Documentation of Uncertainties in Experimental Cross Sections for EXFOR

Abstract

Documentation of uncertainties and covariances in experimental nuclear reaction cross sections has been assessed. Following consideration of the importance of covariances for nuclear data in various nuclear applications, and presentation of a simple numerical example to demonstrate this point, the minimum basic concepts (mean, covariance, standard derivation, partial uncertainties, micro- and macro-correlation coefficients) are introduced. A deterministic approach to propagating the covariances in primary measured parameters (e.g., counts) to the derived cross sections is discussed, using a neutron-induced activation cross section measurement as an example. Finally, various approaches to documentation (publication, compilation) of experimental cross sections to facilitate their use in future evaluations are mentioned.