|GeV excess paper by Stecker et al
|Page 1 of 1|
|Author:||strong [ Fri Jun 01, 2007 2:08 pm ]|
|Post subject:||GeV excess paper by Stecker et al|
This is a comment on Stecker, Hunter and Kniffen
The GeV excess is claimed to be instrumental, based on the fact that it appears
all over the sky. This uniformity was in fact already shown by Strong et al. ApJ 613,962
What is puzzling in this paper is that the EGB is rederived and claimed consistent
with Sreekumar et al (as opposed to Strong etal ApJ 613,956 ) despite the fact that
they are claiming the EGRET response is faulty at GeV energies - the response used
by Sreekumar et al. ! This seems self-contradictory.
|Author:||tporter [ Fri Jun 01, 2007 2:30 pm ]|
It is a confusing paper. I am not sure why GLAST will be using the EGRET data for calibration (?). GLAST will probably only be using data from PAMELA for this, as far as I know.
|Author:||digel [ Tue Jun 05, 2007 1:48 am ]|
The analysis in this paper makes me uneasy - maybe because it seems kind of breezy. Maybe space was limited - the posting doesn't make clear whether this was submitted/accepted in a journal, and I have not asked Floyd et al. yet - but a more detailed presentation is needed for such an important conclusion, especially one for which a cause has not been identified.
I suspect that in Figure 1 no accounting is made for the varying solid angles of the 0.5 x 0.5 deg pixels - so effectively regions of the sky at higher latitudes count more, although of course they also tend to have lower intensities.
Also the dashed line indicating agreement between model and observation does not really run through the centroid of the distribution for fluxes less than ~0.4e-4 cm-2 s-1 sr-1. On average it looks like the observed flux (after scaling) is ~0.03e-4 cm-2 s-1 sr-1 or so higher in this range. This is not much, but it is greater than the extragalactic isotropic intensity (0.015e-4 cm-2 s-1 sr-1) that was apparently significant enough to be added to the model intensity.
Regarding the model intensities, the 0.015e-4 cm-2 s-1 sr-1 intensity added to the model does not result in any offset from the origin on the "Model Flux" axis. I guess that this implies that the model prediction has a lot of small negative intensities, not what I expected for a part of the plot that should be dominated by the high-latitude sky.
Anyone have an explanation for the 'notch' or upturn in the contours near ~0.05e-4 cm-2 s-1 sr-1 - that is, why the distribution of observed flux suddenly gets broader at low model intensities? Also, is it to be expected that the dispersion of the observed intensities should have a more or less constant value, independent of intensity? I'd have guessed that at higher intensities, the statistical fluctuations would be less, but this kind of plot is hard to interpret.
One more comment: The paper does not indicate that any modification was made to the inverse Compton component of the Hunter et al. (1997) model. My understanding was that the IC calculation was been based on what is now considered to be too small a scale height for electrons, which meant that the intensities at high latitudes and higher energies probably would tend to be underestimated.
|Author:||strong [ Tue Jun 05, 2007 2:03 am ]|
|Post subject:||another view of the matter|
Meawhile Brian Baughman et al. come to a different conclusion based on
They find the GeV excess is actually larger when the instrumental response is corrected.
|Author:||strong [ Tue Jun 05, 2007 5:03 am ]|
|Post subject:||fig 1 of Stecker etal|
Further to Seths comments on Fig 1 of Stecker etal,
it is actually an attempt to contour something which is evidently
very noisy - so it's hard to deduce much from it since it's all broken up.
A scatter plot would have been better, like in our EGB paper.
|Author:||imos [ Tue Jun 05, 2007 5:23 pm ]|
The paper looks a bit funny: it is called "The likely cause of the EGRET GeV anomaly and its implications," but does not name the "likely cause" neither discusses the "implications." Yet, the paper does not provide any new information compared to the one published by Esposito et al. in "In-Flight Calibration of EGRET on the Compton Gamma-Ray Observatory", ApJS 123 (1999) 203. Undoubtedly, the EGRET is a very complicated instrument, but it is a very vague statement to say that (see Conclusions) "Although a detailed reanalysis of the calibration data is impossible at this time, we have shown that plausible systematic uncertainties in the calibration of the EGRET sensitivity for gamma-ray energies above 1 GeV can readily account for the universal anomalous excess flux." See also a discussion in Moskalenko et al. arXiv: astro-ph/0609768.
On page 2, the authors state "The pion production cross section at these energies is also very well known"
In fact, it is not. At low energies (proton momentum below 3 GeV/c), the data used are from 1960s and 1970s and the error bars easily allow a rescaling of the predictions by a significant factor (see plots in, e.g., Dermer, A&A 157 (1986) 223 or, better, look at the original papers referenced there). At high energies (proton momentum above 7 GeV/c), the scaling approximation is used. In the intermediate region, one can use interpolation (Dermer, A&A 157  223) or event generators (e.g., Kamae et al., ApJ 647  692). However, independently on the tool used they are all tuned to the same data from 1960s and 1970s so that the resulting gamma-ray spectrum does not change much; this does not mean that it is correct. See also a discussion in Strong et al., ApJ 613 (2004) 962.
On page 2, the authors state "as we will show, the anomaly is seen at all galactic latitudes..."
This has been known for more than 3 (4?) years, see Strong et al., ApJ 613 (2004) 962. The authors, however, state that it is their "new" finding.
On page 2, the authors state "The three dimensional cosmic-ray density for the latitude range |b| < 25 was derived from the Galactic plane matter distribution on the assumption of dynamic balance."
I can imagine that this model can give something close to the reality in the region where the pionic emission dominates (e.g. in the Galactic plane), but outside of this region inverse Compton scattering is an important player (e.g. Strong et al., ApJ 613  962; ApJ 537  763). It dominates the pion-decay emission at intermediate and high galactic latitudes. This means that the model gives a wrong description of the gamma-ray flux outside of the plane if the inverse Compton emission is not properly accounted for. The conclusion on the extragalactic background is thus wrong if based on such a model.
On page 3, the authors state "Any uncertainty in the self-veto correction would alter the sensitivity of the detector at energies above ~1 GeV"
The authors do not specify why exactly above 1 GeV. I think, the only reason for that is the discussed "GeV excess" which is observed above ~1 GeV. As I remember correctly, Alex Moiseev Monte Carlo simulations show that the backsplash is/may-be important near/above 10 GeV.
On page 3, the authors state "The all-sky pervasiveness of the apparent excess flux above 1 GeV energy (see above) strongly indicates that this anomaly is intrinsic to the detector."
I am not convinced that this "pervasiveness" points undoubtedly to the detector miscalibration. An existence proof without much of the discussion: there are, at least, two other interpretations of the excess, cosmic ray fluctuations (Strong et al., ApJ 613  962) and dark matter annihilation (de Boer et al., A&A 444  51).
On page 3, the authors state "We find that for different observations of these sources, there were variations in their measured fluxes in excess of 40%. Since the flux from these sources is expected to be constant, this suggests systematic errors caused apparent time variations over different observing periods. Thus, it is not safe to assume there are no systematic errors, especially above 1 GeV."
This is exactly illustrated in Fig 2 of Esposito et al. ApJS 123 (1999) 203 (the reference is not given). Yet, these sources and the diffuse emission itself were used to correct for changing spark chamber efficiency.
On page 3, the authors state "we are unable to explain the GeV anomaly as caused by any single systematic instrument effect that we have studied."
That already contradicts to the title "likely cause"...
On page 4, the authors continue "It is difficult to eliminate any of these possibilities, but no specific evidence has been found that they exist."
Okay, so we do not know the cause of the excess. Why do authors think it is instrumental? It can really be the case, but the data that we have do allow for other interpretations.
On the same page there is a discussion of the importance of EGRET observations to calibrate GLAST LAT. I doubt it.
Finally, if the EGRET excess is instrumental, how is it possible that the authors confirm the extragalactic background by Sreekumar et al. (ApJ 494  523) including the region above 1 GeV ?
|Author:||strong [ Thu Jun 07, 2007 5:17 am ]|
|Post subject:||further thoughts|
So in the end they are just renormalizing the EGRET response assuming their diffuse model is correct. Or in other words, the astrophysics of the Galaxy is known better than the EGRET calibration ! In fact as Igor points out, their model is anyway inadequate for the inverse-Compton emission.
Note that they are actually saying the response is a factor 2 *better* at GeV than previously thought, which would be pretty remarkable if true.
Unfortunately there seems no way to decide one way or the other with current information, there being no standard candle (not even the Crab has a 'simple' spectrum at GeV).
Evidently we have to await GLAST to resolve this, but meantime Brian's work seems to shed some light on the issue.
|Author:||strong [ Wed Jun 27, 2007 7:47 am ]|
|Post subject:||related paper at ICRC|
Baughman has also a paper on this at the upcoming ICRC in Mexico, looking forward to seeing it when it comes online
http://indico.nucleares.unam.mx/contrib ... 9&confId=4
|Author:||imos [ Wed Jun 27, 2007 9:21 am ]|
It is online:
|Page 1 of 1||All times are UTC - 8 hours [ DST ]|
|Powered by phpBB® Forum Software © phpBB Group