Announcements:

Improving documentation:

Propose to expand upon Brían's track reconstruction PDF and include a thorough explanation of JGandalf, JStart and JEnergy, similar to what has already been done for JPrefit;

Bouke will prepare a ~30 min presentation in two weeks about the Jpp reconstruction chains.
The presented material and discussion will be collected afterwards and incorporated into Brían's document.

Mjg: Request to also collect and write up documentation on the simulation chain, with special attention directed towards where each simulation step (e.g. detector response, time slewing) enters the chain and which information is retained in each step.

Rounds

1.) Brían

Work on combining muon and shower PDFs is progressing.
The software currently used for the analysis, defines a shower vertex with an outgoing muon.
A PMT hit is specified at a specific time and location, after which the program evaluates the JPP arrival time PDFs with respect to the hypothesis that the hit corresponds to shower light emitted directly from the event vertex or the hypothesis that the hit was caused by Cherenkov light originating from the muon. The combined (i.e. summed) probability is shown in a 1D histogram with the time-residual calculated w.r.t. the muon Cherenkov hypothesis on the x-axis.

Mjg: How is it possible to plot the shower arrival time PDF using the muon arrival time on the x-axis?

Ronald: Where you define your zero on the x-axis (i.e. w.r.t. the muon Cherenkov hypothesis or w.r.t. the shower-light hypothesis) does not matter. For each hit you can define both arrival times separately. You need to use the correct arrival time to evaluate the PDF corresponding to either hypothesis. However, which one you use to plot the PDF, is up to you and yields the same results up to a constant shift on the x-axis.

Three example plots were shown for different angles between the z-axis and the muon-direction for hits defined on the z-axis, occurring 60 meters away. If the angle between them is equal to the Cherenkov angle (~ 41 degrees), the combined arrival time PDF only shows a single peak. For angles above and below that a double peak is found.

Ronald: The fact that a double peak structure shows up even above 41 degrees separation between the PMT-to-vertex line and the muon direction, suggests that we are currently using an infinite muon track hypothesis. After all, these hits occur behind the muon Cherenkov cone.

Also shown were three plots where the muon is pointed along the z-axis, whilst the direction of the shower vertex w.r.t. the PMT hit was varied. In this case a double peak structure was present in all three situations (above, below and at 41 degrees separation between the z-axis and the PMT-to-vertex line).

Mjg: Why are the heights of both peaks identical? Conferring with the PDF drawing applications in Jpp (e..g. JDrawPDF) may help to alleviate any confusion here.

Finally, two plots were presented which show that the arrival time PDFs are not necessarily 0, when the muon and shower energy are set to 0.

Mjg: For the muon, this is because the dE/dx for a minimal ionizing particle comprises two parts: one relating to energy loss due to ionization and the other one relating to energy loss from Bremsstrahlung. The part relating to Bremsstrahlung is dependent upon the energy of the muon, whilst the part relating to the ionization energy loss is constant. You need to check prior to evaluating the energy loss, whether a track particle is actually above the minimum ionization threshold or not.

2.) Jordan:

The vertex resolution for shower events in the official MC are a lot lower than the ones computed up until now for the toy Monte Carlo (3 cm compared to 30 cm). More on this next time.

Mjg: How is the toy MC defined?

Jordan: We sample the detected number of photo-electrons from a Poisson distribution. The sampling of the shower PDF is done via the corresponding CDF, calculated in ROOT.