Thesis (Ph.D.) - Cairo University - Faculty of Science - Department of Astronomy and Meteorology

The atmospheric transport and dispersion modeling (ATM) of man-made radionuclides (RN) plays a vital role in monitoring and verifying any possible violation of the comprehensive nuclear-test-ban treaty (CTBT). Both forward ATM and backward (adjoint) ATM are used in this field. On the one hand, the forward ATM is used when the RN sources are known (e.g., nuclear power plants, medical isotope facility, etc.) in order to estimate the RN background at CTBT's international monitoring system (IMS) stations. Also, The forward modeling is very important in CTBT's on-site inspection (OSI) for a nuclear test, where it can be used to specify the best location and time of RN sample collection. On the other hand, The adjoint modeling is used as part of inverse modeling in order to estimate the unknown sources of RNs measurements which are observed at one or more of the IMS stations. From a decision maker point of view, these two types of modeling are essential tools in the field of CTBT's verification system, but the existence of many sources of uncertainty (radionuclides measurements, meteorological fields, and modeling errors) represents the main challenge in using ATM model outputs. Therefore, many of the workers in this field continuously emphasize the importance of uncertainty quantification of the ATM models outputs. During the last two decades, many statistical techniques were developed in order to quantify the uncertainties in the meteorological fields (forecasts or simulations) by using the ensemble approacheveloping some statistical methods to quantify the uncertainty of these two types of modeling

Issued also as CD