Thursday, January 16, 2014

Creating the Radio Sky Planisphere

Meetings of the American Astronomical Society have, for the last couple of years, included a ‘schools outreach’ event. Arecibo Observatory participated in this for the first time at the Long Beach meeting in January 2013. For that meetings we had stellar-evolution bookmarks (with graphics by +Rhys Taylor), which the students finished by putting coloured beads on a ribbon, each colour corresponding to an evolutionary stage. Unfortunately the age-range attending turned out to be somewhat older than the target range for the activity, so we didn’t want to repeat it this year. We also wanted to do something more radio-astronomy oriented (we are a radio observatory after all) and that still left the students with something they could bring home afterwards. And we wanted (if possible) to out-do the model pulsars made by the Fermi gamma-ray telescope!

Thus the Radio Sky Planisphere was born.

The idea came from my wife, who has worked in outreach, and the folks at the Visitors Center were enthusiastic, so I set about working out how to make it a reality. There were two major problems to be overcome: firstly, I needed to find out how to calculate the window shape for a planisphere; secondly I needed to get an all-sky radio map in the correct projection for a planisphere. This is the azimuthal equidistant projection, and is fairly simple to work in using polar coordinates: hour angle (or right ascension) is the angle and 90 - dec (or 90 + dec for a projection centred on the south pole) is the radius.

To make the window, I first looked about on the internet, but couldn’t find anything existing so I coded it myself in IDL. This involved solving the spherical trigonometry equations to find the declination and hour angle of the horizon. Fortunately, at the horizon many of the angles are 0 or 90 degrees, so sin and cos reduce to 0 or 1. The equations I ended up with were:

sin(Dec) = cos(Lat)*cos(Az)
cos(HA) = tan(Lat)*tan(Dec)

By sweeping around 360 degrees in Az(imuth) and inputting the Lat(itude), this meant I could solve for Dec(lination) and then for H(our) A(ngle). There were a few other complications, such as making sure I got the negative hour angle solution out as well as the positive, but this was essentially solved. I could make windows, and I could make them for any arbitrary latitude. The IDL code to do this is at I also set this up so multiple windows could be printed on a single planisphere, leaving the selection of the exact latitude to the person cutting it out.

The next challenge was getting the actual map. I decided to use the 408-MHz Bonn all-sky survey (Haslam, Salter, Stoffel and Wilson, 1982) which is available via the Max Planck Institute for Radioastronomy’s surveys page at As an added bonus, Chris Salter - the second author on the survey - works down the corridor from me at Arecibo.

This proved a bit trickier. It turned out that the ‘arc’ projection (which should be what I wanted) was actually returning a straight pixel map, but with the coordinates for the azimuthal equidistant projection! Once I realised this, I was able to download the straight pixel map and then wrote a simple IDL code to convert this into a set of radius, angle coordinate pairs for the polar projection, with a corresponding set of flux values for each coordinate. I could then use IDL’s built-in polar_contour function to plot these, using filled contours and an appropriate colour map to get the radio sky-wheel to put inside my planisphere. In order to get a good spread of colours, I used the built-in hist_equal function to get a histogram-equalized distribution of the flux values. Overall, this ends up looking very similar (except for the projection) to the colour map in Haslam et al.’s paper.

The last step was to plot on some optical sources so people could orientate themselves, and see how these correspond to the radio sky. I chose asterisms/constellations I can actually recognize, so apologies if your favourite one is missing. The ones I went with were the Plough (a.k.a. the Big Dipper) and the nearby stars of Arcturus and Polaris; Cassiopeia; Cygnus and the two other nearby stars from the summer triangle (Altair and Vega); Scorpius; the Southern Cross and Pointers (alpha and beta Centauri); Orion along with Sirius and the Pleiades star cluster; and the Large and Small Magellanic Clouds along with the bright southern star of Canopus. The stars are shown as filled circles, while the ‘fuzzy’ extended objects (Pleiades and the Magellanic Clouds) are shown as open circles. For both stars and extended objects, the diameter is proportional to 5 - magnitude (as brighter stars have lower magnitudes, this means that brighter objects are larger; a magnitude of 5 is about the limit of what is visible with the naked eye).

All that was left was to prettify what was currently rather clinical looking IDL output. I did this in Adobe Illustrator (for the planisphere body) and Photoshop (for the sky wheel). This also involved reducing the size of the sky wheel from around 50 MB for the IDL postscript output to around 500 kB for the Photoshop PDF output - it’s quite unusual for rasterizing a vector image to reduce the size, but that was certainly the case here!

The Radio Sky Planisphere was a great hit at the AAS meeting. We got far fewer schools at the actual outreach event than hoped, as the Maryland public schools were closed due to the freezing weather, but we gave out planispheres to teachers, outreach professionals, and astronomers over the remaining couple of days of the meeting and ended up getting rid of all 250 that we had brought. I also had enquiries about a southern hemisphere version, which I have now also produced.

There are now three versions available for download from Google Drive, along with a ‘Guide to the Celestial Features on the Radio Sky Planisphere’: the AAS meeting version (which can be cut to latitudes between 20 and 60 degrees north), the northern hemisphere version (5 to 65 degrees north), and the southern hemisphere version (5 to 55 degrees south). Note that all of these are sized for US Letter paper with fairly narrow margins, to match the printer we have at the observatory. If you print these yourself, you should scale them to fit your page - as the planisphere body and the sky wheel are in the same document, they should scale by the same amount so everything should still fit.

The AAS meeting version was designed to cover (to within 5 degrees) the US from Puerto Rico (or Hawaii) up to Anchorage in Alaska, therefore also taking in large swathes of the rest of the world along the way, but missing out on places closer to the equator and (of course) the whole southern hemisphere. The northern and southern hemisphere versions between them can be cut to match (to within 5 degrees) every town with a population over 10,000 on the Earth. If anyone wants one for the Antarctic, then please contact me!