Hi Folks, this is Masaki Mori from Tokyo.

There appeared a preprint entitled "A revised catalogue of EGERT gamma-ray sources" by Casandjian and Grenier [arXiv:0806.0113]. It claims the number of significant sources reduces to 188 compared to 271 in the 3rd EGRET catalog (Hartman et al., ApJS 1999). Main difference comes from their galactic diffuse model, which includes dark gas contribution inferred from gamma-ray data [Science 307, 1292 (2005)]. (I saw their preliminary result in 2005 at the Cherenkov2005 workshop poster in Paris: it took 3 years to publish since then.) I would like to hear your thought on this paper. Personally I think it is a matter of definition of the "diffuse source": dark gas blob incorporated in the model [Casandjian and Grenier], or diffuse source standing out from the model [Hartman 1999].

First of all I have to say that the diffuse emission model used by the EGRET team is far from being perfect, which means that EGRET could have missed some sources and could see sources which do not exist. (We will know the truth soon since GLAST is already in the orbit!)
Casandjian and Grenier "diffuse emission model" is not a model but rather a template fit. They took the gas HI, H2, and "dark gas," inverse Compton, and isotropic components, sum them up and fit to the data with free normalization parameters for each component and in each of 3 energy ranges separately. In this fit, they used the gas distribution (H2, HI) in the form of 6 galactocentric rings with free normalization for each ring; this gives 15 free parameters for the diffuse emission model for each energy range. Such a fit implicitly assumes that the cosmic ray spectra (p, e) and fluxes are the same in every part of the Galaxy and thus the emissivity per H atom or per electron is the same everywhere. This is not correct, however. The cosmic ray spectra are different in different places (energy losses, diffusion, and the source distribution) and thus the emissivity is different as well. Note that the interstellar radiation field is also different in different places which affects electron energy losses and gamma-ray emission (inverse Compton). Their model, therefore, gives a wrong brightness in the places where the cosmic ray spectra differ from the average (note also the the observed diffuse emission is the line-of-sight integration of the emissivity times the gas density plus inverse Compton emission). This results in an excess or a deficit in the diffuse emission photons in different places; using this model could yield "new sources" relative to the 3EG catalog or "no detection" in the place of some of the EGRET sources. However, the diffuse emission model of Casandjian and Grenier is perhaps better than the original EGRET model.
Looking at their paper, I see that they also used one of GALPROP models of the diffuse emission with free normalizations of the components (pion decay, IC, bremsstrahlung, isotropic component, and "dark gas" component, but most of the analysis is done using the "Ring" model described above. Note that there is no "dark gas" component in the original GALPROP model and adding this component can change the background model. Besides, any background model is not free from the intrinsic systematic errors associated with the gas distribution derived from velocity information and the assumed rotation curve. For more discussion on the difficulties associated with the determination of the diffuse emission see our paper: http://adsabs.harvard.edu/abs/2007NuPhS.173...44M

_________________
Igor Moskalenko
Stanford University

Thank you for your interest in our work Masaki,
the statement above is not correct. Our model/template does not assume a uniform cosmic-ray spectrum since we divided the gas that fills the Galaxy in galactocentric rings (as well as the interstellar radiation field) as it is done in Galprop. Instead of being calculated, the gas emissivities are fitted to the Egret counts maps which produces a better residual.
The results you saw in 2005 concerned mainly the dark gas and extragalactic gamma-ray sources. The ring procedure was developed to extract correct fluxes and positions for sources in the galactic plane.
Jean-Marc Casandjian

The use of rings does assume that the CR spectra and gamma-ray emissivity are the same within the ring. The true emissivity, however, changes with the distance from the Galactic plane.

_________________
Igor Moskalenko
Stanford University

Thank you for having amended your initial text Igor and for reading our paper. As you said, Glast will tell us if our model is perhaps better than the Egret one