Source Attribution of Atmospheric Radionuclide Emissions from Signatures Embedded Within Varying Background Signals

Intentional and inadvertent atmospheric releases, when sampled, can identify and quantify the nuclear material involved. Recent world events such as the Fukushima-Daiichi disaster, nuclear weapons tests, and accidental/unplanned venting underscore the need to accurately attribute sampled radionuclides to their sources. Complicating factors include an increasing background environment due to sources such as nuclear power plants and, larger still, contributions from medical isotope facilities emitting either continuously, or in production batches. Data-fusion techniques employing several critical isotopes with half-lives ranging from several hours to a few weeks have been developed to characterize their sources. To attribute properly the sources of radionuclides, however, background concentrations, and their variability, must be accurately prescribed since the environmental contribution will differ according to half-life. The identification of an emission source can be problematic if the signal of interest is superimposed on a background signal with fluctuations of comparable magnitude. However, the sensitivity can be greatly enhanced if we use the power spectrum to separate sources from the background. By combining signal processing techniques with numerical weather prediction models, improvements in source attribution are possible.