Models of radionuclides behavior in "soil-plant" chain for decision support systems

Abstract

The main objectives for semi-mechanistic models enhancement are justified in the article. The "soil-plant" chain is an essential part of radioisotopes flows from nuclear accident depositions to human beings. Therefore a model which describes this system should be integrated into decision support systems for liquidation consequences of accidents with releasing radioisotopes into the environment, evaluation effectiveness of measures for radiation protection, and designing hazardous radiation facilities. Such a model must show rather exact forecast results, flexibility and wide application area convenience for practical use, and other properties. Presented now models of radionuclides behavior in "soil-plant" system divided on empiric, mechanistic, and semi-mechanistic. The empirical ones do not take into account the basic mechanisms of changes in the biological availability of radionuclides and their absorption by plants, and require constant updating and refinement of the transition parameters. Mechanistic models are of little use in real life. The last ones best meet the requirements noted above. However, substantial efforts are needed for improving their accuracy, usability, and generalization. This requires integration into data models from existing and planned sensor systems; consideration of additional factors influ-encing the transfer of radionuclides to plants; increasing the level of generalization of models with adjustment to local conditions; the use of machine learning methods to integrate information accumulated in related fields into the model; coverage of more radioactive isotopes; adding an uncertainty estimate to the simulation result; integration of models of radionuclide behavior into geoinformation systems; maintaining a sufficient level of interpretability and visibility of modeling results.