Boyko O. Information technologies of processing of spatially distributed data for river runoff forecasting and wash-off radioactive pollutants

Thesis solved problem of the development of information technologies of processing of spatially distributed data for rivers runoff prediction and wash-off radioactive pollutants from watersheds. Thesis is based on GIS methods and the development of the cascade of interrelated mathematical models of corresponding physical processes. It was shown that usage of distributed models which utilize spatially distributed data would greatly improve predictive performance of river runoff forecast. We developed the river runoff forecast system which is comprised of the processing module of a spatially distributed data (GIS preprocessor) and fully distributed rainfall-runoff model TOPKAPI-IMMS. Model TOPKAPI-IMMS is extended version of TOPKAPI model which is based on combining kinematic wave approach and the topography of the watershed. The latter is described by a Digital Elevation Model (DEM), which subdivides the application domain by means of a grid of squared cells. The integration in space of the non-linear kinematic wave equations results in three non-linear reservoir equations - drainage in the soil, overland flow on saturated or impervious soils and channel flow. In order to better represent the runoff processes modules of evapotranspiration and infiltration were added. Evapotranspiration is described by Penman-Monteith equation and infiltration is based on Green-Ampt infiltration model which was adapted to computational structure of TOPKAPI model. TOPKAPI-IMMS was used to calculate the continuous flow of flood dangerous rivers of Transcarpathia, namely modeling of flood of 1998 for Uzh river watershed. The results have been verified with hydrological observations and evaluated using Nash-Sutcliffe coefficicent. Parallelized algorithm of runoff calcualtion was developed for TOPKAPI-IMMS model. Algorithm is based on spatial decomposition of watershed on independent subwatersheds and allows to perform calculations on high performance computing systems with distributed memory. It was developed two models of watershed radionuclides washoff - RETRACE-T and RETRACE-R. First is based on the computational structure of TOPKAPI-IMMS model and describes processes of exchanges in system "water - surface depositions". Second one uses simplified approach which is based on the two main assumptions: (a) the rate of the radionuclide wash off formed by each elementary grid cell of the watershed is linearly depends on the precipitation rate and the density of the fallout in this cell through the "wash-off" coefficient which is specific for particular radionuclide; (b) Cell washoff is transported without time delay to the nearest river channel cell, i.e. to the grid element of the hydrodynamic river model. It was compared washoff quantities of 137Cs and 90Sr caclulated by both models for the test fallout. RETRACE-R was integrated with the hydrodynamic model of radionuclides propagation in river systems RIVTOX in European DSS for emergency management after nuclear or radiological accidents JRODOS. Models chain was configured for Styr river basin in the framework of JRODOS implementation for Rivne nuclear power plant region in Crisis Center of the State Agency for Nuclear Regulation of Ukraine.