Recent data released by the AMS experiment on the primary spectra and
secondary-to-primary ratios in cosmic rays (CRs) can pose tight constraints to
astrophysical models of CR acceleration and transport in the Galaxy, thereby
providing a robust baseline of the astrophysical background for dark matter
search via antimatter. However, models of CR propagation are affected by other
important sources of uncertainties, notably from solar modulation and nuclear
fragmentation, that cannot be improved with the sole use of the AMS data. The
present work is aimed at assessing these uncertainties and their relevance in
the interpretation of the new AMS data on the boron-to-carbon (B/C) ratio.
Uncertainties from solar modulation are estimated using improved models of CR
transport in the Heliosphere constrained against various type of measurements:
monthly-resolved CR data collected by balloon-born or space missions,
interstellar flux data from the Voyager-1 spacecraft, and counting rates from
ground-based neutron monitor detectors. Uncertainties from nuclear
fragmentation are estimated using semiempirical cross-section formulae
constrained by measurements on isotopically-resolved and charge-changing
reactions. We found that a proper data-driven treatment of solar modulation can
guarantee the desired level of precision, in comparison with the improved
accuracy of the recent data on the B/C ratio. On the other hand, nuclear
uncertainties represent a serious limiting factor over a wide energy range. We
therefore stress the need for establishing a dedicated program of cross-section
measurements at the $\mathcal{O}$(100 GeV) energy scale.Comment: 15 pages, 10 figures, 3 tables - matches published versio