Where to find vector and reconstructed MC data files.
Contents
Flux Files
You can find the latest flux files at http://jnusrv01.kek.jp/jhfnu/internal/nubeam/flux/index.html
These gzipped tar files contain several vector files each.
The 2KM detector is ND1. In order to get the right position you should use 04b.
For SK you can use the flux files in 04a
Nues are inside these files : int(mode/10) = 1 for numus, int(mode/10) = 3 for nues (other modes for anti numus & nues). See efill.kumac from the aforementioned pages (or ask M.F. for efill.C if you like ROOT).
In order to make your own flux files you need to check out jnubeam from :ext:jnudevel@jnusrv01.kek.jp:/home/jnurep/repository . If you want the 04b version use the release tag release-04b. To get the standard flags used by Ichikawa-san use the nubeam.card file. Notes on successfully building jnubeam:
- On fedora core 3 chgen won't build now. Comment this section out of the Makefile and copy a working version into the nubin directory.
- Get a copy of gcalor_unix and put it in the nubeam directory. Do a "make gcalor_unix.o" in the nubeam directory and update the Makefile to point at it.
- make links or copies to chetc.dat and xsnet.dat in your nubin directory.
The resulting nubin executable will be in the nubin directory. To run edit the nubeam.card file and set 'NFRD 1'. This will use 2KM. You can set the beam size with HSIZ and VSIZ. I (CWW) modified the nubeam.sh so that the card file used is a parameter with the line setenv NUBEAM_CARD $2. Then you can run like:
nubeam.sh ntuple_out input.card
The resulting output ntuple file will be called ntuple_out.hbk Call the output what you want and use a suitably copied and edited version of nubeam.card for your input card.
Interaction Vectors
Find the interaction vectors and the ntuples produced with them at http://jnusrv01.kek.jp/~jnurep/vectors/04b/40GeV/. A description of the ntuple files taken from an email from Hayato-san can be found in VectorNtupleDescription.
You can find the vectors converted into Nuance format so it can be read by the 2KM geant4 code in http://www.phy.duke.edu/~cwalter/2km-vector-files
Nue vectors are generated separately. The corresponding files are available from the same pages.
NEW : january 2007 You can find Mitsuka-san's new vectors at Kashiwa. If you don't have a Kashiwa account you can get them from our local copy (nuance format only) : http://www.phy.duke.edu/~fechner/mitsuka-vectors/vectors
Reconstructed Data Files
2KM (@ Kashiwa) --> you need a login... /kam/work/process_2km_k2kcomp/ntuple/nue or numu -- make ntuples available somewhere ?
- SK:
- ZBS before polfit: on suketto, Oobayashi-san's work directory.
- polfit5 : on suketto. M Fechner's work directory /net/suknas1/shares/jnu/fechner/t2k-sk-polfit5
polfit2 : http://jnusrv01.kek.jp/~jnurep/physics/nue Mine-san's work directory.
NEW DECEMBER 2005 :
2KM : http://www.phy.duke.edu/~mfguest/ M. Fechner's work directory. ONLY POLFIT2 applied
SK : http://jnusrv01.kek.jp/jhfnu/internal/physics/skmc/04b/ POLFIT2 AND POLFIT5 applied Normalizations: (all nue files, 1018.4 years at 1E21 pot/yr) (all numu files, 338.49 years at 1E21 pot/yr)
Needed Normalization Factors etc.
Emails from Hayato-san :
Here I put the table of event rates for 04b 40GeV flux
(/10^21pot):
Detector ID Dimensions nu_mu nu_e
X/2 Y/2 /ton /ton
ND1(2km) 400 400 2.67e+03 5.48e+01
ND2(280m) 150 150 1.08e+06 9.32e+03
ND3(280m) 500 50 8.83e+05 8.38e+03
ND4(280m) 50 500 9.00e+05 8.44e+03
ND5(280m) 150 150 1.70e+05 3.34e+03
ND6(280m) 150 400 1.82e+05 3.31e+03
Another email concerning SK (04b !!)
| # of intr./year
|
nu_mu | 2203.27
\bar{nu_mu} | 128.16
nu_e | 45.83
\bar{nu_e} | 6.37
CAUTION : CORRECTED EVENT RATES
It seems an off-by-one bin error occured in the previous calculation. The new event rates (by M. Fechner, confirmed independently by Hayato-san) are here. "Fully oscillated" event rates are useful for nue appearance at SK : they are event rates corresponding to a mu->e oscillation probability equal to 1 at all energies (= numu flux * nue cross section).
NEW (25dec.) : confirmed by Hayao-san using an independant program
| binned method | unbinned method (MF)
nu_mu | 2046.90 | 2078.00 +/- 6.52
\bar{nu_mu} | XXXX | 129.67 +/- 4.12
nu_e | 44.17 | 44.94 +/- 0.46
\bar{nu_e} | XXXX | 7.01 +/- 0.41
fully oscillated events (if all numus were seen as nues) : 2130.87 (binned) , fully oscillated : 2162.03 +/- 6.63 (unbinned).
Slight differences are expected : in the binned case, the cross section is set to its value at the center of each 50 MeV bin between 0 and 10 GeV.
In the unbinned method, it is allowed to vary freely, there are no bins. These "unbinned" event rates are computed over the whole range available in the flux files (0 to 15 GeV).
Hayato-san showed that this is the main source of the difference between the 2 approaches (see below).
NEW (25 dec) :
For reference here are Hayato-san's numbers :
IP | rate (precise) | rate (binned method)
| |
----+-------------------+------------------------------------------
14 | 2076.076 | 2046.896 / 2077.0183 (with E>10GeV)
12 | 44.641 | 44.171 / 44.6457 (with E>10GeV)
-14 | 131.733 | 123.211 / 131.7214 (with E>10GeV)
-12 | 6.373 | 6.229 / 6.3745 (with E>10GeV)
The slight differences in the SK event rates (precise method) come from the use of 2 different sets of flux files. It is clear that
those differences are within statistical errors
At 2KM (100.2t FV) :
| binned method | unbinned method
nu_mu | 247606.38 | 250604.94 +/- 1757.00
\bar{nu_mu} | XXXX | 15258.63 +/- 785.44
nu_e | 5428.555 | 5471.25 +/- 128.63
\bar{nu_e} | XXXX | 819.66 +/- 53.073
fully oscillated : 260763.34 +/- 1790.064 (unbinned).
I used NEUT 4.5.1. Fluxes are 04a(=04b) for SK and 04b for 2KM.
IMPORTANT REMARK using the unbinned fluxes :
- nue/numu = 2.18321 (1 +/- 0.025) % = 2.18 +/- 0.054 % @ 2KM
- nue/numu = 2.163 (1 +/- 0.0107) % = 2.163 +/- 0.023 % @ SK
The main contribution to the error comes from nue flux uncertainties at 2KM, due to the very small geometric acceptance of the 2KM detector in JNUBEAM. More flux files would be needed to improve it.
Another way to look at the normalization for the 2KM det :
Each file contains 50k interactions.
The interaction rate is
1714 intearctions/10^21pot/1cm.
~~~~<- depth(z-direction)
of the detector
(The size of the detector is 800cm x 800cm.)
I think this means : 1714 interactions / 1yr/ 0.64 tons
Parameters of the T2K Beam
5us spill every 3.53 sec 8 bunches x 58 ns 20 Normal Magnets 28 Combined Function SC Magnets Water cooled graphite target 3 horns @ 320 kA 130m long decay pipe He filled iron vessel, water cooled concrete 10^21 pot = 1 year