Coma Seeing

In order to reliabiliy discriminate point sources from extended sources, it is first necessary to estimate the point spread function. GALWORKS performs this task by estimating the "shape" or "sh" (as well as a number of other parameters) for each source (presumably a star) and constructing the ridgeline that characterizes the PSF as a function of time (or equivalently, the in-scan coordinate). The ridge can then be employed for star-galaxy separation.

The "sh" and ridge is computed early on in the 2MAPPS pipeline with the sub-system module "SEEMAN" (which is really a GALWORKS module), from which the information is used in several subsytems, including PROPHOT. Later in the pipe, GALWORKS recomputes the ridgeline -- GALWORKS has more information at hand to derive a more robust ridge or "sh" estimate. GALWORKS typically requires more stars per bin to derive a "sh" estimate (for the ridge) compared to SEEMAN (which derives a "sh" estimate per coadd), and so some differences can be seen when the stellar number density is low (as in the Coma region) and/or when the seeing changes very rapidly. Some examples are given here to demonstrate the difference between the SEEMAN ridge and the GALWORKS ridge, as well as to show the typical seeing conditions for the Coma data.

A detailed analysis of the seeing conditions during the night of 970423 can be found in the document Seeing Analysis for 970423n , which also includes parameter definitions and other useful information relevant to seeing.

Scan 003, night of 970418

A number of points (in green) lie well above the mean ridge line, these are galaxies (the Coma cluster is centered near the scan center, scan coords ~ 8000 to 15000). A few points lie well below the mean ridge line, these are sources affected by bad pixels (masked pixels and otherwise) as well as poor sub-pixel dithering.

The SEEMAN ridge and the GALWORKS ridge are nearly identical for this scan. The mean "sh" value is about 1.1, corresponding to a FWHM ~ 2.9 arcsec, roughly corresponding to an atmospheric seeing of 2.1 arcsec.

Scan 052, night of 970423

This scan shows a very unusual feature in the seeing: the seeing significantly jumps between two or three scans, then returns to its previous value. It is as if a transient (either local or atmospheric) of some sort has passed over the observing dome. What is crucial to note, however, is that the seeing is not well tracked as shown by the Galworks ridgeline (in red), although the SEEMAN ridgeline does better (in white) because is tracks the seeing with slightly higher frequency. The consequence of this action is that sources located in the region of the "transient" phenomenon, will be severly underestimated in their seeing profile -- stars will look like galaxies. A number of false galaxies will be pickup up by GALWORKS. This scan should, in normal production mode, be rejected.