LAB 1: VARIABLE STAR LIGHTCURVES In this lab we will acquaint ourselves with variable star lightcurves and a lightcurvein data. We will use the data from the ROTSE-I prototype telescope that ran from1997-2000 to study gamma-ray bursts. This data, however, was taken in a way thatallows for straightforward identification and study of short period variable stars. In order to analyze astronomical data, one needs to be familiar with a computersystem (‘operating system’ or ‘OS’), and an analysis program. We will rely on theLinux OS, and the IDL programming platform for analysis. Our ultimate goal will be to find variable stars in fields of data taken from ROTSE-Iand to measure the variable period and amplitude for those that are regular. Wewill ultimately use the lightcurve shape and other information to classify thosevariables that we identify. The final deliverable for this lab will be a report,adhering to typical research standards, of the work accomplished.

LAB 1: VARIABLE STAR LIGHTCURVESLINUX 101: BASIC COMMANDS Login: Account: smurotse Password: ‘grb990123’ On Linux command line: pwd: what is current directory or folder ls: list contents of current directory less: view or read a text file specified asargument mkdir: make a new directory with name asargument cd: change directory by specifying a ‘pathname’ as argument cp: copy one file to a new file with adifferent name. Retains old file mv: rename a file. Original file is gone One really useful comment: ‘man’ for ‘manualpages’ for any Unix program pwd ls less .cshrc mkdir plots cd plots cp ./.cshrc cshrc copy.txt mv cshrc copy.txt copy.txt

LAB 1: VARIABLE STAR LIGHTCURVESIDL 101: PROCEDURES AND FILES Log into analysis workstation in Rm 101 ssh -Y [email protected] For this week, use ‘’ though password: ‘grb990123’ Go to the directory where the data structures are: cd phy3368/ your last name ls Read in our data, and examine the contents IDL IDL restore, ‘000409 xtetrans 1a match.dat’ This reads in the file and loads the data structure into memory forexamination. Cameras 1a, 1b, 1c, 1d for each user.R. Kehoe - VSP Labs5

LAB 1: VARIABLE STAR LIGHTCURVESDATA STRUCTURES To examine the overall ‘match structure’ IDL help, match, /str Try this with the contents of the ‘match structure’ IDL print, match.jd # Prints ‘Julián dates’ for observations IDL print, match.m[*,1000], match.ra[1000], match.dec[1000] Print magnitudes, RA and Dec for all observations, ‘star’ 1000 IDL print, match.m[120,1000:1100] Print magnitude for observation 120 for stars 1000 thru 1100 Now, print the observation time for the first observation, and magnitude andcoordinates of the first observation of the 1900th source in the match structure Then write down some attributes of your data you will want handy: # observations, # stars, starting and stopping time, field RA (RAC) and Dec (DECC)R. Kehoe - VSP Labs6

LAB 1: VARIABLE STAR LIGHTCURVES7LIGHTCURVES Now select an object number — it doesn’t matter whichone for now. I’ll pick 1900. Run a procedure ‘find burst’ to extract the data and plota lightcurve for this object IDL x find burst(match, 0,0,objid 1900,log ‘firsttry’) IDL ls *.ps IDL evince & Let’s have a look at your plot!R. Kehoe - VSP Labs

LAB 1: VARIABLE STAR LIGHTCURVES8SUPPORTING MATERIAL There is a ‘manual’ written by former students using thisdata for research: GBkhn5T5LEG255ynzC aWDrKADUY The first 3 chapters are useful for this course Also the web-site for ROTSE,, has severalreferences that may be of use at: In particular, Kehoe et al. 2002 (source of the data for thiscourse).R. Kehoe - VSP Labs

LAB 1: VARIABLE STAR LIGHTCURVES9PROJECT REPORT OUTLINES Research is most often communicated in papers publishedin research journals, whether they be academic or tradejournals. In this course, we will implement our report ofthe lab effort in a manner similar to the typical structure ofthese papers. As such, there are five typical sections: Introduction/Motivation: A section reviewing recenttheoretical or experimental questions, with a summaryof the state of experimental results. Detector/Apparatus: Summary of the key aspects of theexperimental setup that yielded the data.R. Kehoe - VSP Labs

LAB 1: VARIABLE STAR LIGHTCURVES10PROJECT REPORT OUTLINES Sections (cont.) Data Samples/Data Reduction: A section describing the nature ofthe data, how it was taken. It is also important to describe the earlier‘data reduction’ pipeline that yielded the physically meaningfulmeasurements in the match structures you will be using. Analysis/Results: Analysis steps, from selection, to backgroundreduction, phasing and classification. Analysis results are provided. Conclusions: The most important final results are summarized here,with whatever evaluations can be made about shortcomings andpotential future improvements or questions. There will also be an initial ‘Abstract’, and a final ‘References’ section.R. Kehoe - VSP Labs


LAB 2: FINDING VARIABLE STAR CANDIDATESCCD PERFORMANCE: UNCORRELATED IMPACTS Dark current Production of free electrons per pixel vs. time - due to thermalenergy present Also results noise per pixel that increases as sqrt(t) Malfunctioning pixels/columns Hot pixels - generally produces wide variations in reportedcharge independent of incident photons Bad columns can be a column that fails to be controlled properly Shows up as a column of very high charge Cosmic raysR. Kehoe - VSP Labs12

LAB 2: FINDING VARIABLE STAR CANDIDATES13CCD PERFORMANCE: CORRELATED IMPACTS Several effects occur such that they scale with the # of photons incidenton a pixel Pixel gain variations These are small changes in the # of electrons produced per #incident photons Can be due to nonuniformities in Si, or in electronics variationscontrolling pixels Gain variation across the field Vignetting: Aperture of telescope blocks out light from edges offield more than center Shutter sticking and other impacts on amount of light ontoprimary mirrorR. Kehoe - VSP Labs

14LAB 2: FINDING VARIABLE STAR CANDIDATESDATA REDUCTION Image must be corrected for correlated and uncorrelatedimpacts Usually done with specialized images for correction ‘Darks’: exposures with closed shutter to measure darkcurrent and flag bad pixels ‘Flats’: exposures with open shutter taken of a uniformillumination field, or medianing many sky fields Correction is then:Imagecorrected(x, y) (Image (Gives us ‘’ files)R. Kehoe - VSP Labsraw(x, y) dark(x, y))/flat(x, y)

LAB 2: FINDING VARIABLE STAR CANDIDATES15SOURCE EXTRACTION Once image is correct, we can find stars Generally a clustering algorithm of some kind: look for patters of illuminatedpixels closely spaced ‘Source extraction’ We use Source Extractor (SEXtractor) for this Some of the light coming to pixels is physical, but from the sky (eg. OH linesthat deexcite) Develop a model of this ‘sky background’ and subtract Look for nearest neighbor pixels above some threshold Yields # counts in a ‘source’ centroid (x,y) coordinate on the CCD (yields our ‘’ files)R. Kehoe - VSP Labs

LAB 2: FINDING VARIABLE STAR CANDIDATES16ASTROMETRY The array of sources arranged in rectangular (x,y) CCD coordinate grid This is a projection of the spherical astronomical coordinates (α,δ) onto aplane How to obtain astronomical coordinates Ie. Astrometry - measure position of sources Compare unique triangles of objects in (x,y) to unique triangles in a knownwell-measured astronomical catalog Eg. USNO catalog Fit to 2D analytic functions transforming (x,y) to (α,δ) Termed ‘matching’ Once each calibrated object list is made, they can be easily matched toeach other since they are all in absolute common coordinate system (α,δ) This is how a ‘match structure’ is made (after next step of photometry)R. Kehoe - VSP Labs

LAB 2: FINDING VARIABLE STAR CANDIDATESPHOTOMETRY Each source has physical coordinates, but still # counts in terms of measure ofbrightness We want brightness in some absolute measure of flux Eg. Magnitudes Now that we have the coordinate transformation We know what all of observed sources are in terms of USNO stars Compare measured # counts to the magnitudes for V or R band in USNO Fit the transformation relating the two Also determine, for field, magnitude at which efficiency to observe goesto 50% — ‘limiting magnitude’ M LIM Yields catalog of (mag, α, δ) from (#counts, x, y) While all stars are used in the comparison, including variable stars, theaverage transformation is dominated by nonvariables and so we can stillsee them Astrometry and photometry yields ‘’ from ‘’R. Kehoe - VSP Labs17

χ2 Σ ((mi – mavg)/ σi)218LAB 2: FINDING VARIABLE STAR CANDIDATESSOURCE SELECTION Picking up where we left off, let’s examine our match structure further to select lightcurves thatappear to be varying We use ‘find burst’, but this time we will use it to select objects that are statistically differentthan constant 3 statistical quantities we can cut on to select variable candidates delta magnitude: the difference in magnitude between the brightest and dimmestmeasurement in a lightcurve: Δm mmax mmin Maximum significance: the significance of the maximum variation: σmax (mmax mmin)2 ϵ2ϵmaxmin Minimum chi-squared: the square of the difference of each measurement from the medianof a lightcurve, divided by the square of the statistical uncertainty on this difference:2χ Σ[(mi mavg)σi]2 Values are summed up over all measurements in a lightcurve and divided by thenumber of measurements minus 2, to give a ‘χ2 per degree of freedom’ You can find more about these in the VSP paper: E. Fagg, et al., Journal of UndergraduateResearch in Physics (2009).R. Kehoe - VSP Labs

LAB 2: FINDING VARIABLE STAR CANDIDATESFINDING CANDIDATE VARIABLES How do we use find burst in this way? IDL x find burst(match, 0.1,3.0,log ‘secondtry’) Where 2nd and 3rd arguments are delta mag and max sig,respectively Examine outputs to see what you get: IDL evince & Examine this file, in all it’s pages. How many objects are found? What happens when you vary delta mag? Max sig? Now let’s add a powerful variable, chi-squared: IDL x find burst(match, 0.1,3.0,minchisq 2.0,log ‘thirdtry’) What happens to the # of candidates now?R. Kehoe - VSP Labs19

LAB 2: FINDING VARIABLE STAR CANDIDATES20FINDING CANDIDATE VARIABLES Now, look at the individual lightcurves for all pages in your .psfiles Can you pick out lightcurves that are physical, and those thatmay be unphysical? Why are there lightcurves that are not physical? What is happening there? What physics could produce the physical lightcurves? You will notice a lot of ‘background’ from apparent variations thathave nothing to do with variables How do we handle these? We will revisit this next time.R. Kehoe - VSP Labs

LAB 2: FINDING VARIABLE STAR CANDIDATES21PROJECT REPORT: INTRODUCTION AND MOTIVATION In conducting research, it is important to establish the scientific questions that aremore important to study. In presenting scientific research, we need to provide areview of these questions and description of how they related to each other, andthe project at hand. These questions might be theoretical, or experimental/observational. In many cases, both are relevant. In this section of your report, youwill: Conduct a literature search on a topic relevant to your research. For instance,the ‘period evolution of close eclipsing systems’ Describe concisely the theoretical problems or questions that are recently inthe literature, eg. What are theoretical predictions of time evolution of period? Also describe the current experimental or observational state of work in thisarea, eg. What sensitivities or surveys have been pursued to find such timeevolution?R. Kehoe - VSP Labs

LAB 2: FINDING VARIABLE STAR CANDIDATESPROJECT REPORT: INTRODUCTION AND MOTIVATION (CONT.) Your 1st section will answer the above questions along the following lines: 2 or more pages Double spaced, 1” margins Not including Abstract, figures or tables, or references It will include at least one paragraph each covering the three main areas: What is the overall topic and why interesting generally What are the theoretical issues with this topic? What is the observational status of this research area? Plus a References Cited section at the end. Please aim to include at least 5references from the literature (not including textbooks) in this draft. Suggested places to look for articles on-line: or Put in some keywords to each and examine the search results This will be the first section of your final report, and this draft is due at mid-term asgiven on the syllabus.R. Kehoe - VSP Labs22


LAB 3: REDUCING BACKGROUNDS24TELESCOPE OPERATIONS Telescopes operated according to a scheme which isorganized by target For ROTSE-I orphans data: Gamma-ray bursts (GRBs) In most cases, specific targets specified before observing Eg. Specific galaxies to measure in the Dark EnergySurvey Not all observations scheduled the same priority Variable stars may be scheduled lower priority than GRBs Affects when observations occur, and their frequencyR. Kehoe - VSP Labs

LAB 3: REDUCING BACKGROUNDS25TELESCOPE OPERATIONS (CONT.) Seasonal and monthly cycles Targets rise and set at different times depending on season Moon up during observing some fraction of each month Moon injects more photons into sky during observations, Varies with phase particularly large wh