UVES data reduction cookbook
Contents
1. 4 Define the order positions The order positions are usually defined by means of the so called order flatfields flatfield exposures obtained with a narrow slit producing thin echelle orders Again the first step is to make a BDF frame from the FTTS file Midas gt SPLIT UVES order FF fits In this example the order FF fits is an order flatfield of the blue arm The output file will be stored in the local directory as order_FF_b bdf We can now create a reference catalog to store all the needed calibration frames To allow a pipeline check on the guess and final order tables alignment we add also in the reference catalogue the guess order table Midas gt crea icat refB cat o346BLUE tbl DO CLASSIFICATION Now it is possible to determine the order positions by giving the following command Midas gt ORDERP UVES order FF b bdf refB cat refB cat which creates an order table a background table and a DRS setup table frame DO CLASSIFICATION meaning 0346 2x1 tbl ORDER TABLE BLUE Order table b346 2x1 tbl BACKGR TABLE BLUE Background table d346_2x1 tbl DRS SETUP BLUE DRS setup table 1346BLUE tbl LINE TABLE BLUE Line Guess table All these tables will be stored in the output catalog refB cat To proceed in the data re duction the line guess table 1346BLUE tbl in our case which we rename as 1346_2x1 tbl to evidence the bin setting and the line reference table thargood 3 tbl should be added to this catalog Midas gt re2. Having a so called format check frame which is a ThAr exposure taken with a very small slit length you are able to generate a line table which can be used as a first guess solution This determination is based on computations of a physical model of UVES For further information iflmidas is an alias to if it does not exist it is useful to create it inmidas j Qd pipeline start Cd pipeline control D set context flames PIPE HOME uves context mid mode 3 0 please use mid mode 3 0 if you have bash UVES data reduction cookbook VLT MAN ESO 13200 4033 3 on the physical model refer to P Ballester and M R Rosa 1997 AASuppl 126 563 As in almost all cases you will start with Midas gt SPLIT UVES format_TAL fits Assume format_TAL fits is a format check frame of the blue arm then SPLIT UVES will translate the input file to MIDAS BDF format file and transform the frame in the way that the wavelength increases from left to right and from bottom to top standard orientation The output file will be stored in the local directory as format_TAL_b bdf In general the command SPLIT UVES frame fits generate the MIDAS format file s frame x bdf x b or 1 u for arm Blue and Red respectively For the following UVES context command we need first to transform a line reference table from FITS to MIDAS format Midas gt INDISK FITS thar tfits thargood 3 tbl Next you pass the transformed calibration frame to PREDICT UVES Midas gt P
3. longname tbl 1 2 Reference documents 1 UVES User Manual VLT MAN ESO 13200 1825 Issue 83 28 08 2008 A Kaufer S D Odorico L Kaper C Ledoux G James H Sana 2 UVES Templates Reference Guide VLT MAN ESO 13200 1567 Issue 83 28 08 2008 A Kaufer C Ledoux G James H Sana 3 UVES Calibration Plan VLT PLA ESO 13200 1123 Issue 83 28 08 2008 A Kaufer R Hanuschik C Ledoux G James 4 UVES Pipeline User s Manual VLT MAN ESO 19500 2964 Issue 8 12 10 2007 P Ballester O Boitquin A Modigliani S Wolf 1 3 Abbreviations and acronyms The following abbreviations and acronyms are used in this document SciOps Science Operations ESO European Southern Observatory MIDAS Munich Image Data Analysis System FITS Flexible Image Transport System VLT Very Large Telescope UVES data reduction cookbook VLT MAN ESO 13200 4033 2 2 Prepare the UVES Midas session 1 Start the FLAMES UVES Midas session flmidas The pipeline environment will automatically be setup by the two procedures Cd pipeline start and d pipeline control Furthermore the FLAMES and the UVES MIDAS context will be initialized 2 Configuration of the display At the Midas prompt type Midas gt CONFIG DISPL Three image displays and two graphic windows will be created This is a standard setup used by the UVES pipeline procedures Internally some other MIDAS keywords supporting the graphics and display handling will be set They will be accessed d
4. Y Y it may happen that due to a big value of the object offset automatically determined during data reduction one sky window is less than 4 pixels wide so that the data reduction procedure automatically switches to one from a default value of two sky extraction window See also the help of the command REDUCE UVES for more information on how to properly set user defined extraction parameters in case of optimal extraction For operational purposes from pipeline version 1 1 1 on we have decided to produce dummy solutions also relative to the extraction window which is automatically suppressed This is to keep the same order in the pipeline data products Such solutions which have no physical meaning are labeled with the prefix dummy in the file name The real solutions are as before the one obtained considering only the good sky extraction window The errorbar image filenames having prefix errm is obtained as the square root of the variance frame varm The variance frame is calculated considering the contribution from the read out noise and from the source In case of average extraction this is calculated per pixel The variance of the flat fielded object is next given propagating the variance for the ratio UVES data reduction cookbook VLT MAN ESO 13200 4033 11 extracted object extracted flat field In case of optimal extraction an input variance is calculated as described above This will follow transformations
5. allows one to use UVES REDUCE to reduce more than one source on the slit interactively determining LOAD IMA LOAD ECH GET CURS and setting SET ECHELLE the values of the three involved parameters De fault setting for P7 is Y Y Y which means automatic determination of the three parameters To use OPTIMAL extraction with multi sources the real keyword OBJSET has to be appro priately set see more in section 13 The command REDUCE UVES will reduce every science frame stored in the input catalog using the appropriate calibration frames from the reference catalog refB cat So for each configura tion there has to be a complete reference set In this example there has to be two sets one for the lower red and one for the upper red arm In principle you are able to mix blue and red arm exposures for the science reduction process Finally all products will be stored in the output catalog sc redB cat The following data products will be created only the products of the lower CCD EEV of the red arm are shown UVES data reduction cookbook VLT MAN ESO 13200 4033 9 in principle they are always the same for the other configurations Filename Format ESO PRO CATG Description r rbf 0 1 bdf 1D wav REDUCED SCI POINT REDL extracted flatfielded wavelength calibrated merged sky subtracted science frame m rbf 0 1 bdf 1D wav MERGED_SCI_POINT_REDL extracted flatfielded wavelength calibrated merged science frame wxb
6. classified products which may be iden tified by the UVES commands These tables control the whole reduction process If you are changing some of the ECHELLE context keywords using SET ECHELLE you should store these changes for later use in a DRS table Midas gt SAVE DRS drs_346 tbl This call will save all ECHELLE keywords in the MIDAS table drs_346 tb1 as MIDAS descrip tors existing DRS tables will be overwritten UVES data reduction cookbook VLT MAN ESO 13200 4033 14 11 Message level The output of information is reduced to a minimum for use of the UVES context within the pipeline infrastructure When using the context in an interactive mode it could be helpful to get some more information In this case you may control the message level by Midas gt VERBOSE OUT VERY Instead of VERY you may use ON or OFF which switches back to the default message level or even more switches all the messages off except for warnings and error messages 12 Automatic preparation of calibration solutions To quickly prepare calibration solutions with default parameter settings it is worth to describe the use of the script uves_popul sh included as part of the distribution in PIPE_HOME uves uves scripts directory which should be included in your local PATH This script can be executed from any shell with the following syntax we suppose to be in the directory where the raw FITS files are located PIPE_HOME uves uves scripts uves_popul sh r
7. guess solution 4 Define the order positions 5 Wavelength calibration 6 Master calibration files 7 Science frame reduction 8 Calibration scripts 9 UVES data display and hardcopy 10 Saving the keyword setup 11 Message level 12 Automatic preparation of calibration solutions 13 Reduction of more than one object source on the slit 14 Reduction of extended sources 15 Session example Blue Data 15 1 Default Displey Initialization es s o osan AN kb E e A 15 2 Predictive Format Determination a a a a a a 15 3 Order Position Determination lt gt lt oe siare daro iras 15 4 Wavelength Calibrations se ss A ea d moeta da STA A KORo 15 5 Master Bias Determination 15 6 Master Flat Determination TaT peente SUECO o as RARA 16 Session example Red Data 161 Default Display Initialization ss s s siose ica a 16 2 Predictive Format Determination ss ica oe ee eS 16 3 Order Position Determination e e a e a a a a a a a a Se bah Fa Fi 12 13 13 14 14 15 18 UVES data reduction cookbook VLT MAN ESO 13200 4033 vi 164 Waveleneth Calibration e versa a A 20 16 5 Master Bias Determination gt s s ea setora setora ie ERA ee pe 21 16 6 Master Flat Determination 21 16 7 Science Reduction 21 UVES data reduction cookbook VLT MAN ESO 13200 4033 1 1 Introduction 1 1 Purpose This document describes the usage of the commands of the MIDAS context UVES which allow the user to perf
8. has to convert the FITS data to MIDAS format 13 Reduction of more than one object source on the slit The UVES pipeline has been designed to do automatic data reduction of point like sources well centered on the slit Some users may be interested to observe and reduce data of more than one object on the slit For this purpose we have upgraded some commands to allow the user to interactively perform a proper extraction This is still possible in a manual session using the UVES context Here we give a small example which may be adapted to the user s needs Let s suppose the user has on the slit two adjacent spectra Let s suppose the user has used the uves popul sh to prepare all the data calibrations During this step the echelle funda mental data tables are also produced yORDER tbl yLINE tbl yB ACKGR tbl y BLUE or REDU REDL Let s suppose these tables the calibration solutions and the science raw data to be reduced are available in our directory here for example indicated with raw_sci_two_sources_x bdf We would give the following commands Midas gt SPLIT UVES raw_sci_two_sources_x bdf Midas gt INIT ECHE y Midas gt LOAD ECHE Midas gt GET CUR With this last command you overplot the detected order trace to have a reference to get the extraction parameters offset slit skywind this last for average extraction were x b or x 1 x u and y BLUE or y REDL y REDU respectively for the BLUE or REDL REDU chip So you can calculate
9. is rotated flat fielded wavelength calibrated and finally merged The command to be used is Midas gt REDUCE SPAT split cat out cat ref cat BckMeasMeth FfMeth MerMeth MerSwitch deltai delta2 split cat is an input image catalog with images to be reduced oriented in the proper way SPLIT UVES for example the one produced as described in the normal data reduction for science data out cat is an output image catalog produced from the pipeline ref cat is the reference catalog for science data reduction produced as described in the section for normal data reduction BckMeasMeth is the background measurement method MIN MED see help subtract background FfMeth is the flat fielding method which can assume values E P or N with the same meanings as the corresponding parameter in the command REDUCE UVES MerMeth is the merging method which can assume values O Optimal A Average N Noappend with similar meaning as for the standard MIDAS command MERGE ECHELLE We suggest not to use REDUCE SPAT after having used on the same data REDUCE UVES In fact after the flat fielding or the background have been applied on the science frame the pipeline sets given descriptors so that it can recognize that the corresponding opera tion does not need to be repeated If the user has already processed a science frame with REDUCE UVES the science frame may be already background and flat field corrected if the flat field correc
10. similar to the flux until in the end after having evaluated the best Gaussian cross order profile coefficients it is evaluated for each X point the chi square between a normalized Gaussian times a variable amplitude to which is added the found background and the actual spectra doing the quadratic sum along the cross order direction and using as weight the input variance For a number of values of amplitudes one gets corresponding values of chi square Assuming that the chi square as a function of the amplitude is a parabola near the minimum one can calculate which is the change in amplitude that generates a unitary increase of the chi square This change can be assumed as an error associated with the amplitude and from this one can get an estimate of the variance associated to the optimal extraction process Clearly this variance value depends on how good is the Gaussian model approximation assumed from the cross order profile Up to version 2 9 7 optimal extraction was having problems apparent strange ripples and patterns within an order on a few pixel scale in particular for high S N data greater than around 50 Those problems were solved in the latest version For very high S N 200 data the user may still want to use the average extraction Remember also that Average extraction as suggested by the name makes an average of the extracted signal along the extraction slit So the intensities of data reduced with optimal and average extractio
11. Determination crea icat refB cat o346BLUE tbl DO CLASSIFICATION SPLIT UVES order_ff346 fits ORDERP UVES order_ff346_b bdf refB cat refB cat rename 1346blue tbl 1346_2x1 tbl add icat refb cat 1346_2x1 tbl add icat refb cat thargood 3 tbl Wavelength Calibration SPLIT UVES wcal346_TAL fits WAVECAL UVES wcal346_TAL_b bdf refB cat refB cat AUTO yes Master Bias Determination CREATE ICAT biasB cat bias346_ fits SPLIT UVES biasB cat split_biasB cat MASTER UVES split_biasB cat refB cat UVES data reduction cookbook VLT MAN ESO 13200 4033 20 15 6 Master Flat Determination Midas gt CREATE ICAT ffB cat ff346 fits Midas gt SPLIT UVES ffB cat split_ff346 cat Midas gt MASTER UVES split ff346 cat refB cat refB cat 15 7 Science Reduction Midas gt SPLIT UVES sc_346 fits Midas gt REDUCE UVES sc_346_b bdf reducedB cat refB cat E O MED 16 Session example Red Data We refer now to a case of RED arm data with wcent 580 nm and 1x1 binning 16 1 Default Display Initialization Midas gt CONFIG DISP 1600 1200 0 6 16 2 Predictive Format Determination Midas gt SPLIT UVES frmtChk580_TAL fits Midas gt INDISK FITS thargood 3 tfits thargood 3 tbl Midas gt PREDICT UVES frmtChk580 TAL 1 bdf thargood_3 tbl Midas gt PREDICT UVES frmtChk580 TAL u bdf thargood 3 tbl 16 3 Order Position Determination Midas gt crea icat refB cat o580REDL tbl DO CLASSIFICATION Midas gt add icat refB cat o580REDU tbl Midas gt SPLIT UVES order
12. E arm data binning 2x1 slit length 8 arcsec the following frames frame DO CLASSIFICATION meaning mf346_2x1_s08_b bdf MASTER FLAT BLUE master FF av346_2x1_s08_b bdf average FF bg346_2x1_s08_b bdf bkg The master Bias is subtracted from the master flatfield default option M of P4 and no dark is subtracted One can also subtract a constant bias level in this case P4 number this bias level number can be determined with STATISTIC IMAGE on different portions of the bias frames An inter order background is also determined and subtracted to the Flat Field frame the parameter P6 sets the method The master flat frame is added to the reference catalog Midas gt MASTER UVES split_ff cat refB cat refB cat 120 This example assumes a constant bias of 120 counts used for the master flatfield creation Furthermore the master flats will be background subtracted which requires appropriate back ground tables and DRS setup tables to be present in the catalog refB cat products of ORDERP UVES 7 Science frame reduction The science reduction for UVES supports different modes controlled via additional parameters ffmode Flatfielding may be done in the pixel pixel space P as well as in the extracted pixel order space E extract The extraction of the object may be performed as a simple average AVERAGE or by the optimal extraction method OPTIMAL The number of rows to be averaged per order is defined by the MIDAS keyword SL
13. EUROPEAN SOUTHERN OBSERVATORY ES Organisation Europ enne pour des Recherches Astronomiques dans l H misphere Austral O Europ ische Organisation ftir astronomische Forschung in der s dlichen Hemisphare ESO European Southern Observatory Karl Schwarzschild Str 2 D 85748 Garching bei Miinchen Very Large Telescope Paranal Science Operations UVES data reduction cookbook Doc No VLT MAN ESO 13200 4033 Issue 83 Date 14 09 2008 C Ledoux A Modigliani G James Prepared ces pee eee Rad EEO ees Cewek a ERS Date Signature G Marconi Approved warn Sao Se IA Date Signature O Hainaut Released REA Date Signature UVES data reduction cookbook VLT MAN ESO 13200 4033 This page was intentionally left blank UVES data reduction cookbook VLT MAN ESO 13200 4033 111 Change Record Issue Rev Date Section Parag affected Reason Tnitiation Documents Remarks Issue 79 04 10 2006 first standalone SciOps version Issue 82 22 05 2008 update for MIDAS pipeline version 2 9 7 Issue 83 14 09 2008 update of references for P83 UVES data reduction cookbook VLT MAN ESO 13200 4033 This page was intentionally left blank UVES data reduction cookbook VLT MAN ESO 13200 4033 Contents 1 Introduction LU AMOS os A A E RS et SRS E Cj ee E 1 2 Reference documents ooo e a 1 3 Abbreviations and acronyms NEEN ee ai 2 Prepare the UVES Midas session 3 Generate the first
14. IT bmeasure The inter order background subtraction is based on the measurements on the grid of background positions For each background position the median MEDIAN or the minimum MINIMUM within a certain window will be used as the measurement at that point In case of very narrow inter order space the minimum method could pro duce better background images as otherwise the measurements could be contaminated by neighboring orders Usually the method MEDIAN gives better results than the MIN IMUM From version 2 0 0 spectra merging can be controlled via parameter P8 which may have 4 com ponents merge_method delta_set_switch deltal delta2 merge_method is the method used to merge spectra OPTIMAL or AVERAGE delta_set_switch is a parameter used to set UVES data reduction cookbook VLT MAN ESO 13200 4033 8 D Default delta setting as was in previous pipeline releases for BLUE arm deltal delta2 3 for RED arm deltal delta2 5 A Automatic setting of deltas Appropriate deltas are chosen for each instrument setting See on line help U User defined deltas In this case are taken the values of deltal and delta2 as specified by the user deltal user specified value of delta used to merge the blue edge of the spectra delta2 user specified value of delta used to merge the red edge of the spectra The possible parameter values are shown in brackets As usual at first you have to trans form your science frame Midas
15. LLE GET CUR with respect to a reference position for example the order trace Be ord trace the position of the order trace Be offset the offset chosen for the extraction this being the center of the extraction slit slit_ext The parameter objset used in the optimal extraction is the distance of the object from the center of the extraction slit slit_ext It is used as first guess to start the Gaussian fit of the object s light distribution cross order profile within the extraction slit Next the optimal extraction algorithm searches automatically for the best object position to achieve within an order an overall best fit of the cross order profile These parameters are related by the following relation offset objset 0bj_trace order_trace The slit of integration will be centered at the position order trace offset The user could get such a formula taking into account the previous information for example doing a plot in which one has for a numerical example an object trace at position 40 an order trace at position 20 and the extraction slit centered at position 25 Let s also suppose that one would like to have an extraction window of 36 pixels We have chosen this sequence as it is simpler and all the parameters are positive With our numbers objset object trace order trace offset 40 20 25 20 40 20 5 15 which is actually what one can measure on a scaled plot Obviously objset and offset have a sign and the situation can change If o
16. REDICT UVES format TAL b bdf thargood 3 tbl The only auxiliary file is a line reference list of a ThAr lamp By measuring the line positions identifying them through the physical model and comparing them with the line reference list this command will finally produce a line table which may work as a first guess solution for IDENT ECHE or WAVECAL UVES see below making the wavelength calibration step automatic In particular this command will produce for a blue frame and central wavelength 346 nm the following files frame DO CLASSIFICATION meaning drs setup BLUE tbl DRS SETUP BLUE calibration table b346BLUE tbl BACKGR TABLE BLUE background table 1346BLUE tbl LINE_TABLE_BLUE line table guess solution o346BLUE tbl ORDER GUESS TAB BLUE order table guess solution At this point the ECHELLE context parameter NBORDI is equal to zero This means that an automatic determination of the orders is performed In principle one could also pass reference frames through a catalogue and give the command Midas gt PREDICT UVES format TAL b bdf predictI cat predict0O cat where predictl cat is an image catalog which must contain the line reference table In this case the output names given above will be present after data reduction in the output catalog predictO cat The line table 1346BLUE tbl is the Guess Solution and will be classified in our case in the catalog as LINE TABLE BLUE UVES data reduction cookbook VLT MAN ESO 13200 4033 4
17. ated f flatfielded x extracted b background subtracted data file The prefix m indicates merged data which are implicitly always wfxb data UVES data reduction cookbook VLT MAN ESO 13200 4033 10 Filename Format ESO PRO CATG Description w_xb_sky_REDL1 bdf 2D wav ord WCALIB_SKY1_REDL extracted wavelength calibrated sky 1 frame wfxb_sky REDL1 bdf 2D wav ord WCALIB_FF_SKY1_REDL extracted flatfielded and wavelength calibrated sky 1 frame m_sky_REDL1 bdf 1D wav MERGED_SKY1_REDL extracted wavelength calibrated flat fielded merged sky 1 frame w_xb_sky_REDL2 bdf 2D wav ord WCALIB SKY2 REDL extracted wavelength calibrated sky 2 frame wfxb_sky_REDL2 bdf 2D wav ord WCALIB FF SKY2 REDL extracted wavelength calibrated flatfielded sky 2 frame m sky REDL2 bdf 1D wav MERGED_SKY2_REDL extracted wavelength calibrated flat fielded and merged sky 2 frame m_sky_REDL bdf 1D wav MERGED_AV_SKY_REDL extracted flatfielded wavelength calibrated merged average of sky 1 and sky 2 frame w xb_rbf 8 bdf 1D wav WCALIB FLAT OBJ REDL extracted wavelength calibrated flatfield of the object w xb1 rbf 8 bdf 2D wav ord WCALIB FLAT SKY1 REDL extracted wavelength calibrated flatfield w_xb2_rbf_8 bdf 2D wav ord WCALIB FLAT SKY2 REDL of the two sky windows Note also that with automatic determination of the offset slit skywind parameters P7 Y
18. aw fmtchk fits raw_orderpos fits raw_wavecal fits raw bias fits raw Hat fits ThAr ReferLineTable fits where the input raw data refers to a coherent instrument setting same instrument arm mode central wavelength and binning the script starts a MIDAS session and produces results in the directory HOME midwork tmpwrk In particular this creates a subdirectory SHOME midwork tmpwrk data which contains all the calibration solutions Moreover in SHOME midwork tmpwrk the files xORDER tbl xLINE tbl xBACKGR tbl will be present to be used as input reference of the MIDAS command INIT ECHELLE x x BLUE or REDL REDU This command sets up all the important keywords of the echelle environment The uves_popul sh is a script which executes the procedure uves_prepcalib prg This procedure executes in series all the main UVES pipeline reduction steps involved in the calibration data analysis Mainly it executes the command CONFIG INSTR on the reference catalog to get set or check mode the detector involved the central wavelength the binning factor if a dichroic is inserted Next it executes the first main data reduction step it runs the physical model to determine the geometrical predicted spectral format a first guess solution for the line dispersion relation and the order tables and generates a first DRS SETUP table if a reference formatcheck frame MASTER FORM x x BLUE or REDL REDU is provided it also does a QC stability check Th
19. er k sigma clipping of residual outliers and in magenta are shown the points used to determine the last fit Good extraction is typically reached when the black points are well fit by the blue ones As the plots show the point distribution has usually a well aggregated parabolic distribution with sloppiness and curvature typically of a small fraction of pixel So in general the fit for position and FWHM is parabolic In case the data distribution is not well aggregated this may happen for particularly low S N data it might happen that the fit gives too big values for the sloppiness and the curvature To prevent such a problem thresholds and checks on these parameters are set in the code so that if the fit is not very good and the parabola fit parameters are wrong the fit is switched first to linear and eventually to uniform The linear or even uniform approximation is a safer approximation than a parabolic one to fit a highly scattered point distribution The user should always check to have a reasonably good fit max scatter in y bin should be 0 1 0 3 bins UVES data reduction cookbook VLT MAN ESO 13200 4033 12 Having a proper master response frame provided by DFO and adding it in the input reference catalogue one could also produce flux calibrated merged spectra having prefix flx_ 8 Calibration scripts In order to make life a bit easier three additional MIDAS procedures exist which may help to fill your calibration database for late
20. ese data ORDTAB DRSTAB are used in the following data reduction step the order position determination ORDERP UVES In this step an Hough Transform is performed to determine the order positions Initially it is given through a SAVINI ECH DRSTAB UVES data reduction cookbook VLT MAN ESO 13200 4033 15 READ command as parameter to DEFINE HOUGH the number of orders as predicted by the physical model previous step In standard configuration settings this number usually coincides with the actual number of orders on the detector On particular non standard configurations due to the presence of some filter along the path it may be possible that the detector illumination may drop to almost zero on some regions and for this reason the detected number of orders is lower than the one predicted by the physical model For this reason the order position procedure always does a quality control check check the standard deviation of the RESIDUAL column in the order table which may have a jump in case the predicted number of orders is greater than the detected one and eventually iteratively decrease the number of orders given as input parameter to DEFINE HOUGH Finally the ORDER BACKGR and DRS_SETUP tables are produced The following step is the wavelength calibration WAVECAL UVES Next the Master bias and the Master Flat frames are created To use the calibration database data in an interactive MIDAS section for example to reduce a science frame one
21. ff580 fits split order cat Midas gt ORDERP UVES split order cat refR cat Midas gt rename 1580REDL tbl 1580L_1x1 tbl Midas gt rename 1580REDU tbl 1580U_1x1 tbl Midas gt ADD ICAT refR cat 1580L_1x1 tbl Midas gt ADD ICAT refR cat 1580U_1x1 tbl Midas gt ADD ICAT refR cat thargood_3 tbl 16 4 Wavelength Calibration Midas gt SPLIT UVES wcal580 TAL fits split wcal cat Midas gt WAVECAL UVES split wcal cat refR cat refR cat AUTO yes UVES data reduction cookbook VLT MAN ESO 13200 4033 21 16 5 Master Bias Determination Midas gt Midas gt Midas gt 16 6 Midas gt Midas gt Midas gt Midas gt 16 7 Midas gt Midas gt CREATE ICAT biasR cat bias580_ fits SPLIT UVES biasR cat split_biasR cat MASTER UVES split biasR cat refR cat Master Flat Determination CREATE ICAT ffR cat ff580 fits SPLIT UVES ffR cat split_ff580 cat MASTER UVES split ff580 cat refR cat refR cat ADD ICAT refR cat mf580_1x1_s08_1 bdf mf580_1x1_s08_u bdf Science Reduction SPLIT UVES sc 580 fits split_sc cat REDUCE UVES split sc cat reducedR cat refR cat E O MED 000
22. g used The products generated by this procedure are the following If pl is the MIDAS procedure parameter specifying the input frame e xb2d_ pl is the background subtracted extracted rotated frame UVES data reduction cookbook VLT MAN ESO 13200 4033 e fxb2d pl is the background subtracted extracted rotated flatfielded frame 19 e w xb2d_ p1 is the background subtracted extracted rotated flatfielded wavelength calibrated frame e mwfxb2d_ p1 is the background subtracted extracted rotated flatfielded wavelength calibrated merged frame In case Noappend the merging option is chosen the procedure generates an image frame per each order with indexed names such as the following mwfxb2d_ p1 0001 mwfxb2d_ p1 00NN where NN is the number of extracted orders This last option may be used if the user has not found proper values of deltal and delta2 parameters 15 Session example Blue Data We refer now to a case of BLUE arm data with wcent 346 nm and 2x1 binning 15 1 Midas gt 15 2 Midas gt Midas gt Midas gt 15 3 Midas gt Midas gt Midas gt Midas gt Midas gt Midas gt 15 4 Midas gt Midas gt 15 5 Midas gt Midas gt Midas gt Default Display Initialization CONFIG DISP 1600 1200 0 6 Predictive Format Determination SPLIT UVES frmtChk346_TAL fits INDISK FITS thargood 3 tfits thargood 3 tbl PREDICT UVES frmtChk346 TAL b bdf thargood 3 tbl Order Position
23. gt SPLIT UVES sc fits split_sc cat We can reduce the data giving the following UVES context MIDAS command Midas gt REDUCE UVES split sc cat sc redB cat refB cat E OPTIMAL MEDIAN In this case we use P4 E meaning that the Flat Fielding is done in the pixel order space during extraction In case of data taken in the far Red wcal 860 to better correct for the fringing effect one could apply the P method i e doing the Flat Fielding before extraction in the pixel to pixel space P5 OPTIMAL means we choose optimal extraction This method has been proven to give good quality for low to medium signal to noise S N ratio science objects For very high S N it is suggested to use the average extraction method The optimal extraction may show quality problems appearing as sudden spikes on the spectra This occurrence can be confirmed also looking at the weight bdf weight image which for a successful extraction should appear uni form with only a few randomly scattered holes corresponding to detection and suppression of cosmic rays If instead significant portion of holes in the weight image with a periodicity are noticed this means that this step has failed and one should use average extraction P6 MINIMUM MEDIAN is the background estimation method See help SUBTRACT BACKGROUND for clarification Proper setting of parameter P7 allows one to chose if used P7 N N N one s own setting respectively for the offset slit skywind setting This
24. light distribution on the detector induced by filters eventually present along the light path the predicted number of orders is greater than the detected one In this case using the value of NBORDI contained in the DRS setup table generated from the physical model would lead to an overestimation of the number of orders and to a wrong solution For this reason in pipeline releases after version 1 0 2 a quality check on the predicted vs detected spectral format has been introduced The expected number of orders can be also set manually SET ECHELLE NBORDI No 5 Wavelength calibration For the wavelength calibration you will need a ThAr lamp exposure a line reference table and first guess solutions order table and line table which allows the automatic mode of the UVES command WAVECAL UVES The interactive mode may be enforced by its mode parameter The ThAr lamp exposure may be obtained using the observation template UVES lt mode gt y lt mode gt blue red dicl dic2 y wave wavefree Again at first you have to transform the original input file by Midas gt SPLIT UVES b346_TAL fits Assuming b346_TAL fits as an exposure of the blue arm the output from this command will be used as the input file of the wavelength calibration command The next UVES context MIDAS command uses for simplicity the input reference catalog name as the output catalog refB cat Midas gt WAVECAL UVES b346 TAL b bdf refB cat refB cat AUTO This comma
25. log Midas gt CREATE ICAT biasB cat bias346_ fits At first the set of input frames have to be transformed into the standard orientation wave length increases from left to right and from bottom to top and into the MIDAS BDF file format by means of Midas gt SPLIT UVES biasB cat split bias cat The transformed data will be stored in the output catalog split bias cat Having prepared the input data one can give the command Midas gt MASTER UVES split bias cat refB cat which produces a master frame for each configuration blue red arm lower and upper part i e you may use a mixed set of input frames e g blue bias frames red bias frames lower and upper part 5 frames each as a result you will get 3 master biases All products will be stored in the output catalog which again for simplicity has the same name as before refB cat This command will produce a master bias frame frame DO CLASSIFICATION meaning mbBLUE 2x1 b bdf MASTER BIAS BLUE Master Bias frame and its name will be added to the reference catalog For the flatfields and as before for simplicity we put all the Flat Field frames in one catalog Midas gt CREATE ICAT ffB cat ff346 fits Midas gt SPLIT UVES ffB cat split ff cat UVES data reduction cookbook VLT MAN ESO 13200 4033 7 The master Flat Field is saved in the usual catalog refB cat Midas gt MASTER UVES split_ff cat refB cat refB cat The products of this step are in our case of BLU
26. n differ approximatively by a factor equal to the slit size in pixel Flux calibrated merged spectra may be generated if a master response frame is added in the input reference catalog In this case using average extraction the products have same units as the ones generated using optimal extraction Because it is difficult to model the light profile coming from the image slicer and to estimate the sky contribution from the image slicer data should be extracted with AVERAGE method and NO sky subtraction For this instrument setting during science spectra extraction the pipeline recognizes if the frame has been created using an Image Slicer and in such a case it is automatically set to the relevant extraction parameters object extraction slit extraction method sky subtraction option Optimal extraction quality has been improved a lot in the latest release We can now prudently say that usually the extraction quality is quite good It is important to check it using the command Midas gt MPLOT CHUN order trace x half size y switch where x BLUE REDL or REDU half size y is half size in bin units of the plots in Y direction and switch can assume values pos ition or fwhm respectively for the plots of the cross order chunk position or FWHM distributions These plots display in black the values of the raw data pos FWHM for each chunk in green the values predicted from the first fit after some pre cleaning of outliers in blue the last fit aft
27. name 1346BLUE tb1 1346 2x1 tbl Midas gt ADD ICAT refB cat 1346 2x1 tbl Midas gt ADD ICAT refB cat thargood_3 tbl The catalog refB cat may now be used as a reference catalog for the next steps of the reduction procedure Instead of using order definition flatfields you may also use standard star exposures It is worth to mention here that this step has been performed without using a reference DRS table and with automatic order detection NBORDI 0 In this case the Hough Transform will determine the orders present on the frame In case of low photon level in part of the order definition frame this step may underestimate the number of orders To have a complete determination of the orders one should visually check that the number of determined orders corresponds to the one present on the frame If not it is better to manually set this number No using the command SET ECHELLE NBORDI No before giving the ORDERP UVES command The UVES pipeline indeed uses the results of the physical model which predicts the geometrical spectral format and thus also the number of orders which should be present on the frame This value is stored in the DRS setup table when the first guess solution UVES data reduction cookbook VLT MAN ESO 13200 4033 5 is generated This method is appropriate for the standard setting and is more robust and uniform than doing an automatic order detection But in case of a non standard setting it may well be that due to non uniform
28. nd performs the wavelength calibration using the following default options parameter value purpose P4 AUTO the previously determined line table from refB cat is used P5 yes the procedure generates the resolution plots P6 Y N performs Y N the wavelength calibration only at order center P7 Y N produced output FITS file P8 see on line help of command this parameter control offset and extraction window of object and sky for wavelength calibration solution This step generates the line tables for each slit window sky object sky which will be stored in the output catalog refB cat and updates the DRS SETUP x x BLUE in our example table UVES data reduction cookbook VLT MAN ESO 13200 4033 6 frame DO_CLASSIFICATION meaning 1346 2x1 1 tbl LINE_TABLE_BLUE1 line table lower sky 1346 2x1 2 tbl LINE TABLE BLUE2 line table object 1346 2x1 3 tbl LINE TABLE BLUES line table upper sky d346_2x1 tbl DRS SETUP BLUE DRS Setup Table As the ThAr line reference list is no a product of the UVES context you have to ensure that the descriptor ESO PRO CATG is set to LINE REFER TABLE READ DESC WRITE DESC 6 Master calibration files Master calibration frames master bias and master flatfields are used for the science reduc tion They are stacked median averages of a set of input frames They are created by the command MASTER UVES To keep the data reduction simple we collect all the bias frames in an image cata
29. ne puts the integration slit below the order trace the offset will be negative Similarly one can have a situation for which objset is negative To make things easier if the extraction window is centered on the object this means objset 0 the parameter offset measures the distance of the object trace from the order trace which is exactly what one would imagine When is all this important Usually one will have only one object in the slit and in such a case offset is automatically determined by offset echelle so that one can take objset 0 as the pipeline does in default mode and the optimal extraction will start to search for the object at the slit center without any problem A more interesting case happens when there is more than one object in the full slit and in particular if the two sources are very close to each other as may happen for traces of lensed quasars or in a binary system Obviously one does not want to use a slit covering both objects UVES data reduction cookbook VLT MAN ESO 13200 4033 17 otherwise the spectral information coming from the two spectra will be mixed moreover doing so the optimal extraction would get crazy trying to fit both traces It is also suggested not to have a small integration window centered on each object objset 0 as probably one would cut off part of the object and or not well estimate the sky In such a case it is better to chose an offset and an integration slit such that the slit include
30. orm a complete science data reduction For a thorough description of the UVES MIDAS based pipeline the reader is referred to the UVES Pipeline User s Manual Issue 8 We assume the user is familiar with the concepts of echelle data reduction and we suggest taking a look at the description of the MIDAS context ECHELLE The UVES context itself is based on the ECHELLE context The development of this context has been done under MIDAS version 98NOVpl2 1 and successive We describe here how to produce master calibration frames order position and background tables as well as line tables used for the re sampling into wavelength space Finally the science reduction command will be introduced All commands used are described in more detail in the help files HELP command name It may be useful to make printouts of these help files which is easily done using the graphical interface of MIDAS help CREA GUI help This cookbook also describes how to prepare calibration solutions either using commands of the context or scripts How to reduce multi object sources on the slit using average or optimal extraction is also described How to do a simple extraction of extended sources is shortly explained Finally session examples to perform data reduction of BLUE and RED data are reported In the following examples of UVES data reduction we suggest the user to adopt as temporary table products in MIDAS format names of maximum 8 characters plus extension tbl e g
31. plete set of calibration frames from the calibration database Midas gt GET CALDB ref346 2x1 cat data caldb UVES data reduction cookbook VLT MAN ESO 13200 4033 13 in order to be well prepared for the science reduction Midas gt REDUCE UVES sc346 2x1 b bdf sc346 cat ref346_2x1 cat E OPT MED For more details and additional options please read the on line help of the commands 9 UVES data display and hardcopy The echelle data may be displayed using the command PLOT UVES For a detailed description please see the help file HELP PLOT UVES Midas gt PLOT UVES extract bdf 1 13 0 100 title extract ps Furthermore a hardcopy utility is provided HARDCOPY PLOT especially for hardcopies of the image displays as the usual hardcopy command COPY DISP only properly works for non covered displays Midas gt HARDCOPY PLOT P ff 346 ps ff 346 bdf This will produce a postscript hardcopy ff 346 ps of the input file ff 346 bdf Addi tionally the main characteristics will be printed at the bottom of the plot By default the hardcopy facility is disabled You may enable it doing Midas gt HARDCOPY PLOT ON Use OFF instead of ON in case you wish to disable hardcopies This may be useful as some UVES commands are sensible to the hardcopy command status 10 Saving the keyword setup The keywords used during the reduction of the data may should be saved in a so called data reduction system setup tables DRS tables These are
32. r science reduction The first command PREPARE CALDB puts all the calibration commands mentioned before together into one script so that one has only to create an input catalog for a certain UVES setting and pass it to this procedure First one collects in an image catalog all the main calibration fits files 1 raw formatcheck frame fits 2 raw order definition flatfield fits 3 ThAr lamp exposure fits 4 raw list of biases fits 5 raw list of flat fields fits And apply SPLIT UVES to get the data in the proper format and orientation Midas gt CREA ICAT raw fits cat fits Midas gt SPLIT UVES raw fits cat raw Split Cat Next one prepares a catalog refer cat containing the ThAr line reference table in the MIDAS format Midas gt INDISK FITS thargood 3 tfits thargood 3 tbl Midas gt CREA ICAT refer cat null DO CLASSIFICATION Midas gt ADD ICAT refer cat thargood_3 tbl Finally one applies the PREPARE CALDB script with the following syntax Midas gt PREPARE CALDB raw_split cat refer cat Assume one has created an input catalog for binned data 2x1 of the central wavelength 346 nm then after having executed the script one will get at the end all the necessary calibra tion solutions which are listed in an output catalog ref346_2x1 cat Use Midas gt SAVE CALDB ref346_2x1 cat data caldb to store the solutions in one s calibration database The third command allows one to retrieve the com
33. rbf 0 1 bdf 2D wav ord WCALIB_SCI_POINT_REDL extracted wavelength calibrated science frame wfxb rbf 0 1bdf 2D wav ord WCALIB FF SCI POINT REDL extracted flatfielded wavelength calibrated frame errmrbf 0 1 bdf 1D wav ERRORBAR SCI POINT REDU standard deviation of reduced science frame var rbf 0 Lbdf 1D pix ord VARIANCE SCI POINT REDU variance of flatfielded extracted science frame reduced debiased inter order background subtracted flatfielded re sampled merged and sky subtracted data merged merged orders no sky subtraction one dimensional wavelength calibrated re sampled extracted orders sky sky contribution is determined from the two sky windows below sky 1 and above the object sky 2 in each order For the optimal extraction the weights used for the object extraction are applied to the averaged sky From pipeline release 1 3 0 on are also created in case of optimal extraction frames with optimally extracted sky These contains in the frame name the sequence opt sky The second column of the table shows the hierarchical FITS header keyword for the product category By means of this keyword the products may easily be identified The MIDAS output catalog uses this keyword as identifier field In case of the upper CCD of the red arm the category extension REDL changes to REDU and for the blue arm to BLUE The prefix of the filenames also immediately shows the different product types w wavelength calibr
34. s one object but not the adjacent and a lot of sky on one side of the object It is not a good idea to have in this case objset 0 and leave the optimal extraction search for the object position as in this particular situation the object will be near the slit border and the algorithm may not be clever enough to find the object For this reason one has to specify objset the starting offset with respect to the slit center Following these indications and choosing an extraction slit size such that at least three pixels are left on each side of the object one can iteratively do optimal extraction of all the sources If all is done correctly in the setting of these parameters one can notice that the object position reported by the optimal extraction at each order varies slightly with the order position and it is quite close to the value s1it_ext 2 objset set from the user It is always a good practice after optimal extraction to use the command MPLOT CHUN to display the trace object positions or FWHM as a function of X and verify that a good fit was obtained It is only necessary to check one trace This means that the extraction slit includes only one object Moreover the magenta points should fit well to the dark ones this being an indication of a good extraction Another interesting test one could do is to get the best combination of parameters to have a reasonable extraction and next satisfying the formula above move the extraction window un
35. the values of off slit skywind skywind in general is a 4 component pa rameter skywind skyw 1 skyw 2 skyw 3 skyw 4 for each of the object to be extracted Midas gt SET ECHE offset off1 slit slit1 skywind skywind1 Midas gt SAVINI ECHE drstab tbl Midas gt REDUCE UVES raw_sci_two_sources_x bdf out cat ref cat E A MED N N N UVES data reduction cookbook VLT MAN ESO 13200 4033 16 Where drstab tbl indicates the drs setup table being used Here we use P7 N N N meaning we have dis activated N the automatic setting respectively of the offset the slit and the sky window used in average extraction values and used the corresponding values set manually SET ECHE We have also saved our setting in the drs setup table In our case we have used average extraction but we could also use optimal extraction Similarly the other object can be reduced It is probably useful to write some more here in case the user would like to extract two sources on the slit using optimal extraction and a manual setting of the extraction parameters For simplicity we start the discussion with the extraction of a single source This case is typically treated in an automatic way by the pipeline The discussion helps to understand the meaning of each relevant parameter also for the more general case in which one has to extract more than one object case in which automatic extraction would fail Be obj_trace the position of the object as measured LOAD IMA LOAD ECHE
36. til the object exits from it At this point the optimal extraction will start to have problems giving warnings like Warning IMASK_COUNTER LESS THAN 10 meaning that only a small number of chunks are left after a k sigma clipping step over position or FWHM values a situation typical of very low S N data even more if no signal is in the extraction window as it can happen at a certain point in the proposed exercise In such last case the plots from MPLOT CHUN and of the extracted spectra will be much worse After such explanations we add only how in practice we could activate such settings using the numbers given Midas gt SET ECHE offset 5 slit 36 Midas gt write key objset r 1 1 15 Midas gt SAVINI ECHE drstab tbl Midas gt REDUCE UVES raw_sci_two_sources_x bdf out cat ref cat E O MED N N N Midas gt MPLOT CHUN order trace y bdf 3 obj Midas gt MPLOT CHUN order trace y bdf 3 fwhm where x b or x 1 x u and y BLUE or y REDL y REDU respectively for the BLUE or REDL REDU chip Here we have also reported the command to check after extraction the quality of the order tracing UVES data reduction cookbook VLT MAN ESO 13200 4033 18 14 Reduction of extended sources The possibility to do simple reduction of extended sources has also been included on the pipeline version 1 0 6 on In this case the source is extracted with a 1 bin extraction slit and a variable offset scanning the full length of the observation slit So the order
37. tion has been performed pixel to pixel thus the user may not be able to choose and do the proper background and flat field methods offered by the REDUCE SPAT command itself Default values for parameter P4 are MED E A One could use as MerMeth O to have a better behaviour in the overlapping region between one order and the next or if not satisfied the Noappend option to have each order in one corresponding image file MerSwitch is the parameter controlling the setting of deltas used in the merging of the spectra It can have values D Default A Auto or U User defined which have the same meaning as the corresponding subparameter of parameter P8 in REDUCE UVES command Deltal controls the amount of overlapping considered in the merging of the blue edge of a spectra delta2 controls the amount of overlapping considered in the merging of the red edge of a spectra Using option A the pipeline will use predefined deltal and delta2 parameter values As those parameters affect significantly the quality of merged spectra we suggest the user to use the U option and choose appropriate values for deltal and delta2 See also on line help of parameter P5 for command REDUCE SPAT Blue data for setting 346 might be better reduced using a NO flat fielding mode because in the short wavelength range of this setting the flat field data might have been taken in non appropriate conditions This rarely happens as LN2 lamps with a proper behaviour are bein
38. uring the reduction process The commands CREATE GRAP or CREATE DISPL should not be used in this context By default CONFIG DISP assumes a 1280x1024 pixel sized monitor In case your monitor is smaller you may reset it using Midas gt CONFIG DISPL 1200 900 where a x and y dimension respectively of 1200 and 900 pixels is assumed You may add a third parameter the fill parameter add 0 8 in order to use only 80 of your terminal 3 In the course of this cookbook you will often be confronted with DRS data reduction system setup tables These are empty tables which control the data reduction process by the use of their descriptors All global keywords of the ECHELLE package are stored in these descriptors In principle DRS tables are classified saved sessions see SAVE ECHELLE These tables guarantee a standardized behavior of the UVES pipeline DRS tables may be created using SAVE DRS In an interactive mode which is described here you may switch off the strict use of DRS tables by setting FORCE_DRS NO 7 In that case the commands use the current setting SHOW ECHE You may control the process by changing keywords using SET ECHE Some keywords are con trolled by the UVES commands For a list of restricted keywords see the help files 4 An example on how to make use of the reference catalogs which are used in almost all commands of the UVES context is described in the following sections 3 Generate the first guess solution
Download Pdf Manuals
Related Search