Nuclide ratios and source identification from high-resolution gamma-ray spectra with Bayesian decision methods

Abstract

Ratios of radionuclide activity concentrations are often used to identify sources of small traces of radioactivity in environmental samples. A new approach is presented to consider statistical uncertainties and their impact on decisions based on nuclide ratios particularly for cases where nuclide ratios are calculated from measurements close to or below detection limit. It is pointed out that decision threshold and detection limit are adequate for decisions on the presence of a single nuclide but not necessarily on decisions on nuclide ratios. A statistical analysis based on a Bayesian approach is presented which allows inference on probabilities of nuclide ratio values taking into account jointly the information available for all radionuclides of interest. Examples are discussed to demonstrate the improvements of this approach in spectrum analysis and in source identification. As an important case the detection of Xenon isotopes ( 135Xe, 133mXe, 133Xe and 131mXe) that are relevant for detecting underground nuclear explosions in the context of the verification of the Comprehensive Nuclear Test-Ban Treaty (CTBT) is discussed. Nuclide ratios are crucial in order to distinguish civil releases from CTBT-relevant sources. A new method is presented for graphical evaluation of the xenon isotope measurements and for assigning potential sources using the statistical approach proposed in this paper.