Application of classification and regression trees for sensitivity analysis of the Escherichia coli O157:H7 food safety process risk model

Microbial food safety process risk models are simplifications of the real world that help risk managers in their efforts to mitigate food safety risks. An important tool in these risk assessment endeavors is sensitivity analysis, a systematic method used to quantify the effect of changes in input variables on model outputs. In this study, a novel sensitivity analysis method called classification and regression trees was applied to safety risk assessment with the use of portions of the Slaughter Module and Preparation Module of the E. coli O157:H7 microbial food safety process risk as an example. Specifically, the classification and regression trees sensitivity analysis method was evaluated on the basis of its ability to address typical characteristics of microbial food safety process risk models such as nonlinearities, interaction, thresholds, and categorical inputs. Moreover, this method was evaluated with respect to identification of high exposure scenarios and corresponding key inputs and critical limits. The results from the classification and regression trees analysis applied to the Slaughter Module confirmed that the process of chilling carcasses is a critical control point. The method identified a cutoff value of a 2.2-log increase in the number of organisms during chilling as a critical value above which high levels of contamination would be expected. When classification and regression trees analysis was applied to the cooking effects part of the Preparation Module, cooking temperature was found to be the most sensitive input, with precooking treatment (i.e., raw product storage conditions) ranked second in importance. This case study demonstrates the capabilities of classification and regression trees analysis as an alternative to other statistically based sensitivity analysis methods, and one that can readily address specific characteristics that are common in microbial food safety process risk models.     

Calculation and uses of mean sediment quality guideline quotients: a critical review

Fine-grained sediments contaminated with complex mixtures of organic and inorganic chemical contaminants can be toxic in laboratory tests and/or cause adverse impacts to resident benthic communities. Effects-based, sediment quality guidelines (SQGs) have been developed over the past 20 years to aid in the interpretation of the relationships between chemical contamination and measures of adverse biological effects. Mean sediment quality guideline quotients (mSQGQ) can be calculated by dividing the concentrations of chemicals in sediments by their respective SQGs and calculating the mean of the quotients for the individual chemicals. The resulting index provides a method of accounting for both the presence and the concentrations of multiple chemicals in sediments relative to their effects-based guidelines. Analyses of considerable amounts of data demonstrated that both the incidence and magnitude of toxicity in laboratory tests and the incidence of impairment to benthic communities increases incrementally with increasing mSQGQs. Such concentration/response relationships provide a basis for estimating toxicological risks to sediment-dwelling organisms associated with exposure to contaminated sediments with a known degree of accuracy. This sediment quality assessment tool has been used in numerous surveys and studies since 1994. Nevertheless, mean SQGQs have some important limitations and underlying assumptions that should be understood by sediment quality assessors. This paper provides an overview of the derivation methods and some of the principal advantages, assumptions, and limitations in the use of this sediment assessmenttool. Ideally, mean SQGQs should be included with other measures including results of toxicity tests and benthic community surveys to provide a weight of evidence when assessing the relative quality of contaminated sediments.     

Hurricane Katrina's impact on New Orleans soils treated with low Lead Mississippi River alluvium

Before Hurricane Katrina flooded 80% of New Orleans, 25 heavily Pb contaminated properties were treated with 15 cm of low Pb Mississippi River alluvium from the Bonnet Carré Spillway (BCS). Four phases of soil collection on the properties included pretreatment (phase 1) median surface soil Pb was 1051 mg/kg (range 5-19 627); after BCS cover (phase 2) the median soil Pb decreased to 6.3 mg/ kg (range 3-18); just before Katrina, a soil collection was partially completed (phase 3); and finally, a post-Katrina collection (phase 4) on all 25 properties. Twenty-three properties were flooded. The objective was to compare Pb changes of phases 2 and 3 pre-Katrina with the post-Katrina (phase 4) soil collection. The post-Katrina soil Pb (phase 4) (median of 16.3, range 5.5-1,155) increased significantly (P-value = 3.7 x 10(-10)) from the clean soil, phase 2 results. Matched soil samples collected in phase 3 before flooding and phase 4 after flooding showed that on initially vacant lots the amounts of Pb were not significantly different (P-value = 0.97); but, on properties with homes, the changes were significant (P-value = 0.03). After catastrophic flooding, the clean soil remained relatively undisturbed; the soil Pb changes were small with increases of median Pb of 12 and 6 mg/kg for vacant lots and properties with homes, respectively. Processes accounting for Pb increases include Pb-based paint abatement on one property, home construction on the vacant lots, and resuspension and deposition of Pb dust. As part of the post-Katrina recovery, the combined benefits of Pb-safe paint abatement or renovation and clean soil cover should outweigh the estimated annual cost of Pb poisoning of children returning to New Orleans.     

Rheological characterization and electrokinetic phenomena of charged whey protein dispersions of defined sizes

A multi-step processing technique produced large colloidal particles from whey proteins, prompting instantaneous thickening upon hydration. Analysis of the rheological characteristics and zeta potential of the modified whey suspensions of defined particle sizes allowed investigation into the role of size on ingredient functionality. Preliminarily, the modified protein powders were sieved to achieve three size ranges, and analyzes were conducted on each of the three distributions and the non-sieved fractions. Following hydration, steady and oscillatory shear analyzes were performed using a controlled stress rheometer to determine rheological characteristics. Intrinsic viscosity was determined with a capillary viscometer and application of the Huggins equation. Zeta potential was calculated from colloidal electrophoretic mobility, measured with a ZetaPALS analyser. After thorough hydration, particle-size analysis revealed a size increase of >1.3 times for each fraction. When analysed on a protein basis, increasing particle size yielded an increase to intrinsic viscosity, flow behavior index, zero shear viscosity, and a decreased zeta potential and consistency coefficient. Knowledge of the interrelationship between zeta potential, rheological properties, and particle size of the modified whey ingredient will further advance an understanding of the functionality of this protein ingredient.     

Contribution of biomass burning to atmospheric polycyclic aromatic hydrocarbons at three European background sites

Radiocarbon analysis of atmospheric polycyclic aromatic hydrocarbons (PAHs) from three background areas in Sweden, Croatia, and Greece was performed to apportion their origin between fossil and biomass combustion. Diagnostic ratios of PAHs implied that wood and coal combustion was relatively more important in the northern European site, while combustion of fossil fuels was the dominant source of PAHs to the two central-southern European background sites. The radiocarbon content (delta14C) of atmospheric PAHs in Sweden ranged between -388 per thousand and -381 per thousand, while more depleted values were observed for Greece (-914 per thousand) and Croatia (-888 per thousand). Using a 14C isotopic mass balance model it was calculated that biomass burning contributes nearly 10% of the total PAH burden in the studied southern European atmosphere with fossil fuel combustion making up the 90% balance. In contrast, biomass burning contributes about 50% of total PAHs in the atmosphere at the Swedish site. Our results suggest that the relative contributions of biomass burning and fossil fuels to atmospheric PAHs may differ considerably between countries, and therefore, different national control strategies might be needed if a further reduction of these pollutants is to be achieved on a continental-global scale.     

Hexabromocyclododecane: current understanding of chemistry, environmental fate and toxicology and implications for global management

Hexabromocyclododecane (HBCD) is a globally produced brominated flame retardant (BFR) used primarily as an additive FR in polystyrene and textile products and has been the subject of intensified research, monitoring and regulatory interest over the past decade. HBCD is currently being evaluated under the Stockholm Convention on Persistent Organic Pollutants. HBCD is hydrophobic (i.e., has low water solubility) and thus partitions to organic phases in the aquatic environment (e.g., lipids, suspended solids). It is ubiquitous in the global environment with monitoring data generally exhibiting the expected relationship between proximity to known sources and levels; however, temporal trends are not consistent. Estimated degradation half-lives, together with data in abiotic compartments and long-range transport potential indicate HBCD may be sufficiently persistent and distributed to be of global concern. The detection of HBCD in biota in the Arctic and in source regions and available bioaccumulation data also support the case for regulatory scrutiny. Toxicity testing has detected reproductive, developmental and behavioral effects in animals where exposures are sufficient. Recent toxicological advances include a better mechanistic understanding of how HBCD can interfere with the hypothalamic-pituitary-thyroid axis, affect normal development, and impact the central nervous system; however, levels in biota in remote locations are below known effects thresholds. For many regulatory criteria, there are substantial uncertainties that reduce confidence in evaluations and thereby confound management decision-making based on currently available information.     

In-use measurement of activity, energy use, and emissions of a plug-in hybrid electric vehicle

Plug-in hybrid electric vehicles (PHEVs) could reduce transportation air emissions and energy use. However, a method is needed for estimating on-road emissions of PHEVs. To develop a framework for quantifying microscale energy use and emissions (EU&E), measurements were conducted on a Toyota Prius retrofitted with a plug-in battery system on eight routes. Measurements were made using the following: (1) a data logger for the hybrid control system; (2) a portable emissions measurement system; and (3) a global positioning system with barometric altimeter. Trends in EU&E are estimated based on vehicle specific power. Energy economy is quantified based on gasoline consumed by the engine and grid energy consumed by the plug-in battery. Emissions from electricity consumption are estimated based on the power generation mix. Fuel use is approximately 30% lower during plug-in battery use. Grid emissions were higher for CO?, NO(x), SO?, and PM compared to tailpipe emissions but lower for CO and hydrocarbons. EU&E depends on engine and plug-in battery operation. The use of two energy sources must be addressed in characterizing fuel economy; overall energy economy is 11% lower if including grid energy use than accounting only for fuel consumption.     

Relationships between milking frequency, lactation persistency and milk yield in Swedish Red heifers and cows milked in a voluntary attendance automatic milking system

A large dataset comprising output from an automatic milking (AM) system between 1999 and 2006 was examined and a total of 172 cow lactation curves and 68 heifer lactation curves were identified for further analysis. Relationships between milking frequency at different stages of lactation and lactation persistency and total lactation yield were determined. Cows had higher peak and total milk yields than heifers, but heifers had higher persistency (defined as the rate of decline in milk yield between days 100 and 300 post calving). Milking frequency did not differ significantly between cows and heifers in early lactation, but thereafter decreased significantly more in cows than in heifers. The effect of milking frequency on yield characteristics was analysed by comparing the highest and lowest quartiles for milking frequency. High milking frequency in early lactation was consistently associated with increased peak yield. High milking frequency averaged across the whole lactation was associated with increased peak yield in both cows and heifers, and with improved lactation persistency in cows only. This resulted in total lactation yield that was 21% greater in the high quartile cows compared with the low.     

Combining land use information and small stream sampling with PCR-based methods for better characterization of diffuse sources of human fecal pollution

Diffuse sources of human fecal pollution allow for the direct discharge of waste into receiving waters with minimal or no treatment. Traditional culture-based methods are commonly used to characterize fecal pollution in ambient waters, however these methods do not discern between human and other animal sources of fecal pollution making it difficult to identify diffuse pollution sources. Human-associated quantitative real-time PCR (qPCR) methods in combination with low-order headwatershed sampling, precipitation information, and high-resolution geographic information system land use data can be useful for identifying diffuse source of human fecal pollution in receiving waters. To test this assertion, this study monitored nine headwatersheds over a two-year period potentially impacted by faulty septic systems and leaky sanitary sewer lines. Human fecal pollution was measured using three different human-associated qPCR methods and a positive significant correlation was seen between abundance of human-associated genetic markers and septic systems following wet weather events. In contrast, a negative correlation was observed with sanitary sewer line densities suggesting septic systems are the predominant diffuse source of human fecal pollution in the study area. These results demonstrate the advantages of combining water sampling, climate information, land-use computer-based modeling, and molecular biology disciplines to better characterize diffuse sources of human fecal pollution in environmental waters.