1: %\documentclass[12pt,preprint]{aastex} %original

2: %\documentclass[apjl]{emulateapj}

3: %\documentclass[article]{aastex}

4: \documentclass{elsart}

5: 

6: 

7: %\bibliographystyle{aas} %original

8: 

9: \usepackage{graphicx}

10: \usepackage{amsmath}

11: \usepackage{epstopdf}

12: 

13: \usepackage{epsfig} %<-----------------not revtex standard - added by RMC

14: 

15: \usepackage{amssymb}

16: 

17: 

18: \newcommand{\gev}{\textrm{GeV}}

19: \newcommand{\ghz}{\textrm{GHz}}

20: \newcommand{\mhz}{\textrm{MHz}}

21: \newcommand{\pos}{e^{+}}

22: \newcommand{\elec}{e^{-}}

23: \newcommand{\epm}{e^{\pm}}

24: \newcommand{\degree}{^{\circ}}

25: %\newcommand{\etal}{\emph{et al.}}

26: \newcommand{\spectralIndex}{S_{\nu}\propto\nu^{\alpha}}

27: \newcommand{\solarmass}{M_{\odot}}

28: \newcommand{\cang}{CANGAROO}

29: \newcommand{\gam}{$\gamma$}

30: %\newcommand{\p0}{$\pi^0$}

31: \def\3EG{{3EG J1746-2851}}

32: \def\p0{{$\pi^0$}}

33: \def\1018{{$10^{18}$}}

34: \def\cry{{cosmic ray}}

35: \newcommand{\be}{\begin{equation}}

36: \newcommand{\ee}{\end{equation}}

37: \newcommand{\bea}{\begin{eqnarray}}

38: \newcommand{\eea}{\end{eqnarray}}

39: \newcommand{\bmu}{\begin{multline}}

40: \newcommand{\emu}{\end{multline}}

41: \newcommand{\nn}{\nonumber}

42: \newcommand{\V}{\vert}

43: \newcommand{\delslash}{/\!\!\!\partial}

44: \newcommand{\sae}{Sgr A East\,}

45: \def\MES{{mass eigenstate}}

46: \def\neu{{neutrino}}

47: \def\dm{{$\delta m^2  $}}

48: \def\dmatm{{$\delta m^2_{\small{ATM}}  $}}

49: \def\dmsun{{$\delta m^2_{\odot}  $}}

50: \def\dmsub{{$\delta m^2_{\small{SUB}}  $}}

51: \def\numu{{$\nu_\mu  $}}

52: \def\nue{{$\nu_e  $}}

53: \def\nutau{{$\nu_\tau  $}}

54: \def\nus{{$\nu_s  $}}

55: \def\RX{{SNR RXJ1713.7-3946}}

56: \def\msun{{\,M_\odot}}

57: \def\lsun{{\,L_\odot}}

58: \def\simlt{\lower.5ex\hbox{$\; \buildrel < \over \sim \;$}}

59: \def\simgt{\lower.5ex\hbox{$\; \buildrel > \over \sim \;$}}

60: \def\um{{\,\mu\rm m}}

61: \def\cm{{\rm\,cm}}

62: \def\km{{\rm\,km}}

63: \def\au{{\rm\,AU}}

64: \def\pc{{\rm\,pc}}

65: \def\kpc{{\rm\,kpc}}

66: \def\mpc{{\rm\,Mpc}}

67: \def\sec{{\rm\,s}}

68: \def\yr{{\rm\,yr}}

69: \def\gm{{\rm\,g}}

70: \def\kms{{\rm\,km\,s^{-1}}}

71: \def\mdot{{\rm\,\msun\,yr^{-1}}}

72: \def\gms{{\rm\,g\,s^{-1}}}

73: \def\gcm3{{\rm\,g\,cm^{-3}}}

74: \def\ncm3{{\rm\,cm^{-3}}}

75: \def\kelvin{{\rm\,K}}

76: \def\erg{{\rm\,erg}}

77: \def\kev{{\rm\,keV}}

78: \def\ev{{\rm\,eV}}

79: \def\hz{{\rm\,Hz}}

80: \def\>{$>$}

81: \def\<{$<$}

82: \def\bsl{$\backslash$}

83: \def\refbook#1{\refindent#1}

84: 

85: 

86: %\lefthead{Clay and Crocker}                      

87: %\righthead{High Energy Cosmic Ray Diffusion}              

88: 

89: 

90: 

91: 

92: 

93: 

94: 

95: 

96: 

97: 

98: 

99: %\renewcommand{\refname}{}   %allows one to define own title for bibli%ography

100: 

101: \received{}

102: \begin{document}

103: 

104: \begin{frontmatter}

105: 

106: 

107: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

108: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

109: 

110: \title{Exploring the High-Energy Cosmic Ray Spectrum with a Toy Model of Cosmic Ray Diffusion}

111: 

112: 

113: \author{Roger Clay \& Roland M. Crocker }

114: \address{School of Chemistry and Physics\\

115: University of Adelaide\\

116: 5005, Australia\\

117: roger.clay,roland.crocker@adelaide.edu.au}

118: 

119: \date{\today}

120: 

121: 

122: 

123: \begin{abstract}

124: 

125: 

126: We introduce a static toy model of the cosmic ray (CR) universe in which cosmic ray propagation is taken to be

127: diffusive and cosmic ray sources

128: are distributed randomly with a density the same as that of local $L_*$ galaxies, $5 \times 10^{-3}$ Mpc$^{-3}$. 

129: These sources ``fire" at random

130: times through the history of the universe but with a set expectation time for the period between bursts.

131: Our toy model model captures much of the essential CR physics 

132: despite its simplicity and,

133: moreover, broadly reproduces CR phenomenology

134: for reasonable parameter values and without extreme fine-tuning. 

135: Using this model we investigate -- and find tenable -- the 

136: idea that the 

137: Milky Way may itself be a typical high-energy cosmic ray source. 

138: We also consider the possible phenomenological implications of the magnetic CR horizon

139: for the overall cosmic ray spectrum

140: observed at Earth.

141: Finally, 

142: we show that anisotropy studies should most profitably focus

143: on cosmic rays detected at energies above the so-called GZK cut-off, $\sim 6 \times 10^{19}$ eV.

144: 

145: \end{abstract}

146: 

147: \begin{keyword}

148: % keywords here, in the form: keyword \sep keyword

149: 

150: % PACS codes here, in the form: \PACS code \sep code

151: \PACS 

152: \end{keyword}

153: \end{frontmatter}

154: 

155: %\maketitle

156: 

157: \section{Introduction}

158: 

159: In the following study, we invoke a toy model of cosmic ray diffusion, in a static universe, 

160:  from a random ensemble

161: of extragalactic sources distributed throughout a cube with 2400 Mpc sides and the Earth at the center.

162: We assume a constant source density equal to $5 \times 10^{-3}$ Mpc$^{-3}$.

163: This is approximately

164: the local density of Milky-Way-like galaxies 

165: as determined in \cite{Aublin2006} on the basis of the ratio

166: of the local star formation rate density and the Galactic star formation rate

167: (also see \cite{Loeb2002}).

168: Each such source is taken to ``fire" at random times through its history, but with a

169: well-defined expectation for the time between such firings 

170: ($t_{\rm wait} \equiv 1 $ Mpc/$c \simeq 5 \times 10^6$ year).

171: 

172: The power density of extragalactic cosmic ray sources can

173: be estimated by requiring it support the observed, high-energy spectrum

174: against losses.

175: On the basis of normalizing to the observed spectrum

176: at $10^{19}$ eV 

177: this has been determined to be around $5 \times 10^{44}$ erg yr$^{-1}$ Mpc$^{-3}$ \cite{Waxman1995,Blasi2004}.

178: With this input,

179: the typical energy into CRs from one of the outbursts we model must be $\sim 5 \times 10^{53}$ erg,

180: an amount of energy much too large to be associated with a single supernova but

181: only $1 - 10 \%$ of the energy released in 

182: a period of Seyfert activity or star-bursting in a ``typical" galaxy 

183: (see \cite{Veilleux2005} and \cite{Bland-Hawthorn2003} and references therein). 

184: We discuss the naturalness of these various scales further below.

185: 

186: Cosmic rays from our assumed extragalactic sources are taken to diffuse through

187: a purely turbulent magnetic field.

188: We assume that our modeled CRs are dumped {\it directly} into the extra-cluster space.

189: This allows the derivation, as a function of magnetic field amplitude and coherence length, 

190: of an upper limit (at any given energy) 

191: on the physical distance to the {\it magnetic horizon}

192: beyond which CRs could not have originated 

193: given the time available (given the age of the universe

194: and energy loss processes) to diffuse through the intervening fields.

195: 

196: We aim to sample field strengths representative of those found in 

197: extra-cluster space.

198: {\it Inside} clusters many lines of evidence now point to magnetic field amplitudes at the

199: few $\mu$G level out to distances of $\sim$ Mpc from cluster cores \cite{Clarke2001}, quite large enough

200: to affect the propagation of CRs to the highest energies.

201: 

202: Unfortunately, constraints on extra-cluster magnetic field strengths are not particularly strong.

203: At the super-cluster scale of $\lesssim$ 10 Mpc

204: there is evidence for magnetic field amplitudes

205: at the few $\times 0.1 \ \mu$G level \cite{Kim1989}.

206: For fields extending over cosmological distances

207: (i.e., on scales exceeding those

208: pertinent to

209: the filaments and walls of large scale structure,

210: $\sim$ 50 Mpc \cite{Widrow2002})

211: upper limits are in the range 1-10 nG for coherence lengths in the range

212: 1-50 Mpc, on the basis of 

213:  examination of the rotation measures of

214: distant QSOs \cite{Blasi1999}.

215: Finally, we note that regular field components  -- neglected in our model --   

216: would tend to push out the magnetic horizon in certain directions (and pull it in elsewhere) 

217: given the phenomenon of CR drift in such fields.

218: The regular component is, however, expected to be very low 

219: in extragalactic space so that CR transport should be dominated, as we shall assume, 

220: by phenomena

221: associated with the turbulent field component \cite{Parizot2004}.

222: 

223: To sample, then, the reasonable parameter space 

224: we investigate field amplitudes of 1, 10 and 100 nG.

225: We also consider 

226: two coherence lengths:

227: 10 kpc 

228: and 1 Mpc. 

229: These values bracket the range of scales from

230: galactic to cluster-size

231: and, therefore, the extragalactic magnetic field

232: coherence ``length" must fall within this range.

233: We neglect any possible evolution of all these quantities 

234: in our modelling.

235: 

236: Additionally, we model CR diffusion away from a single

237: local source -- of the same average power as an extragalactic source -- 

238: that is located at a distance equal to our separation from the center of the Galaxy, 

239: $\sim 8.5 $ kpc.  

240: The purpose of simulating the  additional, local source is to examine the consequences of the assumption

241: that the Milky Way (MW) be, itself, a typical CR source.

242: This idea, labeled the {\it holistic source model} by Aublin \cite{Aublin2006}, has

243: been examined by a number of authors (see, e.g., \cite{Milgrom1996,Loeb2002,Dermer2002,Aublin2006}).

244: 

245: Note that, if, by hypothesis, the MW is 

246: (or has contained) a typical CR source, 

247: then it must be capable of producing -- at least on occasion -- CRs at energies

248: up to and exceeding $10^{19}$ eV.

249: As a corollary to this statement, because CRs at such energies will not be greatly deflected

250: away from rectilinear propagation over Galactic length scales by the Galactic magnetic field,

251: the fact that there are no well-established anisotropies associated with Galactic structures

252: (Galactic plane, Galactic center) at these energies sets a minimum time scale to the time

253: since the last firing of the MW CR source cite{Giler1983,Giller2000}. 

254: Of course, this is not to say that we necessarily do not detect $\gtrsim 10^{19}$ eV CRs from past

255: firings of the putative MW source, only that any we do detect have been diffusion processed to

256: the extent that they are almost equally likely to come from any region of the sky. 

257: 

258: We do

259: not know the properties of the {\it local} magnetic field  

260: (immediately outside our

261: own Galaxy) but we assume that it is {\it unlikely} to have a field

262: strength as low as is found at a great distance from any galaxy

263: cluster.  

264: %Where such fields have been measured within clusters of

265: %galaxies (Coma, Kronberg; Abell Clusters, Clark et al.) they have

266: %been at microgauss levels.  

267: We asume, therefore, 

268: that the magnetic field in the nearby interior of the cluster

269: within which our Galaxy is located is at or above

270: 0.1 $\mu G$. 

271: We chose to sample field strengths of 0.1,0.2, and 10 $\mu {\textrm G}$ in our modelling,

272: the latter representing an extreme value (a direct extension of that typical for the

273: Galactic plane field strength). 

274: The local field is assumed to have a spherically-symmetric structure

275: and a coherence length of 10 kpc.

276: We emphasise that the local field introduced here

277: does {\it not} represent the field typical for the Galactic disk (which is known to have

278: both turbulent and regular components and a typical total amplitude of few $\mu{\textrm G}$ near the Earth), 

279: but rather the larger, cluster-scale magnetic field within which the

280: entire Galaxy is situated.

281: The effect of confinement in this latter field can be

282: calculated by following the procedure described by \cite{Clay2002}

283: but is neglected here as it does not affect the

284: spectrum at energies of $10^{18}$ and above with which we concern ourselves.

285: 

286: 

287: In summary our physical picture is 

288: one in which at energies 

289: above the spectral upturn in the cosmic ray spectrum at $\sim 3 \times 10^{18}$ eV (the {\it ankle})

290: there is overlap between the CR diffusion spheres around individual galaxies 

291: and we measure a flux of CRs that is the same as anywhere else (including inter-galactic space)\cite{Aublin2006}, 

292: i.e., universal.

293: In contrast, at considerably lower energies we measure at earth a flux of CRs that

294: originates from the Milky Way and

295: is over-abundant with respect to the universe-at-large. 

296: At the lower end of the region of concern to us ($\sim 10^{17}$ eV), 

297: the Galactic CR spectrum is in approximate steady-state in our model because

298: the diffusion time from the Galactic center to us through the local field is longer than

299: the expected time between firings of this local source. 

300: At the upper end of the Galactic spectrum

301: (but below the region where the extragalactic flux becomes dominant)

302: this condition is not satisfied, however, and

303: the spectrum is time-dependent. One must tune to the time of

304: the last Galactic CR outburst to arrive at a  spectrum

305: consistent overall with observations. 

306: 

307: %As presaged above,

308: %demanding that modeled anisotropies are at or below empirical limits,

309: %also 

310: %we obtain a handle

311: %on this time (which will, in turn, depend on the other assumptions that

312: 

313: %In demanding that the modeled spectrum resemble that observed 

314: %-- and,  

315: %have gone into the model).

316: 

317: 

318: 

319: Our picture, then,

320: is conventional in the sense that the

321: ankle

322: is associated with the transition from dominance by the Galactic source

323: to dominance by the extragalactic sources. 

324: We show below, however, that the assumption that 

325: the Milky Way be a typical CR source ameliorates a fine-tuning problem

326: implicit in this interpretation of this structure, viz.

327: {\it why is the ankle placed where it is such that we can observe an extragalactic flux but this flux does not

328: dominate the Galactic CR flux?}

329: 

330: Finally, we note in passing here

331: that alternative recent models (see \cite{Berezinsky2007} and references therein)

332: that would posit a transition at considerably

333: lower energies, $\sim 10^{17.5-18}$ eV, and explain the ankle as the result of 

334: a dip in the spectrum of extragalactic protons because of their

335: Bethe-Heitler pair production collisions on the CMB, would seem

336: to suffer an even more extreme fine-tuning problem.

337: This, namely, is the matching  between the normalizations

338: of the Galactic and extragalactic components

339: required so that the transition 

340: is indicated by no strong spectral feature at all or,

341: at least, one as weak as the spectral downturn represented by the

342: so-called second knee.

343: 

344: \section{Diffusive Transport}

345: 

346: In this work we assume that CR transport can be described as a purely diffusive process.

347: Following, e.g., \cite{Gaisser1990}, 

348: denoting the density of protons at position $\mathbf{x}$ and with energy between

349: $E_p$ and $E_p + d E_p$ and at time $t$ by $N_p(E_p,\mathbf{x},t)$, 

350: the proton transport equation with acceleration, 

351: convection, and collision losses and gains all neglected can be written:

352: \begin{equation}

353: {\dot N} \, = \, \nabla \cdot (D \nabla N) \, + \, Q \, ,

354: \label{eqn_transport}

355: \end{equation}

356: where $Q$ is an explict source of particles and $D$ is the diffusion coefficient which, formally, relates the current of particles

357: (in our case protons) to a spatial gradient in the density of such particles \cite{Gaisser1990}.

358: 

359: The Green's function for Eq.(\ref{eqn_transport}) -- 

360: which, physically, gives the probability for finding a particle, that was injected

361: at the origin, at a position $\mathbf{r}$ after a time $t$ -- is

362: \begin{equation}

363: G(\mathbf{r},t) \, = \, \frac{1}{8 (\pi \, D \, t)^{3/2}} \exp\left( \frac{-\mathbf{r}^2}{4 \, D \, t} \right) \, .

364: \label{eqn_greens}

365: \end{equation}

366: We use Eq(\ref{eqn_greens}) to determine the CR flux due to each CR outbursts, at some given time previous to now,

367: from each simulated CR source at (random) position $\mathbf{r}_{\rm source}$. 

368: 

369: The lower limit on the diffusion coefficient is given by the Bohm case wherein

370: the scattering distance is equal to the gyroradius implying that

371: \be

372: D_{\rm Bohm}(E_p, B) \, = \, \frac{c \, r_{\rm gyro}}{3} \, ,

373: \ee

374: where the gyro-radius is, in general given by

375: \begin{equation}

376: r_{\rm gyro}(p,Z,B) = \frac{p \, c}{Z \, B} 

377: \simeq {\rm 1 pc} \, \left( \frac{E}{\rm PeV} \right) 

378: \, \left( \frac{Z \, B}{\rm \mu G} \right)^{-1}\, ,

379: \label{eqn_gyro}

380: \end{equation}

381: (where $Z$ is the CR's charge in units of the charge on the proton and $p$ its 

382: momentum\footnote{Note that in this work we only model the propagation of protons so that

383: $Z \ = \ 1$ in equation \ref{eqn_gyro} and throughout.}).

384: 

385: A more realistic behavior for the diffusion coefficient -- that we employ in our modeling -- is as

386: parameterized on the basis of the numerical work of Parizot

387: \cite{Parizot2004} who considers particle transport in a purely

388: turbulent field with a Kolmogorov spectrum:

389: \be

390: D(E_p) \, = D_\star \left[\left(\frac{E}{E_\star}\right)^{\frac{1}{3}} 

391: + \left(\frac{E}{E_\star}\right) + 

392: \left(\frac{E}{E_\star}\right)^2\right]\, .

393: \ee

394: where $E_\star$ is implicitly defined via

395: \be

396: r_{\rm gyro} (E_\star) \equiv \frac{\lambda_{\rm coh}}{5},

397: \ee

398: and

399: \be

400: D_\star \equiv \frac{1}{4} c r_{\rm gyro} (E_\star) \, .

401: \ee

402: In these equations $\lambda_{\rm coh}$ denotes the coherence length of the magnetic field.

403: 

404: As implicit in the above equations,

405: the investigations of \cite{Parizot2004} reveal that for particle energies such that the gyro-radius

406: considerably exceeds the coherence length of the magnetic field, the scattering length

407: scales as $E^2$. 

408: At the other extreme, for low energies such that $r_{\rm gyro} \ll \lambda_{\rm coh}$ and for a 

409: Komogorov spectrum of magnetic field turbulence, the scattering length scales as $E^{1/3}$.

410: Finally, when $r_{\rm gyro} \sim \lambda_{\rm coh}$ a Bohm-like scaling is approximately followed

411: for a limited energy range

412: with the scattering length $\propto E$ (though note the coefficient of proportionality never

413: actually falls to the Bohm value: $D \gtrsim 3 D_{\rm Bohm}$; \cite{Parizot2004}).

414: 

415: \section{Energy Loss in Extra-Galactic Space}

416: 

417: In propagating over cosmological distances, cosmic ray protons lose energy through red-shifting and via inelastic collisions

418: with background light fields (through both Beth-Heitler pair-production and resonant photo-pion production),

419: the cosmic microwave background most significantly at energies $\gtrsim 10^{19}$ eV 

420: \cite{Greisen1966,Zatsepin1966}. 

421: We account for the modification induced by these effects on input spectra via a

422: parameterization of the energy-dependent attenuation length

423: calculated by \cite{Geddes1996} and presented in their fig.1.

424: Note that we neglect protons down-shifted in energy through their interactions,

425: an approximation that works reasonably for power-law spectra $\propto E^{-2}$

426: or steeper such as we investigate here.

427: 

428: \section{Results}

429: 

430: %In our modelling of the cosmic ray spectrum based on the ideas described above,

431: %we selected cosmic ray source positions randomly out to distances

432: %of 1.2 Gpc, with a source (galactic) density close to that

433: %observed locally.  These source positions were initially assumed to contain

434: %sources which underwent random outbursts of fixed cosmic ray

435: %intensity at a mean rate of one outburst per $3\times 10^6$

436: %years.  Each outburst contributed to our local cosmic ray flux to

437: %an extent determined by its outburst epoch and its distance,

438: %using a model based on simple diffusion (see e.g. \cite{Gaisser1990} pp

439: %123 - 126) with the diffusion coeficient noted above. Allowance

440: %was made for energy loss processes using the energy dependent

441: %attenuation length by \cite{Geddes1996}, also as noted above.

442: 

443: We plot some representative results of our spectral modeling in

444: figure \ref{fig_plot1.eps}.  

445: In this figure, we show

446: spectra derived from a randomly-placed ensemble of

447: extragalactic proton sources.  

448: The cosmic

449: rays have propagated through turbulent magnetic fields (all with a coherence length of 10 kpc) and

450: of

451: strength 1, 10, 30, and 100 nG and had $E^{-2}$ source spectra at injection.

452: Diffusive propagation is a slow process and the magnetic horizon

453: is increasingly important towards lower particle energy  and in the case

454: of strong magnetic fields.  This effect is evident in the

455: figure. We note that for the strongest magnetic field pictured there is a complicated

456: interaction between the GZK and magnetic horizon effects which results in not only

457: the expected systematic shift of the peak of the $E^3$-weighted spectrum to the right but also an

458: overall attenuation of this weighted spectrum.

459: 

460: The effect of changing the magnetic field coherence length 

461: can, again in concert with the GZK attenuation,

462:  also result in complicated phenomenology as shown in figure 2.

463: %In particular, changing this length from 10 kpc to 1 Mpc, while keeping 

464: %the magnetic field amplitude fixed, slows down diffusion at these 

465: %energies and moves the peak of the $E^3$-weighted spectrum to the right while also broadening

466: %the distribution.

467: We noted above that the ratio of the gyroradius to the coherence length determines the scattering properties of 

468: particles in the turbulent field. For a 10 nG field strength the proton gyroradius is $\sim$Mpc

469: at $10^{19}$ eV and we therefore expect an appreciable dependence of the spectrum on the assumed 

470: coherence length at these energies.

471: 

472: The Milky Way galaxy is of unique importance as a CR source but, as presaged above, in our

473: model it can be added in the same way as the random galaxies.

474: This is also shown in the figure, although we now have to

475: explicitly specify the time of the most recent outburst as, in

476: this case, that outburst can have a dominating effect.  

477: The figure shows results for a magnetic field

478: of 0.2 $\mu{\textrm G}$

479: with the most recent local outburst six million years ago -- a case which reproduces the data tolerably well.

480: We have found that, as expected,  ``dialing-up" the local field strength or

481: taking a more recent time for the last local outburst

482: increases the amplitude of the 

483: local component relative to the extragalactic contribution.

484: 

485: Note that our procedure, using only

486: assumptions of similarity between all galaxies such as the Milky

487: Way and a common mean time between identical outbursts of three

488: million years, provides a selection of spectra which are not

489: dissimilar to those observed and commonly assumed to be Galactic and

490: extragalactic. Apart from selecting within the modest range of

491: plausible field parameters shown in the figure, there is no

492: arbitrary normalisation between the spectral components.  

493: 

494: If we select results for a 10 nG intergalactic with 1 Mpc coherence length

495: field and a 0.2 $\mu \textrm{G}$ local field,

496: we get a combined

497: spectrum plausibly similar to that which is observed

498: especially considering the scale of the uncertainties introduced by

499: the imprecisely-known input parameters (particularly, the source density and expectation

500: time between firings) and the admitted crudity of our model

501: (which assumes all sources have the same time-integrated power, a single coherence

502: length for the fields rather than a distribution, etc).

503: Figure 3

504: shows the summed spectrum due to these components

505: and 

506: the  spectrum measured by Auger

507: \cite{VanElewyck2007}.

508: We also emphasise that the error bars on the experimental

509: data only show statistical errors, not systematic, which, in reality

510: can be considerable. 

511: The Auger collaboration \cite{Parizot2007} states that, at present,

512: the statistical and systematic uncertainties within the relevant

513: energy scale are 6\% and 22\% respectively.

514: 

515: 

516: Also plotted in Figure 3 is the extragalactic flux expected 

517: for the same 10 nG field but with the 

518: opposite extremum of coherence length, viz. 10 kpc.

519: In such a field structure

520: it can be ascertained that CR diffusion becomes too fast

521: at energies in the vicinity of the GZK ``cut-off" energy for the highest 

522: flux to be reproduced. The spectrum, therefore, dies away too quickly at high energy.

523: We emphasise that this cut-off in the spectrum in not 

524: simply given by the GZK effect 

525: 

526: In general, one can see that in order to reproduce the observed spectrum a dip in the range

527: 18.5 $\gtrsim$ log($E_{CR}$/eV) $\gtrsim$ 19.0 is required. 

528: This sets a constraint on the extragalactic field strength in the range 10-100 nG, at least within

529: our toy model.

530: For field intensities $\gtrsim 100$ nG the dip in the overall (extragalactic + Galactic sources)

531: is too large.

532: At field intensities $< 10$ nG there is insufficient flux above $10^{19}$ eV 

533: to reproduce the observed spectrum.

534: 

535: Again, at least within our toy model, we can constrain the time

536: of the last outburst of the MW source.

537: If, as we assume, there is a 

538: halo magnetic field local to the intra-cluster space around the Galaxy 

539: then, for a local field intensity of 10 $\mu$G, the last outburst

540: was earlier than 30-300 million years ago (lest the MW source signal

541: completely swamp all other sources 

542: meaning that the observed dip in the spectrum cannot be reproduced).

543: Of course, this is much longer than the expected time between outbursts from the MW

544: if it is an ordinary CR source and, therefore, represents an extreme fine-tuning indicating that

545: this choice of local field intensity is disfavored.

546: For a 1 $\mu$G field, the last outburst is required to be more than 30 million years ago, 

547: again a fine-tuning given the 3 million year expectation time.

548: Finally, for a 0.1 $\mu$G field, the last outburst should be in the range 1-10 million years

549: and we find that, at least from the point of view of naturalness, 

550: that this range of amplitude would seem to be favored by our model

551: (we chose a 0.2 $\mu$G field).

552: 

553: \subsection{Anisotropy}

554: 

555: In the diffusive regime, we can investigate the relative

556: anisotropy due to each of the assumed sources.  

557: Our analysis employs

558: simple diffusion ideas as expressed, for example, in \cite{Allan1972}, 

559: in which the anisotropy is given in magnitude by the

560: ratio of the scattering mean free path (or gyroradius under some

561: circumstances) to the source distance.  Our results for the

562: spectrum model just described give a low anisotropy of below 1\%

563: between 0.2 EeV and 30 EeV.  However, at energies at which the

564: GZK cut-off is important, the anisotropy then rises rapidly with

565: energy and is above 10\% when 100 EeV is reached.  The detail of

566: this energy dependence will clearly depend on the real local

567: distribution of galaxies.  However, the important, and perhaps

568: not surprising, point is that the fact that the GZK effect limits

569: the distance to observable sources makes the actual spatial

570: distribution of sources observable.  In a conventional model for

571: AGN sources of the highest energy cosmic rays, this result is

572: obviously intuitive, but that is not so for our picture of bursts

573: from conventional galaxies.

574: 

575: \section{Discussion -- Naturalness of Inferred Scales}

576: 

577: Bland-Hawthorn and Cohen \cite{Bland-Hawthorn2003} have, on the basis of 

578: infra-red, radio, and X-ray

579: observations, inferred the 

580: existence of a large-scale, bipolar wind out of the Galactic center.

581: This wind and other structures seen on large scales -- in particular, the North Polar Spur, 

582: an X-ray/radio loop that extends from the Galactic plane all the way to 

583: $b = +80^\circ$ \cite{Sofue2000} -- support, in turn, the notion

584: that the Galactic center \cite{Melia2001} is host to explosive outbursts 

585: of total energy $\sim 10^{55}$ erg that

586: occur every $\sim 10$ million years or so (see \cite{Bland-Hawthorn2003}

587: and references therein).

588: Sanders \cite{Sanders1981} long ago predicted a similar time scale and energetics for

589: intermittent activity of the Galactic center and similar spiral-galaxy nuclei

590: with a period of Seyfert luminosity ($L > 10^{43}$ erg s$^{-1}$) of duration $\sim 10^5$ years expected

591: every $\sim 10^7$ years.

592: In Sanders' picture this activity is driven by

593: the intermittent accretion of gas in giant molecular clouds on to the galaxy's central

594: black hole. 

595: In contrast, Bland-Hawthorn and Cohen prefer an interpretation of their observations in terms of star-bursting. 

596: In fact, activity of the central black hole and nuclear star bursting may be closely inter-related. 

597: For instance, Nayakshin and Cuadra \cite{Nayakshin2005} favor a picture in which, some millions of years ago,

598: our Galaxy was ``robbed" of the chance for truly bright AGN activity because

599: the gaseous fuel that would otherwise have been available to power this activity was 

600: driven away by a nuclear star-burst. 

601: Nevertheless, in their picture, the MW achieves a Seyfert luminosity.

602: 

603: Regardless of the particulars, similar

604: energetics and timescales for periodic Galactic center activity emerge from 

605: all of the above and, assuming $\sim 5\%$

606: of the total energy released in each event ends up in high-energy protons, the

607: energetics and outbursting timescales we infer

608: for the putative local CR source match these scales nicely.

609: In this  connection, we note that

610: the scale of our inferred Galactic center CR ``explosions"

611: are well inside the contraints on CR flux implied by the non-detection

612: of Li I and B I in the Sgr A molecular cloud \cite{Lubowich1998}.

613: 

614: It is also interesting to note that Giller and co-workers \cite{Giler1983,Giller2000}

615: have long advocated the idea that CR outbursts from the Galactic center 

616: may account for much of the observed, high energy CR spectrum.

617: Further supporting our general picture and our favored parameters,

618: Giller's work has tended to support 

619: a periodicity of up to 10 million years for the GC outbursts in order that anisotropy

620: upper limits are obeyed.

621: 

622: Another interesting result is that the energy density in the favored field strength 

623: for the local (halo) field, 0.2 $\mu$G, is $\sim \times 10^{-3}$ eV cm$^{-3}$

624: which is close to the energy density represented by the 

625: cosmic ray spectrum above the knee in the spectrum.

626: 

627: \section{Conclusion}

628: 

629: We have demonstrated a toy model that can reproduce the broad phenomenology of the 

630: ultra-high energy cosmic ray spectrum for reasonable parameter values. 

631: Our model invokes a random distribution of extragalactic sources 

632: with a space density equal to that observed in the local universe for Milky Way-like 

633: galaxies that ``fire", i.e. explosively inject, an $E^{-2}$ spectrum of cosmic

634: ray protons every three million years on average. 

635: By construction, our model also incorporates a local source located at a distance of 8.5 kpc,

636: the approximate distance to the Galactic center.

637: From our modeling, we favor a local field (but external to the Milky Way disk) of

638: an amplitude in the $\sim$ 0.1 $\mu$G range with the most recent firing of the 

639: Milky Way source to be at $\sim 6$ million years, a comfortable multiple  

640: (2) of the average temporal separation between firings of the average source 

641: (which we assume the local source to be).

642: As far as the extragalactic contribution to the spectrum goes, our modeling favors an 

643: average intergalactic field of 10 nG amplitude and an average coherence length for this field

644: toward the longer end of the allowable range, viz. 1 Mpc.

645: 

646: We note that these parameter choices are probably not unique 

647: nor do they perfectly reproduce the observed 

648: spectrum. 

649: %The extragalactic CR component, in particular, seems to benefit

650: %from a ``by eye"

651: %renormalization of $\sim 1.5$ in matching the data from Auger.

652: %Such a renormalization could comfortably be effected by an upwards revision of the 

653: %assumed source density or firing rate.

654: %In summary, 

655: Given the crudity of our model, however, and the uncertainty in the input parameters,

656: we think we have demonstrated the tenability of the broad ideas our model instantiates

657: 

658: A more detailed model than that under consideration would, in particular, take into account the known distribution of

659: local galaxies and might, in addition, assign a variable cosmic ray-power

660: to each of these known galaxies correlated with, say, the inferred mass of the central black

661: hole or with some other correlate of past cosmic ray activity.

662: Such considerations

663: are potentially

664: of particular relevance in the case of Andromeda, the closest large galaxy to the

665: Milky Way.

666: This object is located at a distance of only $\sim$ 800 kpc, significantly closer than the expected distance

667: to the closest Galaxy in our model, viz. 

668: $\sim (5 \times 10^{-3} \, \textrm{Mpc}^{-3})^{-1/3} \simeq 6$ Mpc.

669: It also hosts a supermassive black hole of $\sim 10^8 \solarmass$ \cite{Bender2005}, substantially

670: larger than the $\sim 4 \times 10^6 \solarmass$ black hole located at the Galactic center

671: and may, therefore, have been a significantly more powerful CR emitter that the MW 

672: in its history.

673: %On the other hand, our extreme ignorance about the large scale magnetic

674: %field geometry between local galaxies 

675: %would still introduce large and unavoidable uncertainties in this more detailed computation. 

676: Such considerations will be addressed in a future work.

677: 

678: 

679: \begin{figure*}[ht]

680: \epsfig{file=plot1.eps,height=8cm,angle=0}

681: \caption{Energy-cubed weighted fluxes for 

682: a local source in a local field of 200 nG with last outburst 6 Myr ago

683: (shown as the solid (blue) curve)

684: plotted against the following extragalactic cases (all calculated assuming

685: a 10 kpc coherence length for the extragalactic magnetic field structure):

686: dotted (green) -- 100 nG intergalactic field;

687:  long dash (purple) -- 30 nG field;

688: short dash (orange) -- 10 nG;

689: dot-dash (red) -- 1 nG;

690: data points -- Auger spectrum with statistical errors \cite{VanElewyck2007}

691: }

692: \label{fig_plot1.eps}

693: \end{figure*}

694: 

695: \begin{figure*}[ht]

696: \epsfig{file=plot2.eps,height=8cm,angle=0}

697: \caption{Energy-cubed weighted fluxes for 

698: a local source in a local field of 200 nG with last outburst 6 Myr ago

699: (shown as the solid (blue) curve)

700: plotted against the following extragalactic cases (all calculated assuming

701: a 10 nG amplitude for the extragalactic magnetic field structure):

702: short dash (orange) -- 10 kpc coherence length;

703: dotted (green) -- 1 Mpc coherence length;

704: data points -- Auger spectrum with statistical errors \cite{VanElewyck2007}

705: }

706: \label{fig_plot2.eps}

707: \end{figure*}

708: 

709: \begin{figure*}[ht]

710: \epsfig{file=plot3.eps,height=8cm,angle=0}

711: \caption{Modeled Galactic + extragalactic fluxes

712: weighted by energy-cubed.

713: In both cases the Galactic spectrum is calculated assuming

714: a 0.2 $\mu$G local field and a most-recent firing 6 Myrs ago.

715: The extragalactic fluxes are for the following cases:

716: solid (blue) -- 10 nG intergalactic field with 1 Mpc coherence length;

717: long-dash (purple) -- 10 nG intergalactic field with 10 kpc coherence length;

718: data points -- Auger spectrum with statistical errors \cite{VanElewyck2007}.

719: }

720: \label{fig_plot3.eps}

721: \end{figure*}

722: 

723: \section{Acknowledgements}

724: The authors acknowledge enlightening conversations Todor Stanev.

725: RMC is supported at the University of Adelaide by Ray Protheroe

726: and Ron Ekers'

727: Australian Research Council's Discovery funding scheme grant

728: (project number

729: DP0559991). 

730: 

731: \newpage

732: 

733: \begin{thebibliography}{00}

734: 

735: %\bibitem%[Aharonian et al.(2006)]

736: %{Aharonian2006}

737: %F. Aharonian et al. 2006, Nature, 439, 695

738: 

739: %\bibitem{Abbasi2005} Abbasi, R.~U., et al.\ 

740: %2005, Astroparticle Physics, 23, 157 

741: 

742: \bibitem{Allan1972} 

743: Allan, H.~R.\ 1972, Astrophysics Letters, 12, 

744: 237 

745: 

746: \bibitem{Aublin2006} 

747: Aublin, J., \& 

748: Parizot, E.\ 2006, A\&A, 452, L19 

749: 

750: \bibitem{Bender2005} Bender, R., et al.\ 

751: 2005, ApJ, 631, 280 

752: 

753: \bibitem{Berezinsky2007} Berezinsky, V., 

754: Gazizov, A., \& Grigorieva, S.\ 2007, ArXiv Astrophysics e-prints, 

755: arXiv:astro-ph/0702488

756: 

757: \bibitem{Bland-Hawthorn2003} 

758: Bland-Hawthorn, J., \& Cohen, M.\ 2003, ApJ, 582, 246 

759: 

760: \bibitem{Blasi1999} Blasi, P., Burles, S., \& 

761: Olinto, A.~V.\ 1999, ApJL, 514, L79 

762: 

763: \bibitem{Blasi2004} Blasi, P., \& de 

764: Marco, D.\ 2004, Astropart.Phys., 20, 559 

765: 

766: \bibitem{Clarke2001} Clarke, T.~E., Kronberg, 

767: P.~P.,B{\" o}hringer, H.\ 2001, ApJL, 547, L111 

768: 

769: \bibitem{Clay2002} Clay, R.~W.\ 2002, P.A.S.A., 19, 228 

770: 

771: \bibitem{Deligny2004} Deligny, O., 

772: Letessier-Selvon, A., \& Parizot, E.\ 2004, Astropart.Phys., 21, 609 

773: 

774: \bibitem{DeMarco2003} de Marco, D., Blasi, 

775: P., \& Olinto, A.~V.\ 2003, Astropart.Phys., 20, 53 

776: 

777: \bibitem{Dermer2002} Dermer, C.~D.\ 2002, ApJ, 574, 

778: 65 

779: 

780: \bibitem{Gaisser1990} Gaisser, T.~K.\ 1990, 

781: Cambridge and New York, Cambridge University Press, 1990, 292 p.,  

782: 

783: \bibitem{Geddes1996} Geddes, J., Quinn, 

784: T.~C., \& Wald, R.~M.\ 1996, ApJ, 459, 384 

785: 

786: \bibitem{Greisen1966} Greisen, K.\ 1966, Phys.Rev.Let., 16, 748 

787: 

788: \bibitem{Giler1983} Giler, M.\ 1983, Journal of 

789: Physics G Nuclear Physics, 9, 1139 

790: 

791: \bibitem{Giller2000} Giller, M., \& 

792: Zieli{\'n}ska, M.\ 2000, Nuclear Physics A, 663, 852 

793: 

794: \bibitem{Kim1989} Kim, K.-T., Kronberg, 

795: P.~P., Giovannini, G., \& Venturi, T.\ 1989, Nature, 341, 720 

796: 

797: \bibitem{Lemoine2005} Lemoine, M.\ 2005, Phys.Rev.D, 71, 

798: 083007

799: 

800: \bibitem{Loeb2002} Loeb, A., \& Waxman, 

801: E.\ 2002, astro-ph/0205272 

802: 

803: \bibitem{Lubowich1998} Lubowich, D.~A., 

804: Turner, B.~E., \& Hobbs, L.~M.\ 1998, ApJ, 508, 729 

805: 

806: \bibitem{Melia2001} Melia, F., \& Falcke, 

807: H.\ 2001, ARA\&A, 39, 309

808: 

809: \bibitem{Milgrom1996} Milgrom, M., \& Usov, 

810: V.\ 1996, Astropart.Phys., 4, 365 

811: 

812: \bibitem{Nayakshin2005} Nayakshin, S., \& 

813: Cuadra, J.\ 2005, A\&A, 437, 437

814: 

815: \bibitem{Parizot2004} Parizot, E.\ 2004, Nuclear 

816: Physics B Proceedings Supplements, 136, 169

817: 

818: \bibitem{Parizot2007} 

819: Parizot, E., \& for the Auger Collaboration 2007, ArXiv e-prints, 709, 

820: arXiv:0709.2500

821: 

822: \bibitem{Sanders1981} Sanders, R.\ 1981, Nature, 294, 427 

823: 

824: \bibitem{Sofue2000} Sofue, Y.\ 2000, ApJ, 540, 224 

825: 

826: \bibitem{VanElewyck2007} Van Elewyck, V., \& for the Auger 

827: Collaboration 2007, ArXiv e-prints, 709, arXiv:0709.1422 

828: 

829: \bibitem{Veilleux2005} Veilleux, S., Cecil, 

830: G., \& Bland-Hawthorn, J.\ 2005, ARA\&A, 43, 769 

831: 

832: \bibitem{Waxman1995} Waxman, E.\ 1995, ApJL, 452, 

833: L1 

834: 

835: \bibitem{Widrow2002} Widrow, L.~M.\ 2002, Reviews of 

836: Modern Physics, 74, 775 

837: 

838: \bibitem{Zatsepin1966} Zatsepin, G.~T., 

839: \& Kuz'min, V.~A.\ 1966, Journal of Experimental and Theoretical Physics 

840: Letters, 4, 78 

841: 

842: \end{thebibliography}

843: 

844: \end{document}

845: 

846: \section{Discarded Text}

847: 

848: 

849: We are particularly interested to answer the following question: what is the maximum

850: lineal distance, as a function of detection energy,

851: that a cosmic ray proton's source might be located away from us? Two effects limit the propagation distance of charged CRs: 

852: (i) at energies above the so-called GZK cut-off, pion photo production on the cosmic microwave background provides an

853: efficient energy-loss mechanism for propagating CR protons. 

854: At somewhat lower energies, CR protons lose energy in Beth-Heitler pair-production on this same background light field; 

855: (ii) in propagating through extragalactic magnetic fields, charged cosmic rays scattering on magneitc field inhomogeneities,

856: on sufficiently large scales such motion can be described as a diffusive process where the distance from a point source

857: that the average CR might travel only increases as the square-root of the time elapsed since emissions. Given this slow growth,

858: and the large distances involved, the finite age of the universe (or the time since some emission event) presents 

859: a real limit to the distance to detectable sources  -- a {\it magnetic horizon} in other words. This magnetic horizon is

860: energy-dependent -- becoming increasingly close with decreasing energy -- because the average scattering length of particles

861: traversing extragalactic {\bf B} fields will decrease as a function of energy.

862: 

863: \section{Additional Text}

864: 

865: from Astropart.Phys. 9 (1998) 221-225

866: The angular deviation of ultra high energy cosmic rays in

867: intergalactic magnetic fields

868: R.W. Clay *, S. Cook, B.R. Dawson, A.G.K. Smith, R. Lampard

869: 

870: The magnetic field is assumed to be purely turbulent

871: with a Kolmogorov scale length spectrum. That

872: is, the energy in the magnetic field which is associated

873: with a length scale Li varies as Lf’*. The energy

874: is thus weighted towards the larger turbulence scales.

875: 

876: source SPIN is $\sim$ 2.7. 

877: This is consistent with the findings of other works 

878: (e.g., \cite{DeMarco2003}) which have considered proton-only sources.

879: 

880: ...an interesting effect here is the flattening of input spectra by the 

881: combined action of the assumed exGal B fields 

882: and the bursting nature of the assumed sources...in contrast to the situation one 

883: would have with continuous injection by sources,

884: in our model, the flux at energy $E$ from any particular source will be dominated 

885: by the most recent outburst whose emitted CR's 

886: have had time to diffuse to the position of the Earth. Because diffusion is an 

887: energy-dependent process, however, at sufficiently

888: lower energies the CRs from this outburst will not have had time to reach the Earth

889: 

890: MW -- last outburst must have been at...

891: 

892: In an outbursting model,

893: the steepness of the spectrum below the ankle -- assumed due to the local source --

894: is determined by the injection spectrum and the relative scale of the time since the outburst

895: with respect to the time that, for a given field strength, CRs take to diffuse outwards so that the

896: local density becomes equal to the universal one. 

897: As expected a more recent outburst results in a flatter spectrum.

898: