TO: 2MASS Team 1/30/95 FROM: C. Beichman SUBJECT: Limitations to Photometric Accuracy
The undersampled nature of the 2MASS data leads to a variety
of photometric errors. Two sources of error are described here
as due to: 1) areas of low or zero responsivity around each pixel; 2)
completely dead pixels. The physical basis of problem (1) was described
in a note from Mike Skrutskie. Mike described spot measurements that
indicated that as much as 30% of the physical area of each pixel could
be non-responsive to incident light. Typical science-grade NICMOS
arrays have between 0.2-1% of the pixels which are completely dead.
The first problem results in photometry with a dispersion of
1% under most seeing conditions. The proper choice of algorithm
can eliminate most effects of dead pixels from the final photometry.
Deriving the photometry from the coadded images, instead of from the
raw frames, would result in a 2-5% amplitude bias for 10% of the
sources.