.help Tutorial Aug86 "Echelle Tutorial" .ih TOPICS The echelle tutorial consists of a number of different topics. To obtain help on a particular topic type "tutor topic" where the topic is one of the following: .nf TOPICS topics - List of topics overview - An overview of the echelle reduction process dataio - Reading and writing data tapes tracing - Tracing the positions of the orders extraction - Extracting one or two dimensional spectra references - Additional references all - Print all of the tutorial .fi The topics are kept brief and describe the simplest operations. More sophisticated discussions are available for all the tasks in the printed documentation and through the on-line \fBhelp\fR facility; i.e. "help taskname". .ih OVERVIEW The \fBechelle\fR package provides for the extraction of the orders from two dimensional echelle images into one dimensional spectra. After extraction the one dimensional spectra are wavelength and flux calibrated. The usual flow of the reductions is .ls (1) Read data from tape. .le .ls (2) Set the dispersion axis in the image headers using the task \fBsetdisp\fR. This is required by many of the tasks which follow. .le .ls (3) Trace one or more images to define the positions of the orders within the two dimensional format. .le .ls (4) Extract the orders into one dimensional spectra. .le .ls (5) Use arc calibration spectra to determine wavelength solutions. .le .ls (6) Apply the wavelength solutions to the other spectra and rebin the spectra into linear or logarithmic wavelength intervals. .le .ls (7) Determine flux calibrations using standard star observations. .le .ls (8) Apply the flux calibrations to the other object spectra. .le .ls (9) Save the reductions as FITS images and make plots of the spectra. .le There are many variations on these steps possible with the great flexibility of the reduction tools at your disposal. The most important one to mention is that the orders may be extracted as two dimensional strips in order to apply more complex geometric distortion corrections using the \fBlongslit\fR package. .ih DATAIO To read CCD Camera format tapes use \fBrcamera\fR from the \fBmtlocal\fR package. FITS format tapes are read and written with \fBrfits\fR and \fBwfits\fR from the \fBdataio\fR package. Remember you need to \fBallocate\fR the tape drive before you can read or write tapes and you should \fBdeallocate\fR the tapes when you are through with the tape drive. .nf ec> allocate mta ec> deallocate mta ec> rcamera mta 1-99 ech datatype=r >rcam.log & ec> rfits mta 1-99 ech datatype=r >rfits.log & ec> wfits mta spectra* .fi .ih TRACING The positions of the orders across the image dispersion axis as a function of position along the dispersion axis are determined by the task \fBtrace\fR. There are three steps in tracing an image; defining the initial positions of the orders at one point along the dispersion, automatically determining the positions at other points in steps from the starting point, and fitting smooth curves to the positions as a function of dispersion position. The first and last steps are interactive, at least initially. After the first image other images may be traced noninteractively. Select an image with narrow, strong profiles and run trace: ec> trace imagename When you are asked if you want to edit the apertures respond with "yes". The central cut across the dispersion is graphed. Position the cursor over the first order to be traced and type 'm'. Adjust the width of the extraction aperture with the 'l', 'u', or 'y' keys or specify the lower and upper widths explicitly with ":lower value" or ":upper value". If background subtraction is to be used type 'b' and set the background fitting parameters (see the \fBbackground\fR tutorial) Now mark the remaining orders with the 'm' key. The widths of the previous aperture are preserved for each new aperture. When you are satisfied with the marked apertures type 'q'. The positions of the orders are now traced in steps from the initial point. Once the positions have been traced you are asked whether to fit the traced apertures interactively. Respond with "yes". You will now be asked specifically if the first aperture is to be fit. Respond with "yes" again. The traced positions are graphed along with a fitted curve. You now have many options for adjusting the fit. The most important one is the order which is set by typing ":order value", where value is the desired order, and then 'f' to refit the data. For full information of the options see the help for \fBicfit\fR. When you are satisfied type 'q'. You are then prompted for the next order. The previous fitting parameters will be used so at this point you may want to just answer "NO" to skip the interactive fitting of the other traced orders, though the graphs of the fit will still be displayed. You now have several options about how to define the positions of the orders in your other images. .ls (1) You may apply the tracing to all other observations with no further tracing. This is done by specifying the traced image as the "reference" in the extraction process. .le .ls (2) You may maintain the same shape of the traces and correct for shifts in the positions of the orders across the dispersion by recentering each aperture. This is done with the task \fBapedit\fR or the editing switch during extraction using the first traced image as the reference. The apertures are recenter using the 'c' key. .le .ls (3) Finally, you may retrace other images either from scratch or using the first traced image as the initial reference. In the latter case the tracing may be done noninteractively as a batch process. .le .ih EXTRACTION There are two types of extraction; to one dimensional spectra or to two dimensional strips. The second type of extraction is accomplished by the task \fBstripextract\fR in the \fBtwodspec.apextract\fR package and is used if further reductions using the \fBlongslit\fR package are desired. Normally, however, one ignores the small geometric distortion in which curves of constant wavelength differ slightly from the image dispersion axis. Extraction of the traced echelle orders is performed by the task \fBsumextract\fR. The pixels within each aperture at each point along the dispersion axis are summed to produce one dimensional spectra, one for each order and each extracted image. The sum may be weighted in two ways; "profile" or "variance" weighting. The variance weighting may require that you know the CCD readout noise and photon/ADU conversion. For a description of the weights see the help for \fBsumextract\fR or the paper "The APEXTRACT Package". The spectra may also be cleaned of cosmic rays and bad pixels at the same time and have a background subtracted. The background subtraction parameters must be set when defining the apertures or later using the apedit mode in \fBapedit\fR, \fBtrace\fR, or \fBsumextract\fR. See the tutorial on \fBbackground\fR for further information. Once the extraction parameters have been set simply type ec> sumextract images where images is the list of images to be extracted. If each image has not been traced then a traced reference image should be given. One may correct for shifts relative to the traced image by setting the switch to edit the apertures and then recentering each aperture before extracting. If there is no aperture editing then the extractions may be done as a background or batch process. .ih REFERENCES .ls (1) Pilachowski, C. and J. V. Barnes, \fINotes on the IRAF for Reduction of Echelle/CCD Data\fR, NOAO Central Computer Services, 1986. This document is also available in the \fBIRAF User Handbook , Vol. 2B -- NOAO Cookbooks\fR. .le .endhelp