Extended prediction models for crashes at roundabouts

This paper builds upon the results of previously developed crash prediction models for roundabouts. The originally investigated sample was extended from 90 to 148 roundabouts. Poisson and gamma modelling techniques were used, the latter ones since underdispersion in the crash data was observed. Separate models were fit for crashes with six different types of road users: bicyclists, motorcyclists, passenger and heavy four-wheel vehicles, moped riders and pedestrians. A further distinction was made between single-vehicle and multiple-vehicle crashes.The results show that the overall number of crashes is more or less proportional to the number of motorized vehicles. The mean number of single-vehicle crashes per passing vehicle is lower on busier roundabouts. Confirmation is found for the existence of a ‘safety-in-numbers’ effect for different types of road users. Three-leg roundabouts tend to perform worse than roundabouts with four or more legs. More crashes seem to occur at roundabouts with bypasses for traffic in some direction. Larger central islands correlate with more single-vehicle crashes. Moped riders and motorcyclists are strongly overrepresented in both single-vehicle and multiple-vehicle crashes whereas bicyclists are clearly overrepresented in multiple-vehicle crashes. Roundabouts with cycle paths perform better than roundabouts with other types of cycle facilities, particularly in comparison with roundabouts with cycle lanes close to the roadway.     

A rule-based screening environmental risk assessment tool derived from EUSES

Within the context and scope of the forthcoming European Union chemical regulations (REACH), there is a need to be able to prioritise the chemicals for evaluation. Therefore, a simple, pragmatic and adequately conservative approach for the identification of substances of very low or no immediate concern at an early stage is presented. The fundamental principles and basic concepts are derived from the EU Technical Guidance Document and EUSES, and are translated into an easy-to-use rule-based system. For this development, the effect on risk characterisation ratios (RCRs) of the key environmental parameters in EUSES was quantified (taking into account several standardised chemical release scenarios). Using statistical analysis, ranges were identified for each key parameter, within which the end result of the assessment was not significantly affected. This information was then translated into a lookup table from which environmental risk characterisation ratios can be directly read as a function of a few parameters.     

Modelling soil series data to facilitate targeted habitat restoration: a polytomous logistic regression approach

Habitat restoration at a landscape scale is becoming increasingly important in environmental management. In this context, geographical information systems are well suited as they can store and integrate many of the abiotic and biotic criteria used to assess the ecological worth of a site. However, this capacity can be limited by the availability or suitability of spatial data sets. A classic example of the latter case is the National Soil Map of England and Wales, which groups soils of a varied nature into associations. Consequently the national soil map has proved to be a poor predictor of habitat suitability. Using polytomous logistic regression we put forward a method for separating soil associations into their constituent soils within the Chilterns Natural Area. This approach used soil association, aspect, slope and relative height as variables for this analysis. Whilst the model's performance is likely to have been limited by the accuracy of the soil association data set, a predictive accuracy of between 47 and 65% is sufficient to facilitate better targeting of habitat restoration when combined with other abiotic factors such as climate and topography.     

Use of genetic algorithms to select input variables in decision tree models for the prediction of benthic macroinvertebrates

Predicting freshwater organisms based on machine learning is becoming more and more reliable due to the availability of appropriate datasets, advanced modelling techniques and the continuously increasing capacity of computers. A database consisting of measurements collected at 360 sampling sites in non-navigable watercourses in Flanders was applied to predict the absence/presence of benthic macroinvertebrate taxa by means of decision trees. The measured variables were a combination of physical–chemical (temperature, pH, dissolved oxygen concentration, conductivity, total organic carbon, Kjeldahl nitrogen and total phosphorus), structural (granulometric analysis of the sediment, width, depth and flow velocity of the river) and two ecotoxicological variables. The predictive power of decision trees was assessed on the basis of the number of Correctly Classified Instances (CCI). A genetic algorithm was introduced to compare the predictive power of different sets of input variables for the decision trees. The number of input variables was reduced from 15 to 2–8 variables without affecting the predictive power of the decision trees significantly. Furthermore, reducing the number of input variables allowed to ease the identification of general data trends.     

Temporal and spatial relationships in fine particle strong acidity, sulphate, PM1O, and PM2.5 across multiple canadian locations

The Canadian Acid Aerosol Measurement Program (CAAMP) was established in 1992 to gain a better understanding of the atmospheric behaviour of fine particle strong acidity (“acid aerosols”) and to facilitate an assessment of the potential health risks associated with acid aerosols and particles in general. During 1992. 1993 and 1994, annular denuder and filter measurements were taken at four sites in Ontario, two in Quebec, three in the Atlantic Provinces and one in the greater Vancouver area. Mean fine particle sulphate concentrations (SO₄²⁻) were highest in southern Ontario (annual average ranged from 40–70 nmol m⁻³), lowest at a site in the Vancouver area (average = 16 nmol m⁻³) and second lowest in rural Nova Scotia. However, mean fine particle strong acid concentrations (H⁺) were geographically different. The highest mean concentrations were at the east coast sites (annual average of up to 30 nmol m⁻³). Acidities were lower in areas where the fine particle acidity experienced greater neutralization from reaction with ammonia. This included the major urban centres (i.e. Toronto and Montréal) and areas with greater amounts of agricultural activity, as in rural southern Ontario. On average, ambient concentrations of fine and coarse particle mass were larger in the urban areas and also in areas where SO₄²⁻ levels were higher. All the particle components were episodic. However, compared to SO₄²⁻ and fine particles (PM_2.5 or PM_2.1, depending upon inlet design), episodes of H⁺ tended to be less frequent and of shorter duration, particularly in Ontario. Saint John, New Brunswick, had the highest mean annual H⁺ concentration, which was 30 nmol m⁻³. H⁺ episodes (24 h concentration > 100 nmol m⁻³) were also the most frequent at this location. The high levels in Saint John were partially due to local sulphur dioxide sources and heterogeneous chemistry occurring in fog, which, on average, led to a 50% enhancement in sulphate, relative to upwind conditions.There was a substantial amount of intersite correlation in the day to day variations in H⁺, SO₄²⁻ , PM_2.5 and PM₁₀ (fine + coarse particles) concentrations, which is due to the influence of synoptic-scale meteorology and the relatively long atmospheric lifetime of fine particles. Sulphate was the most regionally homogenous species. Pearson correlation coefficients comparing SO₄²⁻ between sites ranged from 0.6 to 0.9, depending on site separation and lag time. In many cases, particle episodes were observed to move across the entire eastern portion of Canada with about a two-day lag between the SO₄²⁻ levels in southern Ontario and in southern Nova Scotia.     

Stress and team performance: principles and challenges for intelligent decision aids

Advances in computer technology and artificial intelligence are providing powerful capabilities for developing intelligent decision aids (IDAs) to support operating teams in managing complex systems. Early system designs, which acted as independent computer advisors, constrained human decision making, were brittle and encountered problems of user acceptance. As a result, there has been a shift towards making IDAs more cooperative in assisting users to make decisions under stress. Although there is a growing body of research on how to design IDAs as collaborate team players, this article provides additional insights by focusing on the competencies manifested by efficient teams in adapting to stress. To this extent, a survey of how teams adapt their decision-making strategies, cooperation patterns and team structure has provided a good basis for proposing design principles for collaborative IDAs. The article concludes with some challenges for further developments in information technology and research needs in the area of teamwork under stress.     

Regional climate sensitivity of wetland environments in Rwanda: the need for a location-specific approach

Wetlands are sustaining large communities of people in Rwanda where 10 % of its land surface consists of many local wetlands. Sustainable future management of these numerous wetlands requires a reliable inventory of their location and a dynamic quantitative characterization that allows assessment of their climate change sensitivity. The aim of this study was to assess the importance of climatic factors for determining wetland location at different regional scales. Wetland locations were analyzed and statistically modeled using their location factors with logistic regression. Wetland location probability was determined using topographic (elevation, slope), hydrological (contributing area) and climatic (temperature and rainfall) location factors. A wetland location probability map was made that demonstrated a calibration accuracy of 87.9 % correct at national level compared to an existing inventory, displaying even better fits at subnational level (reaching up to 98 % correct). A validation accuracy of 86.2 % was obtained using an independently collected dataset. A sensitivity analysis was applied to the threshold values used as cutoff value between wetland/non-wetland, demonstrating a robust performance. The developed models were used in a sensitivity scenario analysis to assess future wetland location probability to changes in temperature and rainfall. In particular, wetlands in the central regions of Rwanda demonstrate a high sensitivity to changes in temperature (1 % increase causes a net probable wetland area decline by 12.4 %) and rainfall (+1 % causes a net increase by 1.6 %). This potentially significant impact on wetland number and location probability indicates that climate-sensitive future planning of wetland use is required in Rwanda.     

Mobilization of aluminum by the acid percolates within unsaturated zone of sandstones

The area of the Black Triangle has been exposed to extreme levels of acid deposition in the twentieth century. The chemical weathering of sandstones found within the Black Triangle became well-known phenomenon. Infiltration of acid rain solutions into the sandstone represents the main input of salt components into the sandstone. The infiltrated solutions–sandstone percolates–react with sandstone matrix and previously deposited materials such as salt efflorescence. Acidic sandstone percolates pH?3.2–4.8 found at ten sites within the National Park Bohemian Switzerland contained high Al-tot (0.8–10 mg?L^?1) concentrations and high concentrations of anions SO₄ (5–66 mg?L^?1) and NO₃ (2–42 mg?L^?1). A high proportion (50–98 %) of Al-tot concentration in acid percolates was represented by toxic reactive Alⁿ⁺. Chemical equilibrium modeling indicated as the most abundant Al species Al³⁺, AlSO₄ ⁺, and AlF²⁺. The remaining 2–50 % of Al-tot concentration was present in the form of complexes with dissolved organic matter Al-org. Mobilization and transport of Al from the upper zones of sandstone causes chemical weathering and sandstone structure deterioration. The most acidic percolates contained the highest concentrations of dissolved organic material (estimated up to 42 mg?L^?1) suggesting the contribution of vegetation on sandstone weathering processes. Very low concentrations of Al-tot in springs at BSNP suggest that Al mobilized in unsaturated zone is transported deeper into the sandstone. This process of mobilization could represent a threat for the water quality small-perched aquifers.     

Anthropometric, socioeconomic, and maternal health determinants of placental transfer of organochlorine compounds

The aim of this study was to relate placental transfer, quantified by the cord to maternal serum concentration ratio (C/M), of five organochlorine pesticides (OCP) hexachlorobenzene (HCB), β-hexachlorocyclohexane (β-HCH), γ-hexachlorocyclohexane (γ-HCH) , p,p'-DDT, p,p'-DDE and 15 polychlorinated biphenyl (PCB) congeners (28, 52, 101, 105, 114, 118, 123⁺¹⁴⁹, 138⁺¹⁶³, 153, 156⁺¹⁷¹, 157, 167, 170, 180, and 189) to anthropometric, socioeconomic, and maternal health characteristics. We included into the study 1,134 births during the period 2002–2004 from two districts in eastern Slovakia with high organochlorine concentrations relative to other areas of the world. Only concentrations >LOD were taken into account. Variables as age, weight and height of mothers, parity, ethnicity, alcohol consumption, illness during pregnancy, smoking during pregnancy, hypertension, respiratory diseases, rheumatoid arthritis and diabetes mellitus, and birth weight were related to C/M. Results of regression analyses showed that C/M was predicted by several factors studied. Positive associations were observed for gestational alcohol consumption, fewer illnesses during pregnancy, maternal age, and maternal weight. Caucasians had a greater C/M compared to Romani for wet weight data of congeners 170 and 180 and in contrast C/M for HCB was greater in Romani. Our results show that drinking mothers compared to abstaining expose their fetuses not only to alcohol but to an increased level of several PCB congeners. A straightforward explanation of associations between C/M shifts and factors studied is very difficult, however, with regard to the high lipophilicity of OCPs and PCBs, changes in their kinetics probably reflect lipid kinetics.     

Environmental impact of a coal combustion-desulphurisation plant: abatement capacity of desulphurisation process and environmental characterisation of combustion by-products

The fate of trace elements in a combustion power plant equipped with a wet limestone flue gas desulphurisation (FGD) installation was studied in order to evaluate its emission abatement capacity. With this aim representative samples of feed coal, boiler slag, fly ash, limestone, FGD gypsum and FGD process water and wastewater were analysed for major and trace elements using the following techniques: inductively coupled plasma-mass spectrometry (ICP-MS), inductively coupled plasma-atomic emission spectrometry (ICP-AES), ion chromatography (IC), ion selective electrode (ISE) and atomic absorption spectroscopy (AAS). Mass balances were established allowing to determine the element partitioning behaviour. It was found that, together with S, Hg, Cl, F, Se and As were those elements entering in the FGD plant primarily as gaseous species. The abatement capacity of the FGD plant for such elements offered values ranged from 96% to 100% for As, Cl, F, S and Se, and about 60% for Hg. The environmental characterisation of combustion by-products (boiler slag, fly ash and FGD gypsum) were also established according to the Council Decision 2003/33/EC on waste disposal. To this end, water leaching tests (EN-12457-4) were performed, analysing the elements with environmental concern by means of the aforementioned techniques. According to the leaching behaviour of combustion by-products studied, these could be disposed of in landfills for non-hazardous wastes.