This is a progress report on an effort to investigate any possible bias in the photometry due to a difference in the propfile fit and aperture photometry ("APPHOT") backgrounds.

The background (BG) in the profile photometry fitting is estimated and subtracted from the data before the least-squares fit proper. The profile fit BG is estimated from a preliminary step using an annulus which is the same size as the background annulus used in the aperture photometry routine APPHOT, but the algorithm is not the same and the center of the annulus is also not quite the same. (The source position determined from the profile fit is used for the APPHOT position, and of course this is not available until after the profile fit BG has been determined. The annulus for the profile fit BG is centered on the initial position from the PIXPHOT source extraction algorithm.)

Evidently any systematic difference between the two backgrounds could introduce a bias in the photometry. In order to study this, a special test version of PROPHOT writes the difference in the mean APPHOT and mean profile fit BGs into the output file for further examination. Both BGs are the average over the frames available for the source, and are in units of DU per camera pixel.

                                    000317n


             Scan 072
                           J           H          K
                    # stars         1475        1011       4057
                    mean(del)      -0.0497     -0.2146    -0.1345
                    median(del)    -0.0088     -0.0900    -0.0838
                    mode(del)      -0.05        0.05      -0.15
                    St.Dev(del)     1.489       2.027      1.431

                 Mode in 0.1 bins.  Units of del [ADU/pixel]


             Scan 140
                                     J           H          K
                    # stars        30946       28938      24184
                    mean(del)       0.3321      0.3245     0.3592
                    median(del)     0.3843      0.3839     0.3924
                    mode(del)       0.45        0.55       0.45
                    St.Dev(del)     1.090       1.109      1.003

                 Mode in 0.1 bins.  Units of del [ADU/pixel]

The following are a couple of scans from a night that seemed to have an unusually large change in the K_S-band plot of (StdMag-PSFMag) versus Mag:

                                   990523n

             Scan 047
                                     J           H          K
                    # stars         1314        1391       3856
                    mean(del)      -0.0769     -0.2349    -0.1880
                    median(del)    -0.0464     -0.0879    -0.0103
                    mode(del)      -0.05       -0.15      -0.10
                    St.Dev(del)     0.713       3.075      1.723

                 Mode in 0.1 bins.  Units of del [ADU/pixel]

             Scan 086
                                     J           H          K
                    # stars         5336        5420       4769
                    mean(del)      -0.0956     -0.1487    -0.1228
                    median(del)    -0.0500     -0.1006    -0.0952
                    mode(del)      -0.0500     -0.25      -0.05
                    St.Dev(del)     1.097       1.672      0.778

                 Mode in 0.1 bins.  Units of del [ADU/pixel]

If a mag 9 star is just starting to saturate in READ2, about 35000 DU above the offset floor, and a 15th mag star has S/N of 10, then one sigma corresponds to about 14 DU in the BG level.

From this the effect does not seem to be important.

5 July 2001, WAW.