We model the transport of cosmic ray nuclei in the Galaxy by means of a new numerical code. Differently from previous numerical models we account for a generic spatial distribution of the diffusion coefficient. We found that in the case of radially u
niform diffusion, the main secondary/primary ratios (B/C, N/O and sub-Fe/Fe) and the modulated antiproton spectrum match consistently the available observations. Convection and re-acceleration do not seem to be required in the energy range we consider: $1 < E < 10^3$ GeV/nucleon. We generalize these results accounting for radial dependence of the diffusion coefficient, which is assumed to trace that of the cosmic ray sources. While this does not affect the prediction of secondary/primary ratios, the simulated longitude profile of the diffuse $gamma$-ray emission is significantly different from the uniform case and may agree with EGRET measurements without invoking ad hoc assumptions on the galactic gas density distribution.
In the last months several ballon and satellite experiments improved significantly our knowledge of cosmic rays (CR) spectra at high energy. In particular CREAM allowed to measure B/C, C/O and N/O up to 1 TeV and PAMELA the anti-p/p ratio up to 100 G
eV with unprecedented accuracy. These measurements offer a valuable probe of CR propagation properties. We performed a statistical analysis to test the compatibility of these results, as well as other most significant experimental data, with the predictions of a new numerical CR diffusion package (DRAGON). We found that above 1 GeV all data are consistent with a diffusion scenario in a well defined range of values of the diffusion coefficient energy power index and normalization.
