On The Gamma-Ray Emission From Reticulum II and Other Dwarf Galaxies
FFERMILAB-15-093-A
On The Gamma-Ray Emission From Reticulum II and Other Dwarf Galaxies
Dan Hooper
1, 2 and Tim Linden Fermi National Accelerator Laboratory, Center for Particle Astrophysics, Batavia, IL University of Chicago, Department of Astronomy and Astrophysics Chicago, IL Kavli Institute for Cosmological Physics University of Chicago, Chicago, IL
The recent discovery of ten new dwarf galaxy candidates by the Dark Energy Survey (DES) and thePanoramic Survey Telescope and Rapid Response System (Pan-STARRS) could increase the FermiGamma-Ray Space Telescope’s sensitivity to annihilating dark matter particles, potentially enablinga definitive test of the dark matter interpretation of the long-standing Galactic Center gamma-rayexcess. In this paper, we compare the previous analyses of Fermi data from the directions of thenew dwarf candidates (including the relatively nearby Reticulum II) and perform our own analysis,with the goal of establishing the statistical significance of any gamma-ray signal from these sources.We confirm the presence of an excess from Reticulum II, with a spectral shape that is compatiblewith the Galactic Center signal. The significance of this emission is greater than that observed from99.84% of randomly chosen high-latitude blank-sky locations, corresponding to a local detectionsignificance of 3.2 σ . We improve upon the standard blank-sky calibration approach through the useof multi-wavelength catalogs, which allow us to avoid regions that are likely to contain unresolvedgamma-ray sources. I. INTRODUCTION
Over the past several years, a bright and statisti-cally significant excess of gamma-rays has been reportedfrom the region surrounding the Galactic Center [1–9]. The spectral and morphological characteristics ofthis signal are each in good agreement with that pre-dicted from annihilating dark matter particles with amass of m DM ∼ σv ∼ − cm s − (for the representative case of annihilationsto b ¯ b ). And although the proposed astrophysical explana-tions for this excess have been shown to face considerablechallenges, it is not currently possible to entirely rule outthe possibility that these photons originate from a largepopulation of unresolved point sources [10–12] or from aseries of cosmic ray outbursts [13–15]. In light of this situ-ation, gamma-ray observations of the Milky Way’s dwarfspheroidal galaxies play a critical role, being potentiallyable to provide a confirmation or refutation of the darkmatter interpretation of the Galactic Center excess.Searches for gamma-rays from known dwarf galax-ies [16–18] have yielded stringent constraints on the darkmatter parameter space. They have not yet, however,been sufficiently sensitive to cover the full range of crosssections favored to explain the Galactic Center excess.It has been anticipated that ongoing and planned opticalsurveys will discover a significant number of presently un-known Milky Way dwarf spheroidal galaxies [19–22]. Ifone of more of these objects happens to be nearby and/orcontain a high density of dark matter, it could constitutean important target for the Fermi Gamma-Ray SpaceTelescope, significantly strengthening their sensitivity toannihilating dark matter.Very recently, optical imaging data from the Dark En-ergy Survey (DES) was used to discover nine new dwarf galaxy candidates [23, 24]. Shortly thereafter, yet an-other dwarf candidate (Triangulum II) was discoveredfrom within the data from the Panoramic Survey Tele-scope and Rapid Response System (Pan-STARRS) [25].Of particular interest is the object Reticulum II (alsoknown as DES J0335.6-5403) whose proximity ( ∼ ± pc)make it likely to be a dwarf galaxy (rather than a glob-ular cluster) and a very promising target for gamma-raysearches for annihilating dark matter. Although spec-troscopic follow-up will be required to measure the darkmatter distributions of these systems, it is plausible thatRecticulum II (or perhaps Triangulum II) could providea gamma-ray signal from annihilating dark matter thatis brighter than that from any other known dwarf galaxy.Two independent groups have already reported the re-sults of their analyses of Fermi Gamma-Ray Space Tele-scope data from the directions of Reticulium II and DES’sother new dwarf galaxy candidates. The first of theseanalyses, presented jointly by the Fermi and DES col-laborations, identified a modest gamma-ray excess fromthe direction of Reticulium II, with a test statistic (TS)of 6.7 [26]. Although the text of that paper quotesa p -value for this excess of 0.06, this includes a trialsfactor of approximately 4, intended to account for therange of possible dark matter masses and annihilationchannels that were scanned over. If one instead employsdwarf galaxies to test the dark matter interpretation ofthe Galactic Center excess (for which the spectrum hasbeen previously measured), no such trials factor is re- TS is defined as twice the difference in the global log-likelihoodbetween the null and alternative hypotheses. a r X i v : . [ a s t r o - ph . H E ] M a r quired. The local significance of this signal (correspond-ing to p (cid:39) . ) is approximately 2.4 σ . The other anal-ysis, carried out by Geringer-Sameth et al. , also findsan excess from Reticulium II, and quotes significancesranging from approximately 2.3 σ to 4.7 σ , depending onthe background model procedure and trials factor thatare adopted. The local significance (without trials fordifferent dark matter models) determined using an em-pirical sample of nearby spatial regions (as opposed toassuming that the background is Poisson distributed) isfound to be approximately 2.8 σ [27], or approximately0.4 σ greater that that found by the Fermi and DES col-laborations. The excesses identified in each of these pa-pers is most prevalent at energies between ∼ σ . And althoughspectroscopic follow-up of Reticulum II will be requiredbefore this observation can be used to constrain or inferthe value of the dark matter annihilation cross section,for a plausible range of dark matter profiles, this resultappears to be consistent with dark matter interpreta-tions of the Galactic Center signal and the null resultsfrom other dwarf galaxies. II. FERMI DATA ANALYSIS
In order to calculate the significance of any gamma-rayemission observed from a given dwarf galaxy (or dwarfgalaxy candidate), we examine approximately 6.5 yearsof Fermi-LAT data, using the P7REP photons in theenergy range of 500 MeV to 500 GeV. We exclude eventsarriving with a zenith angle greater than 100 ◦ , as wellas those which do not pass the “Source” photon data se-lection. We also exclude events that were observed whilethe instrument was not in science survey mode, when theinstrumental rocking angle was > ◦ , or when the instru-ment was passing through the South Atlantic Anomaly.For each source, we examine the photons observed withina 10 ◦ x10 ◦ box centered around the location of the source,and divide the photons into 100x100 angular bins and 24evenly spaced logarithmic energy bins. We analyze the MET range: 239557417 - 447078115
FIG. 1: A comparison of the values of the test statisticfound in our analysis (TS HL ) to those given in Fermi’s 3FGLcatalog (TS ). Overall the agreement between these twodeterminations is good, validating our analysis procedure. instrumental exposure throughout the region of interestusing the P7REP_SOURCE_V15 instrument response func-tions. We employ the latest model for diffuse galacticgamma-ray emission ( gll_iem_v05_rev1.fit ) and thelatest isotropic emission template for the Source pho-ton events ( iso_source_v05.txt ), and include all pointsources given in the 3FGL Catalog [28].In our analysis, we follow a prescription as similaras possible to that employed by the Fermi-LAT col-laboration [16, 18, 26]. Specifically, we first set theglobal normalization of background sources and the dwarfspheroidal over the entire 500 MeV – 500 GeV energyrange, utilizing the Fermi-LAT gtlike code and the
MINUIT algorithm. In this phase, we seed the dwarfspheroidal spectrum as a simple power-law with an in-dex of -2.0. We then fix the normalization of all back-ground components, and employ the pyLikelihood pack-age to scan the flux of the dwarf spheroidal in each energybin, calculating the delta-log-likelihood ( ∆ LG ( L ) ) as afunction of the source flux. In order to calculate the TSfor a given dark matter model, we minimize the total log-likelihood summed over all energy bins after constrainingthe photon flux by the spectral shape of the dark mattermodel.To validate the results of this method, we perform twotests. First, we randomly select 100 Fermi-LAT 3FGLpoint sources [28] with | b | > ◦ , a “curve significance”smaller than 2 (indicating consistency with a power-lawspectrum) and a TS smaller than 100 in the energy rangeof 300 MeV to 100 GeV. In order to compare our resultsto those given in the 3FGL catalog, we employ the abovetechnique with the following modifications. We restrict FIG. 2: The fraction of “blank sky” locations with a teststatistic (TS) larger than a given value, as empirically deter-mined for a collection of 1905 randomly selected sky locationsconstrained to lie at a galactic latitude | b | > ◦ and at least ◦ ( ◦ ) from point-like (extended) 3FGL sources [28]. For theblue curve, no additional requirements are placed on the blanksky locations. For the red curve, the blank sky locations usedare additionally required to lie no closer than . ◦ from anysource listed in the BZCAT, CRATES, CGRaBS, or ATNFcatalogs (see Sec. IV). The shaded region surrounding eachcurve represents the poisson errors on this determination. Ingenerating this figure, we have adopted a spectral shape cor-responding to a 49 GeV dark matter particle annihilating to b ¯ b (corresponding to the best-fit mass for the Galactic Centergamma-ray excess [9]). our analysis to four years of Fermi-LAT data, evaluatean energy range of 300 MeV – 100 GeV in 20 energy binsutilizing 8 energy bins per decade except for the final bin(which was extended to an energy of 100 GeV), and wescan the likelihood fits using power-law, rather than darkmatter motivated, spectral shapes. In Fig. 1 we show thedistribution of the TS calculated in our analysis (TS HL )compared to that obtained by the Fermi-LAT collabora-tion (TS ) in the same energy range. We find that ourTS values are, on average, slightly (13.5%) lower thosereported in the 3FGL. We attribute this primarily to thefact that we normalize the background by fitting over a10 ◦ × ◦ region, rather than over the entire sky. Thedashed curve in Fig. 1 represents the best-fit gaussian ofthis distribution, with a mean of -0.135 and a standarddeviation of 0.176.Secondly, we apply the “blank-sky” null-test employedin previous dwarf spheroidal studies. Specifically, we se-lect 1905 sky locations with | b | > ◦ , which are 1 ◦ re-moved from any 3FGL source and 5 ◦ removed from anyextended 3FGL source. In this case, we employ the full MET range: 239557417 - 365467563
FIG. 3: The log-likelihood fit of Reticulum II in 24 energybins spanning 500 MeV to 500 GeV. The upper limits corre-spond to 2 σ confidence in each energy range. The white linecorresponds to the best fit from a 49 GeV dark matter particleannihilating to b ¯ b . b ¯ b (corresponding to the best-fit value ofthe mass for the spectrum of the Galactic Center ex-cess [9]). In Fig. 2 we show the resulting distribution ofour blank-sky test locations. While the existence of sys-tematic errors in the modeling of the gamma-ray back-ground drives this distribution far from that expectedfrom Poisson variations, the result is in good agreementwith all previous studies. In this figure, we show resultscorresponding to the case in which no additional require-ments are placed on the blank sky locations (blue), andto when the blank sky locations used are further requiredto lie no closer than . ◦ from any source listed in the BZ-CAT, CRATES, CGRaBS, or ATNF catalogs (red). Thiswill be discussed in more detail in Sec. IV. III. RESULTS
In Fig. 3, we show the delta-log-likelihood ( ∆ LG ( L ) )distribution for our analysis of Fermi data from the direc-tion of Reticulum II. As in both Ref. [26] and Ref. [27],we find an excess of events in the bins covering approx-imately ∼ b ¯ b (thebest-fit mass for the Galactic Center excess [9]), we finda value of TS=17.4 from Reticulum II, corresponding toa significance of 3.2 σ (see Fig. 2). If we do not imposethis choice of the dark matter mass, but rather allow themass to float as a free parameter, the value of the TSincreases only slightly (to 18.1), illustrating the compat-ibility between this signal and that observed from the log ( J ) TS (Point-Like) TS (NFW-Like)Dwarf Name Distance (kpc) Latitude ( ◦ ) Ref. [29] Ref. [30] m DM = 49 GeV any m DM m DM = 49 GeV any m DM Reticulum II 32 (30) -49.7 – – 17.4 18.1 – –Tucana II 58 (69) -52.4 – – 1.44 1.82 – –Indus I 69 (100) -42.1 – – 0.0 0.0 – –Horologium I 87 (79) -54.7 – – 0.09 0.17 – –Phoenix II 95 (83) -59.7 – – 0.0 0.55 – –Eridanus III 95 (87) -59.6 – – 0.0 0.53 – –Pictoris I 126 (114) -40.6 – – 0.0 0.0 – –Grus 1 (120) -58.8 – – 0.0 0.40 – –Eridanus II 330 (380) -51.6 – – 0.0 0.61 – –Triangulum II 30 -23.8 – – 0.0 0.0 – –Canis Major 7 -8.0 – – 0.0 0.0 – –Segue 1 23 50.4 19.5 ± . +0 . − . ± . +0 . − . . +1 . − . ± ± . +0 . − . ± . +0 . − . ± . +0 . − . ± . +0 . − . ± . +0 . − . ± . +0 . − . ± . +0 . − . ± . +0 . − . ± . +0 . − . ± . +0 . − . ± . +1 . − . ± . +0 . − . . +0 . − . ± . +0 . − . ± . +0 . − . ± . +0 . − . TABLE I: The distance, galactic latitude, J -factors, and test statistic (TS) of any gamma-ray excess from DES’ nine newlydiscovered dwarf galaxy candidates (top), the new Pan-STARRS dwarf candidate (middle), and the 25 previously known MilkyWay dwarf galaxies (bottom). For each of the DES dwarf candidates, we list distances as reported in Ref. [23] (and as reportedin Ref. [24]). For Triangulum II, we list the distance as reported in Ref. [25]. For the known dwarf galaxies, distances are asgiven in Refs. [29, 30]. Although we list only central values for distances, the error bars on these quantities are typically onthe order of ± (10 − . The J -factors are averaged over a . ◦ radius around each dwarf, as reported by Refs. [29] and [30],respectively, and are given in units of GeV cm − . The TS values listed for each dwarf assume either a spectral shape thatcorresponds to a 49 GeV dark matter particle annihilating to b ¯ b (corresponding to the best-fit mass for the Galactic Centergamma-ray excess [9]) or allowing the dark matter mass to be a free parameter. For those dwarf galaxies constrained by stellarkinematics, we also show the TS values found when the source is treated as a spatially extended object, corresponding to anNFW halo with a scale radius equal to central value reported in Refs. [18, 29]. The most significant detection is from the newlydiscovered and nearby dwarf galaxy candidate Reticulum II. Galactic Center.In Table I, we list the TS values found in our analy-sis for each of the previously known Milky Way dwarfspheroidal galaxies, and for the ten newly discovereddwarf galaxy candidates. Values are given assuming ei-ther a spectrum corresponding to the best-fit mass forthe Galactic Center excess, or for any dark matter mass.For a Galactic Center-like spectrum, Reticulum II yieldsthe highest significance (TS=17.4), followed by Willman1 (3.94), Hercules (3.09), Sagittarius (2.13), Tucana II(1.44), and Segue 1 (1.07). Other than that from Reticu- lum II, no statistically significant excesses are observed.Also given in Table I are the values of the J -factorsfor each dwarf galaxy with sufficient kinematic informa-tion (from spectroscopic data) to obtain a determination.This quantity is defined as follows: J = (cid:90) ∆Ω (cid:20) (cid:90) los ρ dl (cid:21) , (1)where ∆Ω is taken to be a circle of . ◦ radius around thegiven dwarf, ρ describes the dark matter density profileof the dwarf, and the second integral is performed over FIG. 4: The test statistic (TS) as a function of the assumed scale radius of the (NFW) dark matter halo (in degrees, tan θ s ≡ r s /D ) for five selected dwarf galaxies. The blue (green) curves correspond to a spectral shape for the best-fit darkmatter mass (for m DM = 49 GeV) annihilating to b ¯ b . the observed line-of-sight (los). We provide the J -factorvalues as reported by two groups: Martinez et al. [29] andGeringer-Sameth et al. [30]. In general, the dark matterprofiles of the classical dwarfs are well constrained by stel-lar kinematics, resulting in relatively well determined J -factors. In contrast, the ultra-faint dwarfs (Segue 1, UrsaMajor II, Willman 1) contain far fewer stars, and exhibitmuch larger J -factor error bars. Deeper measurements,capable of detecting more numerous faint stars, will ulti-mately improve this situation. Although no spectroscopicinformation exists for any of the ten new dwarf galaxycandidates, we expect such follow-up measurements tooccur in the near future.For those dwarfs with profiles constrained by stellarkinematics, we also list in Table I the values of the TSfound when the source is treated as a spatially extendedobject, rather than as a point-like source. In particular,we adopt an NFW-profile for these systems, with a scaleradius equal to the central value reported in Refs. [18, 29].In Fig. 4, we plot the TS as a function of the halo’s scaleradius (in degrees, tan θ s ≡ r s /D ) for five of the dwarfgalaxies under consideration. No strong evidence for (oragainst) spatial extension is observed. The significances of Willman 1 and Ursa Major II marginally increase ifan extended halo is assumed, while the significances ofReticulum II and Sagittarius marginally decrease. IV. CONTROLLING BACKGROUNDS WITHMULTI-WAVELENGTH SOURCE CATALOGS
In Ref. [31], it was pointed out that Fermi’s sensitivityto dark matter annihilation in dwarf spheroidal galaxiescould be increased by taking into account informationavailable in multi-wavelength source catalogs. In partic-ular, a significant fraction of the highest TS points inthe “blank sky” correspond to the locations of unresolvedblazars, radio galaxies, and starforming galaxies. Bymaking use of only regions of the “blank sky” which arenot near sources listed in multi-wavelength catalogs, it ispossible to reduce the contamination from such sources.In Fig. 2, the TS distribution of the high-latitudeblank-sky is shown without utilizing multi-wavelength in-formation (blue), and after avoiding all locations locatedwithin . ◦ of any source listed in the Roma-BZCATMulti-Frequency Catalog of Blazars (BZCAT) [32], theCombined Radio All-Sky Targeted Eight-GHz Survey(CRATES) catalog [33], the Candidate Gamma-RayBlazar Survey (CGRaBS) catalog [34], or the AustraliaTelescope National Facility (ATNF) pulsar catalog [35](red). The application of this cut significantly reducesthe fraction of the sky with large TS values.To take this multi-wavelength information into ac-count, we apply the following procedure in our analysis.For a given dwarf galaxy (or dwarf galaxy candidate), wecheck the catalogs described in the previous paragraphfor any sources located within . ◦ . If any are found, were-run our analysis, including in the background modela source at that location. We then take the new TS ofthe dwarf, and see how many locations on the blank skyyield a higher TS, when a background source is includedat the location of the nearby catalog source. We then usethe fraction of high-TS blank-sky locations to calculatethe p -value and significance of any dwarf galaxy excess.This procedure is most important in the case of Retic-ulum II, which is located 0.44 ◦ from the source CRATESJ033553-543025. Given the large number of sources con-tained in these catalogs, this is not a particularly sur-prising (we estimate a probability of ∼
20% that at leastone source would reside within . ◦ ). The analysis withthe extra background source at the CRATES locationresulted in a TS of 13.4 from Reticulum II (a modestreduction from 17.4). After re-running our analysis onall “blank-sky” locations with TS>13.4, including in thebackground model sources at the locations of the multi-wavelength catalog sources, we find that only 3 out of1905 of the blank-sky locations yielded a more significantexcess. This corresponds to a p − value of 0.001575 and adetection significance of 3.2 σ (compared to 3.0 σ , whichis found if no multi-wavelength information is utilized).If we include the poisson error bars around the numberof 3 blank-sky locations, the corresponding significancecovers the range of 3.0 to 3.4 σ .As we were finalizing this paper, it was pointed outto us that the faint radio source PMN J0335-5406 in theParkes-MIT-NRAO catalog is located . ◦ from thelocation of Reticulum II. The PMN catalog contains amuch larger number of sources (50,814) than the catalogsemployed in our analysis, and it is not particularly sur-prising that a such source resides near Reticulum II (ornear any other location in the high-latitude sky). Specifi-cally, at the declination of Reticulum II, the PMN surveycontains approximately 17 sources per square degree, cor-responding to a 17% chance that a catalog source exists The presence of this source was pointed out to us by Eric Carlson. We note that the spectral shape absorbed by the CRATES sourceis very hard, and quite unlike that of the gamma-ray emissionfrom typical radio sources. We consider it unlikely that thissource contributes significantly to the gamma-ray flux observedfrom the direction of Reticulum II. We thank Alex Drlica-Wagner for bringing this source to ourattention.
FIG. 5: The test statistic (TS) over the region of the skysurrounding Reticulum II. Although the excess gamma-rayemission is localized to a region of approximately ∼ . ◦ ra-dius, it is not possible to spatially discriminate ReticulumII (black square) from the location of the radio source PMNJ0335-5406 (blue oval). within a 0.1 ◦ ROI around Reticulum II. In Fig. 5, weplot the value of the TS found by our analysis for re-gions of the sky in the vicinity of Reticulum II. Althoughwe can localize the excess gamma-ray emission to a re-gion of approximately ∼ . ◦ radius, it is not possible tospatially discriminate Reticulum II (black square) fromthe location of the PMN source (blue oval). The PMNsource exhibits continuous radio emission that is consis-tent with a blazar origin. When re-running our analysisfor a power-law gamma-ray spectrum (as is observed fromblazars) we find a best-fit spectral index of 2.04, yieldingTS=13.1. In comparing this to the TS value found earlierin this study, we conclude that a dark matter-like spec-trum is preferred to a blazar-like power-law at the level ofapproximately ∼ σ . Spectroscopic observations of thissource would be valuable, as they could aid in determin-ing its nature and inform us as to its likely gamma-rayluminosity and spectrum. V. A SELF-CONSISTENT INTERPRETATION
In this section, we consider the excess observed fromReticulum II, along with the lack of significant detectionsfrom other Milky Way dwarf galaxies, and ask whetherthese results are mutually consistent. Focusing on thecase in which annihilating dark matter is responsible forthe Galactic Center excess ( m DM = 49 GeV, for thecase of annihilatiions to b ¯ b ), the lack of significant ex-cess emission from the known dwarf galaxies constrains σv < ∼ . × − cm /s [16]. And while this constraintis compatible with dark matter interpretations of theGalactic Center excess, the normalization of the Galac-tic Center signal implies that the cross section is unlikelyto be smaller than this value by more than a factor of afew. More specifically, if we allow the overall normaliza-tion and the scale radius of the Milky Way’s dark matterhalo profile to vary within the range allowed by dynami-cal constraints ( ρ local = 0 . − . GeV/cm , r s = 8 − kpc [36, 37]), we find consistency with an annihilationcross section as small as ∼ × − cm /s. From thisperspective, the prospects for the future detection of agamma-ray signal from one or more dwarf galaxies ap-pears encouraging.For an annihilation cross section at the upper limit ofRef. [16] ( σv (cid:39) . × − cm /s), the normalizationof the signal from Reticulium II requires approximately log J (cid:39) . –20.1. Noting the empirical (and approx-imate) relationship between the distances and J -factorsof ultra-faint dwarfs, Ref. [26] points out that ReticuliumII might be expected to have a somewhat smaller value, log J ∼ . , although even a value as high as of 20.1would not be a particuarly significant outlier. The ne-cessity of a large J -factor for Reticulum II (if its gamma-ray excess is from annihilating dark matter) can also beseen from the results of our analysis, as shown in Ta-ble I. Roughly speaking, the predicted value for the TS ofa given dwarf is proportional to its gamma-ray flux, andthus to its J -factor. The modest TS values observed fromSegue 1 and Ursa Major II suggest significantly lower J -factors for these systems than for Reticulum II. Giventhis situation, we eagerly await the spectroscopic follow-up of Reticulium II. If the gamma-ray excess from thissource in fact originates from annihilating dark matter,we should anticipate a large value for its J -factor, likelyin excess of ∼ . GeV cm − . VI. ARE STATISTICAL FLUCTUATIONSSUFFICIENT TO EXPLAIN THE DIFFERENCESBETWEEN THE RESULTS FOUND USING PASS7 AND PASS 8 DATA?
At face value, our determination of TS=18.1 from thedirection of Reticulum II appears to be in conflict withthe more modest value of 6.7 quoted by the Fermi Col-laboration [26]. The most significant difference betweenthese two analyses is in the data sets that are being con-sidered: our analysis makes use of the publicly avail-able Pass 7 data, whereas the Fermi Collaboration pa-per utilizes the more recent Pass 8 data set. Intrigu-ingly, a similar discrepancy can be seen in a compar-ison of the Fermi’s Collaboration’s Pass 7 and Pass 8studies of known dwarf galaxies. In particular, whereasthe Fermi Collaboration’s Pass 7 analysis revealed an ex-cess from three ultra-faint dwarfs (Segue 1, Ursa MajorII, and Willman 1), at a level of TS ≈ [18], no ex-cess was observed according to their more recent Pass8 paper [16]. This would-be conflict between the FermiPass 7 and Pass 8 dwarf papers is somewhat surprising inlight of the fact that these two analyses make use of data taken over significantly overlapping time periods (overthe first four years of Fermi’s mission; the more recentanalysis adds two more years of data to this set). TheFermi Collaboration points out, however, that after tak-ing into account the new event selection associated withthe transition from Pass 7 to Pass 8, the overlap betweenthese two data sets is not particularly large; only ∼ ≈
10 excesswas the result of a statistical fluctuation, we estimatethat there was an approximately ∼
10% chance that themore recent data set would yield TS < (after takinginto account differences in effective area and exposuretime). So in this respect, we concur with the conclusionof the Fermi Collaboration.In order to compare our results directly to thoseof Ref. [18], we reanalyzed the three ultra-faint dwarfspheroidal galaxies (Segue 1, Willman 1, and Ursa Ma-jor II) using the P7V6 Reprocessed data, employingin this test only four years of data and utilizing onlythose events which pass the P7V6REP “Clean” selectioncut (and the gll_iem_v05.fit and iso_clean_v05.txt models for the diffuse and isotropic emission, respec-tively), and examine the data a in 14 ◦ × ◦ regionaround each ultra-faint dwarf in 0.1 ◦ spatial bins. As-suming a spectrum corresponding to a 25 GeV dark mat-ter particle annihilating to b ¯ b , we find that these threedwarfs acquire TS values of 1.18, 2.43, and 2.43, respec-tively. If we model these sources as extended objects(NFW-like) rather than as point sources, we find TS val-ues of 0.89, 3.05, and 2.73. In either case, our TS valuesare similar to but somewhat lower than the TS ≈
10 re-ported in Ref. [18]. Comparing this to the results shownin Table I, however, we see little indication that the addi-tion of 2.5 years of data has reduced the overall TS fromthese three ultra-faint dwarfs (Ursa Major II’s TS fell,while that of Willman 1 increased by a similar amount).In light of this somewhat confusing situation, we eagerlyawait the public availability of Pass 8 data.
VII. SUMMARY AND CONCLUSIONS
In this article, we have revisited the gamma-ray emis-sion from known Milky Way dwarf galaxies, and fromthe dwarf galaxy candidates recently discovered in thedata from DES and Pan-STARRS. Of particular interestare the new dwarf candidates Reticulum II and Triangu-lum II, which are each located at a distance of only ∼ MET range: 239557417 – 365817602 rection of Reticulum II identifies an excess of gamma-rayswith a local statistical significance of 3.2 σ . This is slightlyhigher than, but not dissimilar to, that reported in theprevious studies of other groups [26, 27]. We also confirmthat Reticulum II’s gamma-ray excess is most prominentat energies between ∼ γ -ray emission from the directionof Triangulum II.Looking forward, spectroscopic follow-up of ReticulumII and the other new dwarf galaxy candidates will be im-portant for interpreting this data. In order for this excessto be compatible with the lack of significant gamma-raydetections from other dwarf galaxies (most importantly,Segue 1 and Ursa Major II), Reticulum II must contain ahigh density of dark matter, corresponding to J > ∼ . GeV cm − . A measurement of Reticulum II’s J -factorthat is much smaller than this value would place seri-ous doubt as to any dark matter interpretation of its excess. Additional data from Fermi will also have muchto bear on this question. With 50% more data, such ascould be acquired over the next few years, we estimatethat the detection of Reticulum II’s gamma-ray emissioncould exceed TS=25, corresponding to approximately 4 σ significance, and comparable to the threshold for mem-bership in Fermi’s point source catalogs. Acknowledgements.
We would like to thank An-drea Albert and Alex Drlica-Wagner for helpful com-ments and discussions, as well as Eric Carlson for point-ing out the proximity of the nearby CRATES source.DH is supported by the US Department of Energy un-der contract DE-FG02-13ER41958. Fermilab is operatedby Fermi Research Alliance, LLC, under Contract No.DE- AC02-07CH11359 with the US Department of En-ergy. TL is supported by the National Aeronautics andSpace Administration through Einstein Postdoctoral Fel-lowship Award No. PF3-140110. [1] L. Goodenough and D. Hooper (2009), 0910.2998.[2] D. Hooper and L. Goodenough, Phys.Lett.
B697 , 412(2011), 1010.2752.[3] D. Hooper and T. Linden, Phys.Rev.
D84 , 123005(2011), 1110.0006.[4] K. N. Abazajian and M. Kaplinghat, Phys.Rev.
D86 ,083511 (2012), 1207.6047.[5] C. Gordon and O. Macias, Phys.Rev.
D88 , 083521(2013), 1306.5725.[6] D. Hooper and T. R. Slatyer, Phys.Dark Univ. , 118(2013), 1302.6589.[7] K. N. Abazajian, N. Canac, S. Horiuchi, and M. Kapling-hat, Phys.Rev. D90 , 023526 (2014), 1402.4090.[8] T. Daylan, D. P. Finkbeiner, D. Hooper, T. Linden,S. K. N. Portillo, et al. (2014), 1402.6703.[9] F. Calore, I. Cholis, and C. Weniger (2014), 1409.0042.[10] I. Cholis, D. Hooper, and T. Linden (2014), 1407.5625.[11] D. Hooper, I. Cholis, T. Linden, J. Siegal-Gaskins, andT. Slatyer, Phys.Rev.
D88 , 083009 (2013), 1305.0830.[12] J. Petrovic, P. D. Serpico, and G. Zaharijas, JCAP ,023 (2015), 1411.2980.[13] J. Petrovic, P. D. Serpico, and G. Zaharijas, JCAP ,052 (2014), 1405.7928.[14] E. Carlson and S. Profumo, Phys.Rev.
D90 , 023015(2014), 1405.7685.[15] F. Calore, I. Cholis, C. Evoli, D. Hooper, T. Linden, andC. Weniger, in preparation (2015).[16] M. Ackermann et al. (Fermi-LAT) (2015), 1503.02641.[17] A. Geringer-Sameth, S. M. Koushiappas, and M. G.Walker (2014), 1410.2242.[18] M. Ackermann et al. (Fermi-LAT Collaboration),Phys.Rev.
D89 , 042001 (2014), 1310.0828.[19] C. He, K. Bechtol, A. P. Hearin, and D. Hooper (2013),1309.4780.[20] E. J. Tollerud, J. S. Bullock, L. E. Strigari, and B. Will- man, Astrophys.J. , 277 (2008), 0806.4381.[21] J. R. Hargis, B. Willman, and A. H. G. Peter, Astro-phys.J. , L13 (2014), 1407.4470.[22] B. M. Rossetto et al. (Dark Energy Survey Collabora-tion), Astron.J. , 185 (2011), 1104.4718.[23] K. Bechtol et al. (The DES Collaboration) (2015),1503.02584.[24] S. E. Koposov, V. Belokurov, G. Torrealba, and N. W.Evans (2015), 1503.02079.[25] B. P. Laevens et al. (2015), 1503.05554.[26] A. Drlica-Wagner et al. (Fermi-LAT, DES) (2015),1503.02632.[27] A. Geringer-Sameth, M. G. Walker, S. M. Koushiappas,S. E. Koposov, V. Belokurov, et al. (2015), 1503.02320.[28] F. Acero et al. (Fermi-LAT Collaboration) (2015),1501.02003.[29] G. D. Martinez (2013), 1309.2641.[30] A. Geringer-Sameth, S. M. Koushiappas, and M. Walker,Astrophys.J. , 74 (2015), 1408.0002.[31] E. Carlson, D. Hooper, and T. Linden, Phys.Rev.
D91 ,061302 (2015), 1409.1572.[32] E. Massaro, P. Giommi, C. Leto, P. Marchegiani,A. Maselli, et al. (2008), 0810.2206.[33] S. E. Healey, R. W. Romani, G. B. Taylor, E. M. Sadler,R. Ricci, et al., Astrophys.J.Suppl. , 61 (2007), astro-ph/0702346.[34] S. E. Healey, R. W. Romani, G. Cotter, P. F. Michelson,E. F. Schlafly, et al., Astrophys.J.Suppl. , 97 (2008),0709.1735.[35] G. Hobbs, R. Manchester, A. Teoh, and M. Hobbs, IAUSymp. , 139 (2004), astro-ph/0309219.[36] F. Iocco, M. Pato, G. Bertone, and P. Jetzer, JCAP ,029 (2011), 1107.5810.[37] R. Catena and P. Ullio, JCAP8