In the framework of the Fourth Workshop, a model inter–comparison exercise is proposed with respect to environmental impact assessment. The purpose of the exercise is to study the compatibility of the performance of similar dispersion models over complex topographies. Contributors were asked to use the same meteorological input over complex terrain topography and to perform dispersion calculations in the surroundings of a small power plant located at the foot of a 1426 m high mountain. A statistical analysis based on the bootstrap method is employed to yield the bounds of a compatible result.

Lagrangian particle dispersion models (LPDMs) are increasingly used for nuclear applications. In this paper, a brief description of four of the commonly used LPDMs, namely, FLEXPART, System for Prediction of Environmental Emergency Dose Information, Numerical Atmospheric-dispersion Modelling Environment, and Dispersion over Complex Terrain is given together with examples of their application in different nuclear-related fields. These include nuclear risk studies at the European level, national emergency response systems, international real-time prediction/control systems, and source term analyses.

The Meteorological Pre-Processor (MPP) of the Decision Support System RODOS acts as interface between the incoming meteorological data from stations and/or prognostic models and the Atmospheric Dispersion Models (ADMs) used for predicting the spread of the accidentally emitted radionuclides. The MPP includes a diagnostic Wind Field Model (WFM) to ensure mass conservation of the calculated wind field. Its output is usable by simple and complex ADMs and it is applicable for highly complex topography and from micro- to meso-scales. The MPP has been tested for both real and artificial flow fields and it has been optimized to function with very short execution times and to give the most reasonable results under all terrain complexity and atmospheric stability conditions. DIPCOT (DIsPersion over COmplex Terrain) is a Lagrangian Puff / Particle model that has been implemented in RODOS to simulate radionuclides atmospheric dispersion over complicated terrain. For this purpose, it uses a certain number of fictitious puffs/particles which are assumed to move with the mean wind flow plus a random velocity component to simulate turbulent diffusion. The calculation of the gamma radiation dose rates in air due to the radioactive plume is calculated by a very fast method that takes into account the inhomogeneous 3-dimensional cloud shape. DIPCOT has been evaluated by comparisons to widely used real-scale experimental data sets: Copenhagen, Prairie Grass, Indianapolis and Mol. The integration of the above models greatly enhances the applicability of the RODOS system.

In previous work [Kovalets, I., Andronopoulos, S., Bartzis, J.G., Gounaris, N., Kushchan, A., 2004. Introduction of data assimilation procedures in the meteorological pre-processor of atmospheric dispersion models used in emergency response systems. Atmospheric Environment 38, 457–467.] the authors have developed data assimilation (DA) procedures and implemented them in the frames of a diagnostic meteorological pre-processor (MPP) to enable simultaneous use of meteorological measurements with numerical weather prediction (NWP) data. The DA techniques were directly validated showing a clear improvement of the MPP output quality in comparison with meteorological measurement data. In the current paper it is demonstrated that the application of DA procedures in the MPP, to combine meteorological measurements with NWP data, has a noticeable positive effect on the performance of an atmospheric dispersion model (ADM) driven by the MPP output. This result is particularly important for emergency response systems used for accidental releases of pollutants, because it provides the possibility to combine meteorological measurements with NWP data in order to achieve more reliable dispersion predictions. This is also an indirect way to validate the DA procedures applied in the MPP. The above goal is achieved by applying the Lagrangian ADM DIPCOT driven by meteorological data calculated by the MPP code both with and without the use of DA procedures to simulate the first European tracer experiment (ETEX I). The performance of the ADM in each case was evaluated by comparing the predicted and the experimental concentrations with the use of statistical indices and concentration plots. The comparison of resulting concentrations using the different sets of meteorological data showed that the activation of DA in the MPP code clearly improves the performance of dispersion calculations in terms of plume shape and dimensions, location of maximum concentrations, statistical indices and time variation of concentration at the detectors locations.

The DETRACT computational system, consisting of a meteorological processor and a Lagrangian particles atmospheric dispersion model, is applied to the 'Hanford Purex Scenario' – an accidental 3 1/2 days of lasting release of radioactive 131I from the stack of the Hanford (USA) Purex Chemical Separations Plant (2–5 September 1963) – for evaluation purposes. During the release, the source intensity and the wind direction varied. The variables used for the model evaluation were daily-averaged and time-integrated concentrations in air, according to the available observations. Graphical and statistical means (factor-of-2, of-5 and of-10) were applied for the evaluation. The obtained results (18%, 45%, and 53% for the daily concentrations and 38%, 63%, and 81% for the time-integrated concentrations, respectively) are discussed in view of the computational system suitability for use in complex situations of accidental atmospheric releases of hazardous contaminants, including radioactive pollutants, and its sensitivity to the spatial density of the meteorological observations.

The stochastic dispersion model DIPCOT was validated against data from wind tunnel experiments where heat, used as the tracer, was released in a quasi two-dimensional turbulent flow. Dispersion simulations, based on Langevin equation, were performed using measurements and the model predictions were quantitatively compared to the experimental data. Two approaches of the drift term in Langevin equation were examined, utilising the first three and the first four moments of turbulent velocity, respectively. The overall behaviour of the model is quite satisfactory. The form of the drift term affected the predictions; the introduction of the fourth moment resulted in better estimations of the maximum concentrations. However, the mean performance of the model remained stable, irrespective of the drift term formulation.

INTRODUCTION The “Hanford Scenario” refers to an acute accidental release of radioactive I from the stack of the Hanford (USA) Purex Chemical Separations Plant that occurred between 2 and 5 of September 1963 (BIOMASS, 1999). From the environmental impact point of view it is a very interesting case study for the evaluation of computational systems that simulate the atmospheric dispersion, deposition and passage to food chain of radioactive pollutants. It is a very challenging case too, because of the surrounding topography is complex, while the release is variable in time and lasts for three and a half consecutive days with changing wind direction and atmospheric stability: frequently temperature inversions occur at night and break during the day, resulting in unstable and turbulent conditions.

The paper presents an approach to the treatment and analysis of long-range transport and dispersion model forecasts. Long-range is intended here as the space scale of the order of few thousands of kilometers known also as continental scale. The method is called multi-model ensemble dispersion and is based on the simultaneous analysis of several model simulations by means of ad-hoc statistical treatments and parameters. The models considered in this study are operational long-range transport and dispersion models used to support decision making in various countries in case of accidental releases of harmful volatile substances, in particular radionuclides to the atmosphere. The ensemble dispersion approach and indicators provide a way to reduce several model results to few concise representations that include an estimate of the models’ agreement in predicting a specific scenario. The parameters proposed are particularly suited for long-range transport and dispersion models although they can also be applied to short-range dispersion and weather fields.

Is atmospheric dispersion forecasting an important asset of the early-phase nuclear emergency response management? Is there a 'perfect atmospheric dispersion model'? Is there a way to make the results of dispersion models more reliable and trustworthy? While seeking to answer these questions the multi-model ensemble dispersion forecast system ENSEMBLE will be presented.

The data collected during the long-range European tracer experiment (ETEX) conducted in 1994, are used to estimate quantitatively the ensemble dispersion concept presented in Part I. The modeling groups taking part to the ENSEMBLE activities (see, Part I) repeated model simulations of the dispersion of ETEX release 1 and the model ensemble is compared with the monitoring data. The scope of the comparison is to estimate to what extent the ensemble analysis is an improvement with respect to the single model results and represents a superior analysis of the process evolution.

Efstratios Davakis, Spyros Andronopoulos, George A. Sideridis, Eleftherios G. Kastrinakis, John G. Bartzis, Stavros G. Nychas Department of Chemical Engineering, Aristotle University of Thessaloniki, Greece, Environmental Research Laboratory, Institute of Nuclear Technology and Radiation Protection, NCSR “Demokritos”, Greece, Department of Energy Resources Management Engineering, Aristotle University of Thessaloniki, Greece.

In this paper the Lagrangian particle dispersion model DIPCOT is evaluated by simulating two dispersion experiments over highly complex topographies (the TRANSALP90 and the ETEX experiments). The resulting predicted concentrations were compared with the experimental ones using some well-known performance indices and various types of plots. Furthermore the model results are compared to the statistical performance of other dispersion models that have been applied at the ETEX experiment. The effect of different concentration calculation methods on the predicted concentrations is examined, by comparing the results of five methods: a box counting method with fix dimensions, and four types of concentration estimation density kernels. According to the statistical evaluation, the agreement between the predicted and the measured concentration is affected by the concentration calculation method; the overall behaviour of the model, however, is reasonably good.

INTRODUCTION The stochastic (Lagrangian) particle atmospheric dispersion model DIPCOT (Davakis et. al, 2001, Davakis et. al, 2000) is evaluated by simulating a mesoscale (TRANALP campaign) and a long-range (European Tracer Experiment-ETEX) dispersion experiments. The effect of the method of calculation of the pollutant concentration is also examined, by comparing the results of three methods: a box counting method with variable box dimensions and two Gaussianshaped density kernels. The evaluation procedure is based on the statistical and graphical comparison of the predicted concentrations against the observed ones, using some well-known performance indices and various types of plots.

The land-surface parameterization scheme BATS was incorporated into a diagnostic atmospheric modeling system, using an unstructured prismatic grid, to investigate the capability of a diagnostic tool, designed for emergency response, to assign the land-cover impact on the boundary layer structure and pollutant dispersion. In this framework, two applications were made over a complex alpine region with 14 different landuses. In the first application (case A), the effects of the snow and vegetation processes on the ground-atmosphere heat exchanges (such as snow melt, fractional ground shading, vegetation and snow-cover-dependent albedo, existence of transpiring plant surfaces, plant water budget including foliage and stem water storage, transpiration limitations by stomatal resistance and soil dryness) were taken into account, while in the second (case B), they were omitted. The results were compared with data from the TRANSALP experiment. Considerable differences were found between cases A and B concerning the sensible heat flux, the atmospheric stability and the mixing layer height while the pollutant concentration maxima were significantly underestimated in the second case. The localization of these effects by a diagnostic modeling system, with a minor increase in the computing time, is considered significantly important for the improvement of the accuracy and reliability of the results in case of emergency response.

To assess the ability of a model to simulate atmospheric dispersion, its performance must be tested using data from real field experiments. This paper presents the results of a validation study of the atmospheric dispersion model Dispersion over Complex Terrain (DIPCOT) against data from the Indianapolis field experiment. Three different modules of the dispersion code are examined a puff model, a random walk model based on random displacement, and a stochastic model based on the Langevin equation. The results of the three modules are statistically and qualitative compared with the maximum arcwise concentrations and the near centreline concentrations.

This study presents the performance of the Lagrangian particle dispersion model DIPCOT-II (DIsPersion over COmplex Terrain) at the Kincaid field experiment. The validation of the model is performed on the basis of the maximum arcwise and near centreline concentrations using the bootstrap re-sampling procedure and the variation of the model residuals.

In the framework of the 'Fifth workshop: Harmonization within atmospheric dispersion modelling for regulatory purposes', a model intercomparison exercise was performed which aimed to assess (i) the environmental impact, and (ii) the compatibility of the performance of similar dispersion models under severe weather and topography conditions. The contributors were asked to use the same meteorological input data and perform dispersion calculations in the surroundings of the As Pontes power plant in northwest Spain for four individual days. The predicted concentrations are compared with observed near ground concentrations measured at 17 receptors located around the source. The models are also compared with one another to examine the degree of similarity of their predictions. A statistical analysis, based on the bootstrap re-sampling method, is employed to yield the bounds of a compatible result.

The advantageous utilization of triangular prismatic grid for flow simulation over irregular geometries is widely recognized. Such grid is here utilized to diagnose the atmospheric conditions and pollutant dispersion over complex inhomogeneous surfaces. One case with an extremely complex surface is resolved with regular prismatic grid. A second case, including parts of flat and mountainous areas, is resolved with completely unstructured grid. The atmospheric calculations rely on field data. The detailed surface description achieved through the triangular mesh allows facing lack of observations at significant sites, while it permits the selective use of the measurements for the determination of the local atmospheric conditions. The simulations are compared against observed pollutant concentrations. A quite satisfactory agreement is obtained, indicating that diagnostic models using improved surface resolution can provide reasonable results over complex terrain. This is considered very important since the use of diagnostic models is strongly suggested in emergency response cases.

An estimation of SO2 dispersion over complex terrain is attempted, using the diagnostic wind field generated from limited meteorological data. This effort aims to face the realistic requirement of emergency-response in case of industrial plants settled in poorly instrumented areas. The combination of three numerical procedures is utilized: (1) detailed topographical simulation; (2) meteorological diagnosis using experimental data and accounting for terrain influences; and (3) dispersion calculations based on the diagnosed conditions. For the meteorological diagnosis a conditional exploitation of the measurements is made, allowing the inclusion of topographical effects. This is achieved by recognizing different terrain features, such as plains, valleys, slopes and coasts. This allows the division of the terrain into areas expected to follow the same with—or different from—the station, flow regime. Application is made for a power plant at Megalopolis valley, Greece. The hourly meteorological measurements from one surface station are used, with synoptic information from weather charts. The SO2 dispersion from three buoyant sources is calculated for several days. The obtained daily average concentrations at nine locations are compared with observational values, using the bootstrap resampling evaluation method.