Radionuclide migration in forest ecosystems--results of a model validation study

The primary objective of the IAEA's BIOMASS Forest Working Group (FWG) was to bring together experimental radioecologists and modellers to facilitate the exchange of information which could be used to improve our ability to understand and forecast radionuclide transfers within forests. This paper describes a blind model validation exercise which was conducted by the FWG to test nine models which members of the group had developed in response to the need to predict the fate of radiocaesium in forests in Europe after the Chernobyl accident. The outcomes and conclusions of this exercise are summarised. It was concluded that, as a group, the models are capable of providing an envelope of predictions which can be expected to enclose experimental data for radiocaesium contamination in forests over the time scale tested. However, the models are subject to varying degrees of conceptual uncertainty which gives rise to a very high degree of divergence between individual model predictions, particularly when forecasting edible mushroom contamination. Furthermore, the forecasting capability of the models over future decades currently remains untested.     

Prediction of the fate of radioactive material in the South Pacific Ocean using a global high-resolution ocean model

We investigate the release of radioactive contaminants from Moruroa Atoll in a global high-resolution off-line model. The spread of tracer is studied in a series of simulations with varying release depths and time-scales, and into ocean velocity fields corresponding to long-term annual mean, seasonal, and interannually varying scenarios. In the instantaneous surface release scenarios we find that the incorporation of a seasonal cycle greatly influences tracer advection, with maximum concentrations still found within the French Polynesia region after 10 years. In contrast, the maximum trace is located in the southeast Pacific when long-term annual mean fields are used. This emphasizes the importance of the seasonal cycle in models of pollution dispersion on large scales. We further find that during an El Ni?o/Southern Oscillation (ENSO) event reduced currents in the region of Moruroa Atoll result in increased concentrations of radioactive material in French Polynesia, as direct flushing from the source is reduced. In terms of the sensitivity to tracer release time-rates, we find that a gradual input results in maximum concentrations in the near vicinity of French Polynesia. This contrasts the instantaneous-release scenarios, which see maximum concentrations and tracer spread across much of the South Pacific Ocean. For example, in as little as seven years radioactive contamination can reach the east coast of Australia diluted by only a factor of 1,000 of the initial concentration. A comparison of results is made with previous studies. Overall, we find much higher concentrations of radionuclides in the South Pacific than has previously been predicted using coarser-resolution models.     

Validation of I ecological transfer models and thyroid dose assessments using Chernobyl fallout data from the Plavsk district, Russia

Within the project “Environmental Modelling for Radiation Safety” (EMRAS) organized by the IAEA in 2003 experimental data of ¹³¹I measurements following the Chernobyl accident in the Plavsk district of Tula region, Russia were used to validate the calculations of some radioecological transfer models. Nine models participated in the inter-comparison. Levels of ¹³⁷Cs soil contamination in all the settlements and ¹³¹I/¹³⁷Cs isotopic ratios in the depositions in some locations were used as the main input information. 370 measurements of ¹³¹I content in thyroid of townspeople and villagers, and 90 measurements of ¹³¹I concentration in milk were used for validation of the model predictions.A remarkable improvement in models performance comparing with previous inter-comparison exercise was demonstrated. Predictions of the various models were within a factor of three relative to the observations, discrepancies between the estimates of average doses to thyroid produced by most participant not exceeded a factor of ten.     

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.     

A validation study for the transport of 134Cs to strawberry

The paper presents results on model validation by field experiment for transport of 134Cs to strawberry. The transfer of 134Cs to herbaceous plants was investigated following a wet deposition after an acute release during 2000. Leaf-to-fruit, soil-to-fruit and direct fruit pathways were examined. The available meteorological and local soil information together with the experimental data were taken into account by the model RUVFRU. The processes are described by first order differential equations. In the case of foliar contamination scenarios measured and calculated results for fruit are in good agreement. However, the results of soil contamination scenarios provide large differences of up to three orders of magnitude between model predictions and experimental values for either fruit or other parts of the plant. The bias could be explained by the underestimation of the interception of the plant at the beginning of the season, in the soil contamination scenario. The model output permits prompt assessment of emergency situations and provides aid making decisions concerning mitigation of the consequences of the accident.     

消浪工程对太湖底泥再悬浮及营养盐释放的影响

为了解863消浪工程对太湖底泥再悬浮及营养盐释放的抑制作用,于2005年7月15～17日在工程区投放悬浮物捕获器测定沉积物的再悬浮通量,并分层采集水样进行水体营养盐浓度的垂向分布研究。16日平均风速3 m/s时,测得的再悬浮通量上层最大值为7.22 g/d·m2,下层最大值为41.8 g/d·m2;17日平均风速5 m/s时,测得的再悬浮通量上层最小值24.7 g/d·m2,下层最小值为48.4 g/d·m2,沉积物再悬浮通量与风浪扰动强度关系密切。对比消浪工程区内外沉积物的再悬浮通量表明,消浪工程能够显著减弱风浪对底泥的扰动,抑制沉积物再悬浮,减轻营养盐的内源释放通量。实验结果还表明,太湖水体悬浮物浓度越高,悬浮物的有机质含量就越低,相应地,单位悬浮物中磷的含量也越低。随着风浪扰动的持续和增强,尽管能够将更多的沉积物间隙水中的溶解性磷带入水体,但是,野外观测中发现水体溶解性的磷含量并未相应增高甚至降低,这可能是由于水体中悬浮物浓度越高,对水体溶解性磷的吸附能力也越高,从而使得水中溶解性磷的含量增高不显著甚至降低。     

Mesoscale modelling of radioactive contamination formation in Ukraine caused by the Chernobyl accident

This work is devoted to the reconstruction of time-dependent radioactive contamination fields in the territory of Ukraine in the initial period of the Chernobyl accident using the model of atmospheric transport LEDI (Lagrangian-Eulerian DIffusion model). The modelling results were compared with available 137Cs air and ground contamination measurement data. The 137Cs atmospheric transport over the territory of Ukraine was simulated during the first 12 days after the accident (from 26 April to 7 May 1986) using real aerological information and rain measurement network data. The detailed scenario of the release from the accidental unit of the Chernobyl nuclear plant has been built (including time-dependent radioactivity release intensity and time-varied height of the release). The calculations have enabled to explain the main features of spatial and temporal variations of radioactive contamination fields over the territory of Ukraine on the regional scale, including the formation of the major large-scale spots of radioactive contamination caused by dry and wet deposition.     

Model simulations of the fate of 14C added to a Canadian shield lake

Carbon-14 was added to the epilimnion of a small Canadian Shield lake to investigate primary production and carbon dynamics. The nature of the spike and subsequent monitoring allowed the investigation of both short-term and longer-term processes relevant to evaluating impacts of accidental and routine releases and of solid waste disposal. Data from this experiment were used in the BIOMOVS II program as a validation test for modelling the fate of the ¹⁴C added to the lake. Four models were used: (1) a simple probabilistic mass balance model of a lake; (2) a relatively complex deterministic model; (3) a complex deterministic model; and (4) a more complex probabilistic model. Endpoints were ¹⁴C concentrations in water, sediment and lake whitefish over a thirteen year period. Each model produced reasonable predictions when compared to the range of the observed data and when uncertainty in model predictions is taken into consideration. The simple lake model did not account for internal recycling of ¹⁴C and, in this respect, its predictions were not as realistic as those of the more complex models for concentrations in water. However, the simple model predictions for the ¹⁴C inventory remaining in lake sediment were closest to the observed values. Overall, the more complex probabilistic model was the most accurate in simulating ¹⁴C concentrations in water and in whitefish but it overestimated ¹⁴C retention in the lake sediments, as did the other complex models. Choice of parameter values for transfer rate to sediment and gaseous evasion are important in influencing model predictions. Although predicted concentrations of ¹⁴C in fish of dynamic models were more accurate than those using equilibrium bioconcentration factors typically used in assessments, large variability in observed ¹⁴C concentrations in whitefish emphasizes the need for a better understanding of the important processes that influence these contaminant concentrations.     

Resuspension and redistribution of radionuclides during grassland and forest fires in the Chernobyl exclusion zone: part I. Fire experiments

Controlled burning of experimental plots of forest or grassland in the Chernobyl exclusion zone has been carried out in order to estimate the parameters of radionuclide resuspension, transport and deposition during forest and grassland fires and to evaluate the working conditions of firemen. An increase of several orders of magnitude of the airborne radionuclide concentration was observed in the territory near the fire area. The resuspension factor for (137)Cs and (90)Sr was determined to range from 10(-6) to 10(-5) m(-1), and for the plutonium radionuclides from 10(-7) to 10(-6) m(-1) (related to the nuclides in the combustible biomass). These values are 2 orders of magnitude lower if they are calculated relatively to the total contamination density (including the nuclides in the soil). The radionuclide fallout along the plume axis is negligible in comparison to the existing contamination. However, the additional inhalation dose for firemen exposed in the affected area can reach the level of the additional external irradiation in the period of their mission. The plutonium nuclides constitute the dominating contribution to the inhalation dose.     

Modelling radioactivity in the Irish Sea: From discharge to dose

In order to support authorised discharges of low level radioactive liquid effluent into coastal regions, mathematical models are required to robustly predict radiological impacts on critical groups of current and proposed changes to liquid discharges. The grid model presented here simulates the long term dispersion and transport of radioactivity discharged from the Sellafield site in Cumbria, UK, and the subsequent exposure of critical groups in Cumbria and across the Irish Sea in Northern Ireland. The fine grid of the model allows a good resolution of the seabed sediment distribution. This benefits the predictions for the last decades of low discharge level, when bed sediment can become a source of contamination by bringing back the legacy of past high discharges. This is highlighted by the dose comparison, where the predicted dose to Cumbria critical group follows well the dose estimated from environmental data during the low discharge level period.     

太湖水体中悬浮物研究

为了全面了解太湖水中悬浮物的物质组成、变化规律,选取了太湖站近10年来各测点的连续观测资料及2次定点观测资料,分析了太湖水体中悬浮物的无机和有机颗粒成份、时空分布、垂直变化、与风速、风向的关系以及与重要光 学参数的关系。研究结果表明：太湖悬浮物中有机物大概占30%;浓度大小的湖区分布大致是：湖心区＞河口区、梅梁湖、贡湖、五里湖＞东太湖;悬浮物浓度的季节变化是：湖心区冬季、春季＞秋季＞夏季,东太湖冬季＞春季、夏季、秋季,其他湖区则没有明显季节变化;垂直分布是：深层＞表层并且底泥悬浮的临界风速在5~6.5 m/s之间;悬浮物浓度的增加是引起湖水透明度降低和光学衰减系数增大的主要原因。     

A comparison of three models of 137Cs transfer in forest ecosystems.

The predictions of three models of 137Cs transfer in forest ecosystems (FOA, LOGNAT and FORESTLAND) were compared. The scenario for the model-model comparison consisted of an acute dry deposition of 137Cs over a coniferous forest. The model predictions were subsequently compared (model-data comparison) with values derived from experimental data measured in forests of the Bryansk region in Russia that were contaminated by the Chernobyl accident and that have similar characteristics to the forests described in the scenario. The predictions of radiocaesium levels in the litter-soil layer, berries, needles, wood, whole tree and moose made with the models were in relatively good agreement with each other (within a factor of 1.4-2.9). The best agreement was observed for berries and moose and the worst for wood. There was also good agreement between the model predictions for the same variables and the experimental data (within a factor of 1.2 3.2). In this case, the best agreement was observed for the litter-soil layer and the worst for wood and the whole tree. Overall, at least for the studied scenario and for the first 10 years after deposition, any of the models can be used if the final aim is to estimate average concentrations in different forest components. The agreement between the model predictions worsens with time and there were differences in the form of the time dependencies predicted by the models, especially for wood. This may lead to larger differences between the model predictions and the experimental data for times beyond the period for which data were available for comparison (10 years after the deposition).     

Model testing using data on 131I released from Hanford

The Hanford test scenario described an accidental release of 131I to the environment from the Hanford Purex Chemical Separations Plant in September 1963. Based on monitoring data collected after the release, this scenario was used by the Dose Reconstruction Working Group of BIOMASS to test models typically used in dose reconstructions. The primary exposure pathway in terms of contribution to human doses was ingestion of contaminated milk and vegetables. Predicted mean doses to the thyroid of reference individuals from ingestion of 131I ranged from 0.0001 to 0.8 mSv. For one location, predicted doses to the thyroids of two children with high milk consumption ranged from 0.006 to 2 mSv. The predicted deposition at any given location varied among participants by a factor of 5-80. The exercise provided an opportunity for comparison of assessment methods and conceptual approaches, testing model predictions against measurements, and identifying the most important contributors to uncertainty in the assessment result. Key factors affecting predictions included the approach to handling incomplete data, interpretation of input information, selection of parameter values, adjustment of models for site-specific conditions, and treatment of uncertainties.     

Speciation and hydrological transport of metals in non-acidic river systems of the Lake Baikal basin: Field data and model predictions

The speciation of metals in aqueous systems is central to understanding their mobility, bioavailability, toxicity and fate. Although several geochemical speciation models exist for metals, the equilibrium conditions assumed by many of them may not prevail in field-scale hydrological systems with flowing water. Furthermore, the dominant processes and/or process rates in non-acidic systems might differ from well-studied acidic systems. We here aim to increase knowledge on geochemical processes controlling speciation and transport of metals under non-acidic river conditions. Specifically, we evaluate the predictive capacity of a speciation model to novel measurements of multiple metals and their partitioning, under high-pH conditions in mining zones within the Lake Baikal basin. The mining zones are potential hotspots for increasing metal loads to downstream river systems. Metals released from such upstream regions may be transported all the way to Lake Baikal, where increasing metal contamination of sediments and biota has been reported. Our results show clear agreement between speciation predictions and field measurements of Fe, V, Pb and Zn, suggesting that the partitioning of these metals mainly was governed by equilibrium geochemistry under the studied conditions. Systematic over-predictions of dissolved Cr, Cu and Mo by the model were observed, which might be corrected by improving the adsorption database for hydroxyapatite because that mineral likely controls the solubility of these metals. Additionally, metal complexation by dissolved organic matter is a key parameter that needs continued monitoring in the Lake Baikal basin because dependable predictions could not be made without considering its variability. Finally, our investigation indicates that further model development is needed for accurate As speciation predictions under non-acidic conditions, which is crucial for improved health risk assessments on this contaminant.

     

Ventomod: a dynamic model for leaf to fruit transfer of radionuclides in processing tomato plants (Lycopersicon esculentum Mill.) following a direct contamination event

This paper presents results on the calibration and validation of a model (Ventomod) for leaf to fruit transfer of (134)Cs, (85)Sr and (65)Zn in processing tomato plants after leaf contamination. Several models (e.g. FARMLAND) that deal specifically with the transfer of radionuclides to fruits are adaptations of models that were developed for agricultural crops such as leafy green vegetables. "Ventomod" represents a dynamic evaluation model exclusively built for the short-term behaviour of radionuclide depositions. It forecasts the level of radionuclide contamination in ripe processing tomato fruits following an accidental radionuclide release into the atmosphere. A validation of the developed model by data sets from an independent experiment showed that the model successfully reproduced the observed radionuclide distribution and dynamics in tomato fruits. The level of uncertainty was within the normal range of similar assessment models. For a more general use of this model further testing with independent data sets from experiments obtained under different environmental conditions and data from other horticulturally important plant species would be desirable.     

Effect of user interpretation on uncertainty estimates: examples from the air-to-milk transfer of radiocesium

An important source of uncertainty in predictions of numerical simulation codes of environmental transport processes arises from the assumptions made by the user when interpreting the model and the scenario to be assessed. This type of uncertainty was examined systematically in this study and was compared with uncertainty due to varying parameter values in a code. Three terrestrial food chain codes that are driven by deposition of radionuclides from the atmosphere were used by up to ten participants to predict total deposition of ¹³⁷Cs and concentrations on pasture and in milk for two release scenarios. Collective uncertainty among the predictions of the ten users for concentrations in milk calculated for one scenario by one code was a factor of 2000, while the largest individual uncertainty was 20 times lower. Choice of parameter values contributed most to user-induced uncertainty, followed by scenario interpretation. Due to the significant disparity in predictions, it is recommended that assessments should not be carried out alone by a single code user.     

Model testing using data on 137Cs from Chernobyl fallout in the Iput River catchment area of Russia

Data collected for 10 years following the Chernobyl accident in 1986 have provided a unique opportunity to test the reliability of computer models for contamination of terrestrial and aquatic environments. The Iput River scenario was used by the Dose Reconstruction Working Group of the BIOMASS (Biosphere Modelling and Assessment Methods) programme. The test area was one of the most highly contaminated areas in Russia following the accident, with an average contamination density of 137Cs of 800,000 Bq m-2 and localized contamination up to 1,500,000 Bq m-2, and a variety of countermeasures that were implemented in the test area had to be considered in the modelling exercise. Difficulties encountered during the exercise included averaging of data to account for uneven contamination of the test area, simulating the downward migration and changes in bioavailability of 137Cs in soil, and modelling the effectiveness of countermeasures. The accuracy of model predictions is dependent at least in part on the experience and judgment of the participant in interpretation of input information, selection of parameter values, and treatment of uncertainties.