Distributions of shower parameters (2012-2013 vs 2019-2020)

[mireianievas]

Hi,

I have produced a plot summarizing some shower parameters after cleaning/hillas reconstruction (evndisp stage) when one scales with the scaling factors of 2012-2013 (small) and 2019-2020 (large). For T1 and the same 100MHz noise file at Zd=0 deg processed with different scaling factors.

The nlowgain being fully linear is expected since the MC sims are the same. That might be a limitation of the method at high energies if the low gains are not perfectly calibrated. Recent data might not trigger as much the lowgain switch, while the scaled MC might still trigger it often.

Putting a size cut (e.g. >1000) helps a bit to reduce the dispersion of the plots

[mireianievas]

same data but as 1d histograms with the ratios (2019-2020 over 2012-2013a)

[GernotMaier]

Very nice - so in both ntubes and width we see smaller values (maybe change the scale on the width plot to make it more visible). This is expected - and exactly what we want to show with these plots.

nlowgain is as you say given by the simulations, maybe something we should mention as a drawback of this method compared to full simulations.

On 11. Aug 2020, at 18:02, Mireia Nievas-Rosillo notifications@github.com wrote:

same data but as 1d histograms with the ratios (2019-2020 over 2012-2013a)

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or unsubscribe.

---

Dr Gernot Maier CTA Deutsches Elektronen-Synchrotron DESY Ein Forschungszentrum der Helmholtz-Gemeinschaft Platanenallee 6, D-15738 Zeuthen, Germany

Tel.: +49 33 7627 7598 Internet: www.desy.de/cta Besucheradresse: 1X07

[mireianievas]

The impact on the width is really not that significant (besides generating a somehow wide distribution), see the updated plot with the mean and std values

[GernotMaier]

Can I still get a 'zoomed' version of the width distribution? A max width of 2 deg is pretty extreme for a 3.5 deg camera (maybe cut at 1.25?). Would like to see how symmetric the distribution is.

[mireianievas]

Can I still get a 'zoomed' version of the width distribution? A max width of 2 deg is pretty extreme for a 3.5 deg camera (maybe cut at 1.25?). Would like to see how symmetric the distribution is.

I'm not showing the max width of 2 deg. It is the ratio between the width on 2019-2020 and the one of 2012-2013 what was going up to 2 in the histogram (just to show the same range of ratios for all the variables). Here's a zoomed version, showing the parameter ratios in the range centered at the mean [mu-std-0.05, mu+std+0.05].

Remember they are ze00, NOISE=100MHz.

The same for a different set of MCs (ze00, NOISE=400MHz, I would expect to have a deeper impact as more pixels may sink into the noise).

Still pretty well managed, with everything more or less symmetric and nice.

And the same for higher zenith (ze40, NOISE=100MHz)

Interestingly for this case, the width/length distributions are a bit odd. The distributions are centered at 1 still, but there might be some crazy values that move the mean up to 1.31/1.52 and the standard deviations explode. Also the number of tubes has a very skewed distribution, so some of the events are losing pixels. Size also changes and so does

So with this, we learn I think that high zeniths might be completely gone without proper MCs, at least if we do not raise the thresholds up significantly. Same set of MCs as before (ze40, NOISE=100MHz), but this time with a size cut of >500

with that we reduce significantly the outliers and the mean/stds recover a bit.

[GernotMaier]

Why are the Zd40 different then Zd20? We have Gain/Mirrorreflectivity corrections, that shouldn’t depend on the elevation of the telescopes. Not sure I understand it.

On 12. Aug 2020, at 11:51, Mireia Nievas-Rosillo notifications@github.com wrote:

 Can I still get a 'zoomed' version of the width distribution? A max width of 2 deg is pretty extreme for a 3.5 deg camera (maybe cut at 1.25?). Would like to see how symmetric the distribution is.

I'm not showing the max width of 2 deg. It is the ratio between the width on 2019-2020 and the one of 2012-2013 what was going up to 2 in the histogram (just to show the same range of ratios for all the variables). Here's a zoomed version, showing the parameter ratios in the range centered at the mean [mu-std-0.05, mu+std+0.05].

Remember they are ze00, NOISE=100MHz.

The same for a different set of MCs (ze00, NOISE=400MHz, I would expect to have a deeper impact as more pixels may sink into the noise).

Still pretty well managed, with everything more or less symmetric and nice.

And the same for higher zenith (ze40, NOISE=100MHz)

Interestingly for this case, the width/length distributions are a bit odd. The distributions are centered at 1 still, but there might be some crazy values that move the mean up to 1.31/1.52 and the standard deviations explode. Also the number of tubes has a very skewed distribution, so some of the events are losing pixels. Size also changes and so does

So with this, we learn I think that high zeniths might be completely gone without proper MCs, at least if we do not raise the thresholds up significantly. Same set of MCs as before (ze40, NOISE=100MHz), but this time with a size cut of >500

with that we reduce significantly the outliers and the mean/stds recover a bit.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

[mireianievas]

Maybe at Zd40 the images are slightly different (smaller, maybe the magnetic field distorts them differently and maybe they look also different because of perspective) than at Zd20 ?

If the surface brightness of the image changes, then more/less pixels may be killed by the cleaning and the fit may change.

[GernotMaier]

Quite a lot of ‘maybes’. Really not sure what the problem is.

On 12. Aug 2020, at 12:48, Mireia Nievas-Rosillo notifications@github.com wrote:

 Maybe at Zd40 the images are slightly different (smaller, maybe the magnetic field distorts them differently and maybe they look also different because of perspective) than at Zd20 ?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.