Tutorial

The light of a galaxy

In this lesson you measure how the light of a galaxy fades from its centre outwards, using a Hubble Space Telescope image of the galaxy UGC 12517. It takes about thirty minutes on your own, or one class period. You need no programming. Every step is done with the mouse, and every step is also written as a command you can keep.

1. Open the image

   Press Open the example on the welcome screen (or File ▸ Open example). The image goes into buffer 1; a buffer is a numbered slot that holds one image. The example also loads a mask, u12517j, drawn in red over the bright stars and the detector's edges. Masked pixels are left out of every measurement.

   

   Look at the terminal at the bottom: the two lines starting with › are the commands that did this. rd 1 u12517j.fits.gz reads the image into buffer 1, and mask load … loads the mask.

2. Look at the light

   Move the mouse over the image. The status bar shows the pixel's value and its position on the sky (right ascension and declination). The galaxy's centre is tens of times brighter than its outskirts, so with a linear scale only the core shows.

   In the bar above the image, change Scale to Log or Asinh, and tick Histogram. Drag on the image with the right mouse button to change the contrast. Try a colour map in Colours.

   

   QuestionHow far out does the galaxy's light go? Compare the picture with Linear, Log and Asinh. Does the galaxy have an edge?

3. Find the sources

   Open Tools ▸ Source finder and press Find sources (or the button in the top bar). Each cyan ellipse is a star or a galaxy that was detected and measured. Click one to see its row in Results ▸ Sources, and click a column header to sort the catalogue.

   

   Open Results ▸ Plots and plot FLUX_RADIUS (the radius holding half the light) against MAG_AUTO (the magnitude).

   

   QuestionStars are points: the telescope makes them all the same size, whatever their brightness. Which sources in the plot are stars? Why do the faintest sources all look small?

4. Find the centre and the size

   Open Tools ▸ ELLIPROF. Press Pick on the image and click the galaxy's bright core. With Snap to sources ticked, the small red cross jumps to the centre the source finder measured. Then press Pick beside R1 and click where the galaxy fades into the sky. The dashed red circle shows R1, the largest ellipse to fit.

   The example already holds the values of the published fit: X0 = 567, Y0 = 562, R0 = 9, R1 = 347 pixels, 23 ellipses, and a sky level of 3246 subtracted. Press Estimate beside Sky to measure the sky yourself.

   QuestionThe pixels are 0.128″ wide. How large is R1 in arcseconds? The compass shows which way is north. Is the galaxy round, or stretched in one direction?

5. Mask what is not the galaxy

   Other stars and galaxies would add light that does not belong to UGC 12517. Press Mask the found sources in the ELLIPROF tool: every source except the galaxy itself goes into a new mask layer called sources.

   Touch it up by hand with the tools in the top bar: circle, ellipse, box, polygon, brush, eraser. An ellipse, box or polygon waits in magenta until you apply it. Drag it to move it, drag its squares to resize it, and turn it with the round handle or [ ]. Press Enter to apply it, or Esc to cancel. Hold Alt or Shift to unmask. Ctrl+Z undoes.

   

   The layers on the left list every mask in its own colour. Untick one to hide it, and tick the masks to use under Mask with in the ELLIPROF tool.

6. Fit the isophotes

   Press Fit isophotes (Ctrl+E). An isophote is a line of equal brightness, like a contour on a map. ELLIPROF fits an ellipse to each one, from R0 outwards, measuring its brightness, centre, ellipticity and orientation.

   

   The terminal shows the command that ran:

   elliprof 1 X0=567 Y0=562 R0=9 R1=347 NR=23 NITER=10 SKY=3246 RMSTAR MASK=u12517j TOMODEL=3 TORESID=4

   QuestionDo the ellipses keep the same shape and orientation from the centre to the outskirts? What could make them twist?

7. Read the profile

   Open Results ▸ Profile. It plots the brightness of each isophote against its radius. Choose other quantities in the list: the ellipticity, the position angle, the centre, and the cos 3θ and cos 4θ terms, which show whether isophotes are boxy or discy.

   Tick R¼ axis. Many elliptical galaxies follow de Vaucouleurs' law, a straight line when the log of the intensity is plotted against R^¼. Tick R¼ law to fit it. The fit gives the effective radius R_e, which holds half the light, and the scatter about the law.

   

   Tick Table to see the numbers. Save CSV… writes the table and Save PNG… writes the plot as a figure.

   

   QuestionR_e is 126.5 pixels of 0.128″. What is it in arcseconds? If the galaxy is 60 Mpc away, how many kiloparsecs is that? (1″ at 1 Mpc ≈ 4.85 pc.)

8. Look at what is left

   Click the residual buffer on the left: the image minus the model made of the fitted ellipses. A perfect model would leave only noise. Tick Blink above the image, click the image, and press ← and → to switch between the image, the model and the residual.

   

   QuestionWhat does the model miss? Look for dust, a disc, rings, shells or companions hidden in the galaxy's glow. Try [9] = div([1] - [3], [3], 0) in the terminal: the residual as a fraction of the model.

9. Keep and repeat the work

   Every step you took is a command in the terminal. File ▸ Save the commands writes them to a file, and File ▸ Run a command file (or astrovista --run session.txt) repeats the whole measurement. This is how the example in the repository is written:

   rd 1 u12517j.fits
   mask load u12517j u12517j.dmask 1        # the published mask (0 = masked)
   sources 1                                # find the sources
   mask new sources 1
   mask sources sources SCALE=1.5           # every other source, 1.5 Kron ellipses
   elliprof 1 X0=567 Y0=562 R0=9 R1=347 NR=23 NITER=10 SKY=3246 RMSTAR MASK=u12517j TOMODEL=3 TORESID=4

For a class

Exercises

1. Does masking matter? Fit once with only the u12517j mask, and once with sources ticked too. Save both profiles as CSV and compare the outer points. Where do they differ, and why?
2. The sky. Change the sky by ±1% and fit again. Which part of the profile changes most? What does this say about measuring the faint outskirts of galaxies?
3. Stars and galaxies. In Results ▸ Plots, find the stellar sequence. Use Masks ▸ Mask the sources where… with a condition such as FLUX_RADIUS < 2.5 to mask only the stars.
4. Shape. Plot the ellipticity and the position angle against the radius. Is UGC 12517 rounder in the centre or in the outskirts?
5. Arithmetic. Make an image of only the bright pixels with [5] = where([1] > 10000, [1], nan). What fraction of the image is brighter than 10000? (Hint: stats 5.)
6. Your own galaxy. Download a FITS image of a nearby elliptical galaxy from an archive (HST, Legacy Survey, SDSS), open it with File ▸ Open, and repeat the lesson.

The same lesson is in the program: Help ▸ Documentation ▸ A first lesson. Teachers can hand out a command file so every student starts from the same place.

For astronomers

Using it in research

• Reproducible: a session is a plain text file of commands, so it can go into a paper's repository next to the data.
• Standard names: the source finder keeps SExtractor's keywords (DETECT_THRESH, MAG_AUTO, FLUX_RADIUS, …) and writes the segmentation image too.
• Coordinates: X0, Y0 and the mask commands use ELLIPROF's convention, in which the first pixel spans 0 to 1. The status bar also shows the FITS pixel (from 1) and the sky position.
• Files: FITS is recognised by content, whatever the file is called; name.fits[sci] or [sci,2] picks an extension. Masks can be FITS or legacy .dmask files, with 0 = masked.
• Output: CSV tables with aligned columns, PNG figures, DS9 regions, and every buffer as FITS.

Please cite the repository and the underlying tools: see Citing.

Glossary

Words used