DUNEPRISM EVENT RATES UPDATE

LUKE PICKERING DUNEPRISM WEEKLY 2017-10-30

CODE UPDATES

▸ Calculates the plane-averaged flux for each slice at each stop. ▸ Can use these fluxes to perform linear combination fits. ▸ Can be supplied XSecs, detector definition, and stop POT

exposure to determine realistic event rate predictions and associated statistical uncertainties.

Luke Pickering

2017-08-03 2

▸ Added XML Run plans to my DUNEPrismTools:

▸ Define detector stops and measurement slices:

FLUXES

▸ Fluxes for 50x 10cm slices at 4 stops, on axis, 2.5m 20m and 30m displacement. ▸ Neutrino’s fired randomly and uniformly through measurement plane

▸ Rise in normalisation looks odd, checked code against Laura’s, and predictions sent to Mike, seem to match up, will quintuple check again.

Luke Pickering

2017-08-03 3

(GeV)νE0 2 4 6 8

per

PO

T)-2

cm-1

(GeV

νΦ

0

10

20

30

9−10×

GENIE XSECS

▸ Throw (GENIE) events, sort into true FS topologies, calculate σ(Eν). ▸ Will throw 10 times these stats to smooth out curves. ▸ Have nu-e elastic, but didn’t have time to include before meeting… will update.

Luke Pickering

2017-08-03 4

(GeV)νE0 1 2 3 4 5 6 7 8 9 10

-1 A2

cm

-38

) 10

νE(σ

0.51

1.52

2.53

3.54

4.55

Ar40µνGENIE, πCC 0 πNC 0±πCC 1 ±πNC 10πCC 1 0πNC 1πCC N πNC N

PREDICTED EVENT RATES

▸ One multiplied by the other… ▸ Reference POT is 1E21 / year (I think?)

Luke Pickering

2017-08-03 5

(GeV)νE0 2 4 6 8

Even

t Rat

e / G

eV

0

20

40

60

80

310×

Ar40, 5E19/1E19 POTµνGENIE,

, T = 2.05e+05π0.25 m CC 0

, T = 5.69e+03π20.25 m CC 0

, T = 1.25e+05±πCC 1

, T = 1.44e+03±πCC 1

, T = 4.97e+040πCC 1

, T = 5250πCC 1

, T = 8.97e+04πCC N

, T = 858πCC N

LINEAR COMBINATIONS

▸ Off-axis fits still not as great as they could be… ▸ Have nuPrism regularisation from Mark Scott, was going to have a play this week.

Luke Pickering

2017-08-03 6

(GeV)νE0 2 4 6 8

per

PO

T)-2

cm-1

(GeV

νΦ

0

10

20

30

9−10×

(GeV)νE0 2 4 6 8

per

PO

T)-2

cm-1

(GeV

νΦ

0

10

20

30

9−10×

(GeV)νE0 2 4 6 8

per

PO

T)-2

cm-1

(GeV

νΦ

0

10

20

30

9−10×

(GeV)νE0 2 4 6 8

per

PO

T)-2

cm-1

(GeV

νΦ

0

10

20

30

9−10×

LINEAR COMBINATIONS — BUILDING

▸ Added some fit diagnostics, rushed to make these plots, better ones to follow. ▸ Have nuPrism regularisation from Mark Scott, was going to have a play this week.

Luke Pickering

2017-08-03 7

(GeV)νE0 2 4 6 8

per

PO

T)-2

cm-1

(GeV

νΦ

0.2−

0.1−

0

0.1

6−10×

(GeV)νE0 2 4 6 8

per

PO

T)-2

cm-1

(GeV

νΦ

0

50

9−10×

(GeV)νE0 2 4 6 8

per

PO

T)-2

cm-1

(GeV

νΦ

0

0.05

0.1

0.15

0.26−10×

(GeV)νE0 2 4 6 8

per

PO

T)-2

cm-1

(GeV

νΦ

0

10

20

9−10×

(GeV)νE0 2 4 6 8

per

PO

T)-2

cm-1

(GeV

νΦ

0.4−

0.2−

0

0.2

0.4

6−10×

LINEAR COMBINATIONS — EVENT RATES

▸ ‘Correct’ each fitted coefficient for the relative POT exposure of each stop. ▸ N.B. in the on-axis position, the two halves of the FV see fluxes at the

same off-axis angle, so measurement slice stats are collected twice as fast.

▸ Sum measurement slice event rates weighted by fitted coeffs. ▸ Clearly broken, was hacking it together before the meeting — will try and

fix this week. (But work is mostly redundant c.f. Cris.)

Luke Pickering

2017-08-03 8

(GeV)νE0 2 4 6 8

per

PO

T)-2

cm-1

(GeV

νΦ

0

0.05

0.1

0.15

610×

(GeV)νE0 2 4 6 8

per

PO

T)-2

cm-1

(GeV

νΦ

0

0.1

0.2

610×

(GeV)νE0 2 4 6 8

per

PO

T)-2

cm-1

(GeV

νΦ 0

0.1

0.2610×

(GeV)νE0 2 4 6 8

per

PO

T)-2

cm-1

(GeV

νΦ

0

0.2

0.4

610×

FOR UPDATED▸ Add nu-e Elastic event rates: I have the curve, just

didn’t have time to include here.

▸ Redo osc. flux fits and more gaussians and send around updated slides ASAP: ▸ Interest in also RHC flux fits?

Luke Pickering

2017-08-03 9

FOR NEXT WEEK▸ Look into regularisation — Mark Scott sent me some of

the NuPrism fitting code and suggested that I nab the regularisation: stops neighbouring coefficients being wildly different form each other.

▸ Teijin has a number of varied flux simulations — try and run/mock up a correlated flux uncertainty to show off the systematic error propagation: ▸ Fit for linear combination coefficients assuming

nominal ▸ Build combination from flux throws within

correlated uncertainty.

Luke Pickering

2017-08-03 10

THANK YOU

WE APOLOGISE FOR ANY INCONVENIENCE