The origin and anisotropy of high energy cosmic rays
The evidence for the anisotropy of cosmic rays from 10(^11)eV to 10(^20)eV is considered in detail from both a Galactic and extragalactic viewpoint and placed in astrophysical context. The importance of recent measurements of the local interstellar wind (consistent with a direction from (ɑ,δ) = (252º, -16º) and velocity 21 - 23 kms(^-1)) is noted. Cosmic ray streaming along the lines of the local Galactic magnetic field appears to account for the constant observed anisotropy of 0.05% at (1 - 2)hrs R.A. below l0(^14)eV. However, the distribution of cosmic rays does not appear to conform to the expectations of axial symmetry. Observational and statistical aspects of anisotropy are considered with particular reference to harmonic analysis. The results are used for analysis of the collection of data by Linsley and Watson which are shown to be inconclusive for anisotropy measurements in the range 10(^14)eV to 10(^20)eV. The power of using only phase or only amplitude information from collections of measurements is noted. Anisotropy measurements from 10(^17)eV to 10(^20)eV are considered and seen to favour a mixed origin model in which particles above ~ 10(^17)eV are of extragalactic origin. The cosmic ray spectrum above 10(^18)eV can then be interpreted purely in terms of an effect of extragalactic particles. Three models to account for the upturn in the cosmic ray spectrum at high energies are considered. A simple model in which neutrons escape from clusters of galaxies before decaying into protons is found to account for the observed spectrum if a source spectrum of the form j (E) (_x)E(^-2.25) is adopted. A model based on diffusion of particles from the Virgo Supercluster is seen to give a satisfactory fit to the dataif a diffusion coefficient of D(E) = 5 x 10(^33) E(^1/2)(_19) cm(^2)s(^-1) is taken. The model also accounts for the observed amplitude of the anisotropy above 10(^17)eV. A model assuming production of high energy particles in radio galaxies is shown to be less satisfactory than the neutron or diffusion models.