Data sets
There are two types of data sets that ship together with Gepard. One contain original numbers published by experimental collaborations which made the measurements. Another are derived sets where original sets are transformed in some way (e. g. by discrete Fourier transform). Every data set (and point) has the attributes reference and reference2 which should explain where the data comes from and whether it is original or derived.
Even “original” sets are still transformed by Gepard author(s) into Gepard datafile format either by using datafiles provided by collaborations or tables from published papers. If some errors are introduced in this process, experimental collaborations obviously cannot be blamed for this. If you need “official” datafiles, contact the corresponding collaborations.
Here we list most of the available data sets, with some comments on how we chose to use them in our fits.
Todo
Only DVCS data sets are listed here at the moment.
CLAS
CLAS:2001wjj hep-ex/0107043
Stepanyan et al., Phys.Rev.Lett. 87 (2001) 182002
4.25 GeV electron beam. Just one BSA point: $A_{LU}^{sinphi}$($Q^2$ = 1.25 GeV$^2$, $x_B$ = 0.19, $-t$ = 0.19 GeV$^2$) = 0.202 $pm$ 0.028 $pm$ 0.013.
CLAS:2006krx hep-ex/0605012
Chen et al., Phys.Rev.Lett. 97 (2006) 072002
>>> g.list_data([54])
[ 54] CLAS 6 TSA 0605012
CLAS:2006krx arXiv:0711.4805
F.X. Girod et al., Phys.Rev.Lett. 100 (2008) 162002
>>> g.list_data([6, 7, 8, 25])
[ 6] CLAS 744 ALU 0711.4805 private table
[ 7] CLAS 62 ALU 0711.4805 private table
[ 8] CLAS 12 ALU 0711.4805 F.T. by DM
[ 25] CLAS 16 ALU 0711.4805 F.T. by KK
CLAS:2008ahu arXiv:0812.2950
Gavalian et al., Phys.Rev.C 80 (2009) 035206
>>> g.list_data([81])
[ 81] CLAS 12 ALU 0812.2950 Tables 3 and 4
CLAS:2015bqi arXiv:1501.07052
Pisano et al., Phys.Rev.D 91 (2015) 5, 052014
>>> g.list_data(list(range(88,97)))
[ 88] CLAS 166 ALU 1501.07052
[ 89] CLAS 166 TSA 1501.07052
[ 90] CLAS 166 BTSA 1501.07052
[ 91] CLAS 166 ALU 1501.07052 bin averages from Silvia Niccolai
[ 92] CLAS 166 TSA 1501.07052 bin averages from Silvia Niccolai
[ 93] CLAS 166 BTSA 1501.07052 bin averages from Silvia Niccolai
[ 94] CLAS 10 ALU 1501.07052 F.T. by K.K.
[ 95] CLAS 10 TSA 1501.07052 F.T. by K.K.
[ 96] CLAS 20 BTSA 1501.07052 F.T. by K.K.
There is a precursor PRL paper Seder et al. arXiv:1410.6615 [hep-ex] which brings just TSA. Comparing harmonics, results from this PRL are different from what is published in Pisano et al. but phi-space TSA is the same, so one can use just Pisano et al.
[88-90] are data from the paper, while [91-93] have averaged xB, Q2 and t
so they are used to make Fourier transform to get [94-96].
CLAS:2015uuo arXiv:1504.02009
H.S. Jo et al., Phys.Rev.Lett. 115 (2015) 21, 212003
>>> g.list_data([97, 98, 99, 100, 101, 102, 104, 106])
[ 97] CLAS 2640 XLU 1504.02009 CLAS data base E145M2
[ 98] CLAS 2640 XUU 1504.02009 CLAS data base E145M1
[ 99] CLAS 1152 XLU 1504.02009 CLAS data base E145M2, restricted kinematics
[100] CLAS 1152 XUU 1504.02009 CLAS data base E145M1, restricted kinematics
[101] CLAS 48 XLUw 1504.02009 FT analysis with MC error propagation by KK
[102] CLAS 96 XUUw 1504.02009 FT analysis with MC error propagation by KK
[104] CLAS 1152 ALU 1504.02009 CLAS data base E145M1, E145M2
[106] CLAS 48 ALU 1504.02009 FT analysis with MC error propagation by KK
This data is not independent from (compatible) dataset CLAS:2006krx (F.X. Girod et al.). Out of ALU from that paper and XUU and XLU from this paper, only two observables are independent. Some CLASS experts (M. Garcon) advocate ALU, XUU combination because of larger kinematic coverage of ALU and smaller total systematics, while some other advocate XLU, XUU for consistent cuts and corrections applied.
CLAS:2018bgk arXiv:1810.02110
Hirlinger Saylor et al., Phys.Rev.C 98 (2018) 4, 045203
It seems that no actual data is publicly available. This is certainly new data, since beam energy is 5.88 GeV. (Data published in 2015 are with 5.75 GeV.)
CLAS:2022syx arXiv:2211.11274
Christiaens et al., Phys.Rev.Lett. 130 (2023) 21, 211902
>>> g.list_data([149])
[149] CLAS 1535 ALU 2211.11274
First CLAS12 DVCS measurement.
Hall A
JeffersonLabHallA:2006prd nucl-ex/0107005
Muñoz Camacho et al., Phys.Rev.Lett. 97 (2006) 262002
>>> g.list_data([26, 30, 33, 34, 35, 50, 51, 55, 56])
[ 26] HALLA 12 XLUw2C 0607029 DFT analysis with MC error propagation by KK
[ 30] HALLA 4 XwA 0607029 DM's fit
[ 33] HALLA 288 XLU 0607029 private tables
[ 34] HALLA 96 XUU 0607029 private tables
[ 35] HALLA 8 XUUw2C 0607029 DFT analysis with MC error propagation by KK
[ 50] HALLA 288 XLUw 0607029 DFT analysis with MC error propagation by KK
[ 51] HALLA 96 XUUw 0607029 DFT analysis with MC error propagation by KK
[ 55] HALLA 36 XLUw 0607029 private tables
[ 56] HALLA 20 XUUw 0607029 private tables
For KM09b [30] was used. [26, 35] are $C_{I}$ coefficient extractions. [33,34]
are $phi$-dependent original data. [50,51] are KKs DFT transforms of data.
[55,56] are DMs fit to first few harmonics.
Both of these are weighted with BH propagators.
For global fits we used [50] (1st sin harmonic) as BSSwpoints and [51]
(0th and 1st cos harmonic) as BSDwpoints.
There is also derived asymmetry data ALU = XLU/XUU by DM:
>>> g.list_data([103, 105])
[103] HALLA 96 ALU 0607029 private table
[105] HALLA 96 ALU 0607029 FT analysis with MC error propagation by KK
JeffersonLabHallA:2015dwe arXiv:1504.05453
Defurne et al., Phys.Rev.C 92 (2015) 5, 055202
>>> g.list_data(list(range(107, 124)))
[107] HALLA 120 XUU 1504.05453 Kin2, Table VII
[108] HALLA 120 XUU 1504.05453 Kin3, Table VIII
[109] HALLA 120 XLU 1504.05453 Kin1, Table IX
[110] HALLA 120 XLU 1504.05453 Kin2, Table X
[111] HALLA 120 XLU 1504.05453 Kin3, Table XI
[112] HALLA 120 XUU 1504.05453 KinX2, Table XII
[113] HALLA 120 XUU 1504.05453 KinX3, Table XIII
[114] HALLA 120 XLU 1504.05453 KinX2, Table XIV
[115] HALLA 120 XLU 1504.05453 KinX3, Table XV
[116] HALLA 20 XUUw 1504.05453 FT analysis with MC error propagation by KK
[117] HALLA 15 XLUw 1504.05453 FT analysis with MC error propagation by KK
[118] HALLA 25 XLUw 1504.05453 FT fitting with numerical error propagation by DM
[119] HALLA 40 XUUw 1504.05453 FT fitting with numerical error propagation by DM
[120] HALLA 20 XUU 1504.05453 FT analysis with MC error propagation by KK
[121] HALLA 15 XLU 1504.05453 FT analysis with MC error propagation by KK
[122] HERMES 24 ALU 1206.5683 Table 2
[123] HALLA 20 XUUw 1504.05453 FT analysis with MC error propagation by KK
DM’s BSSw and BSDw include also KINX kinematic ranges, which are not independent measurements.
There may be disagreements DM vs. KK regarding total systematic uncertainty.
[122] doesn’t belong here obviously.
Defurne:2017paw arXiv:1703.09442
Defurne et al., Nature Commun. 8 (2017) 1, 1408
>>> g.list_data(list(range(124, 139)))
[124] HALLA 36 BSD 1703.09442
[125] HALLA 36 BSD 1703.09442
[126] HALLA 96 BSD 1703.09442
[127] HALLA 96 BSD 1703.09442
[128] HALLA 96 BSD 1703.09442
[129] HALLA 36 BSS 1703.09442
[130] HALLA 36 BSS 1703.09442
[131] HALLA 96 BSS 1703.09442
[132] HALLA 96 BSS 1703.09442
[133] HALLA 96 BSS 1703.09442
[134] HALLA 90 BSS 1703.09442
[135] HALLA 18 BSDw 1703.09442 FT analysis with MC error propagation by KK
[136] HALLA 44 BSSw 1703.09442 FT analysis with MC error propagation by KK
[137] HALLA 44 BSSw 1703.09442 FT analysis with MC error propagation by KK
360 XLU and 450 XUU measurements at three beam energies: 3.355, 4.455 and 5.55 GeV.
[137] is extraction where possible c0/c1 enhancement is neglected.
JeffersonLabHallA:2022pnx arXiv:2201.03714
Georges et al., Phys.Rev.Lett. 128 (2022) 25, 252002
Full data available from PhD thesis by Frédéric Georges (Orsay, IPN)
>>> g.list_data([141, 142, 143, 144])
[141] HALLA 1080 XUU Georges:2018kyi Appendix D
[142] HALLA 1080 XLU Georges:2018kyi Appendix D
[143] HALLA 90 XUUw Georges:2018kyi FT analysis with MC error propagation by MC
[144] HALLA 45 XLUw Georges:2018kyi FT analysis with MC error propagation by MC
First 12 GeV Hall A data.
Benali:2020vma arXiv:2109.02076
Benali et al., Nature Phys. 16 (2020) 2, 191-198
>>> g.list_data([138, 139, 140])
[138] HALLA 8 XSintphi doi:10.1038/s41567-019-0774-3 Table 4
[139] HALLA 96 XUU doi:10.1038/s41567-019-0774-3 Table 3
[140] HALLA 16 XUUw doi:10.1038/s41567-019-0774-3 FT analysis with MC error propagation by MC
Neutron DVCS at two beam energies: 4.45 and 5.55 GeV.
HERMES
HERMES:2001bob hep-ex/0106058
Airapetian et al., Phys.Rev.Lett. 87 (2001) 182001
ALU. First DVCS measurement. Just a plot, no published numbers.
Frank Ellinghaus (QCD 2002, unpublished)
>>> g.list_data([29])
[ 29] HERMES 12 ALU Frank E. from DM's notebook
HERMES:2006pre hep-ex/0605108
Airapetian et al., Phys.Rev.D 75 (2007) 011103
$A_C^{cosphi}$ in four $x_B$-$Q^2$-$t$ bins. This data is not used in fits. It is not clear when and how much data is taken. It is probably contained in the next dataset HERMES:2008abz.
HERMES:2008abz arXiv:0802.2499
Airapetian et al., JHEP 06 (2008) 066
>>> g.list_data([31, 65, 66])
[ 31] HERMES 24 AC 0802.2499 Table 1
[ 65] HERMES 12 AUTDVCS 0802.2499 Table 1a
[ 66] HERMES 36 AUTI 0802.2499 Table 1b
$A_C^{cos0phi}$ and $A_C^{cosphi}$. The latter 12 points are used in GLOpoints for KM09a and KM09b fits.
$A_{UT,I}^{sin(phi-phi_S) cos(phi)}$, sensitive to $Im{}m(F_1mathcal{E}-F_2mathcal{H})$ and thus to $mathcal{E}$, is 12 AUTIpoints. As an independent subset (with emphasis on $t$ dependence), first 4 points (alias: AUTIpts) are used in global fits such as KMM12. $A_{UT,I}^{cos(phi-phi_S) sin(phi)}$ is proportional to $Im{}m(2 F_2tilde{mathcal{H}}-F_1tilde{mathcal{E}})$ and is consistent with zero so could be used to put constraint on $tilde{mathcal{H}}$ (alias: AUTICSpts) where advantage w.r.t. $A_{UL}$ is that here we have no DVCS$^2$ pollution.
HERMES:2009cqe arXiv:0909.3587
Airapetian et al., JHEP 11 (2009) 083
>>> g.list_data([5, 32])
[ 5] HERMES 18 ALUI 0909.3587 Table 2
[ 32] HERMES 36 AC 0909.3587 Table 4
$A_{LU}^{sinphi}$ and $A_C^{cosphi}$ are used in KM10 models (all 18 points, not just statistically independent 6).
HERMES:2010dsx arXiv:1004.0177
Airapetian et al., JHEP 06 (2010) 019
>>> g.list_data([52, 53])
[ 52] HERMES 36 TSA 1004.0177 Table 4
[ 53] HERMES 36 BTSA 1004.0177 Table 4
HERMES:2011bou arXiv:1106.2990
Airapetian et al., Phys.Lett.B 704 (2011) 15-23
>>> g.list_data([73, 74])
[ 73] HERMES 39 ALTBHDVCS 1106.2990 Table 2
[ 74] HERMES 65 ALTI 1106.2990 Table 2
This data was not yet used for global fits (was used for local fits with M. Murray). It is consistent with zero.
HERMES:2012idp arXiv:1206.5683
Airapetian et al., JHEP 10 (2012) 042
>>> g.list_data([122])
[122] HERMES 24 BSA 1206.5683 Table 2
Data with recoil detector.
HERMES:2012gbh arXiv:1203.6287
Airapetian et al., JHEP 07 (2012) 032
>>> g.list_data([67, 68, 69, 70, 71, 72])
[ 67] HERMES 72 AC 1203.6287 Table 6
[ 68] HERMES 36 ALUI 1203.6287 Table 5
[ 69] HERMES 18 ALUDVCS 1203.6287
[ 70] HERMES 48 AC Morgan M. combined 1996-2007 data
[ 71] HERMES 24 ALUI Morgan M. combined 1996-2007 data
[ 72] HERMES 12 ALUDVCS Morgan M. combined 1996-2007 data
Final HERMES data on BCA and BSA. First 6 independent $t$-dep points are used for global fits as BCApts (6 0th and 6 1st cos harmonics) and ALUIpts (6 1st sin harmonics). There is also 4-bin version of this data communicated privately by M. Murray.
H1
H1:2001nez hep-ex/0107005
Adloff et al., Phys.Lett.B 517 (2001) 47-58
This data is not used because it is contained withing the next item Aktas:2005ty.
H1:2005gdw hep-ex/0505061
Aktas et al., Eur.Phys.J.C 44 (2005) 1-11
>>> g.list_data([39, 43, 44])
[ 39] H1 8 XGAMMA 0505061 Table 1
[ 43] H1 9 XGAMMA 0505061 Table 3
[ 44] H1 6 XGAMMA 0505061 Table 2
We take Table 1 (cross section differential in t), with separate 4 points for 1996-1997 and 4 points for 1999-2000 data. Tables 2 and 3 are total cross section from same events in dependence on Q2 and W, respectively.
H1:2007vrx arXiv:0709.4114
F.D. Aaron et al., Phys.Lett.B 659 (2008) 796-806
>>> g.list_data([36, 37, 38, 40, 41, 42])
[ 36] H1 12 XGAMMA 0709.4114 Table 3 - upper half
[ 37] H1 12 XGAMMA 0709.4114 Table 3 - lower half
[ 38] H1 24 XGAMMA 0709.4114 Table 4
[ 40] H1 5 XGAMMA 0709.4114 Table 1 - right
[ 41] H1 15 XGAMMA 0709.4114 Table 2
[ 42] H1 4 XGAMMA 0709.4114 Table 1 - left
This data is not used because it is contained withing the next item Aaron:2009ac.
H1:2009wnw arXiv:0907.5289
F.D. Aaron et al., Phys.Lett.B 681 (2009) 391-399
>>> g.list_data([60, 61, 62, 63, 64])
[ 60] H1 4 XGAMMA 0907.5289 Table 1a
[ 61] H1 5 XGAMMA 0907.5289 Table 1b
[ 62] H1 15 XGAMMA 0907.5289 Table 2
[ 63] H1 12 XGAMMA 0907.5289 Table 3a
[ 64] H1 12 XGAMMA 0907.5289 Table 3b
We use Table 3a which is differential DVCS cross section binned in t and Q2.
Todo
Create dataset from BCA data from Table 4 of this reference.
ZEUS
ZEUS:2003pwh hep-ex/0305028
Chekanov et al., Phys.Lett.B 573 (2003) 46-62
>>> import gepard as g
>>> g.list_data(45)
[ 45] ZEUS 6 X 0305028 Table 1
We take data from Table 1 with DVCS cross section in dependence on Q2. Tables 2 and 3 are same data but binned for W dependence. Using Q2 dependence gives us handle on evolution.
ZEUS:2008hcd arXiv:0812.2517
Chekanov et al., JHEP 05 (2009) 108
>>> g.list_data([46, 47, 48, 49])
[ 46] ZEUS 4 X 0812.2517 Table 4
[ 47] ZEUS 6 X 0812.2517 Table 1
[ 48] ZEUS 6 X 0812.2517 Table 2
[ 49] ZEUS 8 X 0812.2517 Table 3
We take data from Table 1 (Q2 dependence of cross section, one might cut low-Q2 points) while Tables 2 (and 3) are same data binned in W (W and Q2). We also take data from Table 4, which is differential cross section in t, extracted from the subset of the above data, so strictly it is not statistically independent.
Combined datasets used for fits
Starting from KMM12 global model, we use just 35 independent
collider points H1ZEUS:
>>> from gepard.fits import H1ZEUS
>>> g.describe_data(H1ZEUS)
npt x obs collab FTn id ref.
----------------------------------------------
8 x XGAMMA H1 N/A 39 hep-ex/0505061
12 x XGAMMA H1 N/A 63 arXiv:0907.5289 [hep-ex]
6 x XGAMMA ZEUS N/A 45 hep-ex/0305028
4 x XGAMMA ZEUS N/A 46 arXiv:0812.2517
5 x XGAMMA ZEUS N/A 47 arXiv:0812.2517
----------------------------------------------
TOTAL = 35