IRF variation with run

[ralphbird]

To get the IRFS I have calculated the average noise, zenith and azimuth for the entire run and then used the VEGAS interpolation. One thing that I would like to check is that the run is not too long for doing this - it might be that we should be using 10 minute bins or ... (I know that Nathan used 8minute bins but he was at LZA so there will be a lot faster variation in the EA with the parameters). This could be done by either producing multiple files for each run or by doing some internal averaging and weighting.

[TarekHC]

Already thinking a bit about this.

Is there any limitation to produce, for instance, N DL3 files per anasum/stage5?

• Subdividing a run between shorter periods (given we have the zenith, azimuth, noise level, etc... per event) only means we need to generate more (smaller) DL3 files, with a finer interpolation of the IRFs (easy, same IRF file). Would need to improve a bit the current code to allow that, but perfectly doable
• The calculation of a correct lifetime per sub-run could be a problem. Not sure how it is currently calculated in ED/VEGAS, but at least from what I remember in MAGIC, we probably want to do it before having only gamma-like events

If I reach the point where I really want to implement this, I will speak with Gernot/Elisa and bother you guys again. Until then, I will try to have a first example of full-enclosure IRFs (averaged over the whole run).

[GernotMaier]

@sona-patel and @tobiaskleiner : both of you did some studies on that. Do you want to post the results?