Next: koords: interactive co-ordinate fitting Up: Karma User Manual Previous: 7 kslice_3d: slicing data

# 8 kshell: looking for expanding shells

This tool allows you to browse your datacube searching for expanding shells. Since most shells are mostly radially symmetric, it is reasonable to azimuthally average concentric annuli around a designated centre. What this means is that the programme allows you to define a centre position on the sky and a maximum radius. It will then step through a range of radii (from zero to the maximum), and at each radius will average the values in one velocity map around the circumference of a circle (hence ``azimuthal average''). This process is repeated for each velocity map, and the result is an image with one axis corresponding to radius and the other velocity.

If you are looking at an expanding shell, for each velocity map the average radius of the emission from the centre is approximately constant. As you progress from one velocity extreme to the other, the radius will change from zero, pass through a maximum at the central velocity and go back down to zero. So, in the computed radius-velocity map, you should see half an ellipse (the ``signature''). If you have slightly missed the centre of the expansion, you will see a double ellipse, so this is a good way of finding the centre. If you are not looking at an expanding shell at all, you just see a mess, so this is also a good way for speculatively searching for shells. The lower the expansion velocity, the more squashed the ellipse becomes (vertically, of course). In fact, if you had a static shell, you would just see a straight line. The power of this tool is that it is both a fast way of browsing for expanding shells (because of their charactistic signature) and that it is insensitive to fragmentation of shells.

To drive this programme, you need a cube file (load with the Cubes button) and an image file (load with the Images button). Actually, the image file is optional: you can compute an image on the fly from the cube by selecting say the 0th moment option from the Image Mode menu, and adjusting the parameters as needed and pressing Apply. The reason you need an image is so that you can search for a possible centre of an expanding shell. The image is displayed in the left window. The image and cube files do not have to lie on the same co-ordinate grid.

Once you have loaded your files, you simply click with the middle mouse button (don't let go yet!) to define the centre of the shell, ``drag'' the mouse so that you see the expanding circle, and release the middle button when you think the radius is sufficiently large such that the entire shell is enclosed. Don't make it larger than is really necessary, since it will take longer for the programme to compute the radius-velocity image. After a short time (fraction of a second to a few seconds for typical cubes), you will see the computed radius-velocity image. Do you see a nice ellipse in the right window? If you move the mouse into the right window, you should see the current position (in sensible units) as well as the averaged intensity at that point. You can use the Export menu above the image to write the image (with co-ordinates) in a variety of formats (choose one of the ``(whole dataset)'' options).

A small tip for people looking at data with very high dynamic range (such as OHIR masers): set intensity scale for the right window to logarithmic (see the zooming section), otherwise you will only see a few bright blobs.

A screen snapshot is available here

Next: koords: interactive co-ordinate fitting Up: Karma User Manual Previous: 7 kslice_3d: slicing data

Richard Gooch
Mon Aug 14 22:25:04 PDT 2006