Biospheric Model Validation Research Articles

Application of RESRAD-OFFSITE to Two Probabilistic Test Scenarios

The use of probabilistic dose and risk analysis will become widespread in radiological assessments because of the insights it provides and because of the availability of fast personal computers. Although the probabilistic analysis for each site will be unique because of the probabilistic inputs selected, their site-specific distributions, and any interrelationship between the inputs, the insights gained during the participation of the RESRAD-OFFSITE computer code in the Biosphere Modelling and Assessment and Biospheric Model Validation probabilistic studies will be useful guides in performing other probabilistic dose and risk analyses. Correlation and regression analysis can be performed on raw or ranked data, and the appropriate analysis can be chosen on the basis of the coefficient of determination. Partial (rank) correlation coefficients and standardized (rank) regression coefficients can be used to identify probabilistic inputs that have a linear (or monotonic) effect on probabilistic outputs and to quantify these effects. A suite of scatter plots can be used to confirm this identification and to also identify the inputs that have nonlinear, nonmonotonic effects on probabilistic outputs.

BIOMOVS II: An international test of the performance of environmental transfer models

The Biospheric Model Validation Study-Phase II (BIOMOVS II) was an international cooperative program that tested the accuracy of predictions of environmental assessment models. Model evaluation was based on calculations made by individual participants for 10 test scenarios that addressed both short- and long-term releases of radioactivity from facilities such as power reactors, solid waste disposal repositories and uranium mill tailings. Model predictions were compared with each other and, where possible, with independent field observations, and reasons were sought for any differences that arose. Qualitative topics were also considered, including development of systematic methodologies for radiological assessments. This paper addresses conclusions arising from the study as a whole. Confidence intervals on predictions and differences between predictions and observations were often less than a factor of 10, although there was much variability among models and scenarios. Model performance depended critically, not only on the formulation and parameter values of the model itself, but also on the experience and assumptions made by the user. The study demonstrated the need to better explain and justify all aspects of model structure and application and to assess all sources of uncertainty. A key recommendation was that assessments should not be undertaken in isolation by one individual using one model.

Data quality and validation of radiological assessment models.

Simulation models must be tested against measurements before their predictions can be viewed as reliable. However, having data to test all pathways in a model is rare, and finding data that were collected for the purpose of model testing is even rarer. Thus modelers must take every opportunity to test portions of their models with data obtained under less than ideal circumstances. With careful preparation of the data and a thorough understanding of the data requirements of the model, the success of the test can be assured. Monitoring data collected after the Chernobyl reactor accident were used for model testing in two international studies, the Biospheric Model Validation Study and the Validation of Environmental Model Predictions. Based on these experiences, this paper describes the data requirements of validation, the validation of models with less than ideal data, the value of good data, the role of modeling in improving the quality of data and, most importantly, ways to improve collection and preparation of data to assist further validation. Although the examples are specific to radiological assessment models, the principles of data collection and preparation herein may be applied to testing all simulation models.

Modeling the resuspension of radionuclides in Ukrainian regions impacted by Chernobyl fallout.

Following the 1986 Chernobyl event, large amounts of radioactive materials were deposited in nearby areas. Concentrations of various radionuclides were measured in air and surface soil. To study the resuspension of radioactive particulate, three different exposure situations were developed on the basis of the collected data under the auspices of the international BIOMOVS II (BIOspheric MOdel Validation Study) project. Modelers were asked to predict seasonal air concentrations and resuspension factors at several locations at different distances from Chernobyl for six successive years following the accident. Measurements of radionuclide deposition on topsoil were provided for each site along with information on soil, vegetation, land use, surface roughness, meteorology, and climate. In this paper, the three exposure situations are described, along with the initial data set provided to the modelers; two modeling approaches used to make the endpoint predictions are also presented. After the model predictions were submitted, the measured air concentrations and resuspension factors were released to the modelers. Generally, the predictions were well within an order of magnitude of the measured values. Time-dependent trends in predictions and measurements were in good agreement with one of the models, which (a) explicitly accounted for loss processes in soil and (b) used calibration to improve its predictive capabilities. Reasons for variations between predictions and measurements, suggestions for the improvement of models, and conclusions from the model validation study are presented.

Summary of the BIOMOVS A4 scenario: testing models of the air-pasture-cow milk pathway using Chernobyl fallout data.

A unique opportunity to test dose assessment models arose after the Chernobyl reactor accident. During the passage of the contaminated plume, concentrations of 131I and 137Cs in air, pasture, and cow's milk were collected at various sites in the northern hemisphere. Afterwards, contaminated pasture and milk samples were analyzed over time. Under the auspices of the Biospheric Model Validation Study (BIOMOVS), data from 13 sites for 131I and 10 sites for 137Cs were used to test model predictions for the air-pasture-cow milk pathway. Calculations were submitted for 23 models, 10 of which were quasi-steady state. The others were time-dependent. Daily predictions and predictions of time-integrated concentrations of 131I and 137Cs in pasture grass and milk for six months post-accident were calculated and compared with observed data. Testing against data from several locations over time for several steps in the air-to-milk pathway resulted in a better understanding of important processes and how they should be modeled. This model testing exercise showed both the strengths and weaknesses of the models and revealed the importance of testing all parts of dose assessment models whenever possible.