3D Field-Scale Environmental Multimedia System Validation of the Dispersion of Benzene for Trail Road Landfill Site and Its Risk Assessment ()
1. Introduction
Early environmental decision making was based on qualitative descriptions of the effects of pollutant dispersion on organisms and the environment, with some reliance on the assumption that the protection of human health would also ensure adequate protection of the environment. Current information and environmental regulations suggest the need for a more quantitative, risk-based approach to decision making for environmental protection. A consultative risk assessment approach is necessary to evaluate the scale of potential hazardous environmental impacts on ecology and human health, such as the Hazardous Waste Identification Rule frameworks [1], the Total Risk Integrated Methodology system [2], the Multimedia Integrated Modeling System [3], the finitesource multimedia, multipathway, and multi-receptor risk assessment [4], and Chemical Hazard Assessment and Risk Management [5]. It has been found that these risk assessments have not taken all factors into consideration. This inadequacy may lead to an incorrect assessment of the risk level. In recent years, much more research, integrated with multimedia models, has been concentrated on human and risk assessments, such as the Monte Carlo method (MCM) [6-8]. This consists of characterizing the risk that a substance poses to human and nonhuman organisms by considering its inherent toxicity and the potential for exposure. In a Monte Carlo analysis, a sample from the distribution of an input parameter is placed in a simulation run to interact in a model with samples from other input parameters. A number of studies have been carried out on the risk assessment of environmental multimedia assessment using the MCM to conduct probabilistic analysis. Although MCM simulation has its limitations (for example, insufficient or imprecise data, naturally, cannot be analyzed by the MCM) [9], there is a small but growing number of multimedia environmental fate models that perform stochastic simulations by including both the uncertainty in chemical parameters and the spatial and temporal variability of the environment [10-12]. Most existing probability analyses in multimedia environmental fate models consider only the uncertainties in chemical properties [13,14]. However, partly because of improvements in computer processing power and the size of the database available, it is now possible to estimate uncertainties using the traditional MCM simulation. For instance, Liu (2007) [7] describes the application of the Multimedia Contaminant Fate, Transport, and Exposure (MMSOILS) model to predict health risk and distributions of those predictions generated using MCMs. With more powerful computers, probability distributions are now used in place of discrete values, and appropriate Monte Carlo analysis is currently the most important technique for quantifying uncertainty in environmental assessments [6-7,15]. However, MCM simulation based on environmental multimedia system (EMS) modeling has seldom been reported.
2. Methodology
The governing equation for the EMS is given as follows [16]:
(1)
where Ca is the contaminant concentration, equal to the mass of contaminant per unit of volume of contaminants
; Vx, Vy, and Vz are the components of the seepage velocity
; Dx, Dy, and Dz are the components of the dispersion coefficient
; R is the total retardation factor (dimensionless); and 
is the effective first-order decay rate constant
.
For the risk assessment, considering the uncertain parameters in the EMS model, hydraulic conductivity, K, bulk density, B or porosity, and θ are the key input variables in the simulations. It has been reported that the mixing coefficient resulted in generally normal distributions. Using the Monte Carlo approach, the values of those parameters are generated from a uniform distribution. The normal generators can be simply expressed as follows:
(2)
where u = hydraulic conductivity, K, or bulk density, B or porosity, θ; 
represents a normal distribution function of 
and
; 
= standard deviation of u; and 
= the mean value of u. After certain sets of random samples for each parameter are generated, the distribution of predicted concentrations for each grid square can be calculated by the EMS model.
The distribution results can then be used to define 5th and 95th percentile concentrations and an uncertainty factor is calculated as the ratio of the 95th and 5th percentile concentrations. Then we should decide the level of risk. In the present thesis, there are two methods used to study the levels of risks or adverse effects associated with multimedia pollutant transportation, the risk quotient (RQ) and probabilistic risk assessment. Specifically, RQ is the ratio of a measured or estimated concentration (PEC) to a benchmark concentration (KEC). RQ is a primary tool to support the environmental evaluation of the use of production chemicals on the basis of available data about these chemicals and platform-related conditions [17,18]. The RQ evaluation is carried out on the basis of a comparison of the predicted environmental concentration with the known environmental criteria, which refer to local environmental guidelines for a species. The RQ factor is calculated as follows:
(3)
The RQ can be viewed as the severity measure of a risk. The higher the value of RQ above 1, the greater the possibility of environmental risk. When RQ > 1, adverse environmental effects may be expected. To evaluate the probability of RQ exceeding 1, the probabilistic distribution for each point of concern resulting from the MCM is taken into account for the quantification of the RQ. For example, the RQ distribution under 95th percentile concentrations, presented below in Section 4, will be performed to show the severity risk levels. Finally, the EMS and the MCM are combined through the probabilistic risk assessment method. Environmental risk associated with the transportation of multimedia pollutants can be expressed as the probability of a pollutant’s concentration (denoted as C) exceeding local environmental guidelines (denoted as Cstandard), i.e., R = P (C > Cstandard), where R denotes risk. Thus, the risk can be quantified as follows [19]:
(4)
The risk simulation calculates numerous scenarios of a model by repeatedly picking values from a probability distribution for the uncertain variables and using those values for the model. These probabilities are propagated through the EMS model, and an output distribution describing the probability of various outcomes is generated (Figure 1).
3. Case Study
3.1. Trail Road Landfill Site
The study site, which includes the Nepean and Trail Road landfills, is located within the Ottawa-Carleton region, which has a population of 750,000. The site, approximately 200 ha, is surrounded by light industry and farmland. Highway 416, Moodie Drive, and Cambrian Road border the site, to the east, west, and north, respectively, and at some distance from the landfills, Barnsdale Road borders the site to the south (see Figure 2) [20-22].
3.2. Model Setting and Topography
The large-scale area is about 4 × 4 km, with the Trail Road Landfill site as the center, as shown in Figure 3. The northwest boundary is Cambrian Road, where there is a large dewatering pond used to catch the local groundwater discharge. The pond water eventually discharges into the Jock River, which is located approximately 1 km to the north. Approximately 500 m from the northern boundary of Trail Road Landfill on the north side of southwest of Trail Road is the Nepean Landfill. Figure 3 shows that surface water runoff flows in a south to southwesterly direction from Trail Road.
Figure 1. An output distribution describing the probability of various outcomes.
Figure 2. Location and map of Trail Road Landfill, Ottawa, Ontario.
3.3. Collection and Estimation of Model Data
Selecting or estimating important parameters has a relatively large potential influence on modeling outputs. These parameters may not be suitable for other sites, where environmental conditions are different. The input parameters related to environmental conditions and the physical properties of this site are given in Table 1 [20-22].
3.4. 3D Simulation Results of Benzene and Comparison with Observed Data at Trail Road Landfill Site
Table 2 shows a comparison of the average concentrations between the modeled and observed current data for different years. All sampled data are from Accutest Laboratories, Ltd., of Ottawa, Ontario. Accutest is certified by the Canadian Association for Environmental Analytical