Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.eso.org/~rfosbury/tmp/nicmoslook.pdf Дата изменения: Tue Oct 28 15:29:57 1997 Дата индексирования: Sat Dec 22 03:35:11 2007 Кодировка: Поисковые слова: http www.badastronomy.com phpbb index.php

SPACE TELESCOPE
European Coordinating Facility

NICMOSlook An Interactive Spectrum Extraction Program for NICMOS Grism Data Manual1 Version 1.1.5
Wolfram Freudling Norbert Pirzkal Robert Thomas Lin Yan August 27, 1997

An HTML version of this manual is available at http: ecf.hq.eso.org nicmos nicmoslook nicmoslook.html
1


2


Contents
1 Introduction 2 Installing NICMOSlook
2.1 2.2 2.3 2.4 2.5 Downloading the software . . Installation Procedure . . . . EnvironmentVariables . . . . Rebuilding the IDL NICMOSlo Customizing NICMOSlook . . .. .. .. ok .. .... .... .... binary .... .... .... .... Library .... . . . . . . . . . . . . .... .... .... ... .... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

5 7
7 7 8 9 9

3 Using NICMOSlook

3.1 Running the Program . . . . . . . . . . . . . . 3.1.1 Startup . . . . . . . . . . . . . . . . . 3.1.2 Input to NICMOSlook . . . . . . . . . 3.1.3 Output of NICMOSlook . . . . . . . . 3.1.4 Viewing the Postscript Spectra Output 3.2 The User Interface . . . . . . . . . . . . . . . 3.2.1 Startup Window . . . . . . . . . . . . 3.2.2 Displaying the Images . . . . . . . . . 3.2.3 Finding Ob jects and Extracting Spectra 3.2.4 Supplementary Pulldown Menus . . . . 3.2.5 Miscellaneous Buttons . . . . . . . . . 3.2.6 Plot Popup Window . . . . . . . . . . 3.2.7 Spectrum Popup Window . . . . . . . 4.1 NICMOSlook . . . . . . . . . . . . . . 4.1.1 Directory Structure . . . . . . . 4.1.2 NICMOSlook Setup Parameters 4.1.3 NICMOSlook Grismspec Entries 4.1.4 NICMOSlook Con gurable FITS 4.2 Debugging . . . . . . . . . . . . . . . . 3

11
11 11 12 12 12 12 12 14 14 17 19 20 21 25 25 26 29 29 30

4 Software Dependencies, System Requirements & Debugging
...... ...... ...... ...... Keyword ...... .... .... .... .... Values ....

25


4

CONTENTS
5.1 5.2 5.3 5.4 5.5 Image Display . . . . . . . . . . . Image Processing . . . . . . . . . Finding Ob jects and Determining Extracting Spectra . . . . . . . . Examining Your Spectra . . . . . ... ... Sizes ... ... ... ... .. ... ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

5 Tutorial

35
35 37 38 40 41

6 Acknowledgments A NICMOSlook Utilities

A.1 Test Image Generator . . . . . . . . . . . . A.1.1 Installation . . . . . . . . . . . . . A.1.2 Running the Test Image Generator A.1.3 Parameter Settings . . . . . . . . . A.2 Additional Utilities . . . . . . . . . . . . . A.2.1 Point Spread Function Generator . A.2.2 Flat eld "Cube" Generator . . . . A.2.3 Reading FITS Tables Into IRAF . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

45 47
47 47 47 49 51 51 52 52

B FITS Input File Format C FITS Output File
C.1 FITS C.1.1 C.2 FITS C.3 FITS

B.1 NICMOS Input FITS File Description . . . . . . . . . . . . . . . . . . . . 55 B.2 NICMOS FITS Header Keyword Requirements . . . . . . . . . . . . . . . 55 File Description . . . . . . . . . FITS Binary Table Description Primary Header Example . . . Table Header Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . .. .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 57 58 60 63 63 63 65 65

55 57

D Setup

D.1 Grism Speci cation File . . D.1.1 Contents . . . . . . . D.1.2 Format . . . . . . . . D.2 Tunable Parameters . . . . . D.3 Con guring FITS Keywords

63

E List of all input les F Output Files of NICMOSlook G Sample Data Files

F.1 Ob ject List . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 F.2 Derived Line Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 F.3 Catalogue Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

67 71 75


Chapter 1 Introduction
A unique capability of NICMOS is the grism mode, which permits slitless spectrometry at low resolution. Typically, a direct image is taken in conjunction with grism images for the wavelength calibration. A quick-look extraction of spectra from a large number of NICMOS grism images requires a convenientinteractive tool which manipulates the pair of images and extracts spectra. NICMOSlook is an IDL program designed for that purpose at the Space Telescope European Coordinating Facility1. NICMOSlook is the interactive counterpart to the Calnic C Facility2, a program which performs the same functionality in a "pipelined" approach. The most common use of NICMOSlook will be for small amounts of data, when users prefer to have full controll of all parameters for individual spectrum extraction, or for cases which Calnic C did not extract spectra in a satisfactory way. Unlike Calnic C, NICMOSlook requires the user to determine the best way to nd an ob ject and provides a number of di erent ways to accomplish this. Similarly, the user decides whether a weighting appropriate for point sources or weighting by the size of the ob ject is used for the extraction of the spectra. The actual extraction of spectra is done employing the same methods and algorithms as Calnic C. Refer to the chapter regarding Algorithms in the Calnic C manual for detailed descriptions. A brief overview of the program's capabilities follows. A direct image and corresponding grism images can be read from FITS les and are displayed. The display options include color tables, zoom factors, and blinking. Basic image manipulation capabilities are provided. They can be used to optimize the identi cation of ob jects. Ob jects are identi ed on the direct image either by user input of coordinates via a le or with the cursor, or by applying a user de ned threshold to automatically nd them. The positions of the ob jects are used to extract spectra from a grism image of the same region. Several options for the weighting are o ered. The wavelength calibration of the extracted spectra is performed using the position of the ob jects and parameterized dispersion relations. After extraction of the spectra, they are corrected for the wavelength dependence of the quantum e ciency of the detector. The ux scale is then computed using the standard
1 2

http: ecf.hq.eso.org http: ecf.hq.eso.org nicmos calnicc calnicc.html

5


6

CHAPTER 1. INTRODUCTION

NICMOS ux calibration data. The extracted spectra can be corrected for contamination from nearby ob jects. All extracted spectra are automatically searched for emission and absorption lines. In addition, the continuum emission is automatically determined. The nal data products are plots on the screen, binary FITS tables and postscript les with the spectra, error estimates, ob ject parameters derived from the direct imaging and details of the spectrum extraction process.


Chapter 2 Installing NICMOSlook
2.1 Downloading the software
The source code for NICMOSlook is available on the WWW as a GNU ZIP1 compressed tar les. After bringing up your favorite Web browser, click on the le of choice and select a directory in which to copy it from the le selection box that appears.

2.2 Installation Procedure
The following instructions assume that a compressed tar le nicmoslook version.tar.gz where version is the version number has been downloaded to a computer with a UNIX operating system running IDL. Speci c UNIX commands are given for the C-shell. Once the compressed tar le is downloaded to a local disk, the installation is as follows: 1. Create a new directory in which to install the program. For example, if you choose to install NICMOSlook in usr local nicmoslook, make sure that the le system has at least 10.0 MB free space and create the directory with: mkdir usr local nicmoslook. 2. Create a new environmentvariable called NICMOSLOOK BASE. This variable should be set to the value of the newly created directory in the step above. For example, if you chose usr local nicmoslook as the location for the installation, NICMOSLOOK BASE should be set to that ENTIRE path name. Create the environmentvariable with:
setenv NICMOSLOOK BASE

Since this variable is necessary for running and rebuilding NICMOSlook, you probably would want to include this statement to your shell's startup script e.g. .cshrc. 7

usr local nicmoslook

1

http: ecf.hq.eso.org nicmos nicmoslook latest.tar.gz


8

CHAPTER 2. INSTALLING NICMOSLOOK
3. Create a new environmentvariable called NLK BIN. This variable is necessary to rebuild the IDL NICMOSlook binary library See section 2.4 and should be set to the bin subdirectory of the NICMOSlook installation: 4. Change to the newly created NICMOSlook directory
cd $NICMOSLOOK BASE setenv NLK BIN $NICMOSLOOK BASE bin

and extract the distribution. Remember that the full path name of the downloaded distribution must be supplied. For example, if you downloaded the NICMOSlook distribution tar le to the directory tmp you would extract it with one of the following commands: zcat tmp nicmoslook version.tar.gz j tar xf 5. NICMOSlook currently comes pre-compiled to run under IDL 5.0. If you are planning to run NICMOSlook under IDL 4.0 then you will need to recompile NICMOSlook. See 2.4 for more information. NICMOSlook is now installed on your system.

2.3 EnvironmentVariables
In addition to the environmentvariables NICMOSLOOK BASE and NLK BIN which must be de ned and whichwere discussed in section 2.2, a few additional variables can be de ned: The two following variables allow the user to specify where NICMOSLOOK will look for input data les and where it will write its output les. If these are not de ned, the CURRENT working directory where NICMOSlook was started is used as the default. NLK DATA Directory NICMOSlook looks for input data les Default directory to write NICMOSlook output les NLK SPECOUT It is recommanded that you add the directory NICMOSLOOK BASE bin to the PATH environmentvariable so that NICMOSlook can be run from any directory. This can be best achieved by adding a line like setenvPATH $PATH:NICMOSLOOK BASE bin to the shell startup script e.g. .cshrc. Alternatively, The Bourne Shell $NICMOSLOOK BASE bin nicmoslook which is used to startup IDL and NICMOSlook may be edited, but this is not recommended. If you decide to modify the $NICMOSLOOK BASE bin nicmoslook script , please bear the following in mind: The fol lowing variables which are set by the $NICMOSLOOK BASE bin nicmoslook shel l script, MUST not be changed by the user:


2.4. REBUILDING THE IDL NICMOSLOOK BINARY LIBRARY
IDL STARTUP NLK CAL NLK SW NLK LIB

9

IDL startup script calibration directory NICMOSlook software directory library directory
NICMOSLOOK BASE bin

The following variable MAY be changed provided that the in the FIRST POSITION in the path:
IDL PATH

is

The IDL search path

2.4 Rebuilding the IDL NICMOSlook binary Library
Any change to the IDL code of NICMOSlook requires a 'recompilation' of NICMOSlook IDL programs in order to remake nicmoslook binary. This must be done while running IDL. To remake this library: 1. cd $NICMOSLOOK BASE idlsrc Make sure that $NICMOSLOOK BASE is set! 2. idl at the Unix prompt 3. @nlk.pro at the idl prompt This executes an IDL .run command on all the IDL program les and will save the results in $NICMOSLOOK BASE bin nicmoslook binary.

2.5 Customizing NICMOSlook
NICMOSlook is mostly controlled by its user interface. In addition, there are a number of setup les and environment variables. The setup les reside in the directory $NICMOSLOOK BASE calib and can be modi ed with a text editor or, with a utilities which are displayed when the user selects the either Edit Setup Params , Edit Grismspec , or Edit FITS Keywords buttons. calnicc.setup Controls behavior of NICMOSlook & Calnic C. Refer to section 4.1.2. grismspec.dat De nes grism characteristics. Read this le carefully before attempting modi cations. Refer to section D.1. fitskeywds.dat De nes FITS keyword values. Refer to section B.2.


10

CHAPTER 2. INSTALLING NICMOSLOOK


Chapter 3 Using NICMOSlook
3.1 Running the Program
3.1.1 Startup
Before attempting to run NICMOSlook make sure that all environment variables section 2.2 and system con g les section 2.5 are set to desired values. There are TWO modes in which NICMOSlook can be run: from the UNIX command line and from within IDL itself. To run from the UNIX command line simply type the following at the UNIX prompt: 1.
nicmoslook

To run NICMOSlook, from within IDL, type: 1. 2.
idl nicmoslook spectrum , direct

At the UNIX prompt At the IDL prompt

The optional parameters spectrum and direct are assumed to be grism and direct images, respectively, that have already been loaded into IDL data structures. The rst method demonstrates how to call NICMOSlook from the command line. The shell script, nicmoslook, sets up additional environmentvariables before calling IDL. All input and output are through the user interface of NICMOSlook. The second method shows how to bring up IDL and then start nicmoslook. When running NICMOSlook from within IDL, some care must be exercised. Advanced IDL users often have an IDL startup le referred to by the environmentvariable, $IDL STARTUP. It is possible that some values set in this le will directly con ict with IDL variables used by NICMOSlook. Read section 2.2 carefully.

11


12

CHAPTER 3. USING NICMOSLOOK

3.1.2 Input to NICMOSlook

Input to NICMOSlook is usually obtained from a dialog box on the user interface. The FITS les needed direct and grisms image are expected to be in the NICMOS FITS format. That is, a standard FITS le with 5 extensions containing the image data. If NICMOSlook encounters a FITS le with no extensions, it simply returns the single image as the science data and 4 images of equal size of zeroed data. Section B.1 describes this format in greater detail.

3.1.3 Output of NICMOSlook

Output from NICMOSlook consists primarily of postscript les containing a graphical representation of each ob ject's spectrum, and a single FITS le with a FITS table extension containing the spectra information for each ob ject detected on the images. However, additional output les can be created on the y for convenience by selecting various buttons on the interface. For a comprehensive list of these les, please refer to section F. Any time output is directed to a le, a dialog box will prompt the user for the desired location.
filnm

tab. ts : Contains one binary table extension for the spectrum information for each ob ject detected. See section C for more details.
n.ps Postscript le for each spectrum on the image, where 'n' is the ob ject number starting with 0 as found on the image. where filnm is the root name the user entered in the dialog box
filnm

The IDL widget plots spectra in a window directly on the screen, with the option to save the plot to a Postscript le. These postscript les can be viewed with any Postscript viewer, suchas ghostview. For example, if you direct the output to a le called ov2 0.ps, you would view this spectrum e.g. with ghostview, with:
ghostview ov2 0.ps

3.1.4 Viewing the Postscript Spectra Output

3.2 The User Interface
3.2.1 Startup Window
The user interface of NICMOSlook is an IDL widget. Figure 3.1 shows the user interface as it appears at the start of the program. This chapter gives a brief overview of the various buttons, switches, and pulldown menus available in NICMOSlook. Each of these items that


3.2. THE USER INTERFACE

13

Figure 3.1: The NICMOSlook user interface after at the start of the program. can accept input are encased in a box . For a detailed example of the process of actually using NICMOSlook to reduce your data, refer to the NICMOSlook tutorial in chapter 5. Starting up NICMOSlook with images already in IDL arrays: Note that image1 MUST be the spectrum, image2 the direct image optional. IDL: nicmoslook, image, header1=header1 Toload an image & header already in IDL memory: IDL: nicmoslook, image, log To load an image and display it on a logarithmic scale: IDL: nicmoslook, image1, header1=header1, image2, header2=header2 Toload 2 images & headers already in IDL memory: If no header is given, then a minimal FITS header will be created. In this case, the user will be prompted for the grism name with a dialog box. To load an image from IRAF, FITS, or SDAS image les, call the program with no arguments. Once the widget appears select the FITS , SDAS or IRAF depending on the type of the input le and press the LOAD FILE button. A window will appear that prompts for the lename. FITS les


14

CHAPTER 3. USING NICMOSLOOK

are assumed to have the extension '. ts', IRAF images '.imh', and SDAS les '.hhh'. If the extension '.Z' or '.gz' is detected in the lename, the les are assumed to be compressed, and will be automatically uncompressed using the appropriate utility. There are 2 images bu ers, one for the direct image, and one for the grism images. Every image is loaded into one of these bu ers. Therefore, at most one direct image and one grism image can be loaded simultaneously.

3.2.2 Displaying the Images

When the program starts, a 512512 separate display window with coordinate display will appear. The most recently loaded image will be displayed in this image. The input image will be resized to t this window. The display window can be resized to any desired shape using standard window resizing techniques of the used window manager. There are anumber of options to manipulate the image display. Changing Contrast and Colors: The image contrast can be changed through the switches labeled Display Scale: Linear , Logarithm , Square Root , and Histogram Eq. . Changing Color Look-up table: The color look-up table and contrast can be changed by pressing the Contrast LUT button. This will pop-up the standard IDL xloadct widget which presentsachoice of look-up tables and display options. Switching Images with Toggle Buttons: Selecting the Spectrum or Direct Image toggle buttons displays the image loaded in that bu er. Switching Images with Blink: Selecting the Blink button pops up a window that allows the user to select the rate in which the two images currently loaded are switched, both in the zoom window and the separate image display. Using the Zoom Window: Clicking any mouse button once while the cursor is within the image display window causes the portion of the image centered on the cursor to be displayed in the zoom window. The zoom magni cation is set through the slider bar below the zoom window. Resetting Images: Clicking the Reset button removes the images from the display and e ectively resets the application to the initial startup state.

3.2.3 Finding Ob jects and Extracting Spectra
Ob jects input le.

After loading both a grism image and a direct image into the tool, there are 2 main menu buttons to produce spectra: Creates an ob ject list from direct image or reads coordinate list from


3.2. THE USER INTERFACE
Spectra These items are thoroughly discussed below.

15

Extracts spectra for selected ob jects with di erentweighting options.

Ob jects The rst step in the spectrum extraction procedure is to de ne the coordinates of the images. This is accomplished with this button. This will bring up submenus with the following options. Find Ob jects Cursor interactive input of ob ject coordinates. Left mouse button clicks in the display window will append those coordinates to the internal ob ject list. The last ob ject location should be speci ed with a right mouse button. This terminates the 'Cursor Mode'. DAOFIND nd ob jects on the DIRECT image independent of what is displayed. and append them to the internal list. Input File read ob ject coordinates, sizes, angles, etc. from input le. See appendix F for a description of the format of this le. Size & Orientation The extraction of spectra can be done by using weights appropriate for point sources, or by determining the weights from the size of the ob jects. The latter is appropriate for extended ob ject and in this case, the size of the ob jects must be determined. This can be done interactively or automatically. Cursor This item allows the user enter four points to mark the ma jor and minor axis of an ellipse surrounding an ob ject on the direct image. After selecting this menu item, click the left mouse button in the graphics window around the ob ject in question. While this is being done, an informative window showing the actual coordinates clicked will appear. From this point on, the ob ject is no longer considered a point source, rather an extended ob ject. Automatic automatically calculate sizes and orientation of ob jects selected. Reset Reset sizes to zero and default type to point source. Mark Ob jects Get a visual representation of an ob ject's size, spectra location, or number. Label ob jects numerically label ob jects on image display. Spectra Location show the position of spectra to be extracted on image display. Mark Sizes show the size and orientation of ob jects. Ob ject List Manipulate list of ob jects.


16

CHAPTER 3. USING NICMOSLOOK

Save to File Save coordinates and sizes of ob jects to an ascii le. This le
can be edited with any ascii editor and read backinto NICMOSlook with Ob jects ! Find Ob jects ! Input le Display List Display coordinates, sizes and orientation of ob jects in the ob ject list. Clear List reset internal list of ob ject coordinates. Spectra This is the option which actually extracts the spectra for the ob jects in the current ob ject list. Items that do not write to le will display a popup window displaying the spectrum. See section 3.4 for further information on this window. Point Source Weighted Produce TinyTim weighted spectra for point sources. Point Source Unweighted Produce unweighted spectra for point sources. Extended Ob ject Weighted Produce weighted spectra for extended ob jects. Extended Ob ject Unweighted Produce unweighted spectra for point sources. The above four menu items al l have the fol lowing 4 menu items! Speci c Ob ject Not Deblended Extract spectra of ob ject selected by user without removing contamination from nearby ob jects. Speci c Ob ject Deblended Extract spectra of ob ject selected by user while attempting to remove contamination from nearby ob jects. All To File - Not Deblended Extract spectra of all ob jects found on image without removing contamination from nearby ob jects. Write the results to a single FITS le and one postscript le for each ob ject. All To File - Deblended Extract spectra of all ob jects found on image while attempting to remove contamination from nearby ob jects. Write the results to a single FITS le and one postscript le for each ob ject. When any of the above All to File ... are selected, a le selection window will appear. Enter the base name of the les that will be written. A single FITS le containing a table for each ob ject, one Postscript le for EACH ob ject, and a catalog le will be written to the directory and le rootname speci ed. Strips produces several spectra by plotting the rows around a selected ob ject and perform a wavelength calibration under the assumption that there are no distortions and rotations of the spectrum. In other words, the tool plots the appropriate sections of the rows without modi cations. Trace At this point, most of the parameters in the NICMOSlook calibration data base grismspec.dat have to be prede ned see D.1. However, the location of the spectra relative to the location of the ob ject on the direct image can be determined automatically with this option. The ob ject has to be de ned and one of it has to be selected. THe program will then search for the position of


3.2. THE USER INTERFACE

17

the spectrum at a location close to the one speci ed in grismspec.dat for the current lter and grism. The user is presented with a plot of the y-positions of the spectrum as a function of x position. Accurate parameters are determined through a t to this plot. The user can choose wether to update grismspec.dat with these parameters. Extraction Images Punched-out Grism Display the "punched-out" grism image: It is often quite informative to see the pixels used for the extraction of a spectrum. This button displays in the graphics window an image whose pixels that are used are "punched out" ie set to the mean background pixel value of the surrounding pixels. Image Background Display pixels used to calculate a background value: The graphics screen will then display an image showing which pixels were used in the calculation of a mean background value for an ob ject. Weights Display weight used in the last extraction: The graphics screen will then display an image showing the weight matrix used in the last spectra extracted

3.2.4 Supplementary Pulldown Menus

IMAGE PROCESSING While NICMOSlook's main use is to extract spectra from already reduced images, a few simple image processing tools are provided so that NICMOSlook can be used as a quick-look of raw data. All tools act on the currently displayed image. All of the following procedures replace the original image in the program's bu er and update the display window. Convert To Count Rate Count-rate conversion is simply dividing the image by the exposure time. The image header is searched for the keyword EXPTIME or DTIME. If not found, an error message is issued. A comment is added to the header. Fix Bad Columns The replacementvalues for bad columns are linearly interpolated from the two nearest non-bad columns. A dialog box prompts the user for the coordinates of the bad columns. The image header is automatically updated. Bias Subtraction Removes the 2-dimensional additive background from the image. Bias image can be loaded from the command line via the appropriate keyword or from an IRAF, FITS, or SDAS le. The bias image must be the same size as the image to be processed. The image header is automatically updated.


18

CHAPTER 3. USING NICMOSLOOK
Dark Subtraction Subtracts dark count image in the same fashion as described above for bias subtraction. Flat Field Compensates for the pixel-to-pixel variation in sensitivity. The image is divided by the at. The at eld image must be normalized. If the image to be processed and the at eld image are not the same size, the task will try to register the at by searching the image header for the starting X and Y positions header keywords 'X0' and 'Y0'. The image header is updated. IMAGE FILTERING A number of image ltering tools are provided to help the detection of ob jects on the direct image. While the main application of these tools is to enhance the direct image, they can also be applied to the grism image. All tools act on the currently displayed image. They overwrite the image in the program's memory. The display window and the image header are updated. Median Useful for removing cosmic ray hits. Smooth Boxcar average. Sobel & Roberts Edge enhancement. Convolution Convolves image with a user-supplied kernel, assumed to be a FITS image. The le selection dialog box appears. PLOTS A number of tools are available to graphically investigate the images. For each of these options a pop-up widget will appear containing the requested plot. This pop-up widget contains options for rescaling the y-axis of the plot, measuring the FWHM of emission features, and writing plots to postscript. The y-axis scaling options include: 1 linear, 2 logarithmic, 3 logarithmic down to ten, and linear from ten to zero. Cross Section Produces a plot of a 'slice' of an image. User must position the cursor in the display window at each end of the desired cross section. Click LEFT on the starting location and click RIGHT on the ending location. The width of the slice is one pixel. Row Sum User must click LEFT on the starting and click RIGHT on ending columns. The plot is produced by direct addition. Column Sum User must click LEFT on the starting and click RIGHT on ending columns. The plot is produced by direct addition. Histogram Produces a histogram of the entire image.


3.2. THE USER INTERFACE

19

ROTATE The spectra on the grism image are assumed to have the orientation as speci ed in the setup le grismspec.dat see appendix D.1. If for some reason the orientation of the image is di erent, it can be rotated into the correct orientation with this option. Another application of this option is to bring the grism image and the direct image into the same orientation if this is not the case for the input images. Note that it is the responsibility of the user to assure that the orientation of the grism image and the direct image are the same. The rotate routines act on the currently displayed image and overwrite the image in the program's memory. The display window and the image header are updated. IMAGE INFO Statistics Computes a basic set of image statistics; mean, median, standard deviation, etc. Encircled Energy Calculates the total number of counts in a user-selected region of the image. The region is de ned by clicking left on the diagonal corners of an imaginary box. Coordinates of the box selected are also displayed. Header Keywords Certain keywords in the FITS header are necessary for calculating various values. This item will display the values of these keywords.

3.2.5 Miscellaneous Buttons

Adding a Comment to a FITS Header: Press the Add CommentTo Header button. A dialog box will appear allowing the user to input a comment to be included in the FITS header of the image currently being displayed. Displaying the FITS Header of Image: Press the Display Header button. A window containing the text of the FITS header of the currently displayed ob ject will appear. Writing Images to Postscript les: Press the Image to PS button. The currently displayed image will be written to a postscript le in the user's current working directory. Saving a Processed Image: Press the SAVE IMAGE button. A dialog box will appear allowing the user to select the name of the output le. The default selection is the input lename, if any. The output format is FITS. Note: This option only saves the currently displayed image, if you havetwo, then this options must selected twice. Editing Setup Parameters: Press the Edit Setup Params button. A screen will appear with all the setup parameters available for editing. Refer to section 4.1.2 for more information.


20

CHAPTER 3. USING NICMOSLOOK
Editing Grismspec File: Press the Edit Grismspec button. A screen will appear with all the grismspec values for the selected lter. Refer section D.1 for more information. Editing FITS Keywords: Press the Edit FITS Keywords button. A screen will appear with all the con gurable FITS keyword values. Refer section B.2 for more information. Getting Help: Press the Help button. A window will appear with helpful information for running NICMOSlook. Quitting NICMOSlook: By pressing the Quit button, the NICMOSlook program is terminated.

3.2.6 Plot Popup Window

Figure 3.2 shows an example of a plot when selected with the one of the Plot menu items.

Figure 3.2: The Popup Window


3.2. THE USER INTERFACE
The following buttons are available on the plot window:

21

Measure FWHM With this button, areas on the plot window can be selected with the mouse. Selection is terminated by using the RIGHT mouse button on the last coordinate of desired region. A gaussian is tted to each selected region and the plot is replaced with panels containing the results. To return to the original plot, push Re-plot . See gure 3.3 for an example. Measure FWHM - output to PS The same procedure as above, but the output will go to a Postscript le instead of the screen. The plot type can be selected with one of these switches: Linear Log Log!linear Replot Redisplay the plot popup window. Plot to PS File Write the plot currently displayed to a Postscript le. Data to File Data currently display on the screen will be written to a data le when this button is selected. Done This button closes the plot popup window and returns control to the main NICMOSlook window.

Plot Popup Window: FWHM

See the above section for a description of the buttons available on this plot popup. When any of the menu items selected from Spectra that do not direct output to a le are selected, a spectrum plot popup like the one in gure 3.4 will appear. The following buttons are available on the spectrum plot window: Replot Redisplay the plot popup window. Plot to PS File Write the plot currently displayed to a Postscript le. Save to FITS Data currently display on the screen will be written to a data le when this button is selected.

3.2.7 Spectrum Popup Window


22

CHAPTER 3. USING NICMOSLOOK

Figure 3.3: The Plot Popup Window: FWHM Show Lambda Flux Select this button and enter the plot window and click on a point. This will bring up a popup that shows the Wavelength and Flux value of the point selected. Response This option divides the current spectrum by a function de ned through a list of wavelength - throughput pairs. The lename of this asci le is de ned in grismspec.dat. Line Params Selecting this button pops up a window containing the derived line parameters from the spectra extraction process. SaveParams Selecting this button pops up a le dialog box prompting for a le name in whichto save the derived line parameters Mark Lines This button allows users to override line boundaries automatically determined in the spectra extraction process. After selecting the button, enter the plot and select the line boundaries with the left mouse button. The right mouse button


3.2. THE USER INTERFACE

23

Figure 3.4: The Spectrum Popup Window is used when selecting the last point on the spectra to use. After the nal point is selected, spectral line search will automatically be run. Set Fit Order A polynomial is automatically t to the continuum emission of each extracted spectrum. Selecting this button pops up a window so that the t order can be entered. This t order will override the t order in the setup parameter le. The spectral line search will be rerun automatically with the new value when the popup's Done button is selected. Response Apply a response. Done This button closes the spectra popup window and returns control to the main NICMOSlook window.


24

CHAPTER 3. USING NICMOSLOOK


Chapter 4 Software Dependencies, System Requirements & Debugging
4.1 NICMOSlook
NICMOSlook is a Bourne Shell script which calls IDL. All the software is written in IDL and therefore Calnic C requires a valid IDL license. NICMOSlook has been tested so far under the Solaris operating environment.

4.1.1 Directory Structure

The directory structure contains all the necessary calibration les, ts images, etc. to run the currentversion of NICMOSlook. The directory structure of the NICMOSlook Software is in gure 4.1
$NICMOSLOOK_BASE/

bin/

calib/

data/

doc/

idlsrc/

Figure 4.1: Directory Structure of NICMOSlook Software Directory Contents: 1.
$NICMOSLOOK BASE

bin 25


26CHAPTER 4. SOFTWARE DEPENDENCIES, SYSTEM REQUIREMENTS & DEBUGGING nicmoslook : The program Bourne Shell script IDL startup le nicmoslook startup.pro : nicmoslook binary : "Compiled" idl code $NICMOSLOOK BASE calib bckgrismX. ts : Image background for the di erent grisms calnicc.setup : NICMOSLOOK Calnic C default values tskeywds.dat : FITS keywords grismspec.dat : The grism speci cation le nicmosFF. ts : Flat Field le this is 3 dimensional *.response : Response les for each grism list of lter curve *.conv: Convolve les for each grism PSF for each lter weight*. ts : Various weight les for di erent grisms $NICMOSLOOK BASE data : One test image pair $NICMOSLOOK BASE doc : Documentation directory nicmoslook.tex : TeX version of NICMOSlook manual nicmoslook.ps : Postscript version of NICMOSlook manual $NICMOSLOOK BASE idlsrc : IDL source code for NICMOSlook

2.

3. 4. 5.

4.1.2 NICMOSlook Setup Parameters

In the calibration subdirectory mentioned in the section above, the le calnicc.setup contains variety of parameters that can be set to modify the behavior of NICMOSlook. It is possible to modify these parameters while running NICMOSlook. Use the button Edit Setup Params to change the values. This will pop-up the parameter editor. All parameters are described in plain English and can be changed see Figure 4.2. The changes take e ect immediately after hitting the Save button on the editor pop-up. In addition, the values are written back to the le $NICMOSLOOK BASE calnicc.setup. See table D.2 for the default values of the parameters described below.
ADCGAIN

Analog to digital gain in electrons ADU. This value will be used as a default IF no keyword ADCGAIN is found in the image header.
BADPIX THRESH

Pixel coordinates that contain ob ject information are compared with a bad pixel map. A ratio of total pixels to the number of the bad pixels is calculated and compared to this parameter. If the ratio is higher than this value, the ob ject is removed from the ob ject list.


4.1. NICMOSLOOK
BG PIXELS

27

To calculate a background estimate for an ob ject during spectral extraction, the mean value of the pixels surrounding the edge of the ob ject is calculated. This parameter determines the width of the region used in this calculation.
BLEND FACTOR

In the spectral line search process, 2 lines are considered to be the same line if the separation of the peaks is smaller than the sum of the line widths multiplied with this factor. In that case, the region of the two lines are joined and a new Gaussian is tted over the entire wavelength range of the line.
CONT FACTOR

This factor determines the region for which the average of the continuum "longwards" of the line with the longest wavelength and "shortwards" of the line with the shortest wavelength is computed. The "long" region is the peak of the line with the longest wavelength added to the of that line multiplied with this factor. The "short" region is the peak of the line with the shortest wavelength minus the of that line multiplied with this factor.
DAO THRESH

Threshold used to nd ob jects on the grism image outside of the regions where spectra are predicted from the ob jects found on the direct image. The unit of this threshold is the rms of the noise in the grism image after removing known spectra. When calculating the deblending error, this value is substituted for deblending errors that are of smaller value.
DET THRESH DEBLEND ERR MIN

Threshold used in the spectral line search to identify lines in units of the rms of the spectrum.
EXTOBJ THRESH

The ob ject detection program SExtractor classi es all ob jects detected as either point source or extended ob ject. This classi cation is in the form of a probability value: the closer this value is to 0.0 the more likely the ob ject is an extended ob ject; conversely, the closer this value is to 1.0, the more likely the ob ject is a point source. This parameter is the "cut-o " point in which all probability values greater are regarded as point sources and all values below, extended ob jects. This is relevant for the weighting of the spectrum extraction.
FIT ORDER1

A polynomial is tted to the continuum in each spectrum. After the rst t, deviating points are removed from the t region and the process is iterated. This parameter is the order of the polynomial for the rst N ITERLOW iterations see below.


28CHAPTER 4. SOFTWARE DEPENDENCIES, SYSTEM REQUIREMENTS & DEBUGGING Polynomials are iteratively tted to the continuum in each spectrum. After each t, deviating points are removed from the t region for the next iteration. This parameter is the order of the polynomial for all but the rst N ITERLOW iterations see below.
MAXLINES FIT ORDER2

This parameter determines the maximum number of lines which can be detected in any spectrum.
MAXOBJS

For computing resource purposes, this number determines how many ob jects on an image set are to be examined.
MAXPIX

For computing resource considerations, the maximum size of input images processed by Calnic C in pixels.
MIN NPOINTS

This parameter is the minimum number of consecutive points which deviate from the continuum t in the spectrum which are considered as a spectral line.
MIN PIX NO

Minimum line width for a line to be considered as real.
MIN SIGNSE

Detected lines whose signal-to-noise ratio is less than this value are rejected.
N ITERATIONS

The t to the continuum in the spectrum is discontinued after this number of iterations.
N ITERLOW

The t of a polynomial to the continuum in the spectrum is done in two steps. For the rst iterations, a low order polynomial is tted in order to identify strong lines. After several iterations in which deviant points are rejected and a new polynomial is t, the order of the t is increased. This parameter determines the number of iterations before the next higher t order is used.
O THRESH1

Close spectra on the grism image potentially contaminate each other. This parameter is the maximum distance of 2 spectra on the grism image for them to be considered being exactly aligned. The parameter is in units of the sum of the line widths of both spectra.


4.1. NICMOSLOOK
O THRESH2

29

Close spectra on the grism image potentially contaminate each other. This parameter is the maximum distance of 2 spectra on the grism image for which a contamination is expected and Calnic C attempts to "deblend".
O THRESH3

Close spectra on the grism image potentially contaminate each other. This parameter is the maximum distance of 2 spectra on the grism image for which a contamination is considered to be still possible.
REJ THRESH SZ FACTOR

Currently not used. Width of a spectrum in units of the 2nd order moment of the ob ject on the direct image.
DEBUG LEVEL

This parameter determines the verbosity of Calnic C's output messages. The higher its value, the more information is printed to the debugging output le. This parameter allows for detailed debugging messages. Refer to section 4.2 for more information.
NOFLATFIELD

This parameter allows users to turn o or on at elding before spectral line extraction commences.
NOBACKGND

This parameter allows users to turn o or on background subtraction before spectral line extraction commences.

4.1.3 NICMOSlook Grismspec Entries

Another le in the calibration subdirectory, grismspec.dat, contains the lter speci cations for the direct image and grism lters. It is possible to modify these parameters while running NICMOSlook. Use the button Edit Grismspec to change the values. This will pop-up the parameter editor. All parameters are described in plain English and can be changed see Figure 4.3. The changes take e ect immediately after hitting the Save button on the editor pop-up. In addition, the values are written back to the le $NICMOSLOOK BASE grismspec.dat. See section D.1 for more details.

4.1.4 NICMOSlook Con gurable FITS Keyword Values

The le, fitskeywds.dat, contains entries for FITS keywords used by NICMOSlook. Since it is likely that di erent data sets will use di erentkeyword values for specifying the same


30CHAPTER 4. SOFTWARE DEPENDENCIES, SYSTEM REQUIREMENTS & DEBUGGING item of interest to NICMOSlook, exibility is needed. For example, some FITS les might use the keyword OPT ELEM to specify the optical component used for an observation. Other FITS les on the other hand, might use the keyword FILTER to specify the same optical component! It is possible to edit the value of the keywords used by NICMOSlooksee Figure 4.4. Use the button Edit FITS Keywords to change these values. The changes take e ect immediately after hitting the Save button on the editor pop-up. See section D.3 for the default values.

4.2 Debugging
In the course of processing, NICMOSlook writes diagnostic information to the shell window from which it was started. In the unlikely event NICMOSlook does not behave in the manner expected, the parameter DEBUG LEVEL can be increased from its default value. When this parameter is set higher, more detailed messages are displayed. This value can be set by selecting the Edit Setup Params button.


4.2. DEBUGGING

31

Figure 4.2: The NICMOSlookParameter Editor.


32CHAPTER 4. SOFTWARE DEPENDENCIES, SYSTEM REQUIREMENTS & DEBUGGING

Figure 4.3: The NICMOSlook Grismspec Editor.


4.2. DEBUGGING

33

Figure 4.4: The NICMOSlook FITS Keyword Editor.


34CHAPTER 4. SOFTWARE DEPENDENCIES, SYSTEM REQUIREMENTS & DEBUGGING


Chapter 5 Tutorial
This chapter gives you a basic guide on how to use NICMOSlook. The detailed descriptions of algorithms used can be found in the appropriate chapter of the Calnic C user's manual. Before you begin, you should make sure that IDL is installed with the correct licenses on your machine. Then ollow the instructions in section 2.2 to install the latest version of NICMOSlook. After the installation process is complete, you are ready to run NICMOSlook by simply typing nicmoslook at the UNIX command line. Refer to section 3.1 for a alternative methods of starting NICMOSlook. In this tutorial, all commands you type in from your keyboard are type-writer font, and commands you clickon your screen are encased byboxes. This command will automatically start IDL and two windows will appear: one is the main NICMOSlook window which contains all of the data processing tasks; the other window is an IDL image display window. See Figure 5.1. All the functional buttons in the main NICMOSlook window can be activated by pressing the left button on your mouse unless stated otherwise. Before you start using various functions, read the help le by pressing the button Help on the main NICMOSlook window. This le contains detailed description of every task in NICMOSlook. To close the windows, press Done

5.1 Image Display
To start the data processing, press Load in the main NICMOSlook window. A separate small window will appear with directory names and the les in each of the directories. See Figure 5.2. Select the correct directory and the image le which you want to load into NICMOSlook. For example, you can click on subdirectories data , a le obj10 dir.fits, then click Ok . This will load the FITS format image ob j10 dir. ts into the IDL image display window. In order to produce a quick spectrum extraction of a grism spectrum, you need to load both a grism spectrum and the corresponding direct image into NICMOSlook. This is because the direct image provides the information about the locations of your ob jects 35


36

CHAPTER 5. TUTORIAL

on the NICMOS detector, and NICMOSlook needs to know the precise positions of your ob jects before starting tracing and extracting the corresponding grism spectra.

Figure 5.1: The NICMOSlook main window when you rst start up nicmoslook. You can load your direct image and grism spectrum in any order you want. NICMOSlook recognizes your image le from its FITS header. If it is a direct image, NICMOSlook puts it in the direct image bu er; if it is a grism spectrum, NICMOSlook puts it in the spectrum bu er. It should be pointed out that each of these bu ers can hold only one image le at a time. If you do any processing to the image, it overwrites the original image in that bu er. If you need the original image, you have to reload the image from the disk again. If you wanttosave what you have done, you have to press Save Image in the main NICMOSlook window. After you have loaded both the direct image and the grism spectrum, you can click the exclusive menu button labelled Display on the main NICMOSlook window to choose which image you wish to display in the main IDL image window. Select Spectrum to display the grism image, or Direct image to display the direct image. If you wish to examine your image in detail, put the cursor at the position you want in the big IDL image display window, then click the left mouse button. A section of your image centered around the cursor position will appear in the small image display window in the main NICMOSlook window. You can then control the image expansion from this small window. You can also examine your le header by pressing Display Header in the main NICMOSlook window. Again a separate window appears to show you the ts le header. To add any comments to the le header, you can press Add Comment to Header button, then a separate window will appear which says: Enter comment to be added to header. After you are done, press Done . You can check your newly added comment by using Display Header .


5.2. IMAGE PROCESSING

37

Figure 5.2: The window con guration when you start loading images to NICMOSlook. Using Edit Setup Params ,you can change some of the parameters used for the spectrum extraction. For example, if you click in the eld ADCGAIN , you can put in the appropriate gain factor, then press Save Changes & Quit . This not only changes the values for the current session, but also saves the values in the calnicc.setup le. When you start a new session of NICMOSlook, these newly saved setup parameter values will be the default. Of course you could change the setup parameters in the le calnicc.setup in the $NICMOSLOOK BASE calib directory. By pressing Contrast LUT you can adjust the color table. Also in Display Scale , you can choose one of the options for your image display.

5.2 Image Processing
If you press the button Image Processing ,you will see a menu appearing with listed functions NICMOSlook has. For example, if you want to do bias subtraction, you can press Image Processing ! Bias Subtraction , then a separate window will appear which ask you to select the bias image for the processing. Notice at the top of the main NICMOSlook window, there is a small message eld which always contains information about what is happening. It is useful for users to keep an eye on that window. Other functions within "Image Processing" are dark counts subtraction, at elding, count rate conversion, overscan subtraction and xing bad pixels. By pressing Blink you can blink the direct image against the grism spectrum image. Also other functions such as Image Filtering


38

CHAPTER 5. TUTORIAL

and Rotate can be used to smooth and rotate your images. Plots can be used to makea plot of either several raws or columns or even a section of your image. This function is very useful when you want to see if the ob jects in your image are point sources or extended, see details in next section. To check the statistics of the displayed image, you can press the button in the main window Image Info ! Statistics .

5.3 Finding Ob jects and Determining Sizes
Before you start the spectrum extraction, you need to assign the sizes to the ob jects you are interested in. The reason is that the spectrum extraction algorithm was designed to give di erentweights depending on the sizes of the ob jects, and more importantly, the way of calculating the weights is di erent for point sources and for extended galaxies. Thus, you need to decide, rst, if your ob jects are extended or point sources; second, the sizes of your ob jects. The actual procedure is as follows: You should rst de ne your ob ject positions on the direct image. If you press Ob jects , a small menu pops up, which lists several functions. Among these functions, Find Ob jects gives several options. For example, Cursor" means you could de ne your ob ject positions directly by clicking on your grism spectrum image IF you know beforehand the exact positions of your ob jects on NICMOS detectors. Similarly, function Input le" can also provide a list of positions for extracting the grism spectra of your ob jects without any direct image of your ob jects. See Figure 5.3 which shows that you could use input le ob ject5.dat pre-de ned the ob ject positions.

Figure 5.3: This shows you can select the ob ject positions from an input le. If you click on Ob jects ! Find Ob jects ! DAOFIND , you will get a separate window which requires three user adjustable parameters | threshold used for the ob ject


5.3. FINDING OBJECTS AND DETERMINING SIZES

39

detection, minimum and maximum sharpness. See Figure 5.4. Very large minimum and maximum sharpness values will allowyou to only pick up point like sources, for instance, minimum sharpness of 0.8 and maximum sharpness of 100 will allow you to ignore most of the extended galaxies in the test image ob j10 dir. ts. Similarly,if you set small values of minimum and maximum sharpness, for example, minimum sharpness of 0.2 and maximum sharpness of 1.0 will allowyou to only pick up extended galaxies and no point sources in the test image ob j10 dir. ts. Of course the actual values of sharpness maybe di erent from the abovenumbers for your real data images. The appropriate values should be tested with your actual data. To ignore any previous ob ject ndings, you can press Ob jects ! Ob ject List ! Clear List to empty the ob ject le. This will actually empty the le from the computer memory. Ob jects ! Mark Ob jects ! Label will allow to label the ob jects you have found on the image display window.

Figure 5.4: What you should get when you use DAOFIND to search for ob jects. After you have selected your ob jects, either point or extended sources, you have to assign the size values to them. Remember that by default, ob jects found with DAOFIND are considered as point sources. There are several ways to do this. Ob jects ! Size & Orientation gives three options. One is to use CURSOR. When you clickon Cursor and clickon your ob ject, you will have another window which asks you to click on the ma jor and minor axes of your extended ob ject. After you click at the appropriate places, the x, y positions of the ma jor and minor axes will be shown in this window too. This option is sometimes not easy to use when your galaxy is small on the NICMOS detector. As described before, Plots can be useful in this situation since it can giveyou some ideas of Full Width Half Maximum FWHM of the cross-section of your ob ject. By pressing Plots ! Rows ,you have a message in the main NICMOSlook windowsaying Position cursor on rst and last raw to sum & click left". After you do that, you will have a separate plot window showing the cross-section. In this window, you can clickon Measure FWHM to roughly measure the ob ject size. See Figure 5.5.


40

CHAPTER 5. TUTORIAL

Figure 5.5: This shows that you can measure the approximate sizes of your ob jects with option Plots . The second option is Ob jects ! Size & Orientation ! Automatic . A separate IDL window appears to ask you which ob ject to determine the size, you can enter the ID number found by Ob jects ! Find Ob jects ! DAOFIND , or clickon All . Then pressing Done will giveyou the automatically calculated sizes. For the detailed algorithm for calculating the sizes of extended sources, you can consult the Calnic C user's manual. After the ob ject sizes are determined, you can visually exam the result by pressing Ob jects ! Mark Ob jects ! Mark Sizes , which draws the ellipses indicating both sizes and orientations of the galaxies in your image. See Figure 5.6. Finally, for some wrongly identi ed extended sources, you can manually reset their status to point sources by pressing Reset , and click Direct in the exclusive menu button labelled Display and then on either Label or Mark Sizes under the menus Ob jects ! Mark Ob jects a fresh view of the marked image will be displayed.

5.4 Extracting Spectra
Clicking Spectrum in the exclusive menu button labelled Display on the main NICMOSlook window will load your spectrum from the spectrum bu er to the IDL image display window. Before extracting any spectrum, you can use Ob jects ! Mark Ob jects ! Spectra Location to mark the spectral centroids on the grism image. See Figure 5.7. You can see the ob jects being marked on the IDL image display window, extended sources with rectangular boxes and point sources with lines through the centroids.


5.5. EXAMINING YOUR SPECTRA

41

Figure 5.6: This shows what it should look like when you mark your ob jects with sizes and ob ject IDs on the direct image. Spectra in the main NICMOSlook window contains several ways of extracting spectra. Each extraction can be weighted or unweighted according to the locations of pixels, and also you can choose to deblend the spectra if your ob jects are too close to each other. Be careful that if you choose point source options to extract the spectra, then the ob ject will be treated as a point source even if it was classi ed as an extended source before. These options over-write the previous classi cation. However, if you choose options for extended sources, ob jects classi ed as point sources will still be treated properly with point source weighting algorithm. There are several options for extracting a spectrum. For example, you can press Spectra ! Extended Ob ject Weighted ! Speci c Ob ject Deblended , a separate idl window will pop up which ask you specify the ob ject ID number. By pressing done, you will have another plot window showing you the extracted spectrum. From this plot window, you can also plot the spectrum in linear scale, or in logarithmic scale or in log-linear scale see the help manual for the detailed explanation. You can also save the spectrum in either PostScript or ASCII format. Then you press Done . To extract the spectra and dump resulting pairs to les, press Spectra ! Extended Ob ject Weighted ! All to File ,, Deblended , a window will pop up and prompt for the output le name where the extracted spectra are stored. By pressing Spectra ! Strips , in a separate plot window nine spectra are shown, each of which represents one pixel row of the 2-D grism spectrum.

5.5 Examining Your Spectra
After you have extracted your ob ject spectrum, you can use several functions to obtain the detailed astrophysical information. For example, if you use option


42

CHAPTER 5. TUTORIAL

Figure 5.7: This shows what it should look like when you mark your ob jects with sizes and ob ject IDs on the grism image. Spectra ! Extended Ob ject Weighted ! Speci c Ob ject Deblended and specify ob ject 3, for instance, you would get another plot window showing you the spectrum, as shown by Figure 5.8. Within this spectrum plot window, you have several options to measure the spectral features. Re-plot will allow you to clear and re-plot the spectrum. Plot to PS File will save the displayed spectrum to a postscript le in your chosen directory. The same is for option Data to FITS File , except the data format is in FITS. Pressing Show Lambda Flux ,you will see a message in the small message window within the spectrum plot window saying Click mouse button at plot location". For example, if you click on the line feature shown in Figure 5.8, another small window will pop up showing the wavelength and ux of the location you just clicked. Notice that wavelength is in units of micron and ux in milli-Jansky. This function allows you to do interactive measurements. Show Line Params gives you the parameters determined for the significant spectral lines, i.e, Line number ID, peak ux in the line, the central wavelength, FWHM, signal-to-noise ratio S N, 2 of the Gaussian t to the line and its integrated line ux. Also you can save the above parameters in a ascii le with Save Line Params . Mark Lines enables you to select the lines you think are missed by the automatic line detection. You can click at the left and the right side of the continuum for each line feature you want to select. The last location should be identi ed with the RIGHT mouse button. This is NICMOSlook's signal to immediately display the new spectrum using the newly delimited lines. With Set Fit Order ,you can choose the order for the polynomial tting of the continuum. After you are happy with all the spectral line measurements, clickon Done to go back to the next line.


5.5. EXAMINING YOUR SPECTRA

43

Figure 5.8: This is what you would get when the extracted spectrum for a speci c ob ject is displayed. You can obtain detailed spectral parameters from the line plot window.


44

CHAPTER 5. TUTORIAL


Chapter 6 Acknowledgments
Anumber of people contributed to the development of NICMOSlook. The user interface is based on the IDL widget stislook which was written by Terry Beck. He kindly made it available to us before its release. Rudolf Albrecht supported the development starting from the rst concepts through its nal implementation. Lin Yan computed the background image. Richard Hook computed the PSFs used for the weighting of the extraction. Hans-Martin Adorf helped with release 1.2 of NICMOSlook. Jeremy Walsh graciously took time to proofread this manual. The algorithms employed in NICMOSlook were repeatedly discussed with Rodger Thompson, Betty Stobie and Dave Axon. Their input and suggestions were invaluable.

45


46

CHAPTER 6. ACKNOWLEDGMENTS


Appendix A NICMOSlook Utilities
A.1 Test Image Generator
In the process of developing NICMOSlookit became clear that it would be necessary to have some test data for debugging the system. An IDL widget program was developed solely for the purpose of de ning test cases that would adequately exercise all the functionality of NICMOSlook. This program will soon be available for public use. Figure A.1 shows the test image generator program.

A.1.1 Installation

Todownload and run this program: Create a directory called testim;change to this directory. Download testim 0.0.tar.gz from the NICMOS Web page to the directory created above. Type: zcat testim 0.0.tar.gz j tar xvf Unpack the source

A.1.2 Running the Test Image Generator
Type: Type:
source testim.setup idl

Set the environment

To Create an Ob ject: Set the parameter values refer to section A.1.3 Select ob ject type Click in either of the two image windows at the center point of the ob ject 47


48

APPENDIX A. NICMOSLOOK UTILITIES

Figure A.1: NICMOSlooktest image generator.


A.1. TEST IMAGE GENERATOR
Press Create Ob ject

49

Press Save to create a NICMOS FITS le data pair. A le name dialog box will prompt the user for an the output lename. The grism image will be written to a le that has the selected lename plus ' gri' appended before the '. ts' extension. Similarly, the direct image will be written to the lename with ` dir' appended. Both les are written to the fits subdirectory that is located in the install directory created in the rst step.

A.1.3 Parameter Settings

In order to facilitate the testing of the many features of NICMOSlookmaximum exibility was a priority in the design of the test image generator. Using the input elds supplied, one is able to generate galaxies and point sources with varying sizes, inclinations, central wavelengths, etc. Also, by setting bad pixels, random noise, and the like, the FITS le that is written will contain the NICMOS FITS le format with this data in the extensions. Below is a comprehensive list of all the input elds for the test image generator: The fol lowing items relate to the object's size, location, and other characteristics:
Alpha

Sets the angle in degrees of an extended ob ject.
Gal Size

The surface luminosity pro le is de ned as being of a point from the center of the galaxy in pixels.
EPS

ex=GalSize,

where x is the distance

Ellipticity of galaxy disk.
X Coord

This is the x-coordinate of the center of the ob ject to be created in pixels. It can be set either bytyping in the number or by clicking at a position on the image display.
Y Coord

This is the y-coordinate of the center of the ob ject to be created in pixels. It can be set either bytyping in the number or by clicking at a position on the image display.
Wavelength

The of the central wavelength of a spectral line relative to the center of the spectrum in pixels.
Spec Width RA

The line width of the spectral line in pixels. The right ascension of the reference point within the image. This information is used to create a WCS.


50
DEC

APPENDIX A. NICMOSLOOK UTILITIES
The declination of the reference point within the image. This information is used to create a WCS. The following items are concerned with the image characteristics and the FITS data le output:
Exposure

The exposure time.
PHOTFNU

The ux scale of the image in
R D Noise

JY



sec=DN

The readout noise in electrons.
Gain

The gain of the detector.
Background

The value of the background level in DN.
Total Flux

The total ux in the created ob ject in DN.
CRPIX1

The pixel coordinate of the reference point within the image. This information is used to create a WCS.
CRPIX2

The pixel coordinate of the reference point within the image. This information is used to create a WCS.
CRVAL1

not used.
CRVAL2

not used. The following are parameters to set ags in the mask plane of the image. They are currently not used.
Rd Sol Err

Line Corr


A.2. ADDITIONAL UTILITIES
Dark Corr Flat Field Bk Ground Bad Pixel Cosmic Ray Defect Pix Sat Pix Miss Data

51

A.2 Additional Utilities
Code for utilities for the generation of test and calibration data is in the ECF IDL directory:
home ns3c idl ecf

A.2.1 Point Spread Function Generator

For PSFS creation, TinyTim must be used. See TinyTim documentation for further details. TinyTim is an interactive program that has two steps. The rst step tiny1 generates a parameter le for the second step. Using tiny1 to generate these parameter les is very tedious and time consuming. An IDL program is available to create these parameter les for each grism. To use the program: 1. Start IDL 2. psfs grismid 3. exit idl Where grismid is the name of the grism.

The next step in PSFS creation is now ready to begin. The following commands are for the UNIX c-shell.


52 prompt foreach foreach? tiny1 $i foreach? end
i `ls *.pf` $i

APPENDIX A. NICMOSLOOK UTILITIES

The above command takes each input le, as input to tiny1 and writes the output to the same le thus overwriting it; which is OK because it is not needed at this point. The following command takes the output from tiny1 and uses it for tiny2. prompt foreach i `ls *.pf` foreach? tiny2 $i foreach? end writes a FITS le for each run. These les must be placed in a single FITS le for use by the Test Image Generator. A second IDL utility has been written for this purpose.
tiny2

1. Start IDL 2. mkf grismid 3. exit idl

Where grismid is the name of the grism.

Look at the code les: psfs.pro and mkf.pro, directly to see what they expect and where output les are written. These steps are meant as an aid and therefore, are not particularly exible.

A.2.2 Flat eld "Cube" Generator
1. Start IDL 2. mkcube 3. exit idl Look at the code le, is written.
mkcube.pro

CalnicC and NICMOSlook expect a FITS le containing 3-dimensional at eld data. Once the 18 at eld ts les are assembled in a directory, IDL may be run to create the 3-dimensional FITS le.

, to see what is expected and where the output le

A.2.3 Reading FITS Tables Into IRAF

It is possible to read the output FITS binary tables into an IRAF onedspec format with spec tab2im.cl. Within IRAF, spec tab2im should be de ned as a task bytypeing


A.2. ADDITIONAL UTILITIES
Subsequently, NICMOS FITS spectra can be converted to IRAF images bytypeing
task spec tab2im *.fits task spec tab2im = spec tab2im.cl

53


54

APPENDIX A. NICMOSLOOK UTILITIES


Appendix B FITS Input File Format
B.1 NICMOS Input FITS File Description
The FITS format is used for storing NICMOS science and calibrator reference data. Multiple images are stored in a single le using FITS Image extensions. The layout of the FITS input data les that NICMOSlook expects is: Primary FITS header Header for Science Data Image for Science Data Header for Error plane Image for Error plane Header for Quality ags Image for Quality ags Header for Sampling Image for Sampling Header for Integration time Image for Integration time

B.2 NICMOS FITS Header Keyword Requirements
55


56 Name

APPENDIX B. FITS INPUT FILE FORMAT
FITS Keyword Requirements Description The following keywords are REQUIRED for NICMOSlookto run. EXTEND States le may contain extensions NEXTEND Number of extensions should be 5 for NICMOS data les XTENSION For the 5 extensions this should be 'IMAGE' EXTNAME Actual data stored in current extension SCI ERR etc EXTVER Extension version not used currently but should be 1 ROOTNAME Rootname of observation set. Used for output le names FILTER Name of the lter grism used in observation The following keywords are not required but some functionality depends on them. ADCGAIN Analog digital conversion gain BACKEST1 Background estimate 1 BACKEST2 Background estimate 2 BACKEST3 Background estimate 3 CRPIX1 x-coordinate of reference pixel CRPIX2 y-coordinate of reference pixel CRVAL1 First axis value at reference pixel CRVAL2 Second axis value at reference pixel CTYPE1 The coordinate type for the rst axis CTYPE2 The coordinate type for the second axis Partial of rst axis coordinate w.r.t. x CD1 1 CD1 2 Partial of rst axis coordinate w.r.t. y CD2 1 Partial of second axis coordinate w.r.t. x CD2 2 Partial of second axis coordinate w.r.t. y EXPTIME Commanded exposure duration PHOTFNU Inverse sensitivity JY*sec DN


Appendix C FITS Output File
C.1 FITS File Description
When any of the submenus that direct all output to a le is chosen under Spectra a single FITS le is created. The rst entry in this le is an extensive primary FITS header that contains general information of the observation as well as setup information for the series of binary table extensions that follow the header. Table C.3 gives a detailed description of this primary header.

C.1.1 FITS Binary Table Description
Following the primary FITS header in the NICMOSlook FITS output le are the binary tables which contain information for each ob ject detected on the image pair. One binary table is written out for each object detected on the image pair. Preceding each table, is an extensive FITS header describing the ob ject whose spectra information is contained in the table. The binary table consists of ve columns which contain an ob ject's, ux in mJy, wavelength in microns, statistical errors, deblending errors, and total errors. The error vectors contain values for each ux wavelength pair for the entire spectrum. An example of a FITS table header is in table C.3 57


58

APPENDIX C. FITS OUTPUT FILE

C.2 FITS Primary Header Example
Name SIMPLE = T BITPIX = 8 NAXIS = 0 EXTEND = T ADCGAIN = 10.0000 DROOTNM = 10.0000 Description Written by IDL: 26-Mar-1997 11:21:10.00

Binary Table follows analog-digital conversion gain analog-digital conversion gain Direct image header information DROOTNAM= 'N33L0101T' rootname of the observation set DPARALEL= 'NO ' indicates if observa taken in parallel DTARGNAM= 'OBJECT1 ' proposers target name DRA TARG= 15.0000 right ascension of the targetdeg J2000 DDEC TAR= 45.0000 declination of the target deg J2000 DCAMERA = 1 Camera in use 1, 2, or 3 DAPETURE= 'NIC1 ' aperture in use DOBSMODE= 'ACCUM ' array readout mode DFILTER = 'F205W ' lter wheel element in beam during obs DBAKEST1= 1.00000 background estimate number 1 DBAKEST2= 0.50000000 background estimate number 2 DBAKEST3= 1.50000 background estimate number 3 DPHOTMOD= 'NICMOS,1,F205W,DN' photometry mode DPHOTFLA= 8.9776100e-19 inverse sensitivity DPHOTFNU= 1.00000 inverse sensitivity JY*sec, DN DPHOTZPT= -21.100000 ST magnitude system zero point DPHOTPLA= 11292.4 Pivot wavelength of the photmode DPHOTBW = 1653.64 RMS bandwidth of the photmode DPOHTPDG= 'GROUND ' photometric calibration tabl pedigree DBAKPDGR= '- ' background model parameters table pedigree


C.2. FITS PRIMARY HEADER EXAMPLE
FITS Primary Header Example: continued

59

Description Grism image header information ROOTNAME= 'N33L0101T' rootname of the observation set PARALELL= 'NO ' indicates if observa taken in parallel PROPOSI = 'UNKNOWN ' Unable to determine from GRISM image PEP EXPO= '- ' PEP exposure identi er including sequence LINENUM = '- ' PEP proposal line number PR INV L= '- ' last name of principal investigator PR INV F= '- ' rst name of principal investigator PR INV M= '- ' middle initial of principal investigator ORIENTAT= 0.00000 position angle of imag y axis deg. e of n SUNANGLE= 0.00000 angle between sun and V1 axis MOONANGL= 0.00000 angle between moon and V1 axis altitude of the sun above Earths limb SUN ALT = 0.00000 FGSLOCK = 'FINE ' commanded FGS lock DATE-OBS= 'now' UT date of start of observation TIME-OBS= 'now' UT time of start of observation EXPSTART= 0.00000 exposure start time Modi ed Julian Date EXPEND = 0.00000 exposure end time Modi ed Julian Date EXPTIME = 600.000 commanded exposure duration EXPFLAG = 'NORMAL ' Exposure interruption indicator CAMERA = 1 Camera in use 1, 2, or 3 APERTURE= 'NIC1 ' aperture in use OBSMODE = 'ACCUM ' array readout mode BACKEST1= 1.00000 background estimate number 1 BACKEST2= 0.50000000 background estimate number 2 BACKEST3= 1.50000 background estimate number 3 FILTER = 'F205W ' lter wheel element in beam during obs PHOTMODE= 'NICMOS,1,F205W,DN' photometry mode PHOTFLAM= 8.9776100e-19 inverse sensitivity PHOTFNU = 1.00000 inverse sensitivity JY*sec, DN PHOTZPT = -21.100000 ST magnitude system zero point PHOTPLAM= 11292.4 pivot wavelength of the photmode PHOTBW = 1653.64 RMS bndwid of the photmode PHOTPDGR= 'GROUND ' photometric calibration tabl pedigree BACKPDGR= '- ' background model parameters table pedigree NEXTEND = 6 Number of table extensions END

Name


60

APPENDIX C. FITS OUTPUT FILE
Name Description XTENSION= 'BINTABLE' Written by IDL:23-Oct-1996 16:57:25.00 BITPIX= 8 NAXIS = 2 Binary table NAXIS1 = 20 Number of bytes per row NAXIS2 = 95 Number of rows PCOUNT = 0 Random parameter count GCOUNT = 1 Group count TFIELDS = 5 Number of columns EXTNAME = '0 ' Ob ject number TFORM1 = '95r ' Length and type of column TTYPE1 = 'Lambda ' Label for column TDIM1 = '95 ' Array dimension for column TFORM2 = '95r ' Length and type of column TTYPE2 = 'Flux ' Label for column TDIM2 = '95 ' Array dimension for column TFORM3 = '95r ' Length and type of column TTYPE3 = 'Stat Err' Label for column TDIM3 = '95 ' Array dimension for column TFORM4 = '95r ' Length and type of column TTYPE4 = 'Debl Err' Label for column TDIM4 = '95 ' Array dimension for column TFORM5 = '95r ' Length and type of column TTYPE5 = 'Tot Err ' Label for column TDIM5 = '95 ' Array dimension for column Sextractor Ob ject Information CLASS = 0.0140434 Ob ject type galaxy or star Ma jor axis of ob ject A OBJ = 0.00000 B OBJ = 0.00000 Minor axis of ob ject XCOORD= 122.913 x-coord of ob ject in pixel COORD= 122.835 y-coord of ob ject in pixel RA OBJ= 0.00000 Right ascension of ob ject deg J2000 Declination of ob ject deg J2000 DEC OBJ = 0.00000 Spectral Line Extraction LINEPK1 = 3495.08 Line peak LINELAM1= 1.58542 Central wavelength LINEWID1= 0.319452 Width of line LINESN1 = 17.7016 Signal to noise CHI2 1 = 0.0376511 Chi-square ILF 1 = 0.0011627 Integrated Line Flux

C.3 FITS Table Header Example


C.3. FITS TABLE HEADER EXAMPLE
FITS Table Header Example: continued Name Description LINEPK2 = 4402.01 Line peak LINELAM2= 1.35420 Central wavelength LINEWID2= 0.927476 Width of line LINESN2 = 22.2950 Signal to noise CHI2 2 = 0.0376511 Chi-square Integrated Line Flux ILF 2 = 0.0011627 LINEPK3 = 4402.57 Line peak LINELAM3= 1.53262 Central wavelength LINEWID3= -0.941618 Width of line LINESN3 = 22.2978 Signal to noise CHI2 3 = 0.0376511 Chi-square ILF 3 = 0.0011627 Integrated Line Flux LINEPK4 = 3568.77 Line peak LINELAM4= 1.99225 Central wavelength LINEWID4= -0.164191 Width of line LINESN4 = 18.0749 Signal to noise Chi-square CHI2 4 = 0.0376511 ILF 4 = 0.0011627 Integrated Line Flux LINEPK5 = 3966.44 Line peak LINELAM5= 2.38475 Central wavelength LINEWID5= -0.681985 Width of line LINESN5 = 20.0889 Signal to noise Chi-square CHI2 5 = 0.0376511 ILF 5 = 0.0011627 Integrated Line Flux NLINES= 6 Actual of lines CONSHORT= 0.00000 Continuum lambda rst line CONLONG = 0.00000 Continuum lambda last line RMSFIT = 0.000294 RMS of last t FORDER = 4 Order of last t MAG BEST= -15.3115 Best magnitude estimate MAG ERR = 0.00190733 Error in magnitude estimate BAKUSED = 1.00000 Background estimate used BACKFIL = 'bckG206W. ts' Background le used END That is all

61


62

APPENDIX C. FITS OUTPUT FILE


Appendix D Setup
D.1 Grism Speci cation File
D.1.1 Contents
The grism speci cation le, grismspec.dat, de nes the lters and grisms NICMOSlook recognizes and the parameters NICMOSlook assumes for each of them. It is found in the $NICMOSLOOK BASE calib directory. The le is read by NICMOSlookand speci es a keyword to read the name of the lter or grism. For each recognized name, it list then: 1. whether the image is a direct image or a grism image 2. in case of a grism image the position of the spectrum in dispersion direction the dispersion relation of the grism the distortion of the spectra the o set of the spectrum relative to the direct image a 3d at eld FITS le to be used for at eld correction a le which de nes the convolution kernel for ob ject detection a FITS le containing image background data a le which de nes the lter response curve 3. the FITS le with the appropriate PSF for the direct image or grism used to identify ob jects.

D.1.2 Format

Parameters in grismspec.dat are listed one entry lter or grism per line. They are written format free and have to be given in the right order. Comment lines can be placed at any place in the le; they are recognized by a ;" in the rst column. 63


64

APPENDIX D. SETUP

Eachentry must start with the keyword in the image header which is used to identify the image. Next, the value for this keyword and the corresponding image type is given d= direct image, s= spectrum. The subsequent information in an entry depends on whether the entry refers to a direct image or a grism image.

Direct Image
For a direct image, the only additional information listed is the name of the FITS le with a PSF used to search for ob jects on the image. The location of the corresponding FITS le has to be the same directory directory as grismspec.dat

Grism Image
For a grism image, the three numbers following the image type de ne the dispersion relation, parameterized with four parameters a0, a1, a2, and a3. They are used to compute the wavelength in microns via = a0+ a1 x + a2 x2 + a3 x3 D.1 where x is the x-position in pixels relative to the center of the direct image. Subsequently, four parameters for the the image distortion in y direction b0, b1, b2, and b3 are listed. They parameterize the distortion as:


= b0+ b1 r + b2 r2 + b3 where
r

r3

D.2

D.3 = x2 + y2 is the radius position in pixels and is the y-deviation in pixels of the spectrum from a horizontal line. The next parameter is the angle  which speci es a possible small global rotation of the spectrum relative to an exact alignment along a row. The angle is measured counterclockwise in units of degrees. Following the angle is the spectra o set eld. This is the number of pixels in the y direction that the grism image is o set from the direct image. The next column speci es the name of the at eld le to be used for extraction of the spectra. The next column lists the name of the ts le with a PSF used to search for ob jects on an image. The le containing the background estimage is listed next, followed by the a le containing the response speci cations for the grims. Refer directly to $NICMOSLOOK BASE calib grismspec.dat for examples.

q


D.2. TUNABLE PARAMETERS

65

D.2 Tunable Parameters
Tunable Parameters default values ADCGAIN 12.83 BADPIX THRESH 50.0 BG PIXELS 2 BLEND FACTOR 2 CONT FACTOR 2.5 DAO THRESH 2 DEBLEND ERR MIN 0.1 DET THRESH 2 EXTOBJ THRESH 0.3 FIT ORDER1 3 FIT ORDER2 4 MAXLINES 24 MAXOBJS 400 MAXPIX 256 MIN NPOINTS 20 MIN PIX NO 10 MIN SIGNSE 7 N ITERATIONS 10 N ITERLOW4 O THRESH1 0.05 O THRESH2 2.0 O THRESH3 3.0 REJ THRESH 3 SZ FACTOR 2 DEBUG LEVEL 101 NOFLATFIELD 1 NOBACKGND 1 ; Analog digital conversion gain ;Percent of allowable bad pixels before ob ject discarded ; Width of region for background estimate in pixels ; Blending factor for merging gauss lines ; Continuum factor ; Brightness threshold for dao nd ; Deblending error minimum ; Spectral line search point detection threshold ; 0.0 = extended ob ject, 1.0 = star ;Polynomial t order for rst run of poly t ;Polynomial t order subsequent runs of poly t ; Maximum number of lines to keep ; Maximum number of ob jects to create for sextractor ; Maximum number of pixels on grism per dimension ; Minimum number of points for tting ; Minimum acceptable line width ; Minimum allowable signal noise value of lines ; Maximum number of t iterations ; Iteration number where 2nd t order param is used ; First overlap threshold for deblending ; Second overlap threshold for deblending ; Third overlap threshold for deblending ; Ob ject rejection threshold ; 2nd order mom factor of extended ob j from sextractor ; Debugging level ;Turn o on 1 0 at elding ;Turn o on 1 0 background subtraction FITS Keywords default values Target A D Gain Filter Inverse Sensitivity TARGNAME ADCGAIN FILTER PHOTFNU Target of observation Analog to Digital conversion gain Filter used in ovservation Unit Conversion ADU - mJy

D.3 Con guring FITS Keywords


66

APPENDIX D. SETUP


67


68

APPENDIX E. LIST OF ALL INPUT FILES

Appendix E List of all input les
Name Description grismspec.dat Descriptions of all available grisms and lters bckG096W. ts Background estimates for grism J bckG141W. ts Background estimates for grism H bckG206W. ts Background estimates for grism K calnicc.setup Tunable parameters for Calnic C default.conv Convolution le for SExtractor default.nnw Neural network weights le for SExtractor default.param List of SExtractor parameters to return default.sex Tunable parameters to modify SExtractor behavior SExtractor output le not used by Calnic C calnicc tmp.cat nicmosFF. ts Flat eld image for eachwavelength of grism cube Weight le for grism J weighted weight G096W. ts weight G096W unweighted. ts Weight le for grism J unweighted weight G141W. ts Weight le for grism H weighted weight G141W unweighted. ts Weight le for grism H unweighted weight G206W. ts Weight le for grism K weighted weight G206W unweighted. ts Weight le for grism K unweighted at. ts For at subtraction dark. ts For dark subtraction bias. ts For bias subtraction G096.response For response G096W.response For response G141.response For response G141W.response For response G206.response For response G206W.response For response tskeywds.dat FITS keyword con guration le F090M.conv Convolution les used by SExtractor for each lter F095N.conv F097N.conv F108N.conv F110M.conv F110W.conv F113N.conv


69 Input les to Calnic C: continued F140W.conv F145M.conv F150W.conv F160W.conv F164N.conv F165M.conv F166N.conv F170M.conv F171M.conv F175W.conv F180M.conv F187N.conv F187W.conv F190N.conv F196N.conv F200N.conv F204M.conv F205W.conv F207M.conv F212N.conv F215N.conv F216N.conv F222M.conv F237M.conv Continuation of convolution les for each lter


70

APPENDIX E. LIST OF ALL INPUT FILES


Appendix F Output Files of NICMOSlook
F.1 Ob ject List
NICMOSlook creates an internal ob ject list for each ob ject detected or otherwise identi ed on the loaded images. This list contains information regarding the location, size ma jor, and minor moments along the axis, and the angle of the ob ject's orientation. It is often convenient to save these values to a data le so, for example, that other images can be loaded without destroying current information. At a later time, the ob ject list can be loaded back in and processing can be resumed where left o . To save the current ob ject list, select the Save to File menu item under the Ob jects ! Ob ject List menus. Below is an example output le of the an output le created in this manner: x Coord y Coord Ma j Mom Min Mom Angle 15.0005 62.9947 0.00000 0.00000 -90.0000 157.000 65.0026 1.38696 0.489883 0.306816 82.9863 67.0088 2.87775 0.747376 -53.1466 To read the le backinto memory, one would use the Input File menu item under the Ob jects ! Find Ob jects menus.

F.2 Derived Line Parameters
By Selecting the Save Line Params button on a spectra's Plot Popup, one can save the derived line parameters to a data le. Below is an example of a le created by choosing this button: Polynomial t order used: 4 Line No. Peak Flux Lambda FWHM 1 0.03 1.01 0.025 2 0.12 0.08 0.020 3 0.09 0.95 0.001 71 SN 20.297 18.302 15.122 CHI2 32.78 30.92 28.16 ILF 6.8508E-04 5.5243E-04 4.8173E-04


72

APPENDIX F. OUTPUT FILES OF NICMOSLOOK

F.3 Catalogue Output
The following shows an example of a human readable catalogue entry. Ob ject Number 0 from observation le .. data direct. ts RA = 0.00000 DEC = 0.00365056 Output FITS le: N33L0101T 0. ts Ob ject Type Probability 1.0 = star; 0.0 = extended ob ject: 0.0140434 Ob ject size 2nd order moments along axes: 13.8394 7.73015 Ob ject angle along ma jor axis:: 43.0521 Total ux of spectrum: 38295.5 Number of lines: 7 Order of last curve t: 4 5 Most signi cant lines: Flux Wavelength FWHM 957.141 1.15657 0.0403234 758.077 1.64862 0.0103599 458.424 1.94372 0.0264251 534.383 2.05612 0.0505358 655.240 2.28973 0.1081202 SN 147.223 11.1825 10.5146 16.7858 30.5714 CHI2 596.2 314.4 154.2 774.5 594.3 ILF 41.1161 22.704 28.51 33.45 12.17

RMS of last polynomial t : 71.2457 Continuum long : 593.520 Continuum short: 252.996 Below is an example of a log le entry. Date Run : Fri Sep 27 10:53:43 1996 Direct image : . data direct. ts Grism image : . data spec. ts Ob jects Number Number Spectra Spectra Spectra Spectra Spectra Found : 3 of extended ob jects : 2 of point Sources : 1 Extracted : 3 with no lines : 0 with 1 lines : 1 with 2 lines : 2 with more than 2 lines: 0

Warnings: none Errors : none


F.3. CATALOGUE OUTPUT

73


74

APPENDIX F. OUTPUT FILES OF NICMOSLOOK


Appendix G Sample Data Files
The directory called:
$NICMOSLOOK BASE data

contains two direct grism pairs of sample data les

ov2 dir.fits ov2 gri.fits ov3b dir.fits ov3b gri.fits

75