Integrated Source Flats

The ideal way to measure the flat-field response would be to observe a single, bright point source 256x256 times, so it was centered on different array pixel each time. The flat would then be constructed by combining the signal from each illuminated pixel from all of the frames. Of course, such a scheme is not practical because it would take lots of time, the atmospheric transparency and seeing variations would cause the star brightness to vary slightly in each exposure, and the sky background could vary.

An alternative way to construct such a flat might be to use the fact that 2MASS will observe millions of sources each night with a wide variety of color and brightness and background levels. If over the course of one or more nights, each pixel sees the same brightness distribution of sources to within some limit (), then all of the frames taken over that time period could be combined to generate a very good flat-field image. For such a procedure to work it would be necessary to show that the distribution of sources seen by each pixel is indeed identical, and to devise the proper way to combine all survey images.

Advantages:

The color of the flat-field is matched perfectly to the average source color. Because this flat is constructed from point source measurements, it accurately measures the point source optical transfer function of the system. This procedure takes advantage of the unique nature of the 2MASS data set, and it requires no distinct flat-field observations.

Disadvantages:

Calculations based on the density of sources predicted by Jarrett's models (Reference 8) suggest that as many as 25-50 nights' data must be combined to obtain uniform illumination of each pixel. This would not be possible of the detector and/or telescope responsivities were found to be unstable over these time scales. Developing such a procedure would take a significant new effort and would severely impact the current 2MAPPS design. Finally, such a procedure remains hypothetical, but could be easily tested off-line when more Protocamera test data has been collected.