The VALOR-SBN repository includes SBN-specific codes and configurations for the VALOR SBN sterile neutrino oscillation fit and related analysis workflows. It is built on top of the common and experiment-agnostic VALOR Software Development Kit (SDK) which includes the bulk of the VALOR analysis code.
Copyright (c) 2010-2021, The VALOR Fitting Group. All Rights Reserved.
Authors (VALOR/SBN): Costas Andreopoulos (Liverpool & STFC/RAL), Dom Barker (Sheffield), Steve Dennis (Cambridge), Thomas Ham (Liverpool), Rhiannon Jones (Liverpool), Tereza Kroupova (UPenn)
For a full list of VALOR SDK authors, see here.
This README, includes key summary information on the evolution and variants of the VALOR SBN oscillation analysis, including information on which VALOR-SBN and VALOR-SDK branches, as well as which data products, are required for reproducing publicly-shown results. This README also lists the naming conventions used for labelling VALOR SBN analyses, analysis configurations and related data products. More detailed information can be found in the VALOR SBN wiki.
Contents
- List of VALOR branches and data products used for the main public results
- Labelling of VALOR analysis variants and of sets of data products
- Information on SBN MC productions used in VALOR
- Information on VALOR/SBN systematic error parameterizations
A number of SBN analysis variants can be supported in the code included in each VALOR-SBN branch. Although each VALOR-SBN branch may work with several different VALOR-SDK branches, the actual SDK branch that was used in the context of each analysis is also listed below, to provide a combination of products which are guaranteed to work and to allow the replication of analysis results.
| SBN branch | SDK branch | Supported SBN analysis variants |
|---|---|---|
For each set of data products required for running an analysis, VALOR uses the following naming scheme:
[sbn analysis type]_[sbn mc production]_[systematic error parameterization] (eg.: sbn_numu_disapp 2019a1 02a1)
where the conventions used for enumerating a) MC productions and b) VALOR/SBN systematic error parameterizations will be summarised at a later section.
These labels can be used for downloading all relevant data products (belonging to that set) using the '%valor_data-fetch-set --data-set=label ...' command.
For more information on the VALOR data distribution scripts, see https://github.com/VALOR-fit/VALOR-SDK/tree/master/config/data_distribution
A list of VALOR SBN data products can be found in: https://github.com/VALOR-fit/VALOR-SDK/blob/master/config/data_distribution/products/sbn.dat
Several analysis variants can be built using the same set of data products. These analysis variants are labelled as:
[set of data products]_[revision] (eg.: see above 1)
These names of analysis variants are used for labelling the main analysis configuration XML files made available in the VALOR-SBN repository. A list and a brief description of our main SBN analysis variants is given below. Analysis variants supported by the version of the code that includes this file as shown in bold.
| Analysis label. | Type | Samples | Fit distributions | MC pro | Syst. | Comment |
|---|---|---|---|---|---|---|
| sbn_numu_disapp_2019a1_01a1_1 | Eff. 3+1 numu disap | numuCC-like (x3 det) | Enu_reco | 2019a1 | 01a1 | |
| sbn_numu_disapp_2019b1_01a1_1 | Eff. 3+1 numu disap | numuCC-like (x3 det) | Enu_reco | 2019b1 | 01a1 | |
| sbn_numu_disapp_2019b2_01a1_1 | Eff. 3+1 numu disap | numuCC-like (x3 det) | Enu_reco | 2019b2 | 01a1 | |
| sbn_numu_disapp_2019b2_02a1_1 | Eff. 3+1 numu disap | numuCC-like (x3 det) | Enu_reco | 2019b2 | 02a1 | |
| sbn_numu_disapp_2019b2_02a2_1 | Eff. 3+1 numu disap | numuCC-like (x3 det) | Enu_reco | 2019b2 | 02a1 | |
| sbn_numu_disapp_prism_1_2019b3_02c2_1 | Eff. 3+1 numu disap | Event samples: 8 SBND numu CC-like samples at different angular bins (0 to 1.6 deg with 0.2 deg increments) + 1 uB numu CC-like sample + 1 ICARUS numu CC-like sample | Enu_reco | 2019b3 | 02c2 | First iteration of SBN Prism analysis in devel |
| sbn_numu_inclusive_2020a1_03d1_1 | Eff. 3+1 nue | numuCC-like (x3 det) | Enu_reco | 2020a1 | 03d1 | |
| sbn_nue_inclusive_2020a1_03d1_1 | Eff. 3+1 nue | nueCC-like (x3 det) | Enu_reco | 2020a1 | 03d1 |
MC productions (flux, interaction, detector simulations), the corresponding uncertainty assignments and internal VALOR representations of the uncertainties, and the reconstruction / data reduction event selection codes are important elements used for characterising each fit. Information about the elements used by each VALOR analysis is outlined below.
VALOR uses the following naming scheme for SBN MC productions and the reconstructed event samples obtained b
[date][physics model][revision]
(eg.: 2020 a 1)
A summary of all SBN MC productions used is given in the table below.
| SBN pro | Flux | GENIE (vrs;tune;err) | SBNCode | Comments |
|---|---|---|---|---|
| 2019a1 | configF | v2.12.10; Default+MEC; proposal | v2.2 | `proposal'; last respin of 2019a still in use while establishing this archive on 04/2020 |
| 2019b1 | configF | v2.12.10; Default+MEC; proposal | v2.2 | `modern'; last respin of 2019b still in use while establishing this archive on 04/2020 |
| 2019b2 | configF | v2.12.10; Default+MEC; proposal | v2.2 / gputnam-workshop (numu); feature/DomBarkerNueSelection (nue) | `modern'; batch of files made available after the 09/2019 SBN analysis workshop |
| 2020a1 | configF | v3.0.6; G18_10a_02_11a; proposal+MEC | gputnam-workshop0320-gen (numu); feature/TomHamNue (nue) | `modern'; batch of files made available after the 03/2020 SBN analysis workshop |
- [1] https://cdcvs.fnal.gov/redmine/projects/sbncode/wiki/Fitter_Validation_NuMu_sample: SBNCode v2.2, 'modern' and 'proposal'
- [2] https://cdcvs.fnal.gov/redmine/projects/sbn-analysis-group/wiki/How_to_access_Workshop_Osc_samples: SBNCode workshop-game samples, modern only from here on
VALOR uses the following naming scheme for its internal parameterizations of the SBN systematic error models (more information on which can be found in the VALOR SBN wiki here). A number of different generations of error parameterizations can exist, each with several versions, each of which can have numerous minor revisons:
[generation][version][revision] (eg.: 01 a 1)
A summary of all error parameterizations implemented in the VALOR SBN analysis is given below.
Each entry includes information on the internal parameterization of detector, flux and neutrino interaction uncertainties and provides link with further details on corresponding SBN systematic error model being parameterized. Each error parameterization can be applied on selected MC productions only, indicated on the table below.
| Err param | Detector | Flux | Interaction | MC Pro | Comments |
|---|---|---|---|---|---|
| 01a1 | n/a | CORLIN{loc,flav,Ereco} parameterization of the 2015a flux error model | CORLIN{Ereco} parameterization of the 2015aP x-section error model | Last version of the 01 parameterization still at use while establishing this table on 04/2020; Used for the 01/2020 Fermilab PAC studies and the associated SBN-OSC WG tech note | |
| 02a1 | n/a | CORLIN{loc,flav,Etrue} parameterization of the 2015a flux error model | CORLIN{Ereco} parameterization of the 2015aP x-section error model | Stopgap implementation towards the full 02 parameterization, implemented at the 03/2020 SBN analysis workshop; Wrt to 01, application of flux parameters was swapped from Ereco to Etrue | |
| 02a2 | CORLIN{det,sample,mode,Ereco}, as specified by the effective 2020a detector error model | as in 02a1 | as in 02a1 | Adding detector effects in 02a1 | |
| 02b1 | n/a | CORLIN{loc,flav,Etrue} parameterization of the 2015a flux error model | UNCORRESP1D parameterisation for all systematics included in the 2015aP x-section error model | Stopgap implementation towards full 02 parameterization; Wrt to 02a, switch x-section systematics from normalisation to non-linear response functions. | |
| 02b2 | CORLIN{det,sample,mode,Ereco}, as specified by the effective 2020a detector error model | as in 02b1 | as in 02b1 | Adding detector effects in 02b1 | |
| 02c1 | n/a | CORLIN{loc,flav,Etrue} parameterization for hadroproduction multi-sim systematics and UNCORRESP1D parameterization of optics and hadronic interaction uni-sim systematics of the 2015a flux error model | UNCORRESP1D parameterisation for all systematics included in the 2015aP x-section error model | Wrt to 03a, take all unisim flux systematics out of the flux covariance matrix and implement non-linear response functions. | |
| 02c2 | CORLIN{det,sample,mode,Ereco}, as specified by the effective 2020a detector error model | as in 02c1 | as in 02c1 | Adding detector effects in 02c1 | |
| 03a1 | n/a | As in 02a1 but with updated modes and sample definitions | As in 02a1 but with updated modes and sample definitions+MEC | ||
| 03b1 | n/a | As in 02b1 but with updated modes and sample definitions | As in 02a1 but with updated modes and sample definitions+MEC | ||
| 03c1 | n/a | CORLIN{loc,flav,Etrue} parameterization for hadroproduction multi-sim systematics and UNCORRESP1D parameterization of optics and hadronic interaction uni-sim systematics of the 2015a flux error model | UNCORRESP1D parameterisation for all systematics included in the 2020a x-section error model | ||
| 03c2 | CORLIN{det,sample,mode,Ereco}, as specified by the effective 2020a detector error model | as in 03c1 | as in 03c1 | Adding detector effects in 03c1 | |
| 03c3 | as in 03c2 | as in 03c2 | as in 03c2 | Adding bias and energy scale systematics to 03c2 and sample migration in exclusive analysis | |
| 03d1 | as in 03c1 | as in 03c1 | as in 03c1 | Adding POT uncertainty which should become part of the 'standard' analysis | |
| 03e1 | as in 03d1 | as in 03d1 | as in 03d1 | Adding 2pct Energy Scale uncertainty (initially implimented in 03c3) which should become part of the 'standard' analysis |
Internal systematic error parameterization methods:
- CORLIN{'space'}: A set of correlated normalization factors (linear response functions), each of which is applied on a single bin of the specified 'space' that can be mapped to subsets of kinematic bins for specific MC templates.
- UNCORLIN{'space'}: As above, but normalization factors are uncorrelated.
- CORRESP1D{'params'}: For each systematic parameter, non-linear 1D response functions are constructed, each of which is acting on all (or a subset of) kinematic bins of selected MC templates. The systematic parameters driving the non-linear response functions are correlated.
- UNCORRESP1D{'params'}: As above, but systematic parameters driving the non-linear response functions are uncorrelated.