From 6fb7681394377400ba9c77d5aaf3a681a0175bb7 Mon Sep 17 00:00:00 2001 From: Anne Heavey Date: Wed, 16 May 2018 16:27:18 -0500 Subject: [PATCH] phys chap updates from Sam --- executive-summary/chapter-science.tex | 5 ++--- executive-summary/chapter-strategies.tex | 8 ++++---- 2 files changed, 6 insertions(+), 7 deletions(-) diff --git a/executive-summary/chapter-science.tex b/executive-summary/chapter-science.tex index ef86d4a..00d15d3 100644 --- a/executive-summary/chapter-science.tex +++ b/executive-summary/chapter-science.tex @@ -77,15 +77,14 @@ \subsection{Context for Discussion of Science Capabilities in this Document} The sections that follow highlight the projected capabilities of DUNE to realize the science program summarized above. These are documented in detail in \href{http://arxiv.org/abs/1512.06148}{Volume 2 -of the DUNE Conceptual Design Report}. Since publication of the CDR in late 2015, the DUNE science -collaboration has undertaken a campaign to develop data analysis tools and strategies to aid +of the DUNE Conceptual Design Report} and in the following section. Since publication of the CDR in late 2015, the DUNE science collaboration has undertaken a campaign to develop data analysis tools and strategies to aid in detector design optimization as well as to obtain a more rigorous understanding of experimental sensitivity. This campaign is in progress as of this writing, and the outcomes will be reported as a component of the DUNE Technical Design Report now in development. Additionally, with currently-operating experiments beginning to reach peak fractional rates of integrated exposure, the rapid evolution of the world-wide experimental landscape in neutrino physics is particularly acute at present. Thus, for the purposes of the present report, the discussion of capabilities here -will reflect what is documented within the CDR unless otherwise noted. +will reflect what is documented within the CDR unless otherwise noted. In addition, the following section describes the status of the simulation and reconstruction strategies used to assess the physics requirements for DUNE. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %\subsection{RYAN'S SECTION HERE} diff --git a/executive-summary/chapter-strategies.tex b/executive-summary/chapter-strategies.tex index 4297b28..df1b329 100644 --- a/executive-summary/chapter-strategies.tex +++ b/executive-summary/chapter-strategies.tex @@ -1,5 +1,5 @@ %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% -\subsubsection{Simulation and Reconstruction} +\subsubsection{Simulation and Reconstruction Strategies} \label{sec:exec-summ-strat-simreco} Liquid argon time projection chambers (\lartpc{}s) provide a robust and elegant method for measuring the properties of neutrino interactions above a few tens of MeV by providing \threed event imaging with excellent spatial and energy resolution. The state of the art in \lartpc event reconstruction and particle identification is evolving rapidly and will continue to do so for many years. The adoption of the common framework \larsoft{}\footnote{\larsoft, \url{http://inspirehep.net/record/1598096/export/hx}.} by several \lartpc experiments facilitates the exchange of tools and ideas. @@ -9,13 +9,13 @@ \subsubsection{Simulation and Reconstruction} \subsubsubsection{Simulation Chain} Simulated events are created in four stages: event generation, {\sc GEANT4} tracking, TPC and \dword{pds} signal simulation, and digitization. The first step is unique to each sample type while the remaining steps are common for all samples. Beam neutrino, atmospheric neutrino, and nucleon decay events are generated using {\sc GENIE} appropriately configured for each. Supernova events are generated using the new low-energy, argon-specific MARLEY generator~\cite{marley}. Cosmogenic events at depth are generated using MUSIC (Muon Simulation Code) \cite{MUSICPaper} and MUSUN (Muon Simulations Underground) \cite{Kudryavtsev:musun}. -The truth particles generated in the event generator step are passed to a {\sc GEANT4}-based detector simulation. Energy depositions are converted to ionization electrons and scintillation photons, with recombination, electron attenuation, and diffusion effects included. The response of the photon detectors is simulated using a ``photon library'' that has precalculated the likelihood for the propagation of photons from any point in the detector to any \dword{pds} element. The response of the TPC induction and collection wires is based on a detailed GARFIELD~\cite{garfield} simulation. Throughout, measurements from test stands or from operating \lartpc experiments are used to establish simulation parameters, where possible. +The truth particles generated in the event generator step are passed to a {\sc GEANT4}-based detector simulation. Energy depositions are converted to ionization electrons and scintillation photons, with recombination, electron attenuation, and diffusion effects included. The response of the photon detectors is simulated using a ``photon library'' that has precalculated the likelihood for the propagation of photons from any point in the detector to any \dword{pds} element. The response of the TPC induction and collection wires is based on a detailed GARFIELD~\cite{garfield} simulation. Throughout, measurements from test stands or from operating \lartpc experiments such as ICARUS, LArIAT, and MicroBooNE are used to establish simulation parameters, where possible. The raw signals on each wire are converted into \dword{adc} versus time traces by convolution with the field response and electronics response. \dword{asic} electronics response is simulated with the BNL SPICE~\cite{spice} simulation. The photon detector electronics simulation separately generates waveforms for each channel of a photon detector that has been hit by photons, with dark noise and line noise added. The raw data are passed through hit finding algorithms that handle deconvolution and disambiguation to produce the basic data used by the downstream event reconstruction. \dword{pds} signals are reconstructed by searching for peaks on individual channels and then forming coincidences across channels. Techniques for matching the correct \dword{pds} signal to TPC signals to reconstruction $t_0$ are being developed, and early results from these tools can be seen in Chapter~\ref{PDS sim physics chapter}. \fixme{not sure where to reference. anne} \subsubsubsection{Reconstruction and Event Identification} -Several approaches to \lartpc reconstruction are under active development in DUNE and in the community at large. En route to the TDR, efforts on all fronts have been supported. One reconstruction path (``TrajCluster'' + ``Projection Matching'') forms \twod trajectory clusters in each detector view and then stitches these together into \threed objects. Resulting objects are further characterized by, for instance, extracting $dE/dx$ information or comparing to electromagnetic shower profiles. An alternative approach is provided by the Pandora reconstrucion package~\cite{Marshall:2015rfa}, in which the reconstruction and pattern recognition task is broken down into a large number of decoupled algorithms, where each algorithm addresses a specific task or targets a particular topology. Two additional algorithms (``WireCell'' and ``SpacePointSolver'') take a different approach and create \threed maps of energy depositions directly by solving a constrained system of equations governed by the geometry of the TPC wires. Finally, several analyses are using deep learning and convolutional neural networks with great success, as these techniques are well suited to the type of data produced by \lartpc{}s. +Several approaches to \lartpc reconstruction are under active development in DUNE and in the community at large. En route to the TDR, efforts on all fronts have been supported. One reconstruction path (``TrajCluster'' + ``Projection Matching'') forms \twod trajectory clusters in each detector view and then stitches these together into \threed objects. Resulting objects are further characterized by, for instance, extracting $dE/dx$ information or comparing to electromagnetic shower profiles. An alternative approach is provided by the Pandora reconstruction package~\cite{Marshall:2015rfa}, in which the reconstruction and pattern recognition task is broken down into a large number of decoupled algorithms, where each algorithm addresses a specific task or targets a particular topology. Two additional algorithms (``WireCell'' and ``SpacePointSolver'') take a different approach and create \threed maps of energy depositions directly by solving a constrained system of equations governed by the geometry of the TPC wires. Finally, several analyses are using deep learning and convolutional neural networks with great success, as these techniques are well suited to the type of data produced by \lartpc{}s. Energy reconstruction is based on electron-lifetime-corrected calorimetry except in the case of muons where energy is determined from track range or (for uncontained muons) multiple Coulomb scattering. Moving forward, more particle-specific energy estimators will be developed. -The output from all reconstruction algorithms is processed into standard ntuple files for use by analysis developers. In the special case of long-baseline oscillation measurements, the CAFAna fitting toolkit developed originally for NOvA is used to combine far detector and near detector information, to assess the impact of systematic uncertainties, and to ultimately produce oscillation sensitivities. +The output from all reconstruction algorithms is processed into standard ntuple files for use by analysis developers. In the special case of long-baseline oscillation measurements, the CAFAna fitting toolkit developed originally for NOvA is used to combine far detector and near detector information, to assess the impact of systematic uncertainties, and to ultimately produce neutrino oscillation sensitivities, discussed next.