1: %% Modified 2005 December 5
2: %%
3: %% This is a sample manuscript marked up using the
4: %% AASTeX v5.x LaTeX 2e macros.
5:
6: %% The first piece of markup in an AASTeX v5.x document
7: %% is the \documentclass command. LaTeX will ignore
8: %% any data that comes before this command.
9:
10: %% The command below calls the preprint style
11: %% which will produce a one-column, single-spaced document.
12: %% Examples of commands for other substyles follow. Use
13: %% whichever is most appropriate for your purposes.
14: %%
15: \documentclass[12pt,preprint]{aastex}
16: %% \documentclass{emulateapj}
17:
18: %% manuscript produces a one-column, double-spaced document:
19:
20: %%\documentclass[manuscript2]{aastex}
21:
22: %% preprint2 produces a double-column, single-spaced document:
23:
24: %%\documentclass[preprint2]{aastex}
25:
26: %% Sometimes a paper's abstract is too long to fit on the
27: %% title page in preprint2 mode. When that is the case,
28: %% use the longabstract style option.
29:
30: %% \documentclass[preprint2,longabstract]{aastex}
31:
32: %% If you want to create your own macros, you can do so
33: %% using \newcommand. Your macros should appear before
34: %% the \begin{document} command.
35: %%
36: %% If you are submitting to a journal that translates manuscripts
37: %% into SGML, you need to follow certain guidelines when preparing
38: %% your macros. See the AASTeX v5.x Author Guide
39: %% for information.
40:
41: \newcommand{\vdag}{(v)^\dagger}
42: \newcommand{\HESS}{H.E.S.S.}
43: \newcommand{\gr}{$\gamma$-ray}
44: \newcommand{\grs}{$\gamma$-rays}
45: \newcommand{\vhe}{V\textsc{HE}}
46: \newcommand{\dg}{\ensuremath{^\circ}}
47: \newcommand{\vitae}{vit\ae{}}
48: \newcommand{\apriori}{\emph{a~priori}}
49: \newcommand{\aposteriori}{\emph{a~posteriori}}
50: \newcommand{\naive}{na\"{\i}ve}
51:
52: %\renewcommand{\topfraction}{.85}
53: \renewcommand{\topfraction}{1}
54: %\renewcommand{\textfraction}{.15}
55: \renewcommand{\textfraction}{0}
56:
57: %% You can insert a short comment on the title page using the command below.
58: \slugcomment{Submitted to the Astrophysical Journal}
59:
60: %% If you wish, you may supply running head information, although
61: %% this information may be modified by the editorial offices.
62: %% The left head contains a list of authors,
63: %% usually a maximum of three (otherwise use et al.). The right
64: %% head is a modified title of up to roughly 44 characters.
65: %% Running heads will not print in the manuscript style.
66:
67: \shorttitle{The GeV-TeV Connection in Galactic $\gamma$-ray sources}
68: \shortauthors{S.~Funk et al.}
69:
70: %% This is the end of the preamble. Indicate the beginning of the
71: %% paper itself with \begin{document}.
72:
73: \begin{document}
74:
75: %% LaTeX will automatically break titles if they run longer than
76: %% one line. However, you may use \\ to force a line break if
77: %% you desire.
78:
79: \title{The GeV-TeV Connection in Galactic $\gamma$-ray sources}
80:
81: %% Use \author, \affil, and the \and command to format
82: %% author and affiliation information.
83: %% Note that \email has replaced the old \authoremail command
84: %% from AASTeX v4.0. You can use \email to mark an email address
85: %% anywhere in the paper, not just in the front matter.
86: %% As in the title, use \\ to force line breaks.
87:
88: \newcommand{\myemail}{funk@slac.stanford.edu}
89:
90: \author{S.~Funk\altaffilmark{1},
91: O.~Reimer\altaffilmark{2},
92: D.~F.~Torres\altaffilmark{3},
93: J.~A.~Hinton\altaffilmark{4}
94: }
95: \altaffiltext{1}{Kavli Institute for Particle Astrophysics and
96: Cosmology (KIPAC), SLAC, CA 94025, USA. funk@slac.stanford.edu}
97: \altaffiltext{2}{Stanford University, W.~W. Hansen Experimental
98: Physics Lab (HEPL) and KIPAC, Stanford, CA
99: 94305-4085, USA. olr@stanford.edu}
100: \altaffiltext{3}{ICREA \& Institut de Ciencies de l'Espai (IEEC-CSIC)
101: Campus UAB, Fac. de Ciencies, Torre C5, parell, 2a planta, 08193
102: Barcelona, Spain. dtorres@aliga.ieec.uab.es}
103: \altaffiltext{4}{School of Physics and Astronomy, University of Leeds,
104: Leeds LS2 9JT, UK. jah@ast.leeds.ac.uk}
105:
106: \begin{abstract}
107: Recent observations with atmospheric Cherenkov telescope systems such
108: as \HESS\ and MAGIC have revealed a large number of new sources of
109: very-high-energy (VHE) $\gamma$-rays from 100~GeV -- 100~TeV, mostly
110: concentrated along the Galactic plane. At lower energies (100 MeV --
111: 10 GeV) the satellite-based instrument EGRET revealed a population of
112: sources clustering along the Galactic Plane. Given their adjacent
113: energy bands a systematic correlation study between the two source
114: catalogues seems appropriate. Here, the populations of Galactic
115: sources in both energy domains are characterised on observational as
116: well as on phenomenological grounds. Surprisingly few common sources
117: are found in terms of positional coincidence and spectral
118: consistency. These common sources and their potential counterparts and
119: emission mechanisms will be discussed in detail. In cases of detection
120: only in one energy band, for the first time consistent upper limits in
121: the other energy band have been derived. The EGRET upper limits are
122: rather unconstraining due to the sensitivity mismatch to current VHE
123: instruments. The VHE upper limits put strong constraints on simple
124: power-law extrapolation of several of the EGRET spectra and thus
125: strongly suggest cutoffs in the unexplored energy range from 10~GeV --
126: 100~GeV. Physical reasons for the existence of cutoffs and for
127: differences in the source population at GeV and TeV energies will be
128: discussed. Finally, predictions will be derived for common GeV--TeV
129: sources for the upcoming GLAST mission bridging for the first time the
130: energy gap between current GeV and TeV instruments.
131: \end{abstract}
132:
133: \keywords{gamma rays: observations; Galaxy: general; (ISM:) supernova remnants}
134:
135: \section{Introduction}
136: In recent years the knowledge of the Galactic VHE $\gamma$-ray sky
137: above 100~GeV has been greatly improved through the detection and
138: subsequent study of many sources, mostly by means of ground-based
139: Imaging Atmospheric Cherenkov telescope systems such as the High
140: Energy Stereoscopic System (\HESS) or the Major Atmospheric Gamma-ray
141: Imaging Cherenkov Observatory (MAGIC). Currently known Galactic VHE
142: $\gamma$-ray emitters include shell-type Supernova remnants
143: (SNRs)~\citep{HESS1713_II, HESS1713_III, HESSVelaJr}, Pulsar Wind
144: Nebulae (PWNe)~\citep{HESSMSH, HESS1825II, HESSKooka}, $\gamma$-ray
145: binaries~\citep{HESSLS5039II, MAGICLSI}, Molecular
146: clouds~\citep{HESSGCDiffuse} and possibly also clusters of massive
147: stars~\citep{HESSWesterlund}. These various source classes were
148: discovered both in pointed observations using \HESS\ and MAGIC as well
149: as in a systematic survey of the inner Galaxy performed with the
150: \HESS\ instrument. The highest energy photons detected from these
151: source classes reach $\sim 100$~TeV~\citep{HESS1713_III}, currently
152: representing the end of the observable electromagnetic spectrum for
153: astrophysical objects. It is natural to investigate the relationship
154: of these TeV sources to sources at lower energies as will be done in
155: this work focusing on Galactic sources. The closest energy band for
156: which data exist is that studied by the Energetic Gamma Ray Experiment
157: Telescope (EGRET) aboard the Compton Gamma-Ray Observatory with an
158: energetic coverage from 100~MeV -- 10~GeV~\citep{EGRET}. The GeV sky
159: has a distinctively different overall appearance compared to TeV
160: energies. In particular focusing on our Galaxy, the most prominent
161: feature of the GeV sky is the dominant diffuse emission from cosmic
162: ray (CR) interactions in the Galaxy, while the TeV sky due to the
163: steeply falling energy spectrum of the diffuse component is dominated
164: by individual sources. However, several prominent $\gamma$-ray sources
165: are known to emit at both GeV and at TeV energies, the Crab Nebula
166: being the most prominent example~\citep{WhippleCrab,EGRETCrab,
167: HEGRACrab, HESSCrab, MAGICCrab}.
168:
169: In this paper the relationship between Galactic EGRET and VHE
170: $\gamma$-ray sources will be assessed in a systematic way. For cases
171: with a positional coincidence between a VHE and an EGRET source (in
172: the following called ``coincident sources'') all currently known
173: Galactic objects will be considered. For cases in which a source is
174: detected only in one band -- the ``non-coincident sources'' -- we
175: focus on the region covered by the \HESS\ Galactic plane survey (GPS)
176: during 2004 and 2005~\citep{HESSScan, HESSScanII} (Galactic longitude
177: $\pm 30^{\circ}$, Galactic latitude $\pm 3^{\circ}$) so that a
178: statistical assessment of the ``non-connection'' can be made. EGRET
179: was unable to perform detailed studies of the $\gamma$-ray sky above
180: 10~GeV, partly due to back-splash of secondary particles produced by
181: high-energy $\gamma$-rays causing a self-veto in the monolithic
182: anti-coincidence detector used to reject charged particles. The
183: upcoming Gamma Ray Large Area Space Telescope (GLAST) Large Area
184: Telescope (LAT) will not be strongly affected by this effect since the
185: anti-coincidence shield was designed in a segmented
186: fashion~\citep{GLASTACD}. Moreover, the effective area of the
187: GLAST-LAT will be roughly an order of magnitude larger then that of
188: EGRET. The GLAST-LAT mission will therefore for the first time fully
189: bridge the gap between the energy range of EGRET and current VHE
190: instruments. Part of the study presented here can be seen as
191: preparatory work for GLAST-LAT studies of sources in the largely
192: unexplored energy band between 10 and 100~GeV.
193:
194: \begin{figure}[ht]
195: \begin{center}
196: \includegraphics[width=0.9\textwidth]{f1.eps}
197: \end{center}
198: \caption{{\bf{Left:}} Integral sensitivities for current, past and
199: future $\gamma$-ray instruments (5-$\sigma$ sensitivity for $E>E_0$
200: multiplied with $E_0$ assuming a spectrum of $E^{-2}$). The solid
201: lines show the nominal instrument sensitivities (for a typical
202: observation time as specified below), the dashed curves show the
203: actual sensitivities for the Inner Galaxy as appropriate for this
204: work. INTEGRAL's (IBIS/ISGRI) sensitivity curve (solid green) shows
205: the sensitivity for an observation time of $10^5$s, a typical value
206: in the Inner Galaxy. The EGRET curves (brown) are shown for the
207: whole lifetime of the mission (periods 1--9) for the Galactic
208: anti-centre (solid) which received the largest exposure time and has
209: a lower level of diffuse $\gamma$-ray emission than the Inner Galaxy
210: and for the position of RX\,J1713.7--3946 (dashed), a typical
211: position in the Inner Galaxy dominated by diffuse $\gamma$-ray
212: background emission. The GLAST curves in red (taken from
213: http://www-glast.slac.stanford.edu/software/IS/glast\_lat\_performance.html)
214: show the 1-year sky-survey sensitivity for the Galactic North pole
215: -- again a position with low diffuse emission (solid), and for the
216: position of RX\,J1713.7--3946 (dashed). The H.E.S.S.\ curves (blue)
217: are shown for a 50-hour pointed observation of a point-like source
218: (solid) and for a 5-hour observation of a somewhat extended source
219: as is typical for the Galactic Plane survey (angular cut of
220: $\sqrt{0.05}$). The MAGIC curve (light blue) represents a 50-hour
221: observation of a point-source. {\bf{Right:}} Energy-dependence of
222: the angular resolution for selected $\gamma$-ray instruments
223: expressed by the 68\%-containment radius of the point-spread
224: function (PSF). As can be seen, the angular resolution of GLAST
225: becomes comparable with current VHE instruments at high energies,
226: whilst at the lower energy end GLAST and EGRET have comparable
227: resolutions.}\label{fig::Sensitivity}
228: \end{figure}
229:
230: From the 2004 and 2005 \HESS\ GPS 22 VHE $\gamma$-ray sources have
231: been reported in the Inner Galaxy. The third EGRET
232: catalogue~\citep{EGRET} represents the companion to the VHE source
233: catalogue above an energy threshold of 100 MeV (with peak sensitivity
234: between 150 and 400 MeV, depending on the $\gamma$-ray source
235: spectrum). It lists 271 sources, 17 of which are located within the
236: \HESS\ GPS region. Whilst the EGRET range currently represents the
237: nearest energy band to VHE $\gamma$-rays, for the very few EGRET
238: sources detected all the way up to $\sim 10$~GeV, there is still an
239: unexplored energy band of roughly one decade before the VHE
240: $\gamma$-ray energy range begins at $\sim 100$~GeV (it should be noted
241: that EGRET does have some sensitivity beyond 10~GeV:
242: \citet{EGRET10GeV} reported the detection of $\sim$ 1500 photons above
243: that energy with 187 of these photons being found within $1^{\circ}$
244: of a source listed in the third EGRET catalogue). Comparing the
245: instrumental parameters of VHE instruments and EGRET there is a clear
246: mismatch both in angular resolution and in sensitivity as can be seen
247: in Figure~\ref{fig::Sensitivity}. In a $\sim 5$ hour observation (as a
248: typical value in the GPS region) \HESS\ is a factor of $\sim 50-80$
249: more sensitive (in terms of energy flux $E^2 dN/dE$) than EGRET above
250: 1~GeV in the Galactic Plane for the exposure accumulated between 1991
251: and 1995 (corresponding to the third EGRET catalogue). Assuming a
252: similar energy flux output in the two different bands this mismatch
253: implies at first sight that \HESS\ sources are not likely to be
254: visible in the EGRET data set. Conversely (again under the assumption
255: of equal energy flux output), VHE $\gamma$-ray instruments should be
256: able to detect the majority of the EGRET sources, as has been
257: suggested in the past. Figure~\ref{fig::EnergyFluxDistribution}
258: compares the energy fluxes $\nu F \nu$ for EGRET sources and \HESS\
259: sources in the inner Galaxy. Clearly, the EGRET sources do not reach
260: down as low in energy flux as the \HESS\ sources, a picture that will
261: change once the GLAST-LAT is in orbit as depicted by the GLAST-LAT
262: sensitivity (dashed line). In reality the \naive\ expectation of equal
263: energy flux output in the GeV and TeV band can easily be wrong in
264: Galactic $\gamma$-ray sources for various reasons: EGRET sources may
265: not emit comparable energy fluxes in the VHE $\gamma$-ray band but
266: rather exhibit cut-offs or spectral breaks in the energy band between
267: EGRET and \HESS\ \citep[this is certainly the case for pulsed
268: magnetospheric emission from pulsars, see for example][]{HESSPulsar}.
269: Furthermore, \HESS-like instruments are typically only sensitive to
270: emission on scales smaller than $\sim 1^{\circ}$. If any of the EGRET
271: sources are extended beyond $1^{\circ}$ without significant
272: sub-structure on smaller scales (not precluded given the poor
273: angular resolution of EGRET), current Imaging Cherenkov instruments may not be
274: able to detect them since these sources would completely fill the
275: field of view (FoV) and be removed by typical background subtraction
276: methods ~\citep[see for example][]{BergeBackground}. Given the
277: upcoming launch of GLAST and the recent \HESS\ survey it seems
278: timely to study the relationship between GeV and TeV emitting sources
279: in more detail.
280:
281: \begin{figure}[ht]
282: \begin{center}
283: \includegraphics[width=0.6\textwidth]{f2.eps}
284: \end{center}
285: \caption{Distribution of integrated energy flux $\nu F \nu$ for
286: sources in the Inner Galaxy discussed here. For EGRET the energy
287: flux between 1~GeV and 10~GeV, for the \HESS\ sources, the energy
288: flux between 1~TeV and 10~TeV is shown. Also shown is the
289: sensitivity prediction for the GLAST-LAT for a typical location in
290: the Inner Galaxy (l=10, b=0).}\label{fig::EnergyFluxDistribution}
291: \end{figure}
292:
293: Section~\ref{sec::analysis} describes the data and analysis methods
294: used in this study, section~\ref{sec::connect} describes the sources
295: detected in both energy bands, and section~\ref{sec::nonconnect}
296: focuses on sources detected in only one of the two energy regimes. In
297: section~\ref{sec::interpretation} astrophysical implications of the
298: study are discussed.
299:
300: \section{Analysis methods}
301: \label{sec::analysis}
302:
303: For the sources discussed in this study locations and source spectra
304: in the EGRET band~\citep{EGRET} and in the VHE $\gamma$-ray band are
305: required. For the inner Galaxy, dedicated upper limits at the specific
306: position of the $\gamma$-ray sources in the respective other band were
307: determined. For the EGRET data these upper limits (at 1~GeV) were
308: derived at the nominal positions of the \HESS\ sources based on a
309: reanalysis of the data used for the production of the third EGRET
310: catalogue, applying the standard EGRET likelihood fitting
311: technique~\citep{Mattox96}. For the \HESS\ data, 2~$\sigma$ upper
312: limits at the nominal position of each EGRET source were
313: estimated. This was done by scaling the flux corresponding to the
314: \HESS-point-source sensitivity in 25 hours (1\% of the Crab) by the
315: square-root of the ratio of 25 hours to the published exposure time at
316: the position of the EGRET source \citep[taken from][]{HESSScanII}.
317:
318: \subsection{Quantifying Positional Coincidence}
319:
320: \begin{figure}[ht]
321: \begin{center}
322: \includegraphics[width=0.9\textwidth]{f3.eps}
323: \end{center}
324: \caption{Map of the \HESS\ GPS region taken from \citet{HESSScanII}.
325: Published \HESS\ sources are marked as squares. EGRET sources are
326: shown with their 95\% positional confidence contours from the
327: 3EG-catalogue. The red (orange) contours and labels denote those 3EG
328: sources for which a \HESS\ source centroid is located within the
329: 95\% (99\%) confidence contour, the blue contours denote the 95\%
330: PUCs for the EGRET sources for which no VHE emission is
331: detected.}\label{fig::Map}
332: \end{figure}
333:
334: Figure~\ref{fig::Map} shows all \HESS\ and all EGRET sources within
335: the HESS GPS region. One property of EGRET and VHE $\gamma$-ray
336: sources becomes immediately apparent: only a minor fraction of the
337: \HESS\ sources coincide within the considerably larger location
338: uncertainty contours of EGRET GeV sources. Given the rather poor
339: angular resolution of EGRET (68\% containment radius of the PSF:
340: $1.5^{\circ}$ at 1~GeV) coupled with typically rather limited photon
341: statistic any systematic assessment of positional matches between
342: EGRET and \HESS\ sources is dominated by the localisation error on the
343: EGRET source position. The likelihood source position uncertainty
344: contour (PUC) as given in~\citet{EGRET} have been used to check for
345: VHE $\gamma$-ray sources within these regions on the sky. While most
346: of the VHE sources are extended, their extension is rather small on
347: the scale of the EGRET positional uncertainty and therefore a source
348: is classified as ``coincident'' if the centre of gravity of the VHE
349: emission is within the EGRET likelihood PUC. For large sources such as
350: e.g.\ the SNR RX\,J1713.7--3946 (HESS\,J1713--395)
351: this approach is clearly an oversimplification, albeit it is the one
352: used at this stage of this study.
353:
354: \begin{table}[ht]
355: \begin{center}
356: \begin{tabular}{c||c|c|c}
357: \hline
358: EGRET & \multicolumn{3}{c}{VHE $\gamma$-ray} \\
359: Source & \multicolumn{3}{c}{Source} \\
360: & Within 68\% & Within 95\% & Within 99\% \\
361: & PUC & PUC & PUC \\
362: \hline
363: & \multicolumn{3}{c}{Within the \HESS\ GPS} \\
364: \hline
365: 3EG\,J1639--4702 & & HESS\,J1640--465 & \\
366: 3EG\,J1744--3011 & & HESS\,J1745--303 & \\
367: 3EG\,J1800--2338 & None & & HESS\,J1800--233 \\
368: 3EG\,J1824--1514 & & & HESS\,J1826--148 \\
369: 3EG\,J1826--1302 & & & HESS\,J1825--137 \\
370: \hline
371: & \multicolumn{3}{c}{Outside the \HESS\ GPS} \\
372: \hline
373: 3EG\,J0241+6103 & & & MAGIC\,J0240+613 \\
374: 3EG\,J0617+2238 & & None & MAGIC\,J0616+225 \\
375: 3EG\,J0634+0521 & & & HESS\,J0632+058 \\
376: 3EG\,J1420--6038 & HESS\,J1420--607 & & \\
377: \hline
378: \end{tabular}
379: \caption{Positionally coincident EGRET and \HESS\ sources within
380: our Galaxy for three confidence levels (68\%, 95\%, and 99\%) of
381: the positional uncertainty of the EGRET source.}
382: \label{tab::position}
383: \end{center}
384: \end{table}
385:
386: The number of spatially coincident sources depends on the EGRET PUC
387: chosen in the investigation. For the \HESS\ GPS-region, no VHE
388: $\gamma$-ray source is located within the 68\% positional confidence
389: contour of any EGRET source. Relaxing the coincidence criterion, two
390: VHE $\gamma$-ray sources are found within the 95\%-confidence contour
391: of EGRET source positions (shown in red in Figure~\ref{fig::Map}) and
392: an additional three VHE $\gamma$-ray sources are located within the
393: 99\%-confidence contours (shown in orange in
394: Figure~\ref{fig::Map}). Outside the \HESS\ GPS-region, no systematic
395: statistical assessment of the non-coincident sources is possible due
396: to the highly non-uniform exposure of the observations with the
397: limited-FoV VHE instruments. Nevertheless, is it relevant to note that
398: four additional coincident sources are found outside the \HESS\
399: GPS-region within the Galactic plane (defined here as as latitude
400: range of $\pm 3^{\circ}$): HESS\,J1420--608~\citep{HESSKooka} in the
401: Kookaburra region is located within the 68\% confidence contour of
402: 3EG\,J1420--6038. The other three coincident Galactic sources are
403: located within the 99\% confidence contours of EGRET sources. The Crab
404: Nebula is not listed in Table~\ref{tab::ConnectCases} although it has
405: been detected by EGRET~\citep{EGRETCrab} as well as by all major VHE
406: $\gamma$-ray instruments~\citep{WhippleCrab, MilagroCrab, HEGRACrab,
407: HESSCrab, MAGICCrab}. The reason for this is that in the 3EG catalogue
408: only the position of the Crab pulsar is given, whereas the PUC of the
409: off-pulse emission (i.e. the Nebula emission) has not been published
410: thus far.
411:
412: These coincident cases are discussed further in
413: section~\ref{sec::connect}. Table~\ref{tab::position} summarises the
414: VHE $\gamma$-ray sources located within EGRET confidence contours
415: inside and outside the \HESS\ GPS region. Within the Galactic Plane
416: survey region 17.1 square degrees (corresponding to 3\% of the total
417: GPS region) are covered by the EGRET 95\%-confidence
418: contours. Randomising the distribution of \HESS\ sources in the region
419: ~\citep[flat in longitude; Gaussian shape in latitude with a mean of
420: --0.2$^{\circ}$ and a width of 0.34$^{\circ}$ as shown
421: in][]{HESSScanII} the probability for spatial coincidence between
422: these two populations can be established. For the 95\% confidence
423: contours $\sim 1.4$ coincidences between the \HESS-source population
424: and the EGRET sources is expected. The probability of detecting 2 or
425: more sources when 1.4 sources are expected by chance is 40\%, i.e. the
426: positional coincidences could well be expected even if the two
427: population of sources are not related. Considering the (smaller) 68\%
428: confidence contours $\sim 0.5$ chance coincidences are expected; the
429: probability for no coincidence when 0.5 are expected is 60\%. For the
430: (larger) 99\% confidence contours the picture is similar: $\sim 2.5$
431: chance coincidences are expected; the probability of finding 5
432: coincidences when 2.5 are expected is $\sim 9\%$. Summarising these
433: numbers, it is well possible within the statistics and properties of
434: the two source classes that all the positional coincidences between
435: \HESS\ and EGRET sources are chance coincidences. The numbers derived
436: here do not strongly suggest common sources between the GeV and the
437: TeV band, although it is (obviously) not precluded that the
438: coincidences found point to real physical associations.
439:
440: \subsection{Determining Spectral Compatibility}
441: \label{sec::spectralmatch}
442: Besides the test for positional coincidence a test of spectral
443: compatibility, based on the simple assumption of a spectral
444: extrapolation by a single power-law between the EGRET and the \HESS\
445: ranges has been performed. To assess the spectral match the quantity
446: $\sigma_{\mathrm{comb}}$ has been defined in the following way:
447: \begin{equation}
448: \label{eq::extra}
449: \sigma_{\mathrm{comb}} = \sqrt{\sigma^2_{\mathrm{3EG}}
450: + \sigma^2_{\mathrm{H.E.S.S.}}}
451: \end{equation}
452: To determine $\sigma_{\mathrm{3EG}}$, the spectral index of the EGRET
453: source has been varied (around the pivot point of the EGRET best fit)
454: until the extrapolation to 1~TeV matches the \HESS\ flux at that
455: energy. This pivot point of the EGRET best fit is the energy at which
456: the error on the index becomes independent from the error on the
457: normalisation. This resulting index is called
458: $\Gamma_{\mathrm{match}}$ and
459: \begin{equation}
460: \label{eq::extra2}
461: \sigma_{\mathrm{3EG}} = (\Gamma_{\mathrm{match}} -
462: \Gamma_{\mathrm{3EG}})/ (\Delta \Gamma_{\mathrm{3EG}})
463: \end{equation}
464: \citep[where $\Gamma_{\mathrm{3EG}}$ and $\Delta
465: \Gamma_{\mathrm{3EG}}$ is the EGRET index and its error taken
466: from][]{EGRET}. Consequently, $\sigma_{\mathrm{3EG}}$ is a quantity
467: that describes by how much the EGRET index has to be altered (with
468: respect to the error on this index) to match the \HESS\ spectrum at
469: 1~TeV. In the same way $\sigma_{\mathrm{H.E.S.S.}}$, is determined by
470: changing the \HESS\ spectral index until the flux matches the EGRET
471: flux at 1~GeV (to avoid biases due to spectral cut-offs at the high
472: end of the \HESS\ energy range the spectra were fitted only below
473: 1~TeV in cases with clear spectral curvature). The two quantities
474: $\sigma_{\mathrm{3EG}}$ and $\sigma_{\mathrm{H.E.S.S.}}$ are finally
475: added in quadrature to yield $\sigma_{\mathrm{comb}}$, describing how
476: well the two spectra can be extended into each other by a linear
477: extrapolation. It should be noted that for the procedure described
478: here, only the statistical (not the systematic) errors on the spectral
479: indices are taken into account. For cases with a source detection only
480: in one band, the same procedure can be applied using the upper limit
481: in the other band (with the obvious difference that only the
482: extrapolation from the source spectrum onto the upper limit can be
483: performed, not the other way around). For cases in which the power-law
484: extrapolation with the nominal source photon index turns out to be
485: lower -- and therefore non-constraining -- to the upper limit the
486: corresponding measure $\sigma_{\mathrm{3EG}}$ or
487: $\sigma_{\mathrm{H.E.S.S.}}$ is set to zero (i.e. the spectra are
488: compatible). In several (but not the majority of) cases the EGRET
489: spectrum can be preferentially fit by a higher order spectral shape
490: (e.g.\ an exponential cutoff or a broken power-law) as will be
491: discussed in section~\ref{sec::nonconnect}.
492:
493: \section{VHE $\gamma$-ray sources with EGRET counterparts}
494: \label{sec::connect}
495:
496: Only a few coincident sources between the GeV and the TeV band have
497: been reported so far. The VHE $\gamma$-ray sources that positionally
498: coincide with EGRET sources are summarised in
499: Table~\ref{tab::position}. Whilst the positional coincidences between
500: EGRET and VHE $\gamma$-ray sources might all be chance coincidences as
501: shown in the previous section, in the following all positional
502: coincidences within the 99\% EGRET PUCs will be considered. Some of
503: the properties of the sources and their respective source classes will
504: be discussed along with an investigation on their spectral
505: compatibility as introduced in the previous section.
506:
507: \subsection{Source Classes}
508:
509: For EGRET sources in the Galactic plane, the only firm identifications
510: with sources at other wavelengths are for pulsars, based on matching
511: radio or X-ray periodicity~\citep{Thomp94}. For many of the remaining
512: Galactic EGRET sources, counterparts have been suggested, but the
513: angular resolution of the instrument and the strong diffuse
514: $\gamma$-ray background in the Galactic plane prevented unambiguous
515: identifications. In VHE $\gamma$-rays, several source classes have
516: been firmly identified as has been discussed e.g.\ in
517: ~\citet{FunkBarcelona}, based on matching morphology, positional
518: coincidence or periodicity. However, the majority of Galactic VHE
519: $\gamma$-ray sources also remain unidentified.
520: Table~\ref{tab::ConnectCases} summarises potential counterparts of VHE
521: sources in the coincident cases. While some of these identifications
522: are rather solid (as e.g.\ in the case of the $\gamma$-ray binaries
523: LS\,5039~\citep{HESSLS5039II} and LSI\,+61\,303~\citep{MAGICLSI}), in
524: most of the other cases the identification of (even) the VHE
525: $\gamma$-ray sources (with relatively small PUC of $\sim 1\arcmin$)
526: lack any evidence of association beyond positional coincidence. In
527: the cases where a firm identification exists, the VHE $\gamma$-ray
528: source can be used to shed light on the nature of GeV source, assuming
529: a physical relationship as shown for the Kookaburra
530: region~\citep{Reimer}. Such studies demonstrate that observations with
531: VHE $\gamma$-ray instruments can provide templates necessary to pin
532: down the nature of unidentified EGRET $\gamma$-ray sources with
533: suggestive but unproven counterparts. With the upcoming advent of the
534: GLAST-LAT instrument this approach will become very useful for
535: associating the GeV emission as measured by a large-aperture
536: space-based $\gamma$-ray instrument with narrow FoV but superior
537: spatial resolution observations of ground-based VHE $\gamma$-ray
538: instruments. Provided that the physical associations discussed in this
539: section and shown in Table~\ref{tab::ConnectCases} are confirmed (as
540: e.g., through more sensitive measurements with GLAST-LAT), three
541: long-suspected classes of Galactic GeV sources (SNRs, PWNe and Binary
542: systems) could finally be conclusively established. In the following
543: these different source classes will be briefly discussed in the
544: context of this study.
545:
546: \begin{table}[ht]
547: \begin{center}
548: \begin{tabular}{l c | l}
549: \hline EGRET source & VHE $\gamma$-ray source & Potential
550: Counterpart\\ \hline \multicolumn{3}{c}{Within the \HESS\ GPS}
551: \\ \hline 3EG\,J1639--4702 & HESS\,J1640--465 & G338.3--0.0
552: (SNR/PWN)\\ 3EG\,J1744--3011 & HESS\,J1745--303 & \\
553: 3EG\,J1800--2338 & HESS\,J1801--233 & W28 (SNR)\\
554: 3EG\,J1826--1302 & HESS\,J1825--137 & G18.0--0.7 (PWN)\\
555: 3EG\,J1824--1514 & HESS\,J1826--148 & LS\,5039 (Binary)\\ \hline
556: \multicolumn{3}{c}{Outside the \HESS\ GPS} \\ \hline
557: 3EG\,J0241+6103 & MAGIC\,J0240+613 & LSI\,+61\,303 (Binary)\\
558: 3EG\,J0617+2238 & MAGIC\,J0616+225 & IC443 (SNR/PWN)\\
559: 3EG\,J0634+0521 & HESS\,J0632+058 & Monoceros \\
560: 3EG\,J1420--6038 & HESS\,J1420--607 & Kookaburra (PWN)\\ \hline
561: \end{tabular}
562: \caption{Coincident sources and potential counterparts to the VHE
563: $\gamma$-ray sources (and hence also to the associated EGRET
564: sources). The counterparts are classified into the source classes
565: shell-type SNRs, PWNe and $\gamma$-ray binaries.}
566: \label{tab::ConnectCases}
567: \end{center}
568: \end{table}
569:
570:
571: \subsubsection{Pulsar wind nebulae}
572: PWNe are currently the most abundant class amongst the identified
573: Galactic VHE $\gamma$-ray sources, it is not therefore surprising that
574: PWN are found as potential counterparts to the coincident sources. The
575: first example for a coincident PWN is HESS\,J1825--137 -- located
576: within the 99\% confidence region of 3EG\,1826--1302. This
577: source~\citep{HESS1825II} is currently the best-known example for an
578: offset $\gamma$-ray PWN and as such represents a prototype for a new
579: class of $\gamma$-ray sources. HESS\,J1825--137 shows a steepening of
580: the energy spectrum with increasing distance from the central
581: pulsar. This property, as well as the observed difference in size
582: between the VHE $\gamma$-ray emitting region and the X-ray PWN
583: associated with the pulsar PSR\,B1823--13~\citep{Gaensler} can be
584: naturally explained by different cooling timescales for the radiating
585: electrons of different energies. In this regard it will be important
586: to study this region with the GLAST-LAT in the GeV band to confirm (or
587: refute) this picture. Another example of VHE $\gamma$-ray PWN is
588: HESS\,J1420--607, one of two likely PWN in the previously discussed
589: Kookaburra region, which is located within the 68\% confidence region
590: of 3EG\,J1420--6038. The Crab Nebula is not listed in
591: Table~\ref{tab::ConnectCases} although being a prominent GeV and TeV
592: source because no position for the Crab off-pulse (nebular) emission
593: has been published for the GeV data. For previously unidentified
594: sources such as HESS\,J1640--465 (G338.3--0.0) an association with the
595: X-ray PWN~\citep{Funk1640} is suggestive but not firmly established at
596: this point.
597:
598: \subsubsection{Shell-type Supernova remnants}
599: Shell-type SNRs constitute another prominent class of VHE $\gamma$-ray
600: sources. However, the two most prominent VHE $\gamma$-ray shell-type
601: SNRs RX\,J1713.7--3946 and RX\,J0852.0--4622 (Vela Jr.) are not
602: prominent GeV emitters even though they are (up to now) the brightest
603: steady VHE $\gamma$-ray sources in the sky after the Crab Nebula. Also
604: Cas~A and RCW~86 have been reported as VHE $\gamma$-ray
605: sources~\citep{CasAHegra, MagicCasA, RCW86HESS} but have not been
606: detected by EGRET. \citet{Sturner, Esposito, Romero, DiegoSNR}
607: assessed the relationship between unidentified EGRET sources at low
608: Galactic latitude and SNRs and found a statistically significant
609: correlation between the two populations at the 4--5 $\sigma$ level,
610: were however not able to firmly and uniquely identify individual SNRs
611: as EGRET sources. The GLAST-LAT will shed more light on the GeV
612: emission in this source as well as in the whole population of Galactic
613: SNRs. By measuring shape and level of the high-energy
614: $\gamma$-ray emission the GLAST-LAT might allow for a distinction
615: between hadronic and leptonic emission models as discussed in
616: section~\ref{sec::interpretation}. Other potential shell-type SNR
617: counterparts related to this analysis are W\,28 (HESS\,J1801--233 and
618: 3EG\,J1800--2338), IC443/MAGIC\,J0616+225~\citep{MAGICIC443}, and the
619: Monoceros Loop SNR (HESS\,J0632+058 and
620: 3EG\,J0634+0521)~\citep{HESSMonoceros}, although in particular in the
621: latter case, the morphology of the VHE $\gamma$-ray source does not
622: lend support to an association with the SNR shell.
623:
624: \subsubsection{$\gamma$-ray binaries}
625: Three binary systems: PSR\,B1259--63/SS\,2883, LS\,5039 and
626: LSI\,+61\,303, have now been established as VHE $\gamma$-ray
627: sources~\citep{HESS1259, HESSLS5039II, MAGICLSI, VeritasLSI}. The
628: latter two of these objects have long been considered as likely
629: counterparts to EGRET sources~\citep{Kniffen_LSI, Tavani_LSI, EGRET,
630: Paredes_LS5039}, however, a definitive identification could not be
631: achieved in the GeV waveband so far. The VHE $\gamma$-ray emission is
632: undoubtedly related to the binary system (as e.g.\ in LS\,5039
633: established through the detection of characteristic periodicity,
634: matching the orbital period of the binary system), strengthening the
635: case that the GeV emission is also associated to these binaries.
636: Recently, the MAGIC collaboration presented evidence for VHE
637: $\gamma$-ray emission from the black-hole X-ray binary Cyg\,X-1,
638: during a flaring state in X-rays~\citep{MAGICCygX}. There is no
639: evidence so far for GeV emission from this object.
640:
641: \subsection{Spectral compatibility}
642:
643: As described in section~\ref{sec::spectralmatch}, a test for
644: compatibility between EGRET and \HESS\ energy spectra based on a
645: single power-law extrapolation has been performed, calculating for
646: each of the coincident cases in the \HESS\ GPS region a measure of
647: spectral mismatch: $\sigma_{\mathrm{comb}}$.
648: Figure~\ref{fig::SpectraConnect} shows the result of these
649: extrapolations. The values for the spectral compatibility parameter
650: $\sigma_{\mathrm{comb}}$ are rather small. The largest value,
651: potentially indicative of a spectral mismatch, is found for the case
652: of the $\gamma$-ray binary association LS\,5039 (3EG\,J1824--1514 and
653: HESS\,J1826--148). However, this value is completely dominated by the
654: small statistical error on the \HESS\ power-law fit below 1~TeV (error
655: on the photon index: $\Delta \Gamma_{\mathrm{stat}} = \pm
656: 0.04$). Taking a typical \HESS\ systematic error of $\Delta
657: \Gamma_{\mathrm{sys}} = \pm 0.2$ on the determination of the photon
658: index $\Gamma$ into account, the GeV and TeV energy spectra in this
659: source match well. Figure~\ref{fig::SpectraConnect} therefore suggests
660: that the energy spectra of sources that show a spatial association can
661: generally be rather well described by a single power-law description
662: across the entire energy range from 0.1~GeV to 1~TeV.
663:
664: \begin{figure}[ht]
665: \begin{center}
666: \noindent
667: \includegraphics[width=0.95\textwidth]{f4.eps}
668: \end{center}
669: \caption{Spectra for the positionally coincident EGRET and \HESS\
670: sources within the \HESS\ GPS region. Sources for which the
671: H.E.S.S.\ source is located within the 95\% confidence level are
672: shown in red, whereas those within the 99\% confidence contour (as
673: give in Table~\ref{tab::position}) are shown in orange. The EGRET
674: ``butterfly'' is determined from the 3EG catalogue~\citep{EGRET},
675: the \HESS\ spectral points are taken from the respective
676: publication. For HESS\,J1826--148 and HESS\,J1825--137, which have
677: significantly curved TeV spectra, only the spectral points below
678: 1~TeV have been fitted and used for the extrapolation. Large values
679: of $\sigma_{\mathrm{comb}}$ indicate mismatches between the spectra
680: at GeV and at TeV energies.}
681: \label{fig::SpectraConnect}
682: \end{figure}
683:
684: \begin{figure}[ht]
685: \begin{center}
686: \noindent
687: \includegraphics[width=0.7\textwidth]{f5.eps}
688: \end{center}
689: \caption{Distribution of $\sigma_{\mathrm{comb}}$. The red histogram
690: shows the distribution for the spectral consistency parameter
691: $\sigma_{\mathrm{comb}}$ of all possible combinations of EGRET
692: sources with \HESS\ sources within the GPS region. The
693: black histogram shows the same distribution for the 5 cases of
694: positional coincidences.}\label{fig::ChanceCoincidence}
695: \end{figure}
696:
697: To estimate the chance coincidence of a spectral consistency the
698: spectra of all 17 EGRET sources and of all 22 \HESS\ sources in the
699: \HESS\ GPS region have been interchanged and ``connected'' to each
700: other (i.e.\ the spectral compatibility of each \HESS\ source has been
701: determined for each EGRET source -- regardless of positional
702: coincidence). The resulting distribution of $\sigma_{\mathrm{comb}}$
703: can be interpreted as the probability density function for
704: $\sigma_{\mathrm{comb}}$ for randomly selected \HESS\ and EGRET
705: sources and is shown in Figure~\ref{fig::ChanceCoincidence} as a red
706: histogram. This distribution should be compared to the measured
707: distribution of $\sigma_{\mathrm{comb}}$ for positionally coincident
708: pairs (black histogram). Even though the distribution for the
709: scrambled sources shows a tail to large values of
710: $\sigma_{\mathrm{comb}}$, most random pairings result in values of
711: $\sigma_{\mathrm{comb}}<5$. A Kolmogorov test yields a probability of
712: 89\% that the two distributions are based on a common underlying
713: distribution. Thus a spectral compatibility based on a power-law
714: extrapolation of a typical (randomly picked) EGRET and a typical
715: \HESS\ source is expected to occur by chance even in the absence of a
716: physical association. This is perhaps not surprising, given that both
717: EGRET as and \HESS\ spectra have typical photon indices of $\sim 2.3$
718: and that \HESS\ measurements occur $\sim 4$ orders of magnitude higher
719: in energy with \HESS\ being 1--2 orders of magnitude more sensitive.
720:
721: \section{Inner Galaxy $\gamma$-ray sources detected in only one band}
722: \label{sec::nonconnect}
723:
724: In this section the remainder (and majority) of sources in the \HESS\
725: GPS region will be discussed. These are the sources which do not have
726: a counterpart in the neighbouring energy band. In
727: Section~\ref{sec::nonconnect::EGRET} EGRET sources without a VHE
728: $\gamma$-ray counterpart will be discussed,
729: section~\ref{sec::nonconnect::VHE} investigates VHE $\gamma$-ray
730: sources without an EGRET counterpart.
731:
732: \subsection{EGRET sources without a VHE $\gamma$-ray counterpart}
733: \label{sec::nonconnect::EGRET}
734:
735: \begin{table}[t]
736: \begin{center}
737: \begin{tabular}{c|c |r}
738: \hline
739: EGRET & H.E.S.S. & $\sigma_{\mathrm{comb}}$\\
740: Source & Upper Limit & \\
741: & ($10^{-12}$ ergs cm$^{-2}$ s$^{-1}$) & \\
742: \hline
743: 3EG\,J1655-4554 & 0.4 & 1.3\\
744: 3EG\,J1710-4439 & 1.5 & 16.3\\
745: 3EG\,J1714-3857 & 0.2 & 1.5\\
746: 3EG\,J1718-3313 & 1.0 & 0\\
747: 3EG\,J1734-3232 & 0.6 & 1.4\\
748: 3EG\,J1736-2908 & 0.3 & 3.5\\
749: 3EG\,J1746-2851 & 0.2 & 15.7\\
750: 3EG\,J1809-2328 & 0.5 & 6.4\\
751: 3EG\,J1812-1316 & 2.1 & 1.0\\
752: 3EG\,J1823-1314 & 0.4 & 0\\
753: 3EG\,J1837-0423 & 0.6 & 0\\
754: 3EG\,J1837-0606 & 0.4 & 5.5\\
755: \hline
756: \end{tabular}
757: \caption{EGRET sources without a VHE $\gamma$-ray counterpart in
758: the \HESS\ GPS region. The \HESS\ differential upper limits
759: (2~$\sigma$) at 1~TeV for a point-source analysis, are derived
760: from the \HESS\ 2004--2005 exposure at the nominal EGRET
761: position as described in the text (under the assumption of a
762: photon index of 2.6).}
763: \label{tab::nonconnect}
764: \end{center}
765: \end{table}
766:
767: Here those EGRET sources are addressed with a 99\%-confidence centroid
768: position region which does not contain a reported VHE $\gamma$-ray
769: source centroid. This sample consist of 12 EGRET detections, with E
770: $> 100$ MeV fluxes ranging between 0.4 and 3.1 $\times
771: 10^{-6}$\,cm$^{-2}$\,s$^{-1}$ and photon indices of the power-law fits
772: between $\sim$1.75 and 3.2. For these 12 EGRET sources, 2~$\sigma$
773: upper limits on the VHE flux at 1~TeV for the nominal EGRET position
774: were determined. This was done by scaling the \HESS\ sensitivity for a
775: 5~$\sigma$ point source detection (1\% of the Crab in 25 h under the
776: assumption of a photon index of 2.6) to the actual exposures as
777: published for the \HESS\ GPS region~\citep{HESSScanII}. As described
778: previously, the spectral compatibility parameter
779: $\sigma_{\mathrm{3EG}}$ was determined according to
780: Equation~\ref{eq::extra2}. For cases in which the EGRET extrapolation
781: with the nominal 3EG photon index undershoots the \HESS\ upper limit,
782: $\sigma_{\mathrm{3EG}}$ is set to zero. The resulting plots are shown
783: in Figure~\ref{fig::EGRETnonconnect}. For Gaussian errors
784: $\sigma_{\mathrm{3EG}}$ represents the probability that the true GeV
785: spectrum would pass through the HESS upper limit.
786:
787: \begin{figure}[ht]
788: \begin{center}
789: \noindent
790: \includegraphics[width=0.7\textwidth]{f6.eps}
791: \end{center}
792: \caption{SED of the EGRET source for which no VHE $\gamma$-ray
793: source was found within the 99\% confidence contour. Sources
794: marked with a square show $\gamma$-ray emission above 10~GeV in
795: the EGRET data as reported by~\citet{EGRET10GeV}, for sources
796: marked with a triangle the EGRET data are better described by
797: either a broken power-law or a power-law with an exponential
798: cutoff as shown in Figure~\ref{fig::OtherFits}.}
799: \label{fig::EGRETnonconnect}
800: \end{figure}
801:
802: In seven of the twelve cases the \HESS\ upper limit does not impose a
803: strong constraint on an extrapolation of the EGRET spectrum
804: ($\sigma_{\mathrm{3EG}} < 1.5$). For the remaining five sources a
805: \HESS\ detection would have been expected, based on a \naive\
806: power-law extrapolation. In particular extrapolations for four of the
807: EGRET sources exhibiting a hard energy spectrum (3EG\,J1710--4439,
808: 3EG\,J1746--2851, 3EG\,J1809--2328, and 3EG\,J1837--0606) are
809: incompatible with the \HESS\ upper limits at levels exceeding
810: $\sigma_{\mathrm{3EG}}>5$. For these cases the VHE $\gamma$-ray data
811: strongly suggest a spectral turnover (cutoff or break) well below the
812: \HESS\ range. Such behaviour is not surprising for some Galactic
813: source classes. For the EGRET-detected pulsars a cutoff in the energy
814: spectrum is seen in many sources \emph{within} the EGRET energy regime
815: (and therefore clearly well below the VHE range). Indeed, for three
816: out of the four EGRET sources for which a spectral change is implied
817: by the \HESS\ non-detection, a pulsar association has been proposed:
818: 3EG\,J1710--4439 was unambiguously identified with
819: PSR\,1706--44~\citep{Thomp94}, 3EG\,J1809--2328 was proposed to be of
820: PWN nature~\citep{Braje02}, and 3EG\,J1837--0606 was suggested as the
821: counterpart of PSR\,J1837--0604~\citep{DAmico01}. The remaining source
822: in the sample for which the spectral extrapolation of the EGRET source
823: is constrained by the \HESS\ upper limit, is the Galactic Centre
824: source 3EG\,J1746--2851. This object is extremely interesting and
825: important, and may be related to the TeV emission detected in this
826: region, however, a proper discussion falls beyond the scope of
827: this paper.
828:
829: It is interesting to note that an analysis of the EGRET data above
830: 10~GeV~\citep{EGRET10GeV} found eleven EGRET sources with evidence for
831: emission above 10~GeV (at a level of less than 10\% probability that
832: the number of photons seen is a fluctuation of the diffuse background
833: emission). Five of these sources are located in the \HESS\ GPS
834: region. These sources are 3EG\,J1655--4554, 3EG\,1710--4439
835: (PSR\,B1706--44, with a 6.1$\sigma$ detection significance above
836: 10~GeV) 3EG\,J1714--3857, 3EG\,J1746--2851, and 3EG\,J1837--0606 (all
837: marked with a white-and-blue square in
838: Figure~\ref{fig::EGRETnonconnect}). Interestingly, all of these
839: sources belong to the class of non-coincident sources, i.e. have no
840: counterpart at VHE $\gamma$-ray energies. The characteristic cut-off
841: energies of these sources are therefore likely confined to the region
842: below $\sim 100$ GeV. This emphatically emphasises the existence of
843: cutoffs within the energetic gap left between the end of the EGRET
844: measurements and the onset of the \HESS\ and MAGIC observations.
845:
846: To further investigate the cutoff hypothesis a spectral analysis of
847: the EGRET energy spectra has been performed by means of higher order
848: representations, as has been reported by ~\citet{Bertsch00,
849: ReimerBertsch01}. The EGRET spectra were fitted with a broken
850: power-law and with a power-law with an exponential cutoff:
851:
852: \begin{equation}
853: \frac{\partial{J}}{\partial{E}}(E,K,\lambda _{1}, \lambda _{2}) = \left\{
854: \begin{array}{ll} K \left(\frac{E}{1 \mathrm{GeV}}\right)^{-\lambda _{1}} (E
855: \leq 1 \mathrm{GeV}) \nonumber\\
856: K \left(\frac{E}{1 \mathrm{GeV}}\right)^{-\lambda _{2}} (E \geq 1
857: \mathrm{GeV})
858: \end{array} \right.
859: \end{equation}
860:
861: \begin{equation}
862: \frac{\partial{J}}{\partial{E}}(E,K,\lambda,E_{c}) = K \left(\frac{E}{300
863: \mathrm{MeV}}\right)^{-\lambda} exp\left(-\frac{E}{E_{c}}\right)
864: \end{equation}
865:
866: The $\chi^2$ of the resulting fits were compared to that for a single
867: power-law and an F-test employed to test if the more complex form was
868: justified. For many $\gamma$-ray sources there is insufficient
869: high-energy data to justify higher order functional fits. However, for
870: four of the 17 EGRET sources considered in this study the F-test
871: strongly suggests a different spectral form (with a chance probability
872: $<0.05$ as discussed in detail in ~\citet{ReimerBertsch01}):
873: 3EG\,J1655--4554 is better fit by a power-law with exponential cutoff,
874: 3EG\,J1710-4439, 3EG\,J1736-2908, and 3EG\,J1746-2851 are best fit
875: with a broken power-law. All of these sources have no positional
876: counterpart at TeV energies (and are marked with triangles in
877: Figure~\ref{fig::EGRETnonconnect}). The different spectral
878: representations are shown in red in Figure~\ref{fig::OtherFits}. It is
879: interesting to note, that out of the four sources mentioned above for
880: which the \HESS\ non-detection strongly suggests a cutoff in the
881: energy spectrum, the two sources with the largest incompatibility
882: measure $\sigma_{\mathrm{3EG}}$ are also characterised by a
883: statistically significant cutoff in the EGRET spectrum. In particular,
884: the previously mentioned source 3EG\,J1746--2851 (Galactic Centre)
885: shows strong indications for an energy break below 10~GeV. The
886: indicated cutoff in some of the EGRET spectra is entirely consistent
887: with the constraining VHE limit based on power-law extrapolation. The
888: prediction that the other two EGRET sources (3EG\,J1809--2328, and
889: 3EG\,J1837--0606) constrained by the \HESS\ upper limits show a cutoff
890: in the energy range between 10~GeV and 100~GeV is therefore well
891: justified and will be tested by upcoming GLAST-LAT observations.
892:
893: \begin{figure}[ht]
894: \begin{center}
895: \noindent
896: \includegraphics[width=0.57\textwidth]{f7.eps}
897: \end{center}
898: \caption{SED at E $> 30$ MeV for the non-coincident cases in which
899: the EGRET spectrum shows significant deviation from a simple
900: power-law form. The previously reported higher order spectral
901: representations are shown in red (exponential cutoff for
902: 3EG\,J1655--4554 and broken power-law for 3EG\,J1710--4439,
903: 3EG\,J1736--2908 and 3EG\,J1746--2851). }\label{fig::OtherFits}
904: \end{figure}
905:
906: \subsection{VHE $\gamma$-ray sources without an EGRET counterpart}
907: \label{sec::nonconnect::VHE}
908: In this section the \HESS\ sources without a catalogued EGRET
909: counterpart are addressed. At all nominal \HESS\ source locations,
910: flux upper limits have been determined from the EGRET data at energies
911: above 1~GeV by means of the EGRET likelihood
912: technique~\citep{Mattox96}. In the determination of the EGRET upper
913: limit, both the Galactic diffuse emission and point-sources exceeding
914: a 5~$\sigma$-detection significance threshold were modelled and
915: subsequently subtracted. The underlying EGRET exposure corresponds to
916: the first four years of the EGRET mission. As previously discussed,
917: the sensitivity of EGRET (in terms of energy flux $E^2 dN/dE$) is
918: considerably worse than the \HESS\ sensitivity so that no EGRET
919: detection of a \HESS\ source is expected under the assumption of equal
920: energy flux -- which might obviously not necessarily be fulfilled in
921: an astrophysical source.
922:
923: \begin{figure}[ht]
924: \begin{center}
925: \noindent
926: \includegraphics[width=0.85\textwidth]{f8.eps}
927: \end{center}
928: \caption{(Part 1) SED at E $> 30$ MeV for the cases in which no
929: EGRET catalogued counterpart source was found for the \HESS\
930: source. The dashed arrow shows the predicted upper limit from a
931: one-year GLAST scanning observation, taking into account the
932: galactic diffuse background. Derived from this is the spectral
933: compatibility parameter $\sigma_{\mathrm{HESS-GLAST}}$ between
934: GLAST and H.E.S.S. assuming a non-detection with GLAST to
935: illustrate that GLAST will be able to probe the power-law
936: extrapolation from VHE $\gamma$-ray energies whereas the existing
937: EGRET upper limits are unconstraining in this
938: regard.}\label{fig::HESSnonconnectA}
939: \end{figure}
940:
941: Methodologically similar to the previous section, the determination of
942: spectral compatibility was performed by extrapolating \HESS-measured
943: VHE spectra to 1~GeV and comparing the resulting flux to the EGRET
944: upper limit at that energy. The spectral compatibility parameter
945: $\sigma_{\mathrm{H.E.S.S.}}$ is determined in a similar way to
946: $\sigma_{\mathrm{3EG}}$. The spectra of \HESS\ sources with
947: significant curvature were only fitted from the threshold energy at
948: $\sim 100$~GeV to 1~TeV. As in previous sections,
949: $\sigma_{\mathrm{HESS-EGRET}}$ describes how well the extrapolated
950: \HESS\ spectrum can be accommodated by the EGRET upper limit. The
951: resulting spectral energy distributions (SEDs) of the non-coincident
952: \HESS\ sources are shown in Figures~\ref{fig::HESSnonconnectA} and
953: \ref{fig::HESSnonconnectB}.
954:
955: \begin{figure}[ht]
956: \begin{center}
957: \noindent
958: \includegraphics[width=0.85\textwidth]{f9.eps}
959: \end{center}
960: \caption{(Part 2) SED at E $> 30$ MeV for the cases in which no
961: EGRET catalogued counterpart source was found for the \HESS\
962: sources. The dashed arrow shows the predicted upper limit from a
963: one-year GLAST scanning observation, taking into account the
964: diffuse emission. Derived from this is the spectral mismatch
965: between GLAST and H.E.S.S.\ assuming a non-detection with GLAST to
966: illustrate the GLAST will be able to probe the power-law
967: extrapolation from VHE $\gamma$-ray energies whereas the EGRET
968: upper limits are unconstraining in this
969: regard.}\label{fig::HESSnonconnectB}
970: \end{figure}
971:
972: In all cases, the values of $\sigma_{\mathrm{HESS-EGRET}}$ are less than
973: or equal to 1, implying that no EGRET upper limit is violated by the
974: \HESS\ extrapolation to 1~GeV, in stark contrast to the results
975: discussed in the previous section. The most interesting case is that
976: of HESS\,J1713--395 (RX\,J1713.7--3946). In this case the power-law
977: extrapolation is at the level of the EGRET upper limit and
978: $\sigma_{\mathrm{HESS-EGRET}} = 1$. The unconstraining nature of the
979: EGRET upper limits is simply a consequence of a lack of instrumental
980: sensitivity at GeV energies, worsened in regions of pronounced diffuse
981: $\gamma$-ray emission such as the \HESS\ GPS region. However, this
982: situation will change significantly in the near future, given the
983: expected sensitivity of the GLAST-LAT as also shown in
984: Figures~\ref{fig::HESSnonconnectA} and \ref{fig::HESSnonconnectB} in
985: which $\sigma_{\mathrm{HESS-GLAST}}$ is calculated for a typical one-year
986: GLAST sensitivity limit in the Inner Galaxy. These numbers suggest
987: that the increased sensitivity of the LAT might render common GeV-TeV
988: studies possible. While the EGRET upper limits are currently
989: insensitive to linear extrapolations of the \HESS\ spectra, the
990: GLAST-LAT will clearly allow for more sensitive studies. It should,
991: however, be noted, that a linear extrapolation between \HESS\ and
992: GLAST-LAT energies most probably represents the ``best-case'' for any
993: such study: physical models typically show spectra that harden towards
994: GeV energies, unless a different emission component/process takes
995: over. It remains to be seen if GLAST will detect emission at
996: comparable energy flux and potentially determine the position of the
997: peak in the SEDs. As discussed previously, the tremendous advantage of
998: the GLAST-LAT over any previous mission is the continuous energy
999: coverage from 30~MeV all the way up into the VHE $\gamma$-ray range at
1000: $\sim 300$ GeV with significantly improved sensitivity and angular
1001: resolution, bridging the current energy gap in which some of the
1002: physically interesting suggested energy cutoffs occur.
1003:
1004: \section{Interpretation}
1005: \label{sec::interpretation}
1006:
1007: \subsection{Sources detected both at GeV and TeV energies}
1008: \label{subsec::Connection}
1009: As previously stated and shown in Table~\ref{tab::ConnectCases}, only
1010: 9 sources exist which can be characterised as coincident Galactic
1011: EGRET and VHE $\gamma$-ray sources at this moment (5 within the inner
1012: Galaxy, 4 outside of the \HESS\ GPS region). Given the large number of
1013: Galactic sources in both GeV and TeV $\gamma$-rays this number is
1014: rather small -- and is indicative of different dominant source classes
1015: in these two energy domains. However, for the few cases where a
1016: positional coincidence may exist some important astrophysical
1017: implications as well as predictions for the upcoming GLAST mission can
1018: be drawn.
1019:
1020: Whilst EGRET and in particular GLAST have sufficiently large FoVs to
1021: be able to efficiently observe the whole sky, the limited FoV of
1022: imaging VHE $\gamma$-ray instruments (typically 4$^{\circ}$ diameter)
1023: allow for only limited sky coverage. However, for known GeV sources
1024: high-angular resolution VHE instruments such as MAGIC and \HESS\ with
1025: tremendously higher photon statistics at high energies can help in the
1026: identification and interpretation of the GeV emission. This approach
1027: has been followed by~\citet{Reimer} for the Kookaburra complex. In
1028: this region of TeV and GeV $\gamma$-ray emission, a re-analysis of the
1029: EGRET data taking advantage of the higher spatial resolution images
1030: from \HESS\ observations, demonstrated that the dominant GeV emission
1031: (3EG\,J1420--6038) is positionally coincident with
1032: HESS\,J1420--607~\citep{HESSKooka}. This EGRET source has been flagged
1033: as confused in the 3EG catalogue~\citep{EGRET} and in the re-analysis
1034: 3EG\,J1420--6038 was found to be partially overlapping with a less
1035: intense second GeV $\gamma$-ray source. This second GeV source --
1036: detected below the nominal detection threshold for EGRET -- is
1037: apparent in a dedicated analysis at approximately 1/3 of the GeV flux
1038: of the dominant source~\citep{Reimer} and is positionally coincident
1039: with the second VHE $\gamma$-ray source in the Kookaburra region,
1040: HESS\,J1418--609~\citep{HESSKooka} (associated with the ``Rabbit''
1041: PWN). This suggestive morphology match between the GeV data and the
1042: \HESS\ data thus helped in the interpretation and identification of
1043: the confused EGRET sources and made a separation into two individual
1044: sources possible. Studies such as this one show how confused GeV
1045: emission regions (in particular in the Galactic plane where the
1046: diffuse $\gamma$-ray background is dominant) may be unravelled using
1047: the GeV emission as measured from a large-aperture space-based
1048: $\gamma$-ray instrument together with narrow FoV but superior spatial
1049: resolution observations provided by ground-based atmospheric Cherenkov
1050: telescopes. This approach seems promising for achieving convincing
1051: individual source identifications in the era of the GLAST-LAT.
1052:
1053: \begin{figure}[ht]
1054: \begin{center}
1055: \noindent
1056: \includegraphics[width=0.65\textwidth]{f10.eps}
1057: \end{center}
1058: \caption{SED for the coincident source HESS\,J1640-465 along with
1059: leptonic IC-models for different magnetic fields and different
1060: ages of the system. The purpose of this figure is to demonstrate
1061: that rather extreme values for the magnetic-field ($\sim 100
1062: \mu$G) or the age of the system ($\sim 10^6$ years) have to be
1063: invoked to fit such a spectral energy distribution in a leptonic
1064: model. These models numerically take into account the
1065: time-evolution of the electron spectrum considering energy losses
1066: and injection of electrons in time-steps much shorter than the age
1067: of the system. Synchrotron and IC losses are calculated following
1068: the formalism in~\citet{Blumenthal}. The injection spectrum for
1069: the electrons was chosen to have a photon index of 2.5, the
1070: Inverse Compton scattering was performed on the CMB only. It
1071: should be noted, that the X-ray flux between 2 and 10~keV detected
1072: from this source is at the level of $10^{-13}$ erg cm$^{-2}$
1073: s$^{-1}$ as determined by~\citet{Funk1640}. This rather low X-ray
1074: flux renders a connection between the \HESS\ and the EGRET source
1075: in any leptonic model extremely difficult as demonstrated by this
1076: figure.}
1077: \label{fig::HESSConnectModelling}
1078: \end{figure}
1079:
1080: On the other hand, the detection of VHE $\gamma$-ray sources with
1081: EGRET (or the GLAST-LAT) may help in the interpretation of the TeV
1082: data and the modelling of the $\gamma$-ray emission mechanism.
1083: Measuring the energy spectrum of a high-energy $\gamma$-ray source
1084: over 5--6 decades in energy should provide rather stringent
1085: constraints on the $\gamma$-ray emission mechanism. The Crab is in
1086: this respect the only good example of a Galactic source for which both
1087: an excellent GeV and TeV coverage exists which in turn helped to
1088: understand the emission mechanism and the magnetic field strength
1089: rather well in comparison to most other $\gamma$-ray sources. With the
1090: advent of the GLAST-LAT many more such sources with a good GeV and TeV
1091: coverage can be expected.
1092:
1093: Figure~\ref{fig::HESSConnectModelling} shows the SED for the
1094: positionally coincident sources 3EG\,J1639--4702 and
1095: HESS\,J1640--465~\citep{HESSScanII, Funk1640}. The figure shows a
1096: rather typical $\gamma$-ray SED for a positionally coincident sources
1097: (see Figure~\ref{fig::SpectraConnect}) with a power-law spectrum at
1098: TeV energies with photon index $2.4\pm 0.15$ and a similar power-law
1099: at GeV energies with photon index $2.5 \pm 0.18$, at an energy flux
1100: level an order of magnitude higher than that at 1~TeV. The EGRET
1101: source 3EG\,J1639--4702 is rather close to the detection significance
1102: threshold (with a TS$^{1/2}$-value of 6.4). Taking this SED as
1103: representative, several scenarios for a common origin of the
1104: $\gamma$-ray emission are considered. For a hadronic model the shape
1105: of this SED can be rather easily fitted, requiring a power-law
1106: distribution of primary hadrons with $dN/dE \propto E^{-\alpha}$, with
1107: $\alpha\approx2.5$ and a maximum particle energy beyond the TeV range.
1108: However, for a simple leptonic model, with the $\gamma$-ray emission
1109: interpreted as inverse-Compton up-scattering of soft photon fields,
1110: matching the shape of the SED requires rather extreme values for the
1111: magnetic field~\citep[given the low level of X-ray synchrotron
1112: emission from this system as reported by][]{Funk1640} or for the age
1113: of the system (given the need to confine the accelerated electrons
1114: within the system). This is demonstrated in
1115: Figure~\ref{fig::HESSConnectModelling} which shows 3 leptonic model
1116: curves. In the generation of these models, the time-evolution of the
1117: electron spectrum due to energy losses was taken into account. These
1118: energy losses were calculated according to the formalism described
1119: in~\citet{Blumenthal}. For high energy electrons the energy-loss
1120: (cooling) timescale $E/(dE/dt)$ is proportional to $1/E$ for losses
1121: predominantly via synchrotron radiation or IC in the Thomson
1122: regime. In this case, for continuous injection of electrons with a
1123: power law spectrum $dN/dE \propto E^{-\alpha}$, a spectral break to
1124: $E^{-(\alpha+1)}$ will occur. The slope of the IC spectrum (again in
1125: the Thomson regime) is given by $\Gamma=(\alpha+1)/2$. In the
1126: idealised case of the Thomson cross-section and a single (thermal)
1127: target radiation field the break energy is given approximately by:
1128: \begin{equation}
1129: E_{\mathrm{break}} \approx 0.4 (t_{\mathrm{source}}/10^{6}
1130: \mathrm{yr})^{-2}((U_{\rm{rad}}+B^{2}/8\pi)/ 1 \rm{eV\,cm}^{-3})^{-2}
1131: (T/ 2.7\,\rm{K}) \,\rm{GeV}
1132: \end{equation}
1133: In all cases shown in Figure~\ref{fig::HESSConnectModelling}, the
1134: time-independent injection spectrum of the electrons was fixed with an
1135: index of 2.5 and a cutoff energy at 100~TeV, with the IC scattering on
1136: the cosmic microwave background (CMB) alone.
1137: The first curve (dashed blue) is derived using values rather typically
1138: assumed for TeV sources: a magnetic field strength $10\mu$G and age of
1139: $10^4$ years. This curve provides an adequate description of the
1140: \HESS\ data, but not the EGRET data due to the characteristic turnover
1141: of the $\gamma$-ray spectrum at lower energies. The other two curves
1142: (solid green and dash-dotted red) are shown to illustrate how the SED
1143: could be accommodated in a leptonic model and thus how the peak of the
1144: IC emission can be pushed into the EGRET range. Taking a typical
1145: Galactic radiation field (which might not be realistic as e.g.\ in
1146: binary system with a massive stellar component) either rather high
1147: magnetic fields (green solid) or rather old sources have to be invoked
1148: (dash-dotted red). The high-magnetic field scenario would, however,
1149: lead to the prediction of a very high X-ray flux. This prediction
1150: contradicts the faint X-ray emission detected from this object (at the
1151: level of $10^{-13}$ erg cm$^{-2}$ s$^{-1}$) as well as in most other
1152: Galactic VHE $\gamma$-ray sources (where the X-ray emission is
1153: typically at the same level or below the VHE $\gamma$-ray energy
1154: flux). To explain the $\gamma$-ray emission of coincident sources
1155: through leptonic IC emission, the sources should thus be rather old to
1156: be able to accumulate enough low energy electrons to explain the high
1157: GeV flux in a typical Galactic radiation field. They should then,
1158: however, either be rather bright X-ray emitters or be very old.
1159:
1160: \begin{figure}[ht]
1161: \begin{center}
1162: \noindent
1163: \includegraphics[width=0.65\textwidth]{f11.eps}
1164: \end{center}
1165: \caption{High-energy SED for the SNR RX\,J1713.7--3946. The black
1166: data points show measurements with H.E.S.S., whereas the blue
1167: circles and red triangles show simulated GLAST data, assuming two
1168: different models (leptonic and hadronic) for the $\gamma$-ray
1169: emission (shown as dashed red and solid blue lines). This simulation
1170: uses the current best estimate of the LAT performance and illustrate
1171: that in principle the GLAST-LAT should be able to detect this
1172: prominent shell-type SNR in a 5-years observation or faster,
1173: depending on the emission mechanism. This figure has been reproduced
1174: from~\citep{FunkICRCSNR}. }
1175: \label{fig::SED1713}
1176: \end{figure}
1177:
1178: VHE $\gamma$-ray sources may be detectable using GLAST even if the
1179: $\gamma$-ray emission is generated by IC scattering on a typical
1180: Galactic radiation field, as demonstrated for the SNR
1181: RX\,J1713.7--3946 where a GLAST detection should shed light on the
1182: heavily debated origin of the TeV
1183: emission~\citep{FunkICRCSNR}. $\gamma$-rays of leptonic origin
1184: (produced by IC) might be distinguishable from those of hadronic
1185: origin (produced by $\pi^0$-decay) through their characteristic
1186: spectral shape, although recent claims have been made that under
1187: certain conditions the leptonic $\gamma$-ray spectra might resemble
1188: those of pionic decays~\citep{Ellison}. Figure~\ref{fig::SED1713}
1189: shows that the GLAST-LAT will have the sensitivity to measure energy
1190: spectra (in 5 years of scanning observations) for both hadronic and
1191: leptonic emission scenarios, illustrating that the LAT energy range is
1192: particularly well suited to distinguish these models. Measuring the
1193: spectral shape of the $\gamma$-ray emission through deep GeV
1194: observations with the GLAST-LAT will play an important role in
1195: interpreting the currently known TeV $\gamma$-ray sources.
1196:
1197: \subsection{The non-connection of GeV and TeV sources}
1198:
1199: For sources where no positional coincidence has been found for the GeV
1200: and TeV domains both instrumental and astrophysical explanations can
1201: be invoked.
1202:
1203: \subsubsection{Instrumental reasons for non-connection}
1204:
1205: The most obvious reason for a non-detection of a TeV source with EGRET
1206: is the sensitivity mismatch. In a typical $\sim 5$ hour observation
1207: \HESS\ has an energy flux sensitivity of about a factor of $\sim
1208: 50-80$ lower than that of EGRET for its entire lifetime (above 1~GeV
1209: in the Galactic Plane). Additionally, with decreasing detection
1210: significance an increasing number of EGRET sources are expected to be
1211: artificial due to source confusion in the Galactic plane and in
1212: particular due to uncertainties from the model chosen to describe the
1213: dominant diffuse $\gamma$-ray emission. The GLAST-LAT will inevitably
1214: shed more light on all persistent EGRET sources, since these will be
1215: rather bright $\gamma$-ray sources for the LAT instrument. However, it
1216: should be noted that the brightest Galactic H.E.S.S.\ sources (such as
1217: RX\,J1713.7--3946) are not going to be very bright GLAST sources as
1218: discussed in the previous section. Certainly, similar to EGRET, the
1219: LAT will (at the lower end of the energy range) suffer from
1220: uncertainties and systematic effects due to intrinsic properties of
1221: the experimental approach and in particular due to the modelling of
1222: the diffuse $\gamma$-ray background, however, at a lower flux level.
1223:
1224: Another instrumental effect that could render a correlation between
1225: GeV and TeV sources unlikely, is the insensitivity of imaging VHE
1226: $\gamma$-ray instruments to very extended sources (radius $>
1227: 1^{\circ}$) without significant sub-structure. The EGRET data do not
1228: put strong constraints on the source extension of a typical source in
1229: the Galactic plane. Source extensions that can be derived from the
1230: data are on the scale of the EGRET PSF, i.e. degree scales. The
1231: angular resolution (and thus the maximum sensitivity) of VHE
1232: $\gamma$-ray instruments on the other hand is of the order of a few
1233: arc minutes. The upper limits for \HESS\ at the positions of EGRET
1234: sources quoted in this study are derived under the assumption of a
1235: point-like source (with a typical size of the source region of less
1236: than $\sim 0.1^{\circ}$ rms width). The sensitivity and thus the upper
1237: limit scales roughly linearly with the source size~\citep{FunkPhD} and
1238: for source sizes in excess of $\sim 1^{\circ}$, the \HESS\ data become
1239: completely unconstraining due to the fact that the source size becomes
1240: comparable with the size of the FoV and no reliable background
1241: estimation can be performed \citep[see][for a description of the
1242: background estimation techniques used]{BergeBackground}. Large-FoV
1243: instruments (with poorer angular resolution) such as
1244: Milagro~\citep{Milagro}, are better suited to detect sources with
1245: intrinsically large sizes in VHE $\gamma$-rays (with sufficiently high
1246: fluxes). However, due to their modest ($\sim1^{\circ}$) angular
1247: resolution, such instruments suffer from problems of source confusion
1248: similar to those of current GeV measurements. Indeed, several of the
1249: recently reported Milagro source candidates are coincident with EGRET
1250: sources~\citep{MilagroPlane}. Hypothesising that EGRET sources
1251: exhibit angular sizes larger than $\sim 1^{\circ}$, Milagro-type
1252: instruments might be better suited to detect large scale emission at
1253: VHE $\gamma$-ray energies. Again, the GLAST-LAT, with its superior
1254: angular resolution to EGRET, will shed more light on the issue of the
1255: intrinsic sizes of GeV sources in the Galactic plane. The constraints
1256: on the power-law extrapolation of EGRET sources by sensitive \HESS\
1257: upper limits as derived in the previous sections are naturally only
1258: valid under the assumption that the VHE counterpart to the EGRET
1259: emission does not exhibit a size much larger than $\sim 1^{\circ}$.
1260:
1261: \subsubsection{Astrophysical reasons for non-connection}
1262:
1263: The non-detection of most TeV sources in the GeV range by EGRET may be
1264: due simply to a lack of instrumental sensitivity. On the other hand,
1265: the lack of TeV counterparts to most bright GeV sources requires the
1266: presence of steepening (or cut-offs) between 10 and 100 GeV in the
1267: spectra of these sources (see section~\ref{sec::nonconnect} and
1268: Figure~\ref{fig::OtherFits}). Steepening in $\gamma$-ray energy
1269: spectra between 10 and 100~GeV can occur for many reasons, the most
1270: prominent of which are discussed briefly below.
1271:
1272: \emph{Acceleration limits}. The maximum energy to which particles are
1273: accelerated in a source may be determined by a balance between the
1274: acceleration and energy loss timescales, or between acceleration and
1275: escape timescales, or simply by the lifetime of the source. In the
1276: limit of Bohm diffusion, the escape time of accelerated particles from
1277: the source can be written as
1278: \begin{equation}
1279: t_{\mathrm{escape}} \sim (r_{\mathrm{source}}/\mathrm{pc})^{2} D_{0}
1280: (E/\mathrm{TeV})^{-\Delta}
1281: \end{equation}
1282: The associated cut-off in the resulting $\gamma$-ray emission may
1283: occur at much lower energies, as in the case of proton-proton
1284: interactions ~\citep[a factor $\sim20$ as shown in][]{Kappes2007}, or
1285: close to the primary particle energy, as in the case of inverse
1286: Compton scattering in the Klein-Nishina limit~\citep{Blumenthal}.
1287:
1288: \emph{Particle transport} may impact on the spectral shape in several
1289: ways. For protons described by a power-law $ J_p(E_p)=K E_p^{-\Gamma}$
1290: the $\gamma$-rays produced in hadronic interactions are expected to
1291: follow a similar power-law spectrum $F_{\gamma}(E_{\gamma})\propto
1292: E_{\gamma}^{-\Gamma}$. Generally, high energy particles escape more
1293: easily leading to a cut-off in the particle and hence $\gamma$-ray
1294: spectrum inside the source. Therefore, due to particle transport, the
1295: spectrum of the protons generating the $\gamma$-rays through hadronic
1296: interactions is not necessarily the same as the one at the
1297: acceleration site. In the case of diffusion the proton spectrum at the
1298: $\gamma$-ray production site can instead be written as
1299: $J_p(E_p,\;r,\;t)=\frac{c}{4\pi} f, $ where $f(E_p,\;r,\;t)$ is the
1300: distribution function of protons at an instant $t$ and distance $r$
1301: from the source. The distribution function satisfies the diffusion
1302: equation~\citep{Ginzburg}.
1303: \begin{equation}\label{difeq}
1304: \frac{\partial f}{\partial t}=\frac{D(E_p)}{r^2} \frac{\partial}{\partial
1305: r} r^2 \frac{\partial f}{\partial r} + \frac{\partial}{\partial
1306: E_p}(Pf)+Q,
1307: \end{equation}
1308: where $P=-dE_p/dt$ is the continuous energy loss rate of the
1309: particles, $Q=Q(E_p,\;r,\;t)$ is the source function, and $D(E_p)$ is
1310: the diffusion coefficient. ~\citet{Atoyan} derived a general solution
1311: for Equation (\ref{difeq}). Hence, as has been emphasised
1312: by~\citet{FelixSN1006}, the observed $\gamma$-ray flux can have a
1313: significantly different spectrum from that expected from the particle
1314: population at the source. In the (expected) case of energy-dependent
1315: diffusion ($D \propto E^{-\Delta}$, with $\Delta$ typically assumed to
1316: lie in the range $\sim 0.3-1.0$) the $\gamma$-ray spectrum will follow
1317: $F_{\gamma}(E_{\gamma})\propto E_{\gamma}^{-(\Gamma+\Delta)}$. The
1318: exact shape of the spectrum will depend on the age of the accelerator,
1319: duration of injection, the diffusion coefficient, and the location of
1320: the target material.
1321:
1322: The influence of convection (lower energy cutoff in primary particle
1323: spectrum) is typically stronger for low energy (GeV) $\gamma$-rays
1324: potentially resulting in a VHE $\gamma$-ray source that has no EGRET
1325: counterpart in cases in which an external accelerator produces primary
1326: hadrons near an active target.\citet{Torres2004} and~\citet{Eva2006}
1327: have recently studied collective wind configurations produced by a
1328: number of massive stars, and obtained densities and expansion
1329: velocities of the stellar wind gas that is the target for hadronic
1330: interactions in several examples, showing that these may be sources
1331: for GLAST and the TeV instruments in non-uniform ways, i.e., with or
1332: without the corresponding counterparts in the other energy band.
1333:
1334: \emph{Particle energy losses} away from the acceleration site may also
1335: produce spectral steepening in a very natural way as discussed earlier
1336: (see section~\ref{subsec::Connection}). In the case where particle
1337: injection is effectively finished (i.e. the injection rate is much
1338: lower than in the past), radiative energy losses may produce a rather
1339: sharp cut-off in the $\gamma$-ray spectrum as e.g. shown in
1340: ~\citep{Funk1640}. For high energy electrons the energy-loss (cooling)
1341: timescale $E/(dE/dt)$ is proportional to $1/E$ for losses dominantly
1342: via synchrotron radiation or IC in the Thomson regime. In this case,
1343: for continuous injection of electrons with a power law spectrum $dN/dE
1344: \propto E^{-\alpha}$, a spectral break to $E^{-(\alpha+1)}$ will
1345: occur. The slope of the IC spectrum (again in the Thomson regime) is
1346: given by $\Gamma=(\alpha+1)/2$. In the idealised case of the Thomson
1347: cross-section and a single (thermal) target radiation field the break
1348: energy is given approximately by:
1349: \begin{equation}
1350: E_{\mathrm{break}} \approx 0.4 (t_{\mathrm{source}}/10^{6}
1351: \mathrm{yr})^{-2}((U_{\rm{rad}}+B^{2}/8\pi)/ 1 \rm{eV\,cm}^{-3})^{-2} (T/
1352: 2.7\,\rm{K}) \,\rm{GeV}
1353: \end{equation}
1354:
1355: \emph{$\gamma$-$\gamma$ pair-production} occurs above a threshold
1356: $\epsilon_{\gamma}\epsilon_{\rm{target}}> 2m_{e}^{2}c^{4}$. For
1357: stellar systems with $\epsilon_{\rm{target}} \sim$ 1 eV, this process
1358: occurs above $\sim$ 500 GeV. Pairs produced in $\gamma$-$\gamma$
1359: interactions may inverse Compton scatter on the same radiation field
1360: -- leading to the development of a
1361: cascade~\citep{Protheroe}. Attenuation on the interstellar IR and CMB
1362: can be neglected below 10 TeV so $\gamma$-$\gamma$ 'cut-offs' are only
1363: expected in compact regions of very high radiation density, for
1364: example within binary stellar systems. These absorption/cascade
1365: 'features' may not represent the end of the $\gamma$-ray spectrum as
1366: emission may recover at energies above the resonance.
1367:
1368: \subsection{Prospects for the GLAST-LAT}
1369:
1370: As demonstrated by Figures~\ref{fig::HESSnonconnectA}
1371: and~\ref{fig::HESSnonconnectB} the GLAST-LAT should be able to detect
1372: several of the VHE $\gamma$-ray sources in the inner Galaxy, assuming
1373: a simple power-law extrapolation of the spectrum from TeV to GeV
1374: energies. However, this power-law assumption may not be valid in
1375: several cases, as discussed in the following for the known TeV source
1376: classes.
1377:
1378: {\emph{Pulsar Wind Nebulae}} are currently the most abundant VHE
1379: $\gamma$-ray sources in the Galactic plane. The most prominent example
1380: is the Crab Nebula~\citep{WhippleCrab, HEGRACrab, MilagroCrab,
1381: HESSCrab, MAGICCrab}. The SED expected of PWNe does not typically
1382: result in significant GeV fluxes: the Crab Nebula, detected throughout
1383: both energy bands, seems to be an exceptional case due to its very
1384: strong magnetic field and relative proximity (2~kpc). Most VHE
1385: $\gamma$-ray PWNe are expected to be dominated by IC emission for
1386: which the energy flux generally turns down at lower energies. The
1387: position of this inverse Compton peak determines detectability for
1388: both GeV and TeV instruments. Also the size and flux of the source
1389: also obviously affect the detectability with GLAST-LAT. In general
1390: the higher the energy of the inverse-Compton peak in these sources,
1391: the lower the chance will be to detect them with GLAST. If a large
1392: fraction of the GeV emission attributed to EGRET Galactic unidentified
1393: sources is related to pulsed magnetospheric emission from pulsars as
1394: opposed to emission from the extended wind nebula then a correlation
1395: between the H.E.S.S.\ and EGRET sources in the Inner Galaxy could be
1396: expected, given that the majority of the H.E.S.S.\ sources in this
1397: region seem to be PWNe associated with energetic
1398: pulsars~\citep{SvenjaPulsars}. However, this expectation may not hold
1399: in general due to diversity of parameters like the beaming geometry or
1400: different conversion efficiency of the pulsar's spin-down power into
1401: the Nebula and into $\gamma$-rays.
1402:
1403: {\emph{Shell-type Supernova remnants}} The two prominent and bright
1404: VHE $\gamma$-ray SNRs (RX\,J1713.7--3946 and RX\,J0852.0--4622) are
1405: not expected to be very bright GLAST-LAT sources. Nevertheless, they
1406: are probably amongst the more easily detectable TeV sources in the
1407: GLAST-LAT band. A detailed simulation of the expected signal from
1408: RX\,J1713.7--3946 shows that it might be detectable in one year of
1409: GLAST-LAT observations depending on the assumed TeV $\gamma$-ray
1410: emission mechanism as shown in the previous section. Morphological
1411: studies in GeV $\gamma$-rays will either have to struggle with
1412: moderate angular resolution at low energies or with low photon
1413: statistic at high energies. However, spectral studies will be
1414: immediately possible following a potential detection as shown in
1415: Figure~\ref{fig::SED1713} for RX\,J1713.7--3946. For RX\,J0852.0--4622
1416: (Vela Junior) the situation is even further complicated by the
1417: close-by bright Vela Pulsar. While both of these prominent
1418: TeV-emitting objects are rather young ($\sim 2000$ years), there is
1419: the potential of older SNRs acting as stronger GeV emitters (but
1420: rather faint TeV sources). In this case the GLAST-LAT might see a
1421: different population of shell-type SNRs than VHE $\gamma$-ray
1422: instruments, namely older SNRs which have accumulated a large number
1423: of lower energy CRs, but for which the higher energy CRs (those that
1424: may give rise to the TeV emission) have already left the acceleration
1425: site. A common detection both with GLAST and VHE $\gamma$-ray
1426: instruments might require a hadronic origin of the $\gamma$-ray
1427: emission rather than an Inverse compton (IC) origin due to the
1428: characteristic turn-over of the IC spectrum at lower energies.
1429:
1430: {\emph{Gamma-ray Binary systems}} host a variety of non-thermal
1431: phenomena. The TeV detected binaries: LS\,5039~\citep{HESSLS5039II},
1432: PSR\,B1259-63~\citep{HESS1259}, LSI\,+61\,303~\citep{MAGICLSI} and
1433: Cyg\,X--1~\citep{MAGICCygX} are currently seen as candidates for
1434: detection at GeV energies. $\gamma$-$\gamma$ absorption in binary
1435: systems may producing anti-correlation of the TeV to GeV radiation
1436: during the orbit of these systems. These orbital modulations are
1437: predicted in basically all models for these systems, irrespective of
1438: the assumptions of a pulsar or a black hole compact object or the
1439: process by which high-energy radiation is emitted~\citep[see
1440: e.g.][]{Dermer2007, Dubus2006, Paredes2006}. Details in predicted
1441: light-curves and spectral evolution in time are however rather
1442: distinctive~\citep{Felix5039, Diego5039}.
1443:
1444: \section{Summary}
1445:
1446: The main results of the study of the relationship between GeV and TeV
1447: sources are:
1448:
1449: \begin{enumerate}
1450: \item There are rather few spatially coincident GeV-TeV sources for
1451: the considered Galactic region.
1452: \item Those few positional coincident GeV-TeV sources could occur by
1453: chance, the chance probability of detecting two coincident sources
1454: within the \HESS\ GPS region is $\sim 40$\%, thus no strong hint for
1455: a common GeV/TeV source population is detected.
1456: \item Spectral compatibility (based on a power-law extrapolation) seems
1457: present for most of the positionally coincident sources, but again,
1458: this is expected to occur by chance (as described in the text) given
1459: the sensitivity mismatch and the different energy bands.
1460: \item Dedicated \HESS\ limits at the position of the EGRET sources are
1461: constraining for a power-law extrapolation from the GeV to the TeV
1462: range for several of the EGRET sources, strongly suggesting cutoffs in
1463: the energy spectra of these EGRET sources in the unexplored region
1464: below 100~GeV. Power-law extrapolation of EGRET spectra seem to be
1465: ruled out for most of the EGRET sources investigated in this study.
1466: \item Dedicated EGRET limits at the position of the \HESS\ sources are
1467: not constraining for a power-law extrapolation from the TeV to the
1468: GeV range. This picture will dramatically change once the GLAST-LAT
1469: with its improved sensitivity over EGRET is in orbit.
1470: \item Several important mechanisms for cutoffs in the energy spectra
1471: of GeV sources have been discussed. There are well motivated
1472: physical reasons why the population of GeV and of TeV sources might
1473: be distinct.
1474: \item If a source can be detected with both GeV and TeV instruments,
1475: the huge energy ``lever arm'' over 5-6 decades in energy will
1476: undoubtedly provide stringent constraints on the $\gamma$-ray
1477: emission mechanism in these Galactic particle accelerators.
1478: \end{enumerate}
1479:
1480: Summarising, the study presented here shows that the GLAST-LAT will
1481: tremendously advance the study of the relationship between GeV and TeV
1482: sources by improving the sensitivity over EGRET by an order of
1483: magnitude and in particular by bridging the currently uncovered energy
1484: range between 10~GeV and 100~GeV.
1485:
1486: \acknowledgments The authors would like to acknowledge the support of
1487: their host institutions. In particular, S.F.\ acknowledges support of
1488: the Department of Energy (DOE) and Stanford University and would like
1489: to thank the whole \HESS\ and the GLAST-LAT collaborations for their
1490: support and helpful discussions on the topic, in particular Werner
1491: Hofmann, Felix Aharonian, Benoit Lott, and Seth Digel. DFT is
1492: supported by the Spanish MEC grant AYA 2006-00530, he acknowledges
1493: Juan Cortina and other members of the MAGIC collaboration for advice
1494: and encouragement. OR is supported by the National Aeronautics and
1495: Space Administration under contract NAS5-00147 with Stanford
1496: University. JAH is supported by an STFC Advanced Fellowship.
1497:
1498: \begin{thebibliography}{}
1499: \bibitem[Abdo et al.(2007)]{MilagroPlane}Abdo, A., et al., 2007, ApJ
1500: 664, 91
1501: \bibitem[Aharonian \& Atoyan(1996)]{FelixSN1006} Aharonian F.A., \&
1502: Atoyan, A.M. 1996, A\&A 309, 91
1503: \bibitem[Aharonian et al.(2001)]{CasAHegra} Aharonian, F.~A., et al., 2001, A\&A 370, 112
1504: \bibitem[Aharonian et al.(2004)]{HEGRACrab} Aharonian, F.~A., et al. ({\it
1505: HEGRA Collaboration}), 2001, ApJ 614, 897
1506: \bibitem[Aharonian et al.(2005a)]{HESSMSH} Aharonian, F.~A., et al.
1507: ({\it \HESS\ Collaboration}), 2005a, A\&A 435, L17
1508: \bibitem[Aharonian et al.(2005b)]{HESSScan} Aharonian, F.~A., et al. ({\it
1509: \HESS\ Collaboration}), 2005b, Science 307, 1938
1510: \bibitem[Aharonian et al.(2005c)]{HESS1259} Aharonian, F.~A., et al.
1511: ({\it \HESS\ Collaboration}), 2005c, A\&A 442, 1A
1512: \bibitem[Aharonian et al.(2006a)]{HESS1713_II} Aharonian, F.~A., et
1513: al. ({\it \HESS\ Collaboration}), 2006a, A\&A 449, 223
1514: \bibitem[Aharonian et al.(2006b)]{HESS1825II} Aharonian, F.~A., et
1515: al. ({\it \HESS\ Collaboration}), 2006b, A\&A 460, 365
1516: \bibitem[Aharonian et al.(2006c)]{HESSKooka} Aharonian, F.~A., et
1517: al. ({\it \HESS\ Collaboration}), 2006c, A\&A 456, 245
1518: \bibitem[Aharonian et al.(2006d)]{HESSLS5039II} Aharonian, F.~A., et
1519: al. ({\it \HESS\ Collaboration}), 2006d, A\&A 460, 743
1520: \bibitem[Aharonian et al.(2006e)]{HESSGCDiffuse} Aharonian, F.~A., et
1521: al. ({\it \HESS\ Collaboration}), 2006e, Science 312, 1771
1522: \bibitem[Aharonian et al.(2006f)]{HESSCrab} Aharonian, F.~A., et al.
1523: ({\it \HESS\ Collaboration}), 2006f, A\&A 457, 899
1524: \bibitem[Aharonian et al.(2006g)]{HESSScanII} Aharonian, F.~A., et
1525: al. ({\it \HESS\ Collaboration}), 2006g, ApJ 636, 777
1526: \bibitem[Aharonian et al.(2007a)]{HESS1713_III} Aharonian, F.~A., et
1527: al. ({\it \HESS\ Collaboration}), 2007a, A\&A 464, 235
1528: \bibitem[Aharonian et al.(2007b)]{HESSVelaJr} Aharonian, F.~A., et al.
1529: ({\it \HESS\ Collaboration}), 2007b, ApJ 661, 236
1530: \bibitem[Aharonian et al.(2007c)]{HESSWesterlund} Aharonian, F.~A., et al.
1531: ({\it \HESS\ Collaboration}), 2007c, A\&A 467, 1075
1532: \bibitem[Aharonian et al.(2007d)]{HESSPulsar} Aharonian, F.~A., et al.,
1533: ({\it \HESS\ Collaboration}), 2007d, A\&A 466, 543
1534: \bibitem[Aharonian et al.(2007e)]{HESSMonoceros} Aharonian, F.~A., et al.,
1535: ({\it \HESS\ Collaboration}), 2007e, A\&A 469, L1
1536: \bibitem[Albert et al.(2006)]{MAGICLSI} Albert, J., et al.
1537: ({\it MAGIC Collaboration}), 2006, Science 312, 1771
1538: \bibitem[Albert et al.(2007a)]{MAGICCrab} Albert, J., et al.
1539: ({\it MAGIC Collaboration}), 2007a, submitted to ApJ
1540: (astro-ph/0705.3244)
1541: \bibitem[Albert et al.(2007b)]{MAGICIC443} Albert, J., et al.
1542: ({\it MAGIC Collaboration}), 2007b, ApJ 664, 87
1543: \bibitem[Albert et al.(2007c)]{MagicCasA} Albert, J., et al.
1544: ({\it MAGIC Collaboration}), 2007c, A\&A in press
1545: (astro-ph/0706.4065)
1546: \bibitem[Albert et al.(2007d)]{MAGICCygX} Albert, J., et al.
1547: ({\it MAGIC Collaboration}), 2007d, ApJ 665, 51
1548: \bibitem[Atkins et al.(2002)]{Milagro} Atkins, R. et al., 2002, NIM A
1549: 449, 478
1550: \bibitem[Atkins et al.(2003)]{MilagroCrab} Atkins, R., et al., 2003,
1551: ApJ 595, 803
1552: \bibitem[Atoyan, Aharonian \& V\"olk(1995)]{Atoyan} Atoyan A.~M.,
1553: Aharonian F.~A., \& V\"olk H.~J. 1995, Phys. Rev. D 52, 3265
1554: \bibitem[Berge, Funk \& Hinton(2007)]{BergeBackground}Berge, D., Funk,
1555: S. \& Hinton, J.~A., 2007, A\&A 466, 1219
1556: \bibitem[Bertsch et al.(2000)]{Bertsch00} Bertsch, D.~L. et al.,
1557: Proc. 5th Compton Symposium, AIP Conf. Proc. 510, 2000, 504
1558: \bibitem[Blumenthal \& Gould (1970)]{Blumenthal} Blumenthal, G.R., \&
1559: Gould, R.J., 1970, RvMP., 42, Number 2, 237
1560: \bibitem[Braje et al.(2000)]{Braje02} Braje, T.~M. et al, 2000, ApJ
1561: 565, L91
1562: \bibitem[Carrigan et al.(2007)]{SvenjaPulsars} Carrigan, S., et al, 2007,
1563: Proc. 30th International Cosmic Ray Conference, Merida, Mexico, 2007
1564: \bibitem[D'Amico et al.(2001)]{DAmico01} D'Amico, N. et al., 2001,
1565: ApJ, 552, L45
1566: \bibitem[Dermer \& B\"ottcher(2007)]{Dermer2007} Dermer, C. D., \&
1567: B\"ottcher, M. 2006, ApJ 644, 409
1568: \bibitem[Domingo-Santamaria \& Torres(2006)]{Eva2006}
1569: Domingo-Santamaria E. \& Torres D.~F. 2006 A\&A 448, 613
1570: \bibitem[Dubus(2006)]{Dubus2006} Dubus G., 2006, A\&A 451, 9
1571: \bibitem[Ellison et al.(2007)]{Ellison} Ellison, D.~C., et al, 2007,
1572: ApJ 661, 879
1573: \bibitem[Esposito et al.(1996)]{Esposito} Esposito, J.~A., Hunter,
1574: S.~D., Kanbach, G., \& Sreekumar, P., 1996, ApJ 461, 820
1575: \bibitem[Fiero et al.(1998)]{EGRETPulsar}
1576: \bibitem[Funk(2005)]{FunkPhD}Funk, S., 2005, PhD thesis,
1577: Ruprecht-Karls-Universit\"at Heidelberg
1578: (http://www.ub.uni-heidelberg.de/archiv/5542)
1579: \bibitem[Funk(2006)]{FunkBarcelona}Funk, S., 2006, ApSS 309, 11
1580: \bibitem[Funk et al.(2007)]{Funk1640} Funk, S., et al., 2007, ApJ 662,
1581: 517
1582: \bibitem[Funk et al.(2007b)]{FunkICRCSNR} Funk, S., et al, 2007b,
1583: Proc. 30th International Cosmic Ray Conference, Merida, Mexico, 2007
1584: \bibitem[Gaensler et al.(2003)]{Gaensler}Gaensler, B.~M., et al 2003,
1585: ApJ, 588, 441
1586: \bibitem[Ginzburg \& Syrovatskii(1964)]{Ginzburg} Ginzburg V.L.,
1587: Syrovatskii S.I. 1964, ``The Origin of Cosmic Rays'', Pergamon
1588: Press, London
1589: \bibitem[Hoppe et al.(2007)]{RCW86HESS} Hoppe, S., et al, 2007,
1590: Proc. 30th International Cosmic Ray Conference, Merida, Mexico, 2007
1591: \bibitem[Hartman et al.(1999)]{EGRET} Hartman, R.~C., et al., 1999,
1592: ApJS 123, 79
1593: \bibitem[Kappes et al.(2007)]{Kappes2007} Kappes, A., et al, 2007, ApJ 656, 870
1594: \bibitem[Khangulyan, Aharonian \& Bosch-Ramon(2007)]{Felix5039}
1595: Khangulyan, D., Aharonian, F., \& Bosch-Ramon, V., submitted to
1596: MNRAS (astro-ph/0707.1689)
1597: \bibitem[Kniffen et al.(1997)]{Kniffen_LSI} Kniffen, D.~A., et al.,
1598: 1997, ApJ 486, 126
1599: \bibitem[Mattox et al.(1996)]{Mattox96} Mattox, J.~M., et al., 1996,
1600: ApJ 461, 396
1601: \bibitem[Moiseev et al.(2007)]{GLASTACD}Moiseev, A., et al., 2007, APh
1602: 27, 339
1603: \bibitem[Nolan et al.(1993)]{EGRETCrab}Nolan, P., et al., 1993, ApJ
1604: 409, 697
1605: \bibitem[Paredes et al.(2000)]{Paredes_LS5039}Paredes, J.~M., et al.,
1606: Science 288, 2340
1607: \bibitem[Paredes, Bosch-Ramon \& Romero(2006)]{Paredes2006} Paredes,
1608: J.~M., Bosch-Ramon, V., \& Romero, G.~E., 2006, A\&A 451, 259
1609: \bibitem[Petry(2001)]{Petry}Petry, D., 2001, ASSL 267, 299
1610: \bibitem[Protheroe, Mastichiadis \& Dermer(1992)]{Protheroe}Protheroe,
1611: R.~J., Mastichiadis, A., \& Dermer, C.~D., 1992, Aph 1, 113
1612: \bibitem[Reimer \& Bertsch (2001)]{ReimerBertsch01} Reimer, O. \&
1613: Bertsch, D.L., Proc. 27th ICRC, 2001, Vol.6, 2546
1614: \bibitem[Reimer \& Funk(2007)]{Reimer} Reimer, O., \& Funk, S., 2007,
1615: Ap\&SS 309, 203
1616: \bibitem[Romero, Benaglia, \& Torres(1999)]{Romero}Romero, G.~E.,
1617: Benaglia, P., \& Torres, D.~F., 1999, A\&A 348, 868
1618: \bibitem[Sierpowska-Bartosik \& Torres(2007)]{Diego5039}
1619: Sierpowska-Bartosik A. \& Torres D. F., 2007, submitted to ApJ
1620: (astro-ph/0708.0189)
1621: \bibitem[Smith et al.(2007)]{VeritasLSI} Smith, A., et al, 2007,
1622: Proc. 30th International Cosmic Ray Conference, Merida, Mexico, 2007
1623: \bibitem[Sturner \& Dermer (1995)]{Sturner} Sturner, S.~J., \& Dermer,
1624: C.~D., 1995, A\&A 293 (1), L17
1625: \bibitem[Tavani et al.(1998)]{Tavani_LSI} Tavani, M., et al., 1998,
1626: ApJ 497, L98
1627: \bibitem[Thompson et al.(1994)]{Thomp94} Thompson, D~.J. et al., 1994,
1628: Nature 359, 615
1629: \bibitem[Thompson, Bertsch \& O'Neal (2005)]{EGRET10GeV}Thompson,
1630: D.~J., Bertsch, D.~L., O'Neal, R.~H., 2005, ApJs 157, 324
1631: \bibitem[Torres et al.(2003)]{DiegoSNR} Torres, D.~F., Romero, G.~E.,
1632: Dame, T.~M., Combi, J.~A., \& Butt, Y.~M., 2003, PhR, 382, 303
1633: \bibitem[Torres, Domingo-Santamaria \& Romero(2004)]{Torres2004}
1634: Torres D.~F., Domingo-Santamaria E., \& Romero G.~E., 2004, ApJ 601,
1635: L75
1636: \bibitem[Wang et al.(2005)]{Wang}Wang, W., et al., 2005, MNRAS 360, 646
1637: \bibitem[Weekes et al.(1989)]{WhippleCrab} T.~C. Weekes et al., 1989,
1638: ApJ 342, 379
1639: \bibitem[Zdziarski(1987)]{Zdziarski}Zdziarski, A.~A., 1987, ApJ 335,
1640: 786
1641: \end{thebibliography}
1642: \end{document}
1643:
1644: %%
1645: %% End of file `gevtev.tex'.
1646: