A more probable explanation for a continuum flash in the direction of a redshift \approx 11 galaxy
Charles L. Steinhardt, Michael I. Andersen, Gabriel B. Brammer, Lise Christensen, Johan P. U. Fynbo, Bo Milvang-Jensen, Pascal A. Oesch, Sune Toft
DDraft version February 23, 2021
Typeset using L A TEX twocolumn style in AASTeX63
A more probable explanation for a continuum flash in the direction of a redshift ≈
11 galaxy
Charles L. Steinhardt,
1, 2
Michael I. Andersen,
1, 2
Gabriel B. Brammer,
1, 2
Lise Christensen,
1, 2
Johan P. U. Fynbo,
1, 2
Bo Milvang-Jensen,
1, 2
Pascal A. Oesch,
1, 2, 3 and Sune Toft
1, 2 Cosmic Dawn Center (DAWN) Niels Bohr Institute, University of Copenhagen, Jagtvej 128, København N, DK-2200, Denmark University of Geneva, Department of Astronomy, Chemin Pegasi 51, 1290 Versoix, Switzerland
ABSTRACTRecent work reported the discovery of a gamma-ray burst (GRB) associated with the galaxy GN-z11 at z ∼
11. The extreme improbability of the transient source being a GRB in the very earlyUniverse requires robust elimination of all plausible alternative hypotheses. We identify numerousexamples of similar transient signals in separate archival MOSFIRE observations and argue that Solarsystem objects — natural or artificial — are a far more probable explanation for these phenomena.An appendix has been added in response to additional points raised in Jiang et al. (2021), which donot change the conclusion. INTRODUCTIONJiang et al. (2020) recently reported a transient sourcein MOSFIRE (McLean et al. 2008) K band slit spec-troscopy of GN-z11 (Oesch et al. 2016), the highest-redshift galaxy yet observed. They interpret this tran-sient as the possible discovery of a UV flash from agamma-ray burst (GRB) at z = 10 . §
2, a MOSFIRE detection of a Solar systemobject or spacecraft is a far more likely explanation. TRANSIENT SIGNALS FROM SOLAR SYSTEMOBJECTS IN MOSFIRE SPECTRATo date MOSFIRE has obtained 81,776 spectroscopicexposures of science targets. If any other exposuresshow similar transients as that reported for GN-z11,then the chance probability of finding such a transientin a single random exposure, regardless of its origin, is > times more likely than a serendipitous GRB de-tection with probability (cid:28) − . In a visual search of12,300 exposures from the Keck
MOSFIRE archive wefind a minimum of 27 single-exposure transients in thefiles summarized in Table 1.Three of these transients appear in a sequence of 12exposures (Fig. 1) taken on 7 Mar 2017 centered on thelensing cluster Abell 1689 (A1689), and two of thoseare seen in a single 120 s exposure (Fig. 2). As withthe GN-z11 flash, this pair of A1689 transient spectrashows telluric absorption features indicating their originabove the atmosphere. Both transients in the pair havethe same spectrum (Fig. 3) suggesting that they arise https://koa.ipac.caltech.edu/UserGuide/mosfire.csv from a single source moving with a minimum angularspeed > . (cid:48)(cid:48) β absorption feature.The steeper slope of the GN-z11 transient in the K band is also consistent with a reflected Solar spectrum(Fig. 3). It is a featureless K band spectrum consistentwith a powerlaw f λ ∝ λ − . ± . , which is significantlysteeper than regularly found for lower-redshift GRB af-terglows that have observed UV-to-optical slopes in therange between − . − . –10 keV in therest-frame (Toma et al. 2011), i.e. the thermal compo-nent peaks in X-ray emission. Jiang et al. (2020) arguedthat a slope of − . a r X i v : . [ a s t r o - ph . H E ] F e b Figure 1.
Twelve raw MOSFIRE spectra of an Abell 1689 field taken in sequence, with each spectrum having an exposuretime of 120 seconds. Only a smaller part of each spectrum is shown. A “flash”, i.e. a trace that only appears in a singleexposure, is seen in the 3rd exposure (
MF.20170307.46851.fits ) in the top slit, and a double flash is seen in the 11th exposure(
MF.20170307.48096.fits ) in the two slits below.
Figure 2.
A double “flash” seen by MOSFIRE on 7 Mar2017 while observing the Abell 1689 field. Shown are twosequential J band difference images of the B − A nod po-sitions where the B exposures were taken just 2.5 minutesapart. The white and black bars indicate the characteristicpositive and negative spectra of a bright alignment star seenin both difference images. The flashes indicated by the bluebars appear in a single 120 s exposure. in the NASA HORIZONS database of natural bodies(e.g., asteroids, comets) and select spacecraft. Based onthe spectral similarity to the reflected Solar spectrumand the angular speed of the moving sources shown in https://ssd.jpl.nasa.gov/?horizons Fig. 4, the transients detected in MOSFIRE are mostlikely high-altitude Earth-orbiting satellites.Jiang et al. (2020) argue that the probability of suchan interloper is very low based on the telescope pointingand assumed distribution of satellite orbits. However, 8additional transients were found among ∼ Hubble Space Telescope images seen fromlow-Earth orbit as to motivate the development of auto-mated software to detect and remove them (Borncamp &Lian Lim 2019)—to say nothing of the constellations oflow-Earth orbit satellites now regularly photo-bombingboth professional and amateur images from the ground(Tyson et al. 2020). Identifying the specific satellitesthat cause the trails in the MOSFIRE exposures sum-marized in Table 1 is beyond the scope of this work, andthe calsky online database Jiang et al. (2020) used tosearch for satellite counterparts to the GN-z11 transientis no longer available to make that specific comparison.Jiang et al. (2020) also argue that time of day andcompactness of the GN-z11 flash are impossible for a f , a r b i t r a r y Scaled Solar Spectral Irradiancesec z = 0 (Meftah+2018) "Flash" f , =-3.5 Figure 3.
Top:
Telluric-corrected J band spectra of the two flashes shown in Fig. 2. The black line shows the Solar spectrumfrom Meftah et al. (2018) normalized to the flash spectra. Lower left:
The flash spectra show an absorption feature consistentwith the wavelength and depth of the Paschen- β line in the Solar spectrum. Lower right:
Expanded near-infrared view of theSolar spectrum, which has a power-law slope β ≈ − . K band (shaded grey region). Flash 1 Flash 2 Flash 3
A1689 Double “Flash”
MF.20170307.48096.fits Triple “Flash”
MF.20121125.24716.fits
Flash 1Flash 2
Figure 4.
Slit mask layouts for two of the MOSFIRE frames with multiple detected flashes in one exposure. The positionof these flashes (blue circles) is consistent with a source moving on a straight trajectory that crosses multiple slits in a singleexposure (blue line). This strongly suggests that such transients occur due to Earth-orbiting satellites often seen in imagingobservations. man-made satellite. However, examples exist with bothproperties. The exposure
MF.20140425.29736.fits , inGOODS-North, shows an H band transient obtainedat UT 08:15 (the GN-z11 transient was observed atUT 08:07). The NAVSTAR 52 (USA 168, NORAD ID27704) satellite, with an orbital inclination angle of 54degrees, is in the approximate vicinity of the GN-z11flash on the night of observations, with a direction oftravel close to east-west. A satellite with a similar or- bit could have produced a compact flash like the onereported. DISCUSSIONThe goal of this analysis is to assign the most probableexplanation to the possible UV flash reported by Jianget al. (2020) in GN-z11. In terms of brightness the ob-servations are not incompatible with being related to aGRB (Kann et al. 2020). However, the spectral shapeis significantly bluer than what is typical for GRB af-terglows, and more similar to those shown in Fig. 2in MOSFIRE observations of Abell 1689, likely naturalSolar system objects or spacecraft.As estimated in Jiang et al. (2020), the probabilitymerely of detecting a GRB is between ∼ − and 10 − under ΛCDM, and the GN-z11 flash does not have theSED of a typical GRB. Although Kann et al. (2020)argue that such an SED is plausible, given current cat-alogs, the probability of finding one is (cid:46) − less likelythan merely finding a GRB. So, the probability of de-tecting a GRB of this type is (cid:46) − .10 − is small enough that many other improbable ex-planations are far more likely. From an archival searchof MOSFIRE observations, the rate of natural bodiesand/or spacecraft causing flashes in dispersed spectrais of order 10 − per exposure, making this explanation (cid:38) times more likely. Due to the rate of theseinterlopers at other observatories, even in the absenceof these examples , this was the most probable explana-tion. At least three of these transients, including the onein the GN-z11 slit, have spectra indistinguishable from a (reflected) Solar spectrum, further supporting Earth-orbiting satellites as their most probable cause.ACKNOWLEDGMENTSCS is supported by ERC grant 648179 ”ConTExt”.The Cosmic Dawn Center (DAWN) is funded by theDanish National Research Foundation under grant No.140. BMJ is supported in part by Independent Re-search Fund Denmark grant DFF - 7014-00017. JPUFacknowledges support from the Carlsberg foundation.This research has made use of the Keck ObservatoryArchive (KOA), which is operated by the W. M. KeckObservatory and the NASA Exoplanet Science Institute(NExScI), under contract with the National Aeronauticsand Space Administration. The authors wish to recog-nize and acknowledge the very significant cultural roleand reverence that the summit of Maunakea has alwayshad within the indigenous Hawaiian community. We aremost fortunate to have the opportunity to conduct ob-servations from this mountain.REFERENCES Blake J. A., et al., 2021, Advances in Space Research, 67,360Borncamp D., Lian Lim P., 2019, in Molinaro M.,Shortridge K., Pasian F., eds, Astronomical Society ofthe Pacific Conference Series Vol. 521, AstronomicalData Analysis Software and Systems XXVI. p. 491Campana S., et al., 2006, Nature, 442, 1008Corbett H., et al., 2020, ApJL, 903, L27Holm S., 1979, Scandinavian journal of statistics, 6, 65Jakobsson P., Hjorth J., Fynbo J. P. U., Watson D.,Pedersen K., Bj¨ornsson G., Gorosabel J., 2004, ApJL,617, L21Japelj J., et al., 2015, A&A, 579, A74Jiang L., et al., 2020, Nature Astronomy, arXiv:2012.06937Jiang L., et al., 2021, arXiv e-prints, p. arXiv:2102.01239Kann D. A., Blazek M., de Ugarte Postigo A., Th¨one C. C.,2020, Research Notes of the American AstronomicalSociety, 4, 247Kramer C. Y., 1956, Biometrics, 12, 307 Li L., et al., 2015, ApJ, 805, 13McLean I. S., Steidel C. C., Matthews K., Epps H., AdkinsS. M., 2008, in McLean I. S., Casali M. M., eds, Societyof Photo-Optical Instrumentation Engineers (SPIE)Conference Series Vol. 7014, Ground-based and AirborneInstrumentation for Astronomy II. p. 70142Z,doi:10.1117/12.788142Meftah M., et al., 2018, A&A, 611, A1Nir G., Ofek E. O., Gal-Yam A., 2021, arXiv e-prints, p.arXiv:2102.04466Oesch P. A., et al., 2016, ApJ, 819, 129ˇSid´ak Z., 1967, Journal of the American StatisticalAssociation, 62, 626Toma K., Sakamoto T., M´esz´aros P., 2011, ApJ, 731, 127Tukey J. W., 1949, Biometrics, 5, 99Tyson J. A., et al., 2020, AJ, 160, 226Uhm Z. L., Zhang B., 2014, Nature Physics, 10, 351
Keck Archive Filename Grating UTGOODS-North
MF.20130114.54832.fits
K 15:13
MF.20130215.56333.fits
K 15:38
MF.20130322.54354.fits
H 15:05
MF.20140425.29736.fits
H 08:15
MF.20150110.55808.fits
K 15:30
MF.20160102.57635.fits
H 16:00
MF.20160128.57006.fits
H 15:50
MF.20170304.45898.fits
J 12:44
MF.20170407.29239.fits † K 08:07Other
MF.20121012.46713.fits
K 12:58
MF.20121125.24716.fits ∗ H 06:51
MF.20121206.49727.fits
H 13:48
MF.20130104.31845.fits
K 08:50
MF.20131128.45592.fits
K 12:39
MF.20131224.25147.fits
J 06:59
MF.20131224.34612.fits
J 09:36
MF.20170304.24709.fits
J 06:51
MF.20170307.46851.fits
J 13:00
MF.20170307.48096.fits (cid:63)
J 13:21
MF.20170416.25674.fits
H 07:07
MF.20170416.51654.fits
H 14:20
MF.20170507.34054.fits
J 09:27
MF.20170508.41251.fits
J 11:27
MF.20170508.41560.fits
J 11:32
MF.20170508.41712.fits
J 11:35
MF.20170508.42020.fits
J 11:40
MF.20170508.49166.fits
J 13:39
Table 1.
Keck archive filenames of MOSFIRE exposureswith identified transient flashes. These files can be down-loaded directly from the Keck Observatory Archive at URLslike MF.20130114.54832.fits. Notes: † GN-z11, Jiang et al.(2020); ∗ Triple flash, Fig. 4; (cid:63)
Double flash, Figs. 1 - 4
APPENDIXFollowing the initial posting of this work, Jiang et al. (2021) raised several additional points, which are addressed here.It should be noted that Jiang et al. use these arguments to bound the probability of an interloper spacecraft at 3 × − .For comparison, Jiang et al. (2020) estimate the corresponding probability of GRB detection as (0 . − × − (this work estimates the probability as (cid:46) − ). Thus, the results in Jiang et al. (2021) are in agreement with thispaper that a spacecraft is the far more probable explanation.However, we disagree with several of the arguments for a lower spacecraft probability raised in Jiang et al. (2021).Jiang et al. argue that the GN-z11 flash is special because it coincides with a known object. However, MOSFIREslit masks are always constructed to take spectra of known objects; in that sense, every object in which a flash isfound would be special. More generally, this is a version of the well-known multiple comparison problem (Tukey 1949;Kramer 1956; ˇSid´ak 1967; Holm 1979). O ff s e t a l o n g s li t , a r c s e c Star 1 difference,
B A nod positionGNz11 Star 2
Figure 5.
Source alignment within the MOSFIRE slits. Three slits are shown in groups of three panels: an alignment star inthe second slit from the top of the detector (“S-29371”; left ), GN-z11 ( center ) and a bright alignment star in the slit just belowGN-z11 (“S-17541”; right ). In each case, the curve shows the average cross-dispersion profile of the difference image createdfrom the exposure containing the transient (nod position “B”) and the exposure preceding it (position “A”). A slice of the 2Dspectroscopic difference image is also shown, and the third, thinnest panel shows the
Hubble
F160W image at the same position,convolved with a 2D Gaussian to approximate the ground-based seeing. The transient in the GN-z11 slit is offset from thegalaxy center by 0 . (cid:48)(cid:48)
7. It is also more extended than the two point sources by a factor of ∼ They also argue that the estimate in this work must be flawed, since otherwise there should be approximately onespacecraft per 5 square degrees. This is a slight overestimate, because a single spacecraft can generate multiple flasheswithin a single observation (Fig. 4). However, with approximately tracked 5400 spacecraft with orbit perigees higherthan 1000 km , one per 5 square degrees is reasonable to within an order of magnitude and the number of artificialobjects that could cause the types of transients discussed here is almost certainly much higher still (e.g., untrackeddebris in high geosynchronous orbits; Corbett et al. 2020; Blake et al. 2021; Nir et al. 2021).Finally, Jiang et al. argue that the flash must be associated with GN-z11 because of its proximity to the center of thegalaxy. A careful analysis of the target alignment within the slits finds that the traces of two alignment stars in slitsseparated by half the instrumental field of view are well aligned with the known target positions while the transienttrace is clearly offset from the galaxy center by ∼ . (cid:48)(cid:48) z = 11. The twostars demonstrate the reliability of the absolute astrometry, and “Star 2” in the slit just below GN-z11 provides anunambiguous relative measurement of the offset. As described by Jiang et al. (2020), the transient is nearly twice asextended as the two point sources. Although the offset and source extent aren’t dispositive of the GRB interpretationof the transient, they would require yet another coincidence to support it. For example, the GRB probability argumentwould need to account for the dramatically lower stellar density at nearly 5 half-light radii ( R e = 0 . ± . Swift . The preferred explanation should instead be the mostprobable one, which both Jiang et al. (2021) and this work agree is a spacecraft.3