Sheshuk
commented
7 months ago @JostMigenda, you're absolutely right, this looks like a mistake I made at initial implementation. @jpkneller also asked me about this factor earlier, and I couldn't remember and check where exactly did I get its. Thanks a lot for your work! We need to fix this factor
(I’m moving this discussion from Slack over into an issue here, so we can keep track of this more easily.)
We currently have an unexplained discrepancy between the
Odrzywolek_2010preSN model in SNEWPY and Fig. 1 in the recent SK preSN paper. Both agree reasonably well for the Patton model, so the discrepancy is likely specific to the Odrzywolek model.I was able to review the script to generate the SK figure and the main difference I see is in the normalization of the
nu_xflux. SNEWPY currently uses a factor of 0.36, which presumably comes from the 36.3% from Table 3 in an earlier Odrzywolek et al. paper? In contrast, the SK paper appears to use 0.19; as does an earlier KamLAND paper (just below equation 3).If I understand correctly, Odrzywolek’s Table 3 says that the 36.3% are the fraction of the total (neutrino plus anti-neutrino) emission that is ν_e. That means ν_μ and ν_τ each contribute 6.85%, to bring the total to 50% (plus another 50% from antineutrinos¹). Then the flux ratio of ν_x to ν_e would be 6.85/36.3≈0.19, i.e., exactly the number used by SK and KamLAND. So I think that the 0.19 is actually correct and we currently have a bug. @Sheshuk, as our preSN expert, can you double-check?
¹ The paper takes into account pair production only, so for each flavour, the number of neutrinos will be equal to the number of antineutrinos.