Abstract:
The propagation of TeV-PeV cosmic rays (CR) in our Galaxy can be described as a diffusive process. We discuss here two effects, with important observational consequences, that cannot be predicted by the diffusion approximation in its usual form. First, we present an explanation for the CR anisotropies observed at small angular scales on the sky. We show that the local magnetic field configuration within a CR mean free path from Earth naturally results in CR flux anisotropies at small and medium scales. Second, we point out that TeV-PeV CRs should be expected to diffuse strongly anisotropically in the interstellar medium on scales smaller than the maximum scale of spatial fluctuations of the field, $\sim 100$ pc. This notably questions the usual assumptions on CR diffusion around sources.

Abstract:
In the energy range from ~ 10^12 eV to ~ 10^15 eV, the Galactic cosmic ray flux has anisotropies both on large scales, with an amplitude of the order of 0.1%, and on scales between ~ 10 and ~ 30 degrees, with amplitudes smaller by a factor of a few. With a diffusion coefficient inferred from Galactic cosmic ray chemical abundances, the diffusion approximation predicts a dipolar anisotropy of comparable size, but does not explain the smaller scale anisotropies. We demonstrate here that energy dependent smaller scale anisotropies naturally arise from the local concrete realization of the turbulent magnetic field within the cosmic ray scattering length. We show how such anisotropies could be calculated if the magnetic field structure within a few tens of parsecs from Earth were known.

Abstract:
We propose a new way to detect individual bright Ultra-High Energy Cosmic Ray (UHECR) sources above background if the Galactic Magnetic Field (GMF) gives the main contribution to UHECR deflections. This method can be directly applied to maps given by experiments. It consists in starting from at least two high energy events above 6x10^19 eV, and looking at lower energy tails. We test the efficiency of the method and investigate its dependence on different parameters. In case of detection, the source position and the local GMF deflection power are reconstructed. Both reconstructions are strongly affected by the turbulent GMF. With the parameters adopted in this study, for 68 % of reconstructed sources, the angular position is less than one degree from the real one. For typical turbulent field strengths of 4 \mu G at the Earth position and 1.5 kpc extension in the halo, one can reconstruct the deflection power with 25 % precision in 68 % of cases.

Abstract:
We present a new method to search for heavy nuclei sources, on top of background, in the Ultra-High Energy Cosmic Ray data. We apply it to the 69 events with energies E>55 EeV published by the Pierre Auger Collaboration. We find a set of events for which the method reconstructs the source near the Virgo galaxy cluster. The probability to have a comparable set of events in some background is ~ 0.7 %. The reconstructed source is located at ~ 8.5 degrees from the active galaxy M87. The probability to reconstruct the source at less than 10 degrees from M87 for data already containing a comparable set of events is ~ 0.4 %. This may be a hint at the Virgo galaxy cluster as an ultra-high energy heavy nuclei source. We discuss the capability of current and near future experiments to test this possibility. Such a scenario gives a self-consistent description of the Auger anisotropy and composition data at the highest energies.

Abstract:
During a supernova explosion, a radiation-dominated shock (RDS) travels through its progenitor. A collisionless shock (CS) is usually assumed to replace it during shock breakout (SB). We demonstrate here that for some realistic progenitors enshrouded in optically thick winds, such as possibly SN 2008D, a CS forms deep inside the wind, soon after the RDS leaves the core, and therefore significantly before SB. The RDS does not survive the transition from the core to the thick wind when the wind close to the core is not sufficiently dense to compensate for the $r^{-2}$ dilution of photons due to shock curvature. This typically happens when the shock velocity is $\lesssim 0.1 {\rm c} \, (\frac{u_{\rm w}}{10\,{\rm km/s}}) (\frac{\dot{M}}{5 \cdot 10^{-4} \, {\rm M}_\odot {\rm /yr}})^{-1} (\frac{r_\ast}{10^{13}\,{\rm cm}})$, where $u_{\rm w}$, $\dot{M}$ and $r_\ast$ are respectively the wind velocity, mass-loss rate and radius of the progenitor star. The radiative CS results in a hard spectrum of the photon flash at breakout, which would produce an X-ray flash. Cosmic ray acceleration would start before SB, for such progenitors. A fraction of secondary TeV neutrinos can reach the observer up to more than ten hours before the first photons from breakout, providing information on the invisible layers of the progenitor.

Abstract:
We propose a new method to search for heavy nuclei sources, on top of background, in the Ultra-High Energy Cosmic Ray data. We apply this method to the 69 events recently published by the Pierre Auger Collaboration and find a tail of events for which it reconstructs the source at a few degrees from the Virgo galaxy cluster. The reconstructed source is located at ~ 8.5 degrees from M87. The probability to have such a cluster of events in some random background and reconstruct the source position in any direction of the sky is about 7 x 10^(-3). The probability to reconstruct the source at less than 10 degrees from M87 in a data set already containing such a cluster of events is about 4 x 10^(-3). This may be a hint at the Virgo cluster as a bright ultra-high energy nuclei source. We investigate the ability of current and future experiments to validate or rule out this possibility, and discuss several alternative solutions which could explain the existing anisotropy in the Auger data.

Abstract:
In this paper, we suggest a new way to identify single bright sources of Ultra High Energy Cosmic Rays (UHECR) on the sky, on top of background. We look for doublets of events at the highest energies, E > 6 x 10^19 eV, and identify low energy tails, which are deflected by the Galactic Magnetic Field (GMF). For the sources which are detected, we can recover their angular positions on the sky within one degree from the real ones in 68% of cases. The reconstruction of the deflection power of the regular GMF is strongly affected by the value of the turbulent GMF. For typical values of 4 microG near the Earth, one can reconstruct the deflection power with 25% precision in 68% of cases.

Abstract:
We investigate the diffusion of cosmic rays (CR) close to their sources. Propagating individual CRs in purely isotropic turbulent magnetic fields with maximal scale of spatial variations Lmax, we find that CRs diffuse anisotropically at distances r <~ Lmax from their sources. As a result, the CR densities around the sources are strongly irregular and show filamentary structures. We determine the transition time t* to standard diffusion as t* ~ 10^4 yr (Lmax/150 pc)^b (E/PeV)^(-g) (Brms/4 muG)^g, with b ~ 2 and g = 0.25-0.5 for a turbulent field with Kolmogorov power spectrum. We calculate the photon emission due to CR interactions with gas and the resulting irregular source images.

Abstract:
We show that the cosmic ray (CR) knee can be entirely explained by energy-dependent CR leakage from the Milky Way, with an excellent fit to all existing data. We test this hypothesis calculating the trajectories of individual CRs in the Galactic magnetic field. We find that the CR escape time $\tau_{\rm esc}(E)$ exhibits a knee-like structure around $E/Z={\rm few}\times 10^{15}$ eV for small coherence lengths and strengths of the turbulent magnetic field. The resulting intensities for different groups of nuclei are consistent with the ones determined by KASCADE and KASCADE-Grande, using simple power-laws as injection spectra. The transition from Galactic to extragalactic CRs is terminated at $\approx 2\times 10^{18}$ eV, while extragalactic CRs contribute sizeable to the subdominant proton flux already for $\gtrsim 2\times 10^{16}$ eV. The natural source of extragalactic CRs in the intermediate energy region up to the ankle are in this model normal and starburst galaxies. The escape model provides a good fit to $\ln(A)$ data; it predicts that the phase of the CR dipole varies strongly in the energy range between $1\times 10^{17}$ and $3\times 10^{18}$ eV, while our estimate for the dipole magnitude is consistent with observations.

Abstract:
We investigate the possibility that the cosmic ray (CR) knee is entirely explained by the energy-dependent CR leakage from the Milky Way. We test this hypothesis calculating the trajectories of individual CRs with energies between $E/Z=10^{14}$ eV and $10^{17}$ eV propagating them in the regular and turbulent Galactic magnetic field. We find a knee-like structure of the CR escape time $\tau_{\rm esc}(E)$ around $E/Z={\rm few}\times 10^{15}$ eV for a coherence length $l_{\rm c} \simeq 2$ pc of the turbulent field, while the decrease of $\tau_{\rm esc}(E)$ slows down around $E/Z\simeq 10^{16}$ eV in models with a weak turbulent magnetic field. Assuming that the injection spectra of CR nuclei are power-laws, the resulting CR intensities in such a turbulence are consistent with the energy spectra of CR nuclei determined by KASCADE and KASCADE-Grande. We calculate the resulting CR dipole anisotropy as well as the source rate in this model.