Environmental risk assessment (in Europe) and equivalent processes including ecological risk assessment (United States) and ecosystem risk assessment (Japan) are collectively referred to in this chapter as ERA. They each involve the scientific analysis and characterization of environmental hazards (frequently chemicals) based on the predicted likelihood and level of exposure, versus the predicted severity of any consequent adverse effects. The discovery of "endocrine disruption" in wildlife due to natural hormones and synthetic mimics (endocrine disrupting chemicals, EDCs) in the last two decades has challenged the established ERA process, because very low exposure levels may illicit adverse effects including nonmonotonic responses in exposed organisms (Section 3.1.2). Accepting that it is impossible to legislate to protect every organism, especially when some natural losses are inevitable (for example, due to climatic stress or predation), the aims of ERA are to ensure the sustainability of populations to protect species and maintain the structure and functioning of ecosystems in which they live.Ultimately, ERA requires clear understanding and communication of risks and objective decision making on the part of risk managers. This is based on an assessment of the most relevant data or facts (assessment endpoints) concerning the predicted behavior, fate, and effects of chemicals in the environment. A number of mathematical models are available to risk assessors, which have enormous potential in helping to relate chemical properties to potential behavior and to understand empirical data concerning exposure, uptake, and effects experienced by individual organisms (see Chapters 6, 7, 11, 13, this volume). Population modeling can offer additional help in linking individual exposure to population effect, extrapolating from laboratory to field, but more importantly by providing a means of communicating risk and uncertainty to risk managers in a form that they can clearly understand and apply. By focusing on ecologically relevant effects and scales, population-level ERA can aid regulatory decision making and help focus finite resources on managing ecologically significant risks (Section 3.1.1). This chapter illustrates how population dynamics modeling can support the established ERA process. Case studies are presented concerning endocrine disruption in fish (Section 3.3).

There is currently uncertainty on the persistence of active pharmaceutical ingredients (APIs) and on their depletion mechanisms in natural surface waters such as rivers, and hence predictions of their fate are often poor. In this study, a beta-adrenergic receptor, propranolol hydrochloride, was selected as a model API to explore the relative significance of direct phototransformation as a potential removal process of hydrophilic APIs in rivers. Phototransformation kinetics of propranolol was measured under simulated solar irradiation in the laboratory, which were then converted to the kinetics applicable in UK and US rivers. The effects of light intensity, light penetration, river size and flow were examined. The extrapolated phototransformation half-lives were applied in the river catchment models of GREAT-ER and PhATE. Results demonstrated that direct phototransformation significantly reduced the predicted environmental concentrations of propranolol in the water phase. Predicted reductions of mean concentrations in the River Aire (UK) were 27% in summer and 3% in winter; and for the US rivers simulated, reductions were 28-68% in summer and 11-41% in winter. The highest reductions were predicted for long rivers with low turbidity and low flow conditions.

Long-term data from the perch population of the north basin of Windermere, UK, were combined with effects data from laboratory toxicity studies of survival, growth and reproduction in other species to assess the likely impact on fish populations of different levels of exposure to nonylphenol (NP) and ethinylestradiol (EE2). A multi-stage Delay Differential Equation fish population model was used to simulate the perch population during two periods when it showed contrasting population structures (1968–1973 and 1983–1988) as the result of a disease outbreak in 1976, and to extrapolate the effects of chemical exposure to EE2 and NP observed in the laboratory to the environment. In the absence of chemical exposure, model simulations predicted population numbers (females) in line with those derived from field survey data. Effects predictions were made for long-term exposure (20 years) to low and high doses of EE2 (1 and 10ngl−1) and NP (1 and 30μgl−1). The sustained high-level exposure of EE2 had a high probability of causing the extinction of a confined fish population such as that in Windermere, however, such an exposure scenario is unlikely. Far greater uncertainty surrounds the prediction of effects due to low-level exposure of fish populations and even though effects may appear to be significant in laboratory studies, such as fecundity lowered by 30%, they may not necessarily translate into significant population effects in the field. This is especially true if life-history data show high natural variability in terms of individual vital rates. Our work suggests that for a decline in the perch population numbers to be significant in Windermere, it would have to be substantially more than 50%. Our model predictions indicated that the post-disease population was generally more vulnerable than the pre-disease population and had a significantly greater probability of declining by 50% following single chemical exposure, but both populations were equally likely to decline following multiple chemical exposure. Rate of population recovery was shown to be a more sensitive measure in terms of differentiating the effects of low and high chemical exposure as well as the vulnerability of populations with contrasting structures, histories or levels of background stress.

Hydrodynamic models of differing scale and complexity were used to estimate spatial patterns of effluent concentration in discharge plumes in the River Esk and the Lower Tees Estuary. The output from the Tees model was used, in conjunction with measurements of toxicity determined in short-term oyster embryo tests, to predict contours/zones of toxicity in the estuary associated with effluent discharges from four chemical processing sites. One of the discharges also combined the input from a municipal sewage treatment works. The models appeared to be effective in predicting patterns of dilution and dispersion of the effluent discharges in the respective receiving environments. Confirmation of the predictive capabilities of the Tees model was achieved by comparing predicted and measured toxicity in different regions of the plumes associated with the four discharges. Differences between predicted and measured toxicity for two of the four discharges were explicable in terms of failure to take account of the effects of real-time wind conditions when test samples were collected or overlap of adjacent discharge plumes. Suggested refinements to the models and measurement of effluent toxicity would further enhance the utility of this approach for determining the extent and significance of the effects of effluent discharges in receiving environments.

The dispersion of gases in complex situations such as the case of buildings in close proximity is a difficult problem, but important for the safety of people living and working in such areas. Computational fluid dynamics (CFD) provides a method to build and run models that can simulate gas dispersion in such geometrically complex situations; however, the accuracy of the results needs to be assessed. As a first step in such an assessment, this study considers the simulation of the dynamics of the basic atmospheric boundary layer using the FLUENT CFD code and the prediction of gas dispersion from a single stack. The CFD results are compared with the predictions from the Atmospheric Dispersion Modelling System (ADMS), a well tested and validated quasi-Gaussian model.When FLUENT was set up to simulate the neutrally stable atmospheric boundary layer, the mean velocity profiles were well predicted and were maintained with downwind distance. The algebraic Reynolds stress turbulence model provided the best predictions for the turbulence kinetic energy (TKE) and dissipation. The dissipation rate was maintained throughout the length of the model domain and, on average, the TKE levels were within 80% of the expected values up to a height of 100 m, but at the ground reduced to 50% of the inlet values. Predictions of TKE using the simpler k-epsilon model turbulence was much poorer. Spread of the gas plume were predicted using an advection-diffusion (AD) method, a Lagrangian particle tracking (LP) method and a large eddy simulation (LES) method. The LP method gave the best results; the horizontal and vertical plume spreads were similar to those predicted by ADMS and ground level and plume centre line concentrations were close to ADMS values. However, some differences were observed with the ground level concentrations rising more rapidly with distance than for ADMS, but reaching similar peak values while the plume centreline concentrations dropped more rapidly than in ADMS. For the AD method the horizontal cross-wind plume spread was significantly lower than expected resulting in higher ground level concentrations than predicted by ADMS, an effect that was attributed to the isotropic formulation of the AD equation in FLUENT. The LES results were intermediate between the AD and LP predictions.Overall, the CFD simulations with the LP method were satisfactory; however, they could not be considered as an appropriate alternative to a model such as ADMS for normal atmospheric dispersion studies because of the much larger run times and the greater complexity of setting up model runs. CFD is more appropriate for applications that involve complex geometry that could not be simulated using ADMS; however, further studies are required to assess the ability of CFD to calculate dispersion in such situations, for instance, around groups of buildings and under a range of atmospheric stability conditions, rather than just the neutral stability considered in this paper. (C) 2003 Elsevier Ltd. All rights reserved.

This study evaluates the applicability and sensitivity of fish population dynamics modeling in assessing the potential effects of individual chemicals on population sustainability and recovery. Fish reproductive health is all increasingly important issue for ecological risk assessment following international concern over endocrine disruption. Life-history data from natural brook trout and fathead minnow populations were combined with effects data from laboratory-based studies, mainly concerning species other than brook trout and fathead minnows, to assess the likely impact of nonylphenol (NP) and methoxychlor (MXC) on brook trout (Salvelinus fontinalis) and fathead minnow (Pimephales promelas) population size. A delay differential equation (DDE) model with a 1-day timestep was used to predict the population dynamics of the brook trout and fathead minnows. The model predicts that NP, could enhance populations by tip to 17% at a concentration of 30 mug l(-1) based oil the results of reduction in survival and increased fecundity from life-cycle toxicity tests, however attempting to allow for growth reduction and its effect on fecundity results in a prediction of a 28% reduction in population numbers. For fathead minnows the DDE model predicts that the same concentration of NP could cause a population reduction of 21%. The differences in these predictions are related to these two species having different life history strategies, which are considered in the parameterization of the model. Post-application concentrations of MXC may peak around 300 mug l(-1) and then decline rapidly with time. Predictions show that such applications could cause a reduction of up to 30% in brook trout populations if the application occurs at the peak of the spawning season on successive years but that the effect would be less than 1% if the spawning season is avoided. Effects on the fathead minnow population size are predicted to be smaller (<4%) even if application occurs during the spawning period. Risk based statistics generated by the population dynamics models, Such as interval decline risk or quasiextinction risk and predicted time to recovery complement traditional effects parameters such as LC50 and LOEC and may ultimately prove to be more useful in risk assessment.

A post-processor of model output, EQUIP (Environmental QUality Information Processor), has been created and applied to effluent discharges to the Tees estuary in the north east of England. EQUIP is not a modelling system, but a statistical post-processor, which uses results from a compressed database of pre-run model scenarios. The system requires the results of a small number of model runs, which are selected using a statistical experimental design. A statistical response surface is used to summarise the model results for the PC database and to reconstruct plume predictions within EQUIP for user selected values of the tidal range and river inflow, or for annual average conditions. The method provides interpolation between model results assuming a smooth variation of the results with variations in the input parameters. The method is implemented on a PC and is much faster than running the full water quality model, producing spatial, temporal and vertical section plots for different states of the tide within a few minutes. In addition, maximum plume movement and tidal average concentrations are computed. Plumes from different outfalls are rapidly combined to give overall concentration fields within the estuary. The system is adaptable for outputs from different types of model, such as atmospheric dispersion models, groundwater models and other non-environmental applications. The methodology is also adaptable to simulate plume uncertainty, if input parameter variability is known.

A random walk model has been used to compute concentration distributions of dispersed oil in the North Sea resulting from produced water discharges. This formed part of a joint study commissioned by Statoil, OLF and BP International with modelling being undertaken by SINTEF and the Brixham Environmental Laboratory. The model has been set up using predicted tidal currents from the Norwegian Meteorological Office 20-km grid three-dimensional Continental Shelf model for the year 1990. Climatology data from the North Sea have been used to define the variation of the thermocline depth at monthly intervals over the year, and wind data from the East Shetland Basin have been used to compute the vertical mixing rates. Peak concentrations of dispersed oil were predicted to be approximately 3 mug l(-1) in the East Shetland Basin, assuming no biodegradation; this value is consistent with the measurements of Stagg et al. (In: Reed, M.. Johnsen, S. (Eds), Produced Water 2, Environmental Issues and Mitigation Technologies, Plenum Press, 1996), where the data were collected in the immediate vicinity of the discharges and the effects of degradation would be expected to be negligible.A matrix of model runs was undertaken with different parameter values for the degradation of dispersed oil, horizontal and vertical mixing, and the mixing across the thermocline. These results were used to estimate the sensitivity of the model and the uncertainty in the predicted concentration fields, The results indicate that for the areas of high concentration, a band approximately 150-km wide stretching up the centre of the North Sea, parameterisation of the vertical mixing is most important for predicting the concentration levels. For areas more remote from the discharges and closer to the land masses., the degradation rate of material has most effect on predicted concentrations. (C) 2001 Elsevier Science Ltd. All rights reserved.

A mathematical model to predict the effect of chemical spills in the Forth estuary in Scotland has been in use for many years. The model, based on the random walk method, predicts chemical concentrations in the estuary waters and estimates the elapsed time before the dilution is sufficient to render the spill harmless (making use of a toxicity measure such as the LC50 or a water quality standard). The model gives a deterministic result without any estimate of the uncertainty. Field studies using tracer dyes to measure the horizontal and vertical mixing rates in the estuary show that these rates vary over time. The literature on turbulent diffusion includes modelling applications using different parameterisations of the mixing process. This paper investigates the uncertainties in predicted concentrations due to model parameterisation of horizontal mixing and due to the variability in the measured mixing rates determined from surveys in the estuary. Estimates of the range of concentrations for a specific spill scenario are presented.The study shows that model formulation and parameter uncertainty are both important factors in estimating the uncertainty in model predictions. The uncertainty caused by the variations with time found in the measured mixing rates is found to be of similar magnitude to the differences in concentration resulting from using three different methods for modelling the horizontal mixing in the estuary. Uncertainties associated with model formulation could be reduced if a small number of longer timescale (e.g. 24 h) dispersion experiments were available. In addition, further data from short-term (similar to 3 h) dispersion experiments would give a better understanding of the distribution of mixing coefficients and how the mixing relates to other parameters such as tidal range and wind speed and direction. (C) 2001 Elsevier Science B.V. All rights reserved.

The potential consequences of a particular gaseous emission usually relate to a specific averaging time. However, the averaging time associated with gas monitoring data in the region of a discharge may differ from the averaging time implicit in a model employed to predict the concentration resulting from that emission. Tables of factors exist for taking account of such differences and recent models include expressions that convert predicted values to the appropriate averaging time. This paper presents the results of a study of the factors used in these regulatory tables, and makes comparisons with the equivalent factors deduced from runs with a gas dispersion model and sonic anemometer observations.

Over the period 1968 to 1999, Brixham Environmental laboratory carried out dye tracing experiments to determine local mixing conditions at coastal and estuarine sires around the United Kingdom. Results from 25 different locations, and from five non-U.K. sites, have been analysed to estimate horizontal and vertical mixing coefficients. Analysis of the data from the 19 sites which had the most complete data sets showed that the minimum lateral dispersion coefficient ranged from 0.003 to 0.42 m(2) s(-1) with a median value of 0.05 m(2) s(-1). Typical vertical diffusion coefficients ranged between 2.0 x 10(-4) to 110.0 x 10(-4) m(2) s(-1) and had a median value of 20.0 x 10(-4) m(2) s(-1). Comparisons of the values obtained have been made with mixing coefficients derived from earlier dye tracer studies in the North Sea, the Irish Sea and off the eastern seaboard of the U.S.A.A plot of minimum lateral dispersion coefficients against the corresponding vertical diffusion coefficients showed that the results fell into two distinct bands, broadly represented by shallow or deep water. Within each band, the lateral mixing coefficients appeared to vary in inverse proportion to the vertical coefficients, as would be expected in shear dispersion in which mixing is restricted by the surface and seabed or pycnocline. Curves representing this variation had different constants of proportionality for the shallow and deep waters. This was ascribed to a reduction in the curvature of the velocity profiles in the deep water areas, where stronger currents may have reduced the degree of stratification. Analysis of the full data set has not identified any generally applicable relationship between the mixing coefficients and tidal currents or wind, however, relationships between these parameters could be found within specific survey areas. For example, lateral diffusion coefficients in the Solent appeared to be correlated with the tidal current amplitude. (C) 2000 Academic Press.

Random walk modelling has been extensively used for pollutant dispersion studies on a timescale of a few tidal cycles. The aim of this study is to assess the potential of random walk modelling for longer timescales and to assess the performance of different modelling formulations for computing the horizontal spread of material in the sea. In order to assess the validity of the longer term predictions, data from large scale dye spreading experiments off Cape Kennedy, Florida (Pritchard et al. 1966) and in the North Sea (RHENO experiment, Weidemann 1973) have been compared with model predictions for the appropriate conditions. Results have shown a need for care in determining the concentrations from the random walk model to ensure a consistent method of predicting the dye patch concentration through time.The three case studies have shown that the model can successfully predict the dye patch dimensions and the dilution of the dye with time, and have demonstrated that the wind induced shear was the dominant factor in the spread. The model has also shown that vertical tidal shear was important for the spread, and under calm conditions would become dominant over the horizontal turbulent mixing after 8 h in high shear regimes (Cape Kennedy, April 1962 and RHENO) and after 90 h in low shear (Cape Kennedy, August 1962 experiments) conditions. In the short term (< similar to 50 h) Fickian (constant) diffusion gave a better representation of the dye patch spread and dilution than the use of a diffusion velocity. However, for longer times the model simulations show that the method of parameterising the horizontal turbulent mixing is not important, but the modelling of wind and tidal shears is critical. (C) 1998 Elsevier Science Ltd. All rights reserved

In January 1995 the construction of a barrage across the estuary of the River Tees was completed. This effectively halved the estuary length and created a freshwater region on the landward side. Prior to construction of the barrage, a detailed survey was undertaken at locations to seaward of its proposed position, and mathematical models were used to predict the expected changes to the flow and mixing regimes. The survey was repeated in the summer of 1995, using identical sampling stations and procedures to those employed previously.Analysis of the physical data suggests that currents in the estuary have become weaker since the barrage was completed, and the spatial acceleration of the current along the estuary has reversed direction. The acceleration reversal is consistent with an observed alteration to the vertical distributions of current on ebb and flood tides; this appears to be responsible for a change in tidal mixing such that much more vigorous mixing now occurs on the flood tide than the ebb in the central estuary. However, except close to the barrage, the tidal average stratification has not changed greatly.

A hydrodynamic model has been constructed for predicting the tidal currents off the south west coast of Singapore Island and has been verified against survey data from the area. A study examining the sensitivity of the model to changes in the input parameters has been undertaken and a statistical experimental design technique has been employed to determine the minimum number of computer runs required to quantify the sensitivity of the model. The uncertainty in predicting the time of the turn of tide is estimated from the statistical results as is the uncertainty in the predicted movement of drogues in both simple and complex tidal flows.The method is extended to estimate the uncertainties in the spread and concentration of a dye patch using a particle tracking random walk model and allowing for the uncertainties in the wind and in horizontal and vertical mixing rates as determined from a series of experimental studies of dye patch spread.

Mathematical models have been set up to describe the dispersing plume formed when liquid effluent is discharged to a ship's wake and the susbequent dilution of the waste field after it has spread over the disposal area. Comparisons were made between the dilution obtainable at the disposal site off the River Tees with that at a site off the Humber. On the basis of the rather limited oceanographic data available, the models indicate that the waste would mix throughout depth more quickly at the Humber site than at the Tees but the shallowness of the water at the Humber would limit the rate of subsequent dilution. Within 48 h of the waste becoming spread over the disposal area, the dilution at the Humber ground was predicted to be 2.2 × 106 times; in the same period the waste discharged off the Tees would be 9.7 × 106 times dilution. Under the stratified conditions which can occur at the Tees disposal area in summer, the estimated dilution would be reduced to 3.7 × 106 times.