a major airway. In order to provide prognostic guidelines to parents who may be considering termination of a fetus with these findings, we have analysed a series of 11 cases diagnosed in our centre over the past 2 years in conjunction with 60 cases from major published series. The data suggest that in the absence of non-immune hydrops fetalis (NIHF) or other anomalies, the outcome for the fetuses is excellent, with over 90 per cent survival. Neither early diagnosis (24 weeks) nor the presence of mediastinal shift is a poor prognostic indicator. In addition, it appears that if NIHF is absent at diagnosis, the chance that it will develop as the pregnancy continues is small (6 per cent). Furthermore, there is a significant (up to 30 per cent) chance that this ultrasound finding will resolve in utero. The development of in utero fetal surgical techniques may be the only hope for those hydropic fetuses who appear to have a dismal prognosis.     

Improved light extinction reconstruction in interagency monitoring of protected visual environments

The U.S. Environmental Protection Agency (EPA) published the Regional Haze Rule (RHR) in 1999. The RHR default goal is to reduce haze linearly from the baseline period of 2000 through 2004 to natural background in 2064. EPA-recommended method for estimating baseline and natural haze uses the Interagency Monitoring of Protected Visual Environments (IMPROVE) light extinction formula. The IMPROVE formula predicts light extinction from measured aerosol chemical concentrations and estimates of the relative humidity multiplier. On average, the IMPROVE formula overpredicts 6156 nephelometer days (24-hr average measured particle light scattering, bsp) of data by 25%. A new IMPROVED method that reconstructs light extinction using a concentration power law model overpredicts these nephelometer days of data by just 2%. Ignoring the 20% lowest light scattering days, this new IMPROVED formula has a 3% underprediction bias over the 4925 highest nephelometer days with light scattering > or =8 inverse megameters. For comparison, the IMPROVE formula has a 12% overprediction bias for the same days. The IMPROVE formula overprediction averages 77%, 27%, 17%, 9%, and -5% broken down by quintile from lowest to highest nephelometer measured light scattering days. The new IMPROVED formula average overprediction is 21%, -5%, -5%, -2%, and 0%. So, agreement between measured and predicted light scattering improves by modifying the current IMPROVE light extinction formula.     

Ozone dynamics and deposition processes at a deforested site in the Amazon basin

The goal of this study is to investigate the impact of deforestation on ozone dynamics and deposition in the Brazilian Amazon basin. This goal is accomplished through i) analyses of ozone levels and deposition rates at a deforested site during the rainy season; and ii) comparisons of these data with similar information derived at a forest. At the pasture site maximum ozone mixing ratios reach 20 parts per billion on a volume basis (ppbv) but about 6 ppbv prevail over the forest. Maximum ozone deposition velocities for pastures can reach 0.7 cm s-1, which is about threefold lower than values derived for forests. Combining ozone abundance and deposition velocities, pasture maximum ozone fluxes reach approximately 0.2 microgram (ozone) m-2 s-1. This flux represents approximately 70% of the deposition rates measured over the forest. Hence, this study suggests that conversion of rainforests to pastures could lead to a net reduction (30%) in the ozone sink in the Amazon.     

Discrimination between nestmate and non-nestmate kin by social wasps (Polistes fuscatus,Hymenoptera: Vespidae)

Summary The threespine stickleback,Gasterosteus aculeatus is a voracious cannibal of both its own eggs and those of conspecific neighbours. Females, but not males, can distinguish their progeny from those of other fish and attacked alien eggs more frequently. In experiments to examine nest raiding, females that initiated raids on nests resulting in cannibalism were the first females to spawn in the reconstructed nest. These results support the hypothesis that cannibalism by females may be adaptive in situations where intense female competition for male guardians occurs.     

Robust benchmark dose determination based on profile score methods

We investigate several methods commonly used to obtain a benchmark dose and show that those based on full likelihood or profile likelihood methods might have severe shortcomings. We propose two new profile likelihood-based approaches which overcome these problems. Another contribution is the extension of the benchmark dose determination to non full likelihood models, such as quasi-likelihood, generalized estimating equations, which are widely used in settings such as developmental toxicity where clustered data are encountered. This widening of the scope of application is possible by the use of (robust) score statistics. Benchmark dose methods are applied to a data set from a developmental toxicity study.     

Incomplete mixing in a small, urban stream

Conservative solute tracer experiments were conducted in Indian Creek, a small urban stream located in Philadelphia, Pennsylvania, USA. Estimated flow rates were between 46 Ls(-1) and 81 Ls(-1), average stream width was 5.5m and average stream depth was 0.2m. Given these dimensions, most researchers would think it reasonable to assume that the stream is completely mixed vertically and horizontally. However, we found that the stream was not vertically completely mixed in a 1.0m deep, 30 m long pool. The limited mixing was demonstrated by the vertical stratification of a tracer cloud which was completely mixed both laterally and vertically across the stream prior to entering the pool. We suggest that the cause of limited mixing is due to a balance between groundwater inflow and transverse dispersion at the cross-section. We show that the unsupported assumption of complete mixing may result in a wide range, and thus increased uncertainty, of the values of stream flow and longitudinal dispersion coefficient estimated from these data. We conclude that the assumption of complete mixing and one-dimensional modeling must be checked against actual field conditions, even in small streams.