Support #3069
Total number of counts computed by ctmodel depends on number of energy bins
| Status: | Closed | Start date: | 11/19/2019 | |
|---|---|---|---|---|
| Priority: | High | Due date: | ||
| Assigned To: | - | % Done: | 0% | |
| Category: | - | |||
| Target version: | - | |||
| Duration: |
Description
I observed that the total number of counts in a given energy range computed by ctmodel depends on the number of energy bins (or the other way around - on the bin sizes themselves?).
I stumbled about it when computing a model map for the energy range of 350-650 GeV using irf background. Computing the model counts with enumbins=1 yielded a total number of counts which was ~18% higher than the one obtained by subdividing the same energy range into sub bins (e.g. with enumbins=3) and stacking along the energy axis.
The attached figure shows the described behavior, exemplarily for a single observation of my use case.
I also see the issue for larger energy ranges (e.g. 1-10TeV) and effective area background (but less dominant).
The overshoot varies for my scenario between 7% and 18% for different individual observations.
May the issue hence be connected to irf integration accuracy?
ctmodel_different_nebins.png (15.3 KB)
Recurrence
No recurrence.
History
#1
Updated by Knödlseder Jürgen over 4 years ago
This is not surprising. ctmodel samples the model at the logarithmic bin centre, hence with a fine sampling things become more accurate. See for example http://cta.irap.omp.eu/ctools/users/tips/index.html#tip-binned.
#2
Updated by Specovius Andreas over 4 years ago
- Tracker changed from Bug to Support
- Status changed from New to Resolved
Choosing apriori multiple sub-bins was a bit unintuitive as I wanted to have a single energy bin in the end but of course, you are totally right!
Thank you for your quick reply!
Set the tracker to Support because there was no bug.
#3
Updated by Knödlseder Jürgen about 4 years ago
- Status changed from Resolved to Closed