New production : declination line 6166 with very high density

[gabemery]

This is to request a new production for the line 6166 using a new high density node spacing. (in a dedicated folder)

In order to keep the same total statistic as the current MC line, while having a uniform event density vs cos(Zd) the total number of simulated shower needs to follow the following formula:

n_per_node(cos_Zd) = N_total / [ ({2 from the Az sides of the lines} / {step_formula}) / {Cos_Zd_range} ] n_per_node(cos_Zd) = N_total / (2/(-0.001+0.0080cos_Zd))/0.59 n_per_node(cos_Zd) = N_total (-0.001+0.0080cos_Zd) /(2 0.59)

Currently the total number of simulated events is N_total = 4400000*51 = 224400000 (4400000 obtained with read_simu_info_merged_hdf5().num_shower on DL1 files, so before shower reuse? The later being 10)

Event distribution with proposed n_per_node :

You can find the positions in the attached file. dec_6166_zenith_azimuth_grid_nearcontinuous.txt

[Voutsi]

Hi @gabemery ,

Currently the total number of simulated events is N_total = 4400000*51 = 224400000

Thats correct for diffuse photons. For protons we are using a scaling with zenith in order to keep a stable ratio of triggered protons / triggered photons. Should I ignore this scaling for this production?

(in a dedicated folder)

Ok, I will then put it in /fefs/aswg/data/mc/DL0/LSTProd2/TrainingDataset/GammaDiffuse/dec_6166_high_density

[moralejo]

Hi, the scaling is likely not optimal, but some sort of scaling is needed. Either we keep the original one or we re-evaluate the scaling (based on the existing production one can get a better one to keep the statistics ratio constant). No strong opinion on my side. Also, having a fixed ratio at trigger level is rather arbitrary. Physics-wise it would be more relevant to have a given ratio of gammas to gamma-like protons, I think (again something that might be studied with the existing production).

[gabemery]

Hi @Voutsi , The formula for the number of event per node is to keep the same density of events along the line (or to same local density as is currently done). But the scaling you use is to correct the reduction of trigger with zenith, right?

In this case I think both can be kept, multiplicatively.

[Voutsi]

Hi @gabemery

yes exactly. Thanks for the heads up, I will launch the production today