Underlying the Selection of Distributions

To develop a probabilistic model, one has to assign probability distributions to model inputs such as degradation rates, partition coefficients, dose-response parameters (or dose-time-response parameters), exposure values, and so on, for a model relating impacts to exposure. This chapter is concerned with several kinds of technical decisions involved in the selection of distributions. 

The simplest situation is represented by most 1-dimensional (ID) models in which the distributions are taken to represent variability, and where there are adequate data to characterize the distributions. More complicated situations may involve ID modeling with data that are inadequate or problematic (e.g., because of availability of only summary statistics), or the inclusion of uncertainties in 2-dimensional (2D) models. 

For distributions that represent variability, initial decisions may relate to the selection of data on which to base distributions. The problem formulation must identify meaningful populations. Ideally, the data are a random sample from the populations of interest in practice, one may be happy to establish that the data are representative. In addition, data should represent a spatiotemporal scale appropriate for the model. 

Having selected an appropriate data set, we must select a type of distribution and fit the distribution to the data, or else use an empirical or other nonparametric distribution. There appears to be some mechanistic basis for the log-normal distribution, for environmental concentrations (Ott 1990, 1995). However, in a given situation there may not be very strong theoretical support for a specific type of distribution, log-normal or otherwise. Alternative distributions may need to be considered based on the quality of fit of the distribution to data. Therefore, it is desirable to have quantitative indices that can be used to compare or rank distributions based on agreement with data. The fit of the log-normal distribution (or whatever distributions we may choose) should be evaluated in particular situations, using graphical as well as statistical procedures. 

An alternative to choice of a parametric distribution is to rely on a nonparametric distribution. The simplest such distribution is the empirical distribution, which assigns equal probability to each datum in a specified dataset. Considerable 

---

The behavior of elements (toxicity, bioavailability, and distribution) in the environment depends strongly on their chemical forms and type of binding and cannot be reliably predicted on the basis of the total concentration. In order to assess the mobility and reactivity of heavy metal (HM) species in solid samples (soils and sediments), batch sequential extraction procedures are used. HM are fractionated into operationally defined forms under the action of selective leaching reagents. 

One of the most rational means for displacing a broad zone is electrolyte desorption under the conditions of decreasing degree of ionization, i.e., when counterions are converted into dipolar ions, uncharged molecules and coions. This conversion corresponds to a sharp decrease in distribution coefficients of the desorbed substance. Hence, the displacement of equilibrium parame ters at a high rate of mass-exchange is one of the methods of selective stepwise chromatography. 

When using any solvent extraction system, one of the most important decisions is the selection of the solvent to be used. The properties which should be considered when choosing the appropriate solvent are selectivity distribution coefficients insolubility recoverability density interfacial tension chemical reactivity viscosity vapour pressure freezing point safety and cost. A balance must be obtained between the efficiency of extraction (the yield), the stability of the additive under the extraction conditions, the (instrumental and analyst) time required and cost of the equipment. Once extracted the functionality is lost and... 

Fig. 1. Schematic illustration of the basic concept of the Doppler-selected TOF technique. The hatched slice on the left represents a Doppler-selection of a given vz- The strip on the Doppler slice (the middle figure) is the ID Vy-distribution measured under the -restriction of a slit in front of the TOF spectrometer. The combination of many Doppler-selected TOF measurements yields the result shown on the right. The lower figures are the corresponding actual data at each stage for the reaction of S(1D) + H2.

The results on olefin isomers (Table Vll) can also be explained by the observation that the constraint index of ZSM-5 is approximately unity under the conditions of this study. Shape selectivity or preferential conversion of straight chain olefins by ZSM-5 cannot be expected at 500 C. Thus, under the conditions of this study, olefin isomer distribution was not significantly affected by deactivated ZSM-5. At temperatures lower than that employed in the present study, it is conceivable that distribution of olefin isomers could be altered by steam deactivated ZSM-5. 

Table 7 shows the yield distribution of the C4 isomers from different feedstocks with specific processing schemes. The largest yield of butylenes comes from the refineries processing middle distillates and from olefins plants cracking naphtha. The refinery product contains 35 to 65% butanes olefins plants, 3 to 5%. Catalyst type and operating severity determine the selectivity of the C4 isomer distribution (41) in the refinery process stream. Processes that parallel fluid catalytic cracking to produce butylenes and propylene from heavy cmde oil fractions are under development (42). 

Evaluation of VOC and SVOC emission potential of individual products and materials under indoor-related conditions and over defined timescales requires the use of climate-controlled emission testing systems, so-called emission test chambers and cells, the size of which can vary between a few cm3 and several m3, depending on the application. In Figure 5.1 the dots ( ) represent volumes of test devices reported in the literature. From this size distribution they can be classified as large scale chambers, small scale chambers, micro scale chambers and cells. The selection of the systems, the sampling preparation and the test performance all depend on the task to be performed. According to ISO, chambers and cells are defined as follows ... 

The distribution (or partition) coefficient, Ka, of a metal cation between an aqueous (aq) and organic (org) phase may also be used to assess the selectivity of a given host for a range of metal cations under standard conditions, using the equilibrium constants (K) for the following processes (Equations 1.17-1.20) (for metal picrate (Pic) salt, water (aq) and water-saturated chloroform (org) phases, 25 °C). 

---