Easy (SME) User Handbookvalenti/sme/SME-Manual.pdfthe marcs2012t02 grid, the Manual folder of the SME distribution contains several PDF les with plots showing the parameter distribution,

Spectroscopy Made “Easy” (SME)User Handbook

Jeff A. Valenti1,Nikolai E. Piskunov2,

Ulrike Heiter3

May 21, 2012

[email protected]@[email protected]

1

Document History

Issue Revision Date Author Comment

D 15 2012-05-21 UH Added description of model atmospheregrids and updated tables with newmarcs2012 grids.

D 14 2012-05-03 UH Updated SME structure overview table.Added descriptions of new fields vmac pro,atm radius, cmod orig, cmod, rchisq, cpu,lib version. Updated description of fieldschisq and atomic.

D 13 2012-05-03 UH Added info on negative vmac

D 12 2012-02-13 UH Added info on new solar atlas nso2011

D 11 2011-12-16 UH Added text on Instrumental Profile

D 10 2011-12-05 UH Revised section on shared library, addi-tional clarification of sme.vrad, startedFAQ chapter

D 9 2011-05-22 UH Minor clarification of sme.vrad andsme.cscale

D 8 2011-03-07 UH Added info/reference for NSO atlas

D 7 2011-02-21 UH Minor update of model atmosphere grid ta-bles

D 6 2011-02-19 UH Major revision of Chapter describing theSME Structure

D 5 2011-02-18 UH Updated sections on sme.feh andsme.abund, including default abundancetable

D 4 2011-02-01 UH Added marcs2010 grid parameters

D 3 2010-10-22 UH Minor revisions related to installation

D 2 2010-07-05 UH Merged plain text and latex versions,updated installation-related parts, addeddocument history, added to svn repository,added section on Line Data from helptextinside IDL code

D 1 2009-06-11 JV+NP Created latex version

D 0 2003-02-21 JV Created plain text version

Contents

1 Introduction 7

1.1 Document Conventions . . . . . . . . . . . . . . . . . . . . . . 7

1.2 Supported Architectures and Operating Systems . . . . . . . . 7

1.3 Vienna Atomic Line Database (VALD) . . . . . . . . . . . . . 8

1.4 Code Heritage . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

2 Installing SME 9

2.1 Obtaining SME . . . . . . . . . . . . . . . . . . . . . . . . . . 9

2.2 Contents of the SME Distribution . . . . . . . . . . . . . . . . 9

2.2.1 Model atmosphere grids . . . . . . . . . . . . . . . . . 10

2.3 Additional Libraries . . . . . . . . . . . . . . . . . . . . . . . 10

2.4 Updating the IDL Path . . . . . . . . . . . . . . . . . . . . . . 11

2.5 Color Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

2.6 Shared Library . . . . . . . . . . . . . . . . . . . . . . . . . . 12

3 Graphical User Interface 16

3.1 Conceptual Overview of SME . . . . . . . . . . . . . . . . . . 16

3.2 Introduction to the SME GUI . . . . . . . . . . . . . . . . . . 17

3.2.1 Starting the SME GUI . . . . . . . . . . . . . . . . . . 17

2

CONTENTS 3

3.2.2 Nested Widgets . . . . . . . . . . . . . . . . . . . . . . 17

3.2.3 Main Menu . . . . . . . . . . . . . . . . . . . . . . . . 18

3.2.4 Help inside SME – “Help” . . . . . . . . . . . . . . . . 19

3.2.5 Exiting SME – “Done” . . . . . . . . . . . . . . . . . . 19

3.2.6 File Selection Widget . . . . . . . . . . . . . . . . . . . 19

3.3 Reading Line Data – “Line Data” . . . . . . . . . . . . . . . . 20

3.4 Defining SME jobs – “Controls” . . . . . . . . . . . . . . . . . 21

3.4.1 Instrumental Profile . . . . . . . . . . . . . . . . . . . 21

3.5 Reading Observed Spectra – “Observations” . . . . . . . . . . 22

3.6 Submitting SME Jobs – “Jobs” . . . . . . . . . . . . . . . . . 22

3.7 Examining SME Job Results – “Examine” . . . . . . . . . . . 22

4 Data Structures 23

4.1 SME Structure – Overview . . . . . . . . . . . . . . . . . . . . 23

4.2 SME Structure – Input Fields . . . . . . . . . . . . . . . . . . 28

4.2.1 sme.version . . . . . . . . . . . . . . . . . . . . . . . . 28

4.2.2 sme.id . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

4.2.3 sme.vrad flag . . . . . . . . . . . . . . . . . . . . . . . 30

4.2.4 sme.cscale flag . . . . . . . . . . . . . . . . . . . . . . 30

4.2.5 sme.accwi . . . . . . . . . . . . . . . . . . . . . . . . . 30

4.2.6 sme.accrt . . . . . . . . . . . . . . . . . . . . . . . . . 30

4.2.7 sme.clim . . . . . . . . . . . . . . . . . . . . . . . . . 30

4.2.8 sme.maxiter . . . . . . . . . . . . . . . . . . . . . . . 30

4.2.9 sme.chirat . . . . . . . . . . . . . . . . . . . . . . . . . 31

4.2.10 sme.nmu . . . . . . . . . . . . . . . . . . . . . . . . . 31

CONTENTS 4

4.2.11 sme.mu . . . . . . . . . . . . . . . . . . . . . . . . . . 31

4.2.12 sme.nseg . . . . . . . . . . . . . . . . . . . . . . . . . 31

4.2.13 sme.wran . . . . . . . . . . . . . . . . . . . . . . . . . 31

4.2.14 sme.species . . . . . . . . . . . . . . . . . . . . . . . . 31

4.2.15 sme.lande . . . . . . . . . . . . . . . . . . . . . . . . . 31

4.2.16 sme.lineref . . . . . . . . . . . . . . . . . . . . . . . . 32

4.2.17 sme.glob free . . . . . . . . . . . . . . . . . . . . . . . 32

4.2.18 sme.gf free . . . . . . . . . . . . . . . . . . . . . . . . . 32

4.2.19 sme.vw free . . . . . . . . . . . . . . . . . . . . . . . . 32

4.2.20 sme.ab free . . . . . . . . . . . . . . . . . . . . . . . . 32

4.2.21 sme.vmac pro . . . . . . . . . . . . . . . . . . . . . . . 32

4.2.22 sme.atmo pro . . . . . . . . . . . . . . . . . . . . . . . 33

4.2.23 sme.n atm . . . . . . . . . . . . . . . . . . . . . . . . . 33

4.2.24 sme.atmo . . . . . . . . . . . . . . . . . . . . . . . . . 33

4.2.25 sme.atm teff . . . . . . . . . . . . . . . . . . . . . . . 33

4.2.26 sme.atm grav . . . . . . . . . . . . . . . . . . . . . . . 33

4.2.27 sme.atm wlstd . . . . . . . . . . . . . . . . . . . . . . 33

4.2.28 sme.atm radius . . . . . . . . . . . . . . . . . . . . . . 34

4.2.29 sme.atm file . . . . . . . . . . . . . . . . . . . . . . . 34

4.2.30 sme.wave . . . . . . . . . . . . . . . . . . . . . . . . . 34

4.2.31 sme.wind . . . . . . . . . . . . . . . . . . . . . . . . . 34

4.2.32 sme.sob . . . . . . . . . . . . . . . . . . . . . . . . . . 34

4.2.33 sme.uob . . . . . . . . . . . . . . . . . . . . . . . . . . 34

4.2.34 sme.obs name . . . . . . . . . . . . . . . . . . . . . . 34

4.2.35 sme.obs type . . . . . . . . . . . . . . . . . . . . . . . 35

CONTENTS 5

4.2.36 sme.iptype . . . . . . . . . . . . . . . . . . . . . . . . 35

4.2.37 sme.ipres . . . . . . . . . . . . . . . . . . . . . . . . . 35

4.2.38 sme.ip x, sme.ip y . . . . . . . . . . . . . . . . . . . . . 35

4.3 SME Structure – Input/Output Fields . . . . . . . . . . . . . 35

4.3.1 sme.teff . . . . . . . . . . . . . . . . . . . . . . . . . . 36

4.3.2 sme.grav . . . . . . . . . . . . . . . . . . . . . . . . . . 36

4.3.3 sme.feh . . . . . . . . . . . . . . . . . . . . . . . . . . 36

4.3.4 sme.vmic . . . . . . . . . . . . . . . . . . . . . . . . . 37

4.3.5 sme.vmac . . . . . . . . . . . . . . . . . . . . . . . . . 37

4.3.6 sme.vsini . . . . . . . . . . . . . . . . . . . . . . . . . 37

4.3.7 sme.vrad . . . . . . . . . . . . . . . . . . . . . . . . . 37

4.3.8 sme.cscale . . . . . . . . . . . . . . . . . . . . . . . . . 38

4.3.9 sme.gam6 . . . . . . . . . . . . . . . . . . . . . . . . . 38

4.3.10 sme.abund . . . . . . . . . . . . . . . . . . . . . . . . . 38

4.3.11 sme.atomic . . . . . . . . . . . . . . . . . . . . . . . . 40

4.3.12 sme.depth . . . . . . . . . . . . . . . . . . . . . . . . . 40

4.3.13 sme.mob . . . . . . . . . . . . . . . . . . . . . . . . . 41

4.3.14 sme.atm type . . . . . . . . . . . . . . . . . . . . . . . 41

4.4 SME Structure – Output Fields . . . . . . . . . . . . . . . . . 41

4.4.1 sme.cintb . . . . . . . . . . . . . . . . . . . . . . . . . 41

4.4.2 sme.cintr . . . . . . . . . . . . . . . . . . . . . . . . . 41

4.4.3 sme.smod orig . . . . . . . . . . . . . . . . . . . . . . 42

4.4.4 sme.cmod orig . . . . . . . . . . . . . . . . . . . . . . 42

4.4.5 sme.final atmo . . . . . . . . . . . . . . . . . . . . . . 42

4.4.6 sme.smod . . . . . . . . . . . . . . . . . . . . . . . . . 42

CONTENTS 6

4.4.7 sme.cmod . . . . . . . . . . . . . . . . . . . . . . . . . 42

4.4.8 sme.jint . . . . . . . . . . . . . . . . . . . . . . . . . . 42

4.4.9 sme.wint . . . . . . . . . . . . . . . . . . . . . . . . . 43

4.4.10 sme.sint . . . . . . . . . . . . . . . . . . . . . . . . . . 43

4.4.11 sme.rchisq . . . . . . . . . . . . . . . . . . . . . . . . 43

4.4.12 sme.chisq . . . . . . . . . . . . . . . . . . . . . . . . . 43

4.4.13 sme.crms . . . . . . . . . . . . . . . . . . . . . . . . . 43

4.4.14 sme.lrms . . . . . . . . . . . . . . . . . . . . . . . . . 43

4.4.15 sme.pfree . . . . . . . . . . . . . . . . . . . . . . . . . 44

4.4.16 sme.punc . . . . . . . . . . . . . . . . . . . . . . . . . 44

4.4.17 sme.pname . . . . . . . . . . . . . . . . . . . . . . . . 44

4.4.18 sme.covar . . . . . . . . . . . . . . . . . . . . . . . . . 44

4.4.19 sme.idlver . . . . . . . . . . . . . . . . . . . . . . . . . 44

4.4.20 sme.md5 . . . . . . . . . . . . . . . . . . . . . . . . . 44

4.4.21 sme.cpu . . . . . . . . . . . . . . . . . . . . . . . . . . 44

4.4.22 sme.lib version . . . . . . . . . . . . . . . . . . . . . . 45

5 Frequently asked questions (FAQ) 46

5.1 Miscellaneous questions . . . . . . . . . . . . . . . . . . . . . . 46

5.1.1 Floating underflow error . . . . . . . . . . . . . . . . . 46

Chapter 1

Introduction

Spectroscopy Made Easy (SME) is a software package that calculates syn-thetic spectra of stars and fits observed spectra.

1.1 Document Conventions

In this document, we denote a widget navigation sequence by a string with">" symbols separating selected menu items. For example, "Controls >

Global Parameters" refers to the result of selecting "Controls" in themain menu of the SME graphical user interface (GUI) and then "Global

Parameters" from the "Controls" menu.

In this document, we call file system containers “directories”, rather than“folders”.

1.2 Supported Architectures and Operating

Systems

You must have an IDL license to use SME. Demand for SME does not yet jus-tify the labor necessary to port SME to a non-proprietary software language.IDL version 5.1 or later should be adequate to run SME.

7

CHAPTER 1. INTRODUCTION 8

SME synthesizes spectra using pre-compiled shared libraries, available forthe following operating systems and architectures:

• Mac OS X on 64-bit Intel processors• Red Hat Linux on Intel processors• Windows on 32-bit Intel processors

We are willing to support other systems, but then we will need temporaryaccess to a C++ compiler on the target system.

1.3 Vienna Atomic Line Database (VALD)

SME requires atomic and/or molecular line data in order to synthesize spec-tra. SME accepts line data in the format returned by the Vienna AtomicLine Database (VALD). The VALD email service is the easiest way to obtainand format line data for SME. To register as a VALD user, see one of thefollowing sites:

• http://vald.astro.univie.ac.at/~vald/php/vald.php

• http://www.astro.uu.se/~vald/php/vald.php

• http://vald.inasan.ru/~vald/php/vald.php

VALD is described in Piskunov et al. (1995, A&AS, 112, 525) and Kupka etal. (1999, A&AS, 138, 119). If you use VALD, please cite these papers.

1.4 Code Heritage

The original release of SME (version 1.0) is described in Valenti and Piskunov(1996, A&A, 118, 595). If you use SME, please cite this paper. NikolaiPiskunov wrote the SME radiative transfer code in C++, assuming a plane-parallel geometry and local thermodynamic equilibrium (LTE). The IDL codewas written by Jeff Valenti and Nikolai Piskunov. Chemical equilibrium isdetermined using atomic and molecular data described in Valenti, Piskunov,& Johns-Krull (1998, ApJ, 498, 851). The SME code continues to evolvewith new code releases about once a year.

http://vald.astro.univie.ac.at/~vald/php/vald.php

http://www.astro.uu.se/~vald/php/vald.php

http://vald.inasan.ru/~vald/php/vald.php

Chapter 2

Installing SME

2.1 Obtaining SME

Create a directory to hold the SME software. The directory name "sme"

will be used in examples in this document. This directory will be an IDLsoftware directory, not a working directory for specific analysis projects.

The SME distribution is bundled together in a single gzipped unix tapearchive (tar) file named sme.tar.gz or similar. Place the tar file in yourSME software directory. Extract the contents using the unix commands"gunzip" and then "tar xvf". The tar file does not include any directorypath information.

2.2 Contents of the SME Distribution

The tar file contains about 70 IDL procedure files (*.pro), shared libraries(sme synth.so.*) for various systems, and a number of data files, listed inTable 2.1. In principle, the large data files may be deleted to save space, butthen the corresponding capability will fail. In general, all the files should beretained.

9

CHAPTER 2. INSTALLING SME 10

2.2.1 Model atmosphere grids

A number of model atmosphere grids are included in the SME distribu-tion. An overview of the grid parameters is given in Table 2.2. “Ku-rucz models” are 1D-LTE plane-parallel model atmospheres taken from Ku-rucz (1993, Kurucz CD-ROM No. 13. ATLAS9 Stellar Atmosphere Pro-grams and 2 km/s grid. Cambridge, Mass.: SAO). “NextGen models”are 1D-LTE plane-parallel model atmospheres described in Hauschildt etal. (1999, ApJ 512, 377). “MARCS models” are 1D-LTE plane-paralleland spherically-symmetric model atmospheres described in Gustafsson et al.(2008, A&A 486, 951), and are distributed on the MARCS website1. Forthe marcs2012t02 grid, the Manual folder of the SME distribution containsseveral PDF files with plots showing the parameter distribution, and modelstructures for selected Teff values.

All models included in SME have microturbulence 2 km/s and all except forthe marcs2012 models have [α/Fe]=0 (i.e. the marcs2008 and marcs2010

models are a mixture of “Standard” and “Alpha poor” models, see MARCSwebsite). The marcs2012 models have “Standard” composition, where [α/Fe]increases from 0 for solar metallicity to +0.4 for [m/H]=−1.0 and below.

2.3 Additional Libraries

SME uses some routines from the IDL Astronomy Users Library, maintainedby Wayne Landsman. You will need this library on your IDL path to usecertain features of SME, for example reading an observed spectrum from aspecific type of FITS file. If the library is not already on your system, youcan download the library from:

http://idlastro.gsfc.nasa.gov/homepage.html

You will also need the MPFIT package of functions for curve fitting underIDL, maintained by Craig Markwardt. If the library is not already on yoursystem, you can download the library from:

http://cow.physics.wisc.edu/~craigm/idl/idl.html

1http://www.marcs.astro.uu.se/

http://idlastro.gsfc.nasa.gov/homepage.html

http://cow.physics.wisc.edu/~craigm/idl/idl.html

http://www.marcs.astro.uu.se/


Keep SME and the additional IDL libraries in separate directories.

2.4 Updating the IDL Path

The new SME software directory must be added to your IDL path, for ex-ample using the unix environment variable IDL PATH. If you do not currentlydefine IDL PATH in your unix startup files, you should add the lines (exam-ple):

setenv IDL_DIR /usr/local/itt/idl

setenv IDL_PATH +$IDL_DIR/lib:/usr/local/lib/idl/astron:

/usr/local/lib/idl/mpfit:~/sme:~/sme/GUI:~/sme/atmospheres

As needed, change the locations of the IDL installation directory, the Astron-omy Users Library, the MPFIT library and the SME software directory. Ifyou already define IDL PATH, then simply add the Astro SME software direc-tory at or near the end of the directory list in the definition of IDL PATH. Ifyou use the IDL Development Environment (IDLDE), specify the path usingthe "Path" tab in "File > Preferences".

Only which.pro is known to conflict with a routine of the same name inanother IDL widely distributed library, namely the IMG library by EricDeutsch. To see if this is a problem for you, type "which, ’which’" atthe IDL prompt. If you are not running the SME version, either reorderdirectories on your IDL path or explicitly compile the SME version in yourIDL startup file.

Use the unix environment variable IDL STARTUP or the "Startup" tab inthe IDLDE "File > Preferences" widget to specify the name of your IDLstartup file ("~/.idlrc", for example).

2.5 Color Setup

The SME GUI option "Examine > Plot Flux Profiles" assumes the dis-play is using a color map or palette, which may not be true by default. If the


SME plot screen contains only reds and black, rather than tan and assortedcolors on a white background, you need to force the use of a color map withthe IDL command:

IDL> device, decompose=0

This must be done before the first plot window or widget is created in an IDLsession. If you normally do not need 16 or 24-bit color in IDL, you can putthis device command in your IDL startup file, using the methods describedin the previous section.

2.6 Shared Library

SME synthesizes spectra by calling an external shared library (sme synth),which must reside in your SME software directory. Pre-compiled sharedlibraries are distributed with SME for different computer architectures andoperating systems in the directory "lib". Currently, supported architecturesare:

Shared Library Operating System and Architecturesme synth.so.darwin.x86 64.64 Mac OS X on 64-bit Intel

(development platform)sme synth.so.linux.x86 64.64 Debian GNU/Linux on 64-bit Intelsme synth.so.linux.x86.32 Debian GNU/Linux on 32-bit Intelsme synth.so.Win32.x86.32 Windows on 32-bit Intel

The appropriate library binary file to use is selected by SME at run-time,based on information stored in the IDL system variable !VERSION (see SMEprocedure sme entrypoints).

For SME versions before r203 (2011-10-09), the pre-compiled librariesreside in the directory "source", and you need to create a symbolic link tothe appropriate shared library in your SME software directory. For example,on a 32-bit Intel system running Linux, use the Unix command:

ln -s source/sme synth.so.linux.x86.32 sme synth.so

The SME shared library will not be called until you submit your first SMEbatch job (see below). At this time, if an error occurs in the IDL routine


CALL EXTERNAL, then the library interface is not working. First, make sureyou are using a supported architecture and operating system, and in the caseof pre-r203 SME versions, that sme synth.so in the SME software directoryis linked properly. If you still have a problem, send the following informationto [email protected]:

• Output from SME, including the entire error message• Output from the IDL command, "help, /struct, !version"

• Detailed operating system description (from "uname -a")

[email protected]


Table 2.1: Data files included in the SME distribution.Directory File Size Relevant GUI Feature or

sme.atmo pro routine nameControls > Model

Atmosphere

atmospheres krz2.sav 16.0 MB > interpkrz2

7611 Kurucz models with parameters given in Table 2.2

atmospheres marcs2012t02.sav 12.0 MB ’interpkrz2’ (not in GUI)4306 MARCS models with parameters given in Table 2.2

atmospheres marcs2012t02cooldwarfs.sav 3.4 MB ’interpkrz2’ (not in GUI)1267 MARCS models with parameters given in Table 2.2

atmospheres marcs2008t2.sav 4.9 MB > interpmarcs20081

2290 MARCS models with parameters given in Table 2.2

atmospheres marcs2010t02p.sav 5.0 MB > interpmarcs2010p1


atmospheres marcs2010t02s.sav 6.3 MB > interpmarcs2010s1


atmospheres nggrid.idl 369 KB > interpng

251 NextGen models with 3000 K ≤ Teff≤ 5000 K

atlases nso.bin2 6.3 MB Observations > Segment of

Kurucz solar atlas

atlases nso2011.bin3 33 MB nonemain (e.g. sme) new pt.bm 13 KB Controls > Abundances1 also sme.atmo pro=’interpkrz2’ with corresponding file name providedin sme.atmogrid file2 Kurucz et al. (1984), “Solar flux atlas from 296 to 1300 nm”, mini-mum wavelength 3003 A, maximum wavelength 12480 A; can be read withrdnso1,w,s,3003,12480,dir=’/SME-dir/atlases’.3 Wallace et al. (2011, ApJS 195, 6), “An Optical and Near-infrared (2958–9250A) Solar Flux Atlas”, minimum wavelength2898 A, maximum wavelength 10000 A; can be read withrdnso2011,w,s,2898,10000,transm=strans,dir=’/SME-dir/atlases’

& scorr=s/strans, where scorr is the spectrum corrected for telluricabsorption lines.


Table 2.2: Parameter coverage of the model atmosphere grids (minimum,maximum, and number of values n). See text for further parameters andreferences. Due to gaps in the grids, the total number of models in each gridis less than that obtained by multiplication of n.

Kurucz modelsparameter min max nTeff [K] 3500 35000 55log g [cm s−2] 0.0 5.0 ≤11[m/H] −5.0 +1.0 ≤19

NextGen modelsparameter min max nTeff [K] 3000 5000 20 (no 4500)log g [cm s−2] 3.5 6.0 6[m/H] −1.0 +0.0 3

MARCS models from 2008 – marcs2008t2 (plane-parallel)parameter min max nTeff [K] 2500/4000 3900/8000 15/17log g [cm s−2] 3.0 5.5/5.0 6/5[m/H] −1.5/−5.0 +1.0 11/16 (incl. +0.05)

MARCS models from 2010 – marcs2010t02p1, marcs2010t02s2

parameter min max nTeff [K] 2500/4000 3900/8000 15/17log g [cm s−2] – p1 3.0 5.5/5.0 6/5log g [cm s−2] – s2 −0.5 3.5 9[m/H] −2.0 +1.0 11

MARCS models from 2012 – marcs2012t02

parameter min max nTeff [K] 2500/4000 3900/8000 15/17log g [cm s−2]3 −0.5 5.0 12[m/H] −5.0 +1.0 15

MARCS models from 2012 – marcs2012t02cooldwarfs

parameter min max nTeff [K] 2500 4000 16log g [cm s−2]1 3.0 5.5 6[m/H] −5.0 +1.0 151plane-parallel, 2spherically-symmetric, mass=1M�3plane-parallel for log g≥ 4, spherically-symmetricwith mass=1M� for log g≤ 3.5

Chapter 3

Graphical User Interface

3.1 Conceptual Overview of SME

Analysis with SME is divided into three tasks: defining a job, running thejob, and examining the results. Each of these tasks is described in detail insubsequent sections.

SME jobs are typically defined using the SME GUI, which is written inIDL (sme.pro). The GUI is used to read line data, set model parameters,read observed spectra, and define fitting masks. Once a request is fullyspecified, the SME GUI is used to create an SME input file with a ".inp"

extension. SME input files are standard IDL save files containing an SMEinput structure named sme.

The SME GUI can be used to submit SME batch jobs, which proceed in-dependently in the background. SME jobs may also be submitted from theIDL prompt. SME job execution is handled by non-interactive IDL code(sme main.pro), which calls an external shared library.

Job execution may take seconds, minutes, hours, or even days, dependingon the number of spectral lines being synthesized and the number of freeparameters. Once a job is done, an SME output file is created with a ".out"

extension. SME output files are standard IDL save files containing an SMEoutput structure named sme. In fact, the SME output structure is just the

16

CHAPTER 3. GRAPHICAL USER INTERFACE 17

SME input structure, with additional fields that contain the results of jobexecution.

The same SME GUI that is used to define jobs (sme.pro) may also be usedto read and plot the results in SME output files. The sme structure in SMEoutput files may also be restored directly into IDL memory and accessed withstandard IDL commands.

SME is often used in an iterative mode, where the results of a completedjob are used to define a new job. The SME GUI is an integrated tool thatsimplifies this process. For example, the results of an initial spectrum syn-thesis may be used to guide interactive specification of a fitting mask beforesubmitting a second job to solve for selected stellar parameters.

3.2 Introduction to the SME GUI

3.2.1 Starting the SME GUI

To activate the SME GUI, simply enter "sme" at the IDL command prompt.This action should create a top level widget titled “SME Tool” with sevenmain menu items (Help, Line Data, etc.) above a message display area. Ifthe SME GUI does not appear, and you get the error message:

Attempt to call undefined procedure/function: ’SME’,

then add the SME software directory to your IDL path, as described inSection 2.4.

The SME software directory is assumed to be the location where IDL firstfinds sme.pro. When the SME GUI is started, the adopted location of theSME software directory is printed in the main message area of the top levelwidget.

3.2.2 Nested Widgets

In response to user input, the top level widget in the SME GUI creates acascade of nested widgets. When a nested widget is created, the SME GUI


transfers input focus to the child widget and disables input to the parentwidget. In most windowing environments, buttons change appearance whenthey lose focus.

Every child widget has one or two buttons with text labels that destroy thechild and return input focus back to the parent widget. "Done" or "OK" but-tons accept switch settings and data entered into the child widget. "Cancel"or "Abort" buttons return control to the parent widget without implement-ing changes requested in the child widget.

Avoid using generic controls provided by the windowing environment to de-stroy SME widgets. In some environments, input focus will not return tothe parent. Instead, use SME widget buttons with text labels, as describedabove.

On the other hand, if all SME widgets are unresponsive, use generic controlsprovided by the windowing environment to destroy SME widgets until inputfocus returns to the IDL command prompt. Then type "reset" to reinitializethe widget handler in IDL, followed by "sme" to restart the SME widget.

3.2.3 Main Menu

Across the top of the main SME widget, seven main menu items are arrangedinto a horizontal toolbar with text labels. Each menu item is described inthe section indicated in the table below.

Menu Item DescriptionHelp 3.2.4Line Data 3.3Controls 3.4Observations 3.5Jobs 3.6Examine 3.7Done 3.2.5

When defining a completely new SME job, enter information by selectingmain menu items from left to right, starting with the “Line Data” item.Within sub-menus and pop-up widgets, enter information from top to bot-


tom. It is possible to enter information in a different order, but in some casesunexpected interactions may occur.

When using an existing SME input or output file as a template for a newSME job, first select "Examine > Read SME Structure from Disk" to readresults from an existing SME file. Then make the desired modifications byselecting menu items from left to right, starting with the “Line Data” item.

3.2.4 Help inside SME – “Help”

Some widgets in SME have a “Help” button that provides context-sensitivehelp. The information is presented in a text widget. Exit the text widget bypressing the wide “Done” button across the top.

3.2.5 Exiting SME – “Done”

To exit the SME GUI, select "Done > Exit". “Exit” is in a sub-menu toguard against accidental selection. The combination "Done > Continue" isa null operation that leaves the main SME widget with input focus.

The SME GUI stores job request information in common blocks, which nor-mally persist for the duration of an IDL session. Thus, job request informa-tion will not be affected by exiting and then restarting the SME GUI. Thisfeature can be useful, but it can also be confusing. To completely reset theSME GUI, exit and then restart IDL.

3.2.6 File Selection Widget

The SME GUI frequently obtains data from disk. Disk files are selected usinga generic file selection widget described here.

At the top of the widget, the current search directory appears in an editabletext box preceded by the label “Path”. Initially, the path is set to the IDLcurrent working directory, but each type of file selection widget independentlymaintains updated path information for the duration of an IDL session.


Immediately below the “Path” specification, the current file selection filterappears in an editable text box preceded by the text label “Filter”. The filteris a space-separated list of wildcarded filename patterns. A "?" matches anarbitrary single character, while an "*" matches an arbitrary string of anylength. The filter is reset to the default value each time a file selection widgetis recreated.

3.3 Reading Line Data – “Line Data”

This section describes the “Line Data” item in the main menu. See alsoSection 1.3.

This interface is used to select one or more line data files, which will

be read, validated, and merged once the "Ok" button is pressed. The default

filter is set so that only files with a ".lin" extension will be selected,

but this behavior can be altered by modifying the contents of the "Filter"

field. The directory tree may be traversed by using the mouse to select an

item in the "Subdirectories" column. Selecting "../" moves you up a level,

while selecting a listed subdirectory (if any) moves you down. A directory

may also be entered explicitly in the "Path" field at the top. All files in

the current directory matching "Filter" will appear in the "Files" column.

Any number of these files may be added to the "Selection" list by clicking

with the left mouse button on the desired file in the "Files" column. Files

may be removed from the list by clicking in the "Selection" column on the

file to be removed. Removing all files from the "Selection" list and

clicking the "Ok" button clears any line data currently being used by SME.

Buttons along the bottom of the widget perform the following actions:

"Add All" - Add all files in "Files" to the "Selection" list

"Remove All" - Remove all files from the "Selection" list.

"View" - Display contents of the file highlighted in "Files" column

"Cancel" - Exit *without* modifying the line data being used by SME,

regardless of the contents of the "Selection" list.

"Ok" - Accept the "Selection" list and process all line data

files as described below.


If "Ok" is selected, files in the "Selection" list are read, validated, and

merged. Files are assumed to be in the ASCII format returned by the Vienna

Atomic Line Database (VALD, Piskunov et al. 1995, A&AS, 112, 525).

See the relevant section of the SME User Handbook (Manual) for instructions

how to register for this extremely useful (and free!) service. Data

successfully read from the selected files will be sorted by wavelength. If

duplicate lines (identical wavelength and lower excitation energy) are

detected, copies are discarded and a diagnostic message is written to the

main SME interface. SME provides several other ways to read, examine,

manipulate, and write line data, including "Controls > Free Parameters",

"Controls > Line Parameters", "Jobs > Review", "Jobs > Save", "Jobs > Submit"

"Examine > Plot > ShowLine", and "Examine > Export".

3.4 Defining SME jobs – “Controls”

This section describes the “Controls” item in the main menu.

3.4.1 Instrumental Profile

The instrumental profile is specified by the type of a function with

which the model spectrum will be convolved. The possible types are

"Gaussian", "Sinc function", and "Table". The function type is

stored in the SME structure as sme.iptype. See the Manual section on

sme.iptype for corresponding string values.

For "Gaussian" and "Sinc function", the function is parametrized by

entering a value for the resolution lambda/(Delta lambda) in the

GUI. This value is a long integer, which is stored in the SME

structure as sme.ipres. For "Table", the profile function is

specified by two double arrays, sme.ip_x and sme.ip_y, containing

relative wavelength and profile values, respectively. In the GUI,

the values are read from an input file, which must have the number


of points in the first line, followed by two columns for sme.ip_x

and sme.ip_y. Click on the "Table" button to view a plot of the

tabulated profile.

The model spectrum will be convolved with a Gaussian or Sinc

function with a FWHM corresponding to wmid / sme.ipres

(wmid ... segment midpoint wavelength), or with the tabulated

profile.

3.5 Reading Observed Spectra – “Observa-

tions”

This section describes the “Observations” item in the main menu.

3.6 Submitting SME Jobs – “Jobs”

This section describes the “Jobs” item in the main menu.

3.7 Examining SME Job Results – “Exam-

ine”

This section describes the “Examine” item in the main menu.

Chapter 4

Data Structures

4.1 SME Structure – Overview

An input sme structure contains input fields and input/output fields. Anoutput sme structure also contains output fields. The concept of input andoutput refers to the Solver. The user passes an input sme structure to theSolver. Based on the contents of the input sme structure, the Solver calculatesone or more synthetic spectra, often fitting an observation. The Solver thenconverts the input sme structure to an output sme structure by modifyinginput/output fields that correspond to free parameters and by appendingoutput fields. The Solver then returns the output structure to the user. TheSolver does not change input fields or input/output fields that correspondto fixed parameters. Table 4.1 lists all sme structure fields with units forphysical parameters, IDL data type, array dimensions, and default valuesassigned when using the GUI.

23

CHAPTER

4.DATA

STRUCTURES

24

Table 4.1: Names of fields (tags) in sme structure, their input/output type, IDL data type, arraydimensions, units for physical parameters, and default values assigned when using the GUI.

Field name IO Data type Multiplicity Units Default valueab free i int [99] – undefinedabund io float [99] – see Table 4.3accrt i float – 0.0001accwi i float – 0.002atm file i string [n atm] – undefinedatm grav i double [n atm] log(cm s−2) read from atm fileatm radius i double [n atm] cm read from atm fileatm teff i double [n atm] K read from atm fileatm type io string – undefinedatm wlstd i double [n atm] A read from atm fileatmo i double [5, number of depth points] see Section 4.2.24 undefinedatmo pro i string – interpkrz2atmogrid file i string – ’krz2.sav’ set in

sme funcatomic io double [8, number of spectral lines] see Section 4.3.11 from line listchirat i float – 0.002chisq o float [number of iterations] – –cintb o float [nmu, nseg] erg/s/cm2/A/sr –cintr o float [nmu, nseg] erg/s/cm2/A/sr –clim i float – 0.01cmod o float [number of wavelength points] erg/s/cm2/A/sr –cmod orig o float [number of wavelength points] erg/s/cm2/A/sr –

CHAPTER

4.DATA

STRUCTURES

25

Table 4.1 – continued from previous pageField name IO Data type Multiplicity Units Default valuecovar o double [number of free par.,number of

free par.]– –

cpu o struct – –crms o float [number of iterations] – –cscale io float see Table 4.2 – undefinedcscale flag i int – see Table 4.2depth io double [number of spectral lines] – from line listfeh io float – undefinedfinal atmo o double [5, number of depth points] see Section 4.2.24 –gam6 io float – 1.0gf free i int [number of spectral lines] – undefinedglob free i string [number of global free param-

eters]– undefined

grav io float log(cm s−2) undefinedid i string – current dateidlver o struct – –ip x i double [number of IP points] – undefinedip y i double [number of IP points] – undefinedipres i numeric – undefinediptype i string – undefinedjint o long [nseg] – –lande i double [number of spectral lines] – from line listlib version o string – –lineref i string [number of spectral lines] – from line listlrms o float [number of iterations] – –

CHAPTER

4.DATA

STRUCTURES

26

Table 4.1 – continued from previous pageField name IO Data type Multiplicity Units Default valuemaxiter i int – 100md5 o string [2] – –mob io int [number of wavelength points] – undefinedmu i double [nmu] – reverse(sqrt(0.5*

(2*dindgen(nmu)+1)/nmu))

n atm i int – undefinednmu i int – 7nseg i int – number of line list inter-

valsobs name i string [number of observation files] – undefinedobs type i int – see Table 4.2pfree o float [number of free par.,number of

iterations]depending onparameter

–

pname o string [number of free parameters] – –punc o double [number of free parameters] depending on

parameter–

rchisq o float [number of iterations] – –sint o double [number of wavelength points

for intensity,nmu]erg/s/cm2/A/sr –

smod o float [number of wavelength points] depending onunits of observedflux

–

CHAPTER

4.DATA

STRUCTURES

27

Table 4.1 – continued from previous pageField name IO Data type Multiplicity Units Default valuesmod orig o float [number of wavelength points] depending on

units of observedflux

–

sob i double [number of wavelength points] from observa-tions

undefined

species i string [number of spectral lines] – from line listteff io float K undefineduob i double [number of wavelength points] same as sob undefinedversion i float – 3.0vmac io float km s−1 undefinedvmac pro i string – undefinedvmic io float km s−1 undefinedvrad io float see Table 4.2 km s−1 0.0vrad flag i int – see Table 4.2vsini io float km s−1 undefinedvw free i int [number of spectral lines] – undefinedwave i double [number of wavelength points] A undefinedwind i long [number of wavelength points] – undefinedwint o double [number of wavelength points

for intensity]A –

wran i double [2,nseg] A from line list intervals

CHAPTER 4. DATA STRUCTURES 28

4.2 SME Structure – Input Fields

Input fields contain fixed attributes of the task that the Solver should per-form. For example, sme.nmu specifies the number of intensity spectra at dif-ferent µ angles that the solver should compute and combine, when producinga flux spectrum for the entire star. The following subsections describe inputfields recognized by the Solver. Not all of the input fields are mandatory fora successful SME run.

4.2.1 sme.version

sme.version indicates the version of SME used to create the SME struc-ture. For example, sme.version may have a value of 2.2. An input sme

structure created by one version of the GUI can be converted to an outputsme structure by another version of the Solver, but the Solver will not changesme.version. Thus, sme.version specifies the format and meaning of inputfields in an sme structure. The Solver uses sme.version to facilitate back-wards compatibility, triggering code in the Solver that handles deprecatedinput fields.

The SME external shared library (the radiative transfer code) has a separateversion numbering, which is handled by the routine sme lib version.

Future versions of SME may add other types of version numbers to describethe equation of state, and the format and meaning of output fields.

4.2.2 sme.id

sme.id is a string that contains the date and time when the SME struc-ture was initially created. For example, sme.id may have a value of Sun

Jan 7 13:07:42 2006. Note that there are two spaces between Jan and 7

because the date may be two characters long. The GUI uses the IDL func-tion systime() to construct the date and time string stored in sme.id. Thestring contains 24 characters that encode day of week (3 characters), space,month (3 characters), space, date (2 digits), space, hours (2 digits), colon,minutes (2 digits), colon, seconds (2 digits), space, year (4 digits). SME doesnot use sme.id, so in principle the format could be altered by the user.


Table 4.2: Flags in SME structure. Meaning of sme.vrad flag,

sme.cscale flag, sme.obs type values.

sme.vrad flag meaning multiplicity-2 fixed global – default Scalar-1 free global Scalar0 solve for each segment Array[sme.nseg]

sme.cscale flag meaning multiplicity-3 fixed global/normalized flux Scalar-2 fixed global/absolute flux Scalar-1 free global Scalar0 fit scalar to each segment Array[sme.nseg]1 fit line to each segment – default Array[2,sme.nseg]

sme.obs type meaning0 none – default1 Segment of Kurucz solar atlas (procedure sme rdnso)2 Segment of an atlas file (procedure sme rdatlas)3 Read from disk file (procedure sme rdobs)4 Read from existing SME structure file


Future versions of SME may add another date field that indicates when theoutput fields were last modified.

4.2.3 sme.vrad flag

Flag for handling radial velocity shift sme.vrad, see Table 4.2.

4.2.4 sme.cscale flag

Flag for handling continuum scale sme.cscale, see Table 4.2.

4.2.5 sme.accwi

Minimum accuracy for linear spectrum interpolation vs. wavelength. Valuesbelow 10−4 are not meaningful.

4.2.6 sme.accrt

Minimum accuracy for sme.sint at wavelength grid points in sme.wint.Values below 10−4 are not meaningful.

4.2.7 sme.clim

If continuum points are not specified in the mask sme.mob (where sme.mob

eq 2), points within sme.clim of the maximum in the observation are con-sidered continuum points and sme.mob is updated.

4.2.8 sme.maxiter

Maximum number of iterations allowed when solving.


4.2.9 sme.chirat

Fractional change in sme.chisq below which convergence is assumed.

4.2.10 sme.nmu

Number of ”equal-area” µ angles at which to calculate specific intensity.

4.2.11 sme.mu

The equal-area midpoints of each equal-area annulus for which specific inten-sities were calculated. µ values for Gaussian quadrature are not conduciveto subsequent disk integration (???).

4.2.12 sme.nseg

Number of distinct wavelength segments, for which synthetic spectra arebeing calculated.

4.2.13 sme.wran

Beginning and ending wavelengths for each spectral segment.

4.2.14 sme.species

Species name for each line in the line list (element symbol and ionizationstage), used for line labels in the GUI plot window.

4.2.15 sme.lande

Effective Lande factor (line parameter, only used in GUI ”Show line” fea-ture).


4.2.16 sme.lineref

Series of reference numbers for each line as returned from VALD (only usedin GUI ”Show line” feature).

4.2.17 sme.glob free

Array of strings, specifying which of the global parameters are free pa-rameters. Possible values are ’TEFF’, ’GRAV’, ’FEH’, ’VMIC’, ’VMAC’,’VSINI’, ’GAM6’, and ’VRAD’ (if sme.vrad flag eq -1), and ’CSCALE’(if sme.cscale flag eq -1).

4.2.18 sme.gf free

Array of flags (0 or 1), specifying for which of the spectral lines the gf valuesare free parameters.

4.2.19 sme.vw free

Array of flags (0 or 1), specifying for which of the spectral lines the van derWaals damping parameters are free parameters.

4.2.20 sme.ab free

Array of flags (0 or 1), specifying for which of the 99 atomic elements theabundances are free parameters.

4.2.21 sme.vmac pro

Name of a procedure used to calculate the macroturbulence from atmosphericparameters.


4.2.22 sme.atmo pro

Name of the procedure used to return the interpolated model atmosphere.See Section 2.2. Undefined if sme.atmo is defined.

4.2.23 sme.n atm

Number of input model atmosphere tables stored in sme.atmo.

4.2.24 sme.atmo

Input model atmosphere table(s). Undefined if sme.atmo pro is defined.atmo[0,*] Mass column density (g cm−2) [sme.atm type eq ’RHOX’]

OR Continuum optical depth [sme.atm type eq ’TAU’]atmo[1,*] Temperature (in K)atmo[2,*] Number density of electrons (cm−3)atmo[3,*] Number density of atoms (but not electrons) (cm−3)atmo[4,*] Mass density (g cm−3)

4.2.25 sme.atm teff

Effective temperatures of input model atmospheres stored in sme.atmo.

4.2.26 sme.atm grav

Surface gravities of input model atmospheres stored in sme.atmo.

4.2.27 sme.atm wlstd

Standard wavelength for continuum optical depth of input model atmo-spheres stored in sme.atmo.


4.2.28 sme.atm radius

Radius of input model atmospheres stored in sme.atmo for spherical models.

4.2.29 sme.atm file

Names of disk files from which the input model atmospheres stored in sme.atmo

were read.

4.2.30 sme.wave

Wavelength scale for observations and/or model flux spectra.

4.2.31 sme.wind

List of indices (into sme.wave, sme.sob, sme.smod, etc.) that mark thelast point in each spectral subinterval, i.e. sme.wave[0:sme.wind[0]] is thewavelength scale for the zeroth subinterval.

4.2.32 sme.sob

Observed spectrum. All subintervals are concatenated, with the boundariesstored in sme.wind.

4.2.33 sme.uob

Uncertainties for observed spectrum points – used to calculate χ2.

4.2.34 sme.obs name

Name of disk file from which the observed spectrum was read. Used forhandling of special case sme.vmac eq -99.


4.2.35 sme.obs type

Flag for observation type, set in procedure sme observations, see Table 4.2.Presumably not used by the Solver.

4.2.36 sme.iptype

Type of profile used for instrumental broadening. Possible values are gauss,sinc, or table. See Section 3.4.

4.2.37 sme.ipres

Resolution λ/∆λ, if sme.iptype is gauss or sinc.

4.2.38 sme.ip x, sme.ip y

Values of relative wavelength and profile for instrumental broadening, ifsme.iptype is table.

4.3 SME Structure – Input/Output Fields

Input/output fields contain model parameters that the solver can adjust,using an observed spectrum as a constraint. If the user lets a parameter(e.g., Teff) be free, then the corresponding field (e.g., sme.teff) in the inputSME structure contains the initial guess. After fitting the observation, theSolver overwrites the initial guess with the final value. If the user leavesa parameter (e.g., Teff) fixed, then the output value is equal to the inputvalue. Sections 4.3.1 to 4.3.11 describe input/output fields that the solver willmodify, if the corresponding parameter is free. The remaining subsectionsdescribe fields which contain auxiliary data modified by the Solver.


4.3.1 sme.teff

sme.teff specifies the stellar effective temperature. For example, sme.teffmight have a value of 5770.28 for the Sun. If the sme.atmo pro field existsin an input sme structure, then the Solver uses sme.teff to interpolate ina grid of model atmospheres. If the sme.atmo pro field does not exist, thenthe Solver ignores sme.teff and obtains the atmosphere from sme.atmo.

4.3.2 sme.grav

sme.grav specifies the base-10 logarithm of the stellar surface gravity. For ex-ample, sme.grav might have a value of 4.437 for the Sun. If the sme.atmo pro

field exists in an input sme structure, then the Solver uses sme.grav to in-terpolate in a grid of model atmospheres. If the sme.atmo pro field does notexist, then the Solver ignores sme.grav and obtains the atmosphere fromsme.atmo.

4.3.3 sme.feh

sme.feh specifies the stellar metallicity, but only if sme.abund contains thesolar abundance pattern and the user is not solving for the abundances ofindividual elements. If sme.abund contains a non-solar abundance patternor if abundances for individual elements are free parameters (which causeschanges in sme.abund), then sme.feh is only a pseudo-metallicity parameterthat differs from the standard [m/H].

The Solver uses sme.feh to scale the abundance pattern in sme.abund forall elements except hydrogen and helium. sme.feh is a base-10 logarithmicfactor, so that a value of 0 yields no scaling and a value of −1 reduces by afactor of ten the abundances of elements heavier than helium.

In some cases, the Solver also uses sme.feh to interpolate in a grid of modelatmospheres. If sme.atmo pro exists and contains the name of a procedurethat uses [m/H] to generate a model atmosphere (usually by interpolation,rather than computation), then the Solver will use sme.feh as a proxy for[m/H]. Otherwise, sme.feh does not affect the model atmosphere. For ex-ample, the procedure ‘interpkrz2’ uses sme.feh to interpolate in a grid of


pre-computed atmospheres. The optimum value of sme.feh from a modelatmosphere perspective may not be the same as the optimum value from anelemental abundance perspective.

4.3.4 sme.vmic

This field specifies the microturbulence parameter used during computationof synthetic spectrum by the external library (Doppler broadening added tothermal broadening). Note: this parameter is not used for the selection of themodel atmosphere (all built-in model atmosphere grids are computed witha microturbulence of 2 km/s). However, an inconsistency for this parameterwill usually lead to negligible errors.

4.3.5 sme.vmac

Radial-tangential macroturbulence parameter (as a rule of thumb, divideGaussian macroturbulence values by 1.4), applied to the synthetic spectrumduring disk integration (function rtint). It may happen that the optimiza-tion procedure moves vmac towards negative values. In that case, the abso-lute value of vmac is used in rtint, but the original negative vmac value isretained, and eventually stored in the output structure. Negative vmac out-put values should be regarded as a warning sign for too high broadening byparameters other than vmac. E.g. the input instrumental resolution (ipres)might be too low, or a fixed input vsini might be too high.

4.3.6 sme.vsini

Projected equatorial rotation velocity, applied to synthetic spectrum duringdisk integration (function rtint).

4.3.7 sme.vrad

Radial velocity shift, applied to synthetic spectrum for each wavelength seg-ment by the Solver. This field is either a scalar (sme.vrad flag eq -2,


-1), or an array variable (sme.vrad flag eq 0), see Table 4.2). When solv-ing for radial velocity, the maximum allowed value is 30 km s−1. For op-timal performance, the observations should be corrected for radial velocitybefore use with SME. The purpose of the wavelength adjustment routine(sme crvmatch) is to measure and correct local uncertainties in the wave-length scale, which might be due to errors in radial velocity or errors inthe transition wavelengths of the dominant lines, or zero point errors in theobserved wavelength scale. Note that the algorithm used to determine the op-timum velocity shift is separate from the one used to optimize the global pa-rameters. The velocity shift algorithm is based on a one-dimensional goldensection search.

4.3.8 sme.cscale

Continuum scaling factor, applied to synthetic spectrum for each wavelengthsegment by the Solver. This field is either a scalar (sme.cscale flag eq -3,

-2, -1) or an array variable (sme.cscale flag eq 0, 1, see Table 4.2). Ifthe observed spectra are normalized to the continuum, sme.cscale flag

should be set to −3 and sme.cscale should be set to 1.0.

4.3.9 sme.gam6

Global correction factor to all van der Waals damping constants. Values of1.5 to 2.5 are sometimes used for iron.

4.3.10 sme.abund

sme.abund contains relative abundances for the first 99 atomic elements (H,He, Li, ..., Es). SME expresses abundances as the fraction of nuclei for eachelement, relative to the total number of nuclei for all elements (not justhydrogen), regardless of whether the nuclei are in ions, atoms, or molecules.There are several other conventions in the astronomical literature, so becareful when making comparisons.


Table 4.3: Default abundance pattern in sme.abund, taken from Grevesse,Asplund & Sauval 2007, Space Sciences Review 130, 105 (these are also thesolar abundances adopted in the MARCS-2008 model atmosphere grid).

H : 0.92 He : −1.11 Li :−10.99 Be :−10.66 B : −9.34C : −3.65 N : −4.26 O : −3.38 F : −7.48 Ne : −4.20

Na : −5.87 Mg : −4.51 Al : −5.67 Si : −4.53 P : −6.68S : −4.90 Cl : −6.54 Ar : −5.86 K : −6.96 Ca : −5.73

Sc : −8.87 Ti : −7.14 V : −8.04 Cr : −6.40 Mn : −6.65Fe : −4.59 Co : −7.12 Ni : −5.81 Cu : −7.83 Zn : −7.44Ga : −9.16 Ge : −8.46 As : −9.75 Se : −8.71 Br : −9.48Kr : −8.79 Rb : −9.44 Sr : −9.12 Y : −9.83 Zr : −9.46Nb :−10.62 Mo :−10.12 Tc :−20.04 Ru :−10.20 Rh :−10.92Pd :−10.38 Ag :−11.10 Cd :−10.27 In :−10.44 Sn :−10.04Sb :−11.04 Te : −9.85 I :−10.53 Xe : −9.80 Cs :−10.97Ba : −9.87 La :−10.91 Ce :−10.34 Pr :−11.46 Nd :−10.59

Pm :−20.04 Sm :−11.04 Eu :−11.52 Gd :−10.93 Tb :−11.76Dy :−10.90 Ho :−11.53 Er :−11.11 Tm :−12.04 Yb :−10.96Lu :−11.98 Hf :−11.16 Ta :−12.21 W :−10.93 Re :−11.81Os :−10.79 Ir :−10.66 Pt :−10.40 Au :−11.03 Hg :−10.91Tl :−11.14 Pb :−10.04 Bi :−11.39 Po :−20.04 At :−20.04

Rn :−20.04 Fr :−20.04 Ra :−20.04 Ac :−20.04 Th :−11.98Pa :−20.04 U :−12.56 Np :−20.04 Pu :−20.04 Am :−20.04


sme.abund[0] contains the relative number of hydrogen nuclei, whereas forother elements, sme.abund[1:98] contains the base-10 logarithm of rela-tive number of nuclei. Table 4.3 gives the default abundance pattern, whichis intended to be solar. The user may change the abundance pattern insme.abund, for example to adopt a different solar abundance pattern. Whensolving for individual elemental abundances, the Solver updates the abun-dance pattern in sme.abund, as the solution progresses. The default abun-dance pattern in sme.abund is normalized, as expressed in Equation 4.1.

sme.abund[0] + total(10 ∧ sme.abund[1:98]) eq 1. (4.1)

However, the abundance pattern will not necessarily remain normalized, ifthe user or the Solver modify sme.abund. Normalization is also sacrificedwhen the Solver applies the pseudo-metallicity offset in sme.feh.

4.3.11 sme.atomic

Line data for all spectral lines (all segments). The 8 entries for each lineare: atomic number, ionization state, wavelength (in A), excitation energyof lower level (in eV), log(gf), radiative, Stark, and van der Waals damp-ing parameters. Two of those can be let free – log(gf) and van der Waals(“Gamma 6”) – and are thus input/output parameters, while the remainingarray elements are fixed input parameters. Note that the first two entries(atomic number and ionization state) are equivalent to the sme.species

field, and are only used in the GUI procedure sme useout in the case whensme.species is undefined.

4.3.12 sme.depth

Line depth at central wavelength. Read from input line list and stored ininput structure. Recalculated by sme main for (final) calculated spectrumand stored in output structure.


4.3.13 sme.mob

Mask for pixels in observed spectrum. 0 = bad point (perhaps a poorlymodelled blend) ignored in fit, 1 = good point used in fit, 2 = continuumpoint used to match model and observed continua. Either input by the useror set automatically by the Solver (function sme crvmatch).

4.3.14 sme.atm type

Indicates whether the independent model atmosphere parameter is the con-tinuum optical depth (’TAU’) or the mass column density (’RHOX’, in gcm−2). This field is an input field when a model atmosphere is given insme.atmo or an output field when sme.atmo pro is used to interpolate in amodel atmosphere grid.

4.4 SME Structure – Output Fields

Output fields contain data generated by the Solver and are appended to theSME structure.

4.4.1 sme.cintb

Blue edge continuum intensities at the µ values in sme.mu, for each wave-length segment.

4.4.2 sme.cintr

Red edge continuum intensities at the µ values in sme.mu, for each wavelengthsegment.


4.4.3 sme.smod orig

Model spectrum flux calculated using initial guesses for free parameters. Allsubintervals are concatenated, with the boundaries stored in sme.wind.

4.4.4 sme.cmod orig

Model continuum flux calculated using initial guesses for free parameters.All subintervals are concatenated, with the boundaries stored in sme.wind.

4.4.5 sme.final atmo

Model atmosphere table interpolated in grid, using final values for free pa-rameters. Same columns as sme.atmo (see Section 4.2.24).

4.4.6 sme.smod

Model spectrum flux calculated using final values for free parameters. Allsubintervals are concatenated, with the boundaries stored in sme.wind.

4.4.7 sme.cmod

Model continuum flux calculated using final values for free parameters. Allsubintervals are concatenated, with the boundaries stored in sme.wind.

4.4.8 sme.jint

List of indices (into sme.wint and sme.sint) that mark the last point ineach specific intensity subinterval, i.e. sme.wint[0:sme.jint[0]] is thewavelength scale for the zeroth subinterval.


4.4.9 sme.wint

Irregularly spaced wavelengths for specific intensities in sme.sint.

4.4.10 sme.sint

Specific intensities on an irregular wavelength grid given in sme.wint.

4.4.11 sme.rchisq

Reduced chi-square for each iteration. It is calculated using the equation forsme.chisq below, but without the sob factor.

4.4.12 sme.chisq

Chi-square weighted by the observed spectrum flux for each iteration. Thisquantity is used by the Solver to determine the best-fit parameters. It iscalculated from the line mask regions of the observed and model spectrumusing the following equation:

χ2 =

∑( sob−smod

uob)2 · sob

nlpts− nfree− nseg

where the sum is over all line pixels, nlpts is the number of line pixels, andnfree is the number of global free parameters.

4.4.13 sme.crms

RMS discrepancy between model and observed continua (where sme.mob eq

2) for all iterations.

4.4.14 sme.lrms

RMS discrepancy between model and observed lines (where sme.mob eq 1)for all iterations.


4.4.15 sme.pfree

Initial, intermediate, and final values of free parameters.

4.4.16 sme.punc

Formal uncertainties in final parameter values. Systematic errors usuallydominate.

4.4.17 sme.pname

Names of parameters that were varied.

4.4.18 sme.covar

Covariance matrix for free parameters.

4.4.19 sme.idlver

Structure containing information on the IDL version used for the calculations.

4.4.20 sme.md5

MD5 checksums on the IDL source code and on the spectrum synthesis sharedlibrary used for the calculations.

4.4.21 sme.cpu

CPU information added in procedure run next inp.


4.4.22 sme.lib version

Version of the spectrum synthesis shared library used for the calculations.

Chapter 5

Frequently asked questions(FAQ)

5.1 Miscellaneous questions

5.1.1 Floating underflow error

Q: I could start SME and process a sample observation. After the message"SME completed the calculations", the SME run ended with an errormessage "% Program caused arithmetic error: Floating underflow".

What is the meaning of this message?

A: “Floating underflow” occurs during multiplication of very small numberswhich create an effective zero. The code replaces this with zero (correctaction) and reports this in the end. This warning can be ignored as it doesnot affect the result in any way.

46

Easy (SME) User Handbookvalenti/sme/SME-Manual.pdfthe marcs2012t02 grid, the Manual folder of the SME distribution contains several PDF les with plots showing the parameter distribution,

Documents