Variance of the Galactic nuclei cosmic ray flux
Measurements of cosmic ray fluxes by the PAMELA and CREAM experiments show unexpected spectral features between 200 GeV and 100 TeV. They could be due to the presence of nearby and young cosmic ray sources. This can be studied in the myriad model, in which cosmic rays diffuse from point-like instantaneous sources located randomly throughout the Galaxy. To test this hypothesis, one must compute the flux due to a catalog of local sources, but also the error bars associated to this quantity. This turns out not to be as straightforward as it seems, as the standard deviation is infinite when computed for the most general statistical ensemble. The goals of this paper are to provide a method to associate error bars to the flux measurements which has a clear statistical meaning, and to explore the relation between the myriad model and the more usual source model based on a continuous distribution. To this end, we show that the quantiles of the flux distribution are well-defined, even though the standard deviation is infinite. They can be used to compute 68% confidence levels, for instance. We also use the fact that local sources have known positions and ages to reduce the statistical ensemble from which random sources are drawn in the myriad model. We present a method to evaluate meaningful error bars for the flux obtained in the myriad model. In this context, we also discuss the status of the spectral features observed in the proton flux by CREAM and PAMELA.