IPP Software Navigation Tools IPP Links Communication Pan-STARRS Links

Changeset 41312 for trunk


Ignore:
Timestamp:
Mar 20, 2020, 10:46:42 AM (6 years ago)
Author:
eugene
Message:

various referee comments

Location:
trunk/doc/release.2015/ps1.analysis
Files:
2 edited

Legend:

Unmodified
Added
Removed
  • trunk/doc/release.2015/ps1.analysis/analysis.tex

    r41311 r41312  
    120120% \end{verbatim}
    121121
    122 The 1.8m Pan-STARRS\,1 telescope is located on the summit of Haleakala
     122The 1.8m Pan-STARRS\,1 telescope (PS1) is located on the summit of Haleakala
    123123on the Hawaiian island of Maui.  The wide-field optical design of the
    124124telescope \citep{2004SPIE.5489..667H} produces a 3.3 degree field of view with
     
    151151details of the telescope, surveys, and resulting science publications
    152152are described by \cite{chambers2017}.
     153
     154Since 2014 March, PS1 has been re-dedicated to a mission of searching
     155for hazardous asteroids, funded by the NASA NEO Program. Additional
     156partners collaborate with the Pan-STARRS team to harvest the transient
     157sources such supernovae and graviational wave counterparts
     158\note{REFS}.  A second Pan-STARRS telescope (PS2), generally matching
     159the PS1 design \citep{Morgan2012} has since been constructed and has
     160been producing science results since early 2018.
    153161
    154162%The Processing Version 3 (PV3) reduction represents the third full
     
    337345image, or a group of related images representing the data read from
    338346\textmod{the multiple chips of a mosaic
    339 camera from} a single exposure.  \textadd{In the IPP sequencing, this step is
    340 called the \ippstage{chip} stage.}  The images are expected to have already
     347camera from} a single exposure.  The images are expected to have already
    341348been detrended so that pixel values are linearly related to the flux.
    342349The gain may be specified by the configuration system, or a variance
     
    347354program \ippprog{ppImage} during the \ippstage{chip} analysis stage.
    348355
     356\textadd{In the standard IPP analysis, the initial stage of processing
     357  is performed in parallel on each of the individual CCDs in the
     358  camera.  This so-called \ippstage{chip} stage analysis includes the
     359  detrending of CCD image as well as the detection and analysis of
     360  sources in the image using the basic version of \ippprog{psphot}.
     361  The next stage of the analysis, the \ippstage{camera} stage,
     362  consists of photometric and astrometric calibration.}
     363
     364  \textadd{After the calibrations are available, the detrended CCD images from
     365  an entire exposure are geometrically transformed to a common pixel
     366  grid in the \ippstage{warp} stage of the pipeline.  The resulting
     367  warped images are generated on a pre-define tessellation of the sky
     368  which starts with projection centers spaced roughly
     369  4\degrees\ across the sky.  Around each of these projection centers,
     370  a large regular pixel grid is defined, and then subdivided along
     371  pixel boundaries into smaller units which are well-matched to the
     372  memory footprint of our processing computesr.  These smaller images,
     373  called `skycells' are defined with 1 arcminute of overlap with their
     374  neighbors to that any modest-sized object can be analysed entirely
     375  on a single pixel grid.  Note that the term skycell is used to
     376  describe the particular subdivision of the sky.  A typical exposure
     377  from the GPC1 camera generates warp images on roughly 70 skycells.
     378  We refer to the specific warped images from an exposure as `warps'.}
     379
     380  \textadd{Multiple warps for the same skycell are combined together in the
     381  \ippstage{stack} stage of the IPP by co-adding the flux to generate
     382  a deep `stack' image.  Alternatively, one warp may be subtracted
     383  from another warp of the same skycell, or a stack image may be
     384  subtracted from a warp image, or indeed from another stack.  These
     385  subtraction operations are used to detect moving and transient
     386  objects within the IPP.  Different variants of \ippprog{psphot} are
     387  used for the source detection and analysis for each of these
     388  different analysis stages.}
     389
    349390The variant called \ippprog{psphotStack} accepts a set of images, each
    350391representing the same patch of sky \textadd{(with pixels aligned)} in
    351 a different \textmod{filter.  This version was used for the analysis
    352   of the deep ``stacks'' (co-added images combining multiple
    353   observations of the same field) produced by the IPP \ippstage{stack}
    354   stage.  Nominally,
    355 the full $grizy$ filter set was used for the analysis} of the PS1 PV3 stack
    356 images, though where insufficient data were available in a given
    357 filter, a subset of these filters was processed as a group.  As
    358 discussed in detail below, the \ippprog{psphotStack} analysis includes the
    359 capability of measuring forced PSF photometry in some filter images
    360 based on the position of sources detected in the other filters.  It
    361 also includes an option to convolve the set of images to a single,
    362 common PSF size across the filters for the purpose of fixed aperture
    363 photometry.
     392a different \textmod{filter.  This version was used in the IPP for the
     393  analysis of the deep ``stacks'' produced by the IPP \ippstage{stack}
     394  stage.  Nominally, the full $grizy$ filter set was used for the
     395  analysis} of the PS1 PV3 stack images, though where insufficient
     396data were available in a given filter, a subset of these filters was
     397processed as a group.  As discussed in detail below, the
     398\ippprog{psphotStack} analysis includes the capability of measuring
     399forced PSF photometry in some filter images based on the position of
     400sources detected in the other filters.  It also includes an option to
     401convolve the set of images to a single, common PSF size across the
     402filters for the purpose of fixed aperture photometry.
    364403
    365404Another variant of \ippprog{psphot} used in the PV3 analysis is called
     
    16381677
    16391678\subsubsection{Fast Ensemble PSF Fitting}
     1679\label{sec:ensemble.fitting}
    16401680
    16411681Before the detailed analysis of the sources is performed, it is
     
    29002940\end{figure*}
    29012941
    2902 \section{Forced Photometry Modes}
     2942\section{Forced Warp Analysis}
    29032943\label{sec:psf.forced.fit}
    29042944
     
    29512991
    29522992The IPP analysis solves this problem by starting with the sources
    2953 detected in the stack images and performing forced photometry on the
    2954 individual warp images used to generate the stack, and then combining
    2955 the resulting measurements to determine a high-quality average value.
    2956 This forced-photometry analysis is performed using the
     2993detected in the stack images and performing \textmod{a highly
     2994  constrained analysis on the individual warp images used to generate
     2995  the stack, and then combining the resulting measurements to
     2996  determine a high-quality average value.  We consider all of these
     2997  measurements to be ``forced'' because of the strong prior
     2998  constraints from the stack.}  This analysis is performed using the
    29572999\ippprog{psphotFullForce} variant of \ippprog{psphot}.
    29583000
    2959 In this program, the positions of sources are loaded from the output
    2960 catalog of the stack photometry.  Candidates PSF stars are
    2961 pre-identified as those stars used to generate the PSF model in the
    2962 stack photometry analysis.  A PSF model is generated for each input
    2963 warp image based on those stars; PSF stars which are excessively
    2964 masked on a particular image are not used to model the PSF.  The PSF
    2965 model is fitted to all of the known source positions in the warp
    2966 images.  Aperture magnitudes, Kron magnitudes, and moments are also
    2967 measured at this stage for each warp.  Note that the flux measurement
    2968 for a faint, but significant, source from the stack image may be at a
    2969 low significance (less than the $5\sigma$ criterion used when the
     3001\subsection{Forced PSF Photometry}
     3002
     3003\textmod{PSF photometry is measured for all objects detected in the
     3004  stack using the positions determined by the stack photometry
     3005  analysis.}  Candidate PSF stars are pre-identified as those stars
     3006used to generate the PSF model in the stack photometry analysis.  A
     3007PSF model is generated for each input warp image based on those stars;
     3008PSF stars which are excessively masked on a particular image are not
     3009used to model the PSF.  The \textadd{normalization of the} PSF model
     3010is fitted to all of the known source positions in the warp images to
     3011determine the PSF fluxes. This measurement is performed simultaneously
     3012for all sources in the image at once using the method described above
     3013(Section~\ref{sec:ensemble.fitting}).
     3014
     3015\textmod{Aperture fluxes, Kron fluxes}, and moments are also measured at
     3016this stage for each warp.  Note that the flux measurement for a faint,
     3017but significant, source from the stack image may be at a low
     3018significance (less than the $5\sigma$ criterion used when the
    29703019photometry is not run in this forced mode) in any individual warp
    29713020image; the measured flux may even be negative due to statistical
    29723021fluctuations.  When combined together, these low-significance
    2973 measurements result in a significant measurement as the signal-to-noise
    2974 increases with the combination of more data.
     3022measurements result in a significant measurement as the
     3023signal-to-noise increases with the combination of more data.
    29753024
    29763025Individual warp images are processed independently with separate
     
    30223071normalization value is determined.  The integrated flux, flux error,
    30233072and the $\chi^2$ value for each grid point are recorded.
     3073\textadd{This analysis is performed on the warp images one object at a
     3074  time with the other objects in the image subtracted according to
     3075  the best model currently known.}
    30243076
    30253077For a given galaxy, the result is a collection of $\chi^2$ values,
     
    30713123For the Pan-STARRS $3\pi$ PV3 analysis, we have measured the full set
    30723124of KSB lensing parameters for all objects with measured second moments
    3073 (i.e.,, excluding saturated stars, suspected cosmic rays, and other
     3125(i.e., excluding saturated stars, suspected cosmic rays, and other
    30743126likely defects) of the data to enable both lensing studies and binary
    30753127/ multiple star searches.  Here we describe the measurements as
    30763128performed within \ippprog{psphot}, reviewing the mathematical
    30773129framework as described by \cite{1995ApJ...449..460K} and
    3078 \cite{1998ApJ...504..636H}.
     3130\cite{1998ApJ...504..636H}.  \textadd{Just like the forced PSF
     3131  photometry and the constrained galaxy models above, this analysis is
     3132  performed by measuring the KSB lensing parameters on the individual
     3133  warp images and averaging these measurements for each object.}
    30793134
    30803135The goal of the KSB technique is to measure the intrinsic ellipticity
     
    30993154$e_1$ values.  An object with a position angle on the 45\degree\ lines
    31003155between the pixel axes will have large positive or negative values of
    3101 $e_2$ and low absoluate values of $e_1$.
     3156$e_2$ and low absolute values of $e_1$.
    31023157
    31033158Note that in our analysis of the second moments, we are applying a
     
    32113266  sm}_\alpha$) and the shear polarizability ($X^{\rm sh}_{\alpha
    32123267  \beta}$, $e^{\rm sh}_\alpha$) for all objects on each of the warp
    3213 images.  We have also selected only the PSF stars from the images and
    3214 interpolated a smoothed version of these parameters to the location of
    3215 the objects, using the grid described above to interpolate the PSF
    3216 parameters.  We also determine the interpolated PSF ellipticities
     3268images.  \textmod{We have also selected only the PSF stars from the images and
     3269interpolated a smoothed version of these parameters from the PSF stars
     3270to the locations of
     3271all objects, using the grid described above to interpolate the
     3272parameters.  We then determine the ellipticities of the stars}
    32173273($e^*_1, e^*_2$) from the equivalent smooth grid.  Thus, for every
    32183274object in the $3\pi$ survey, we are able to report the PSF and object
     
    32413297which generates those image. 
    32423298
     3299\note{Note that this article is limited to the analysis of the
     3300  difference image detections, and that additional work is needed to
     3301  filter real/bogus.  Refer to Denneau et al 2013 PASP for the MOPS analysis.  Refer
     3302  to the Wright et al papers for the SNe classifications (& other
     3303  papers?).  Mention Yuan \& Akerloff 2008.}
     3304
     3305\note{mention the 3 difference image modes (WW, WS, SS)}
     3306
     3307% https://ui.adsabs.harvard.edu/abs/2013PASP..125..357D/abstract 
     3308
    32433309The analysis of the difference image follows the same basic steps as
    32443310other \ippprog{psphot} versions with some minor modifications (see
     
    32473313performed.  The PSF model construction stage is not possible in the
    32483314difference image due to the lack of valid sources.  Instead, the PSF
    3249 model from is generated from the positive image, after PSF-matching
     3315model is generated from the positive image, after PSF-matching
    32503316but before the subtraction is performed.  Because we do not expect to
    32513317have a large number of sources, only a single source detection pass is
     
    33673433repeated measurements of the same exposures). 
    33683434
     3435% PS2 reference:
     3436% https://ui.adsabs.harvard.edu/abs/2012SPIE.8444E..0HM/abstract
     3437
    33693438\acknowledgments
    33703439
  • trunk/doc/release.2015/ps1.analysis/response.txt

    r41310 r41312  
    518518are averaged.
    519519
     520** We agree that the organization of this section could be improved,
     521but we disagree a bit with the proposal.  We consider all of these
     522analyses to be forced in that the fits are highly constrained by the
     523stack priors, with their positions (for PSFs and the lensing anaysis)
     524and most of the structural parameters fixed.  In the extended galaxy
     525analysis, each of the grid steps is completely forced, just like in
     526forced PSF photometry.  By using the term 'forced' to describe this
     527process, we would like to make it clear that, for each grid position,
     528the parameters (except normalization) are completely constrained.
     529
     530We have renamed Section 6 as Forced Warp Analysis and split out the
     531PSF vs extended source analysis sections as recommended.  We added
     532some explanation at the end of the section to explain what we mean by
     533'forced'.
     534
    520535- The general description of the section should end with "variant of psphot",
    521536with the motivation being written such that it applies to both the
    522537averaged forced photometry on stars and the averaged single-epoch fits
    523538for galaxies
     539** see previous
    524540
    525541Sec 6.1 (now 6.2):
     
    532548- The terms "skycell" and "warp image" are first used here without
    533549definition. Are warp images the same as CAMERA and CHIP?
    534 **
     550** Updated Section 2 to outline the relevant processing stages and
     551define 'warp', 'skycell', and 'stack' more cleanly.
    535552
    536553- For the forced photometry on single epoch images, is this a joint
    537554fit for overlapping objects?
    538 Are overlapping galaxy models subtracted first? "The PSF model is
    539 fitted..." -> "The amplitude of the PSF model is fitted as the flux..."
     555** for the galaxy models, no.  The galaxy model fits are measured on
     556each object with the other objects subtracted.  Text added to this
     557effect in Section 6.1
     558
     559Are overlapping galaxy models subtracted first?
     560** yes, see above.
     561
     562"The PSF model is fitted..." -> "The amplitude of the PSF model is fitted as the flux..."
     563** fixed the wording here and also clarified below that we measure
     564Kron and aperture fluxes as well.
    540565
    541566Sec 6.2 (now 6.3):
     
    543568parameters are an average of all of the single-epoch measurements.
    544569This should be stated upfront, and then again at the end of the subsection
    545 clarifying how the PSF systematics are removed. Phrases like
     570clarifying how the PSF systematics are removed.
     571
     572Phrases like
    546573"interpolated PSF ellipticities" are confusing, when I *think* what's
    547574being used are "interpolated star ellipticities".
     575** reworded this to make the analysis clearer
    548576
    549577- Was this lensing code used in any of the GREAT challenge papers,
     
    554582- "ie,," -> "ie,"
    555583" absoluate" -> "absolute"
     584** fixed both
    556585
    557586Sec 7:
     
    561590If these questions are covered in other Pan-STARRS papers, please reference
    562591them here.
     592****
    563593
    564594- A basic piece of information that should be given is whether the differencing
    565595is performed on pairs of images, or single (warp) images compared to
    566596image stacks.
     597****
    567598
    568599- "model from is" -> "model is"
     600** fixed
    569601
    570602Conclusions:
Note: See TracChangeset for help on using the changeset viewer.