4.1. Comparison Between Revised and Previous Flux Density Measurements

Figure 2 shows the ratio of the new RBGS total flux density measurements to the previously published BGS₁ + BGS₂ measurements versus the new total flux density measurements in each IRAS band. In general the largest differences occur at the low end of the range of measured flux densities in all four IRAS bands, with the 12 µm and 25 µm bands showing the most dramatic changes, up to a factor of ± 2 in the few most extreme cases. Much of these differences can be accounted for simply by more mature data processing, which had the greatest effect near the survey lower limits in each of the IRAS bands. At 60 µm and 100 µm, where the measured fluxes were often substantially above the IRAS FSC survey limits, the maximum change is typically a more modest factor of ± 30%. At flux densities more than a factor of 2 above the IRAS FSC limits the "new" and "old" values differ by typically < 15%. There is a noticeable tendency for the revised flux density measurements to be systematically higher, on average ~ 5%, among objects brighter than about 35 Jy at 60 µm, and across the entire range of observed flux densities at 100 µm. This is due to a better understanding of the fact that for many of the extended sources with high signal-to-noise ratios, some of the flux extends beyond the previously adopted f(t) aperture size, and is better measured by using f(z). In addition, the original processing used for BGS₁ + BGS₂ based the choice of whether to use f(t) instead of f(template) only on the coadded scan profile widths (a comparison of FWHM and the width at 25% peak to nominal values observed for pure point sources). However, many galaxies have profile widths which are not significantly broader than what is expected for a point source, yet there is extended emission in a faint "pedestal" can be reliably measured as a statistically significant excess of the f(t) aperture value over the f(template) point source fitted value; see the Appendix (Fig. 16) for further details.

Figure 2. The ratio of new total flux density measurements to original estimates versus the base 10 log of the new total flux density at 12 µm, 25 µm, 60 µm and 100 µm.

All sources with extreme S(new) / S(old) flux ratios in Figure 2 were examined in detail, and they are explained by various improvements in the revised processing. Some objects with S(new) / S(old) < 0.8 are cases where the RBGS flux densities have been estimated by using SCLEAN (see Appendix) or peak values to minimize confusion from companion galaxies in pairs, nearby stars, or cirrus, where in BGS₁ + BGS₂ the f(t) method was used and therefore the flux densities quoted there were contaminated (over estimated). Examples are NGC 5194 and NGC 5195, where SCLEAN was used in RBGS to separate the components of this galaxy pair (M 51) at 12 µm and 25 µm. Another example is IRAS F16516-0948 at 100 µm (S(new)/ S(old) = 0.75), where cirrus confusion has been minimized by using the peak flux estimate in RBGS, and f(t) was contaminated (resulting in an over estimated flux density) in BGS₂. Most objects with S(new)/ S(old) > 1.2 are cases where the RBGS flux selection algorithm resulted in the choice of f(t) or f(z) over f(template), where in BGS₁ + BGS₂ a lower flux density estimate was made for reasons explained in the previous paragraph (also see the Appendix). Examples are NGC 3147 at 12 µm (S(new) / S(old) = 2.10) and MCG +07 -23 -019 at 25 µm (S(new) / S(old) = 1.92). Other extreme ratios are due simply to differences between the f(t) results obtained using the final (PASS3) IRAS archive calibration versus the earlier versions utilized in BGS₁ + BGS₂; an example is NGC 4565 at 60 µm (S(new) / S(old) = 0.79). Most of the remaining outliers in Figure 2 are explained by the use of the SCANPI f(z) measurement for all objects smaller than 25 arcminutes in size, where in BGS₁ + BGS₂ flux measurements from Rice et al. (1983; 1988) were always used when available. Examples are NGC 134 at 60 µm (S(new) / S(old) = 1.32) and NGC 4631 at 100 µm (S(new) / S(old) = 0.77). As discussed in Section 2, comparison of the Rice et al. measurements with SCANPI f(z) measurements for galaxies with optical sizes less than 25 arcminutes showed relatively uniform scatter in the residuals, indicating that the use of SCANPI f(z), when significantly larger than f(t), is the best choice for these objects to insure uniformity and consistency in the calibration with the rest of the RBGS objects.

Figure 3 shows the ratio of total flux density to the peak flux density in the coadded scans at 12 µm, 25 µm, 60 µm and 100 µm. The f(peak) value is used here rather than f(template) because the latter measurement does not exist for objects in which the point source template (PSF) fit failed, while for pure point sources f(total) f(template) and thus the ratio S(total) / S(peak) is very close to unity. This figure illustrates the degree to which point-source fitted measurements in the IRAS PSC and IRAS FSC underestimate the total flux densities for objects in the RBGS. There are likely numerous errors in the literature concerning the infrared flux densities and infrared colors of galaxies, due to the fact that some users of IRAS data have not fully appreciated the fact that most bright infrared galaxies in the local universe, as represented here in the RBGS, are marginally extended or resolved by IRAS.

Figure 3. Ratio of the best estimate of the total flux density to the peak value in the coadded scan vs the base 10 log of the new total flux density at 12 µm, 25 µm, 60 µm and 100 µm. Only a few objects with very large ratios are outside the plot region; the selected range in the ratio was chosen to show details for the largest number of points.

A summary of the percentages of sources that were found to be extended in each of the IRAS bands in given in Table 5. The most notable result is that at 60 µm and 100 µm, where S/N is highest and distinctions can be reliably made, there are significantly more resolved or marginally resolved objects than previously thought: 48% in the RBGS versus 45% as previously reported in the BGS₁ + BGS₂ at 60 µm, and 30% in the RBGS versus 23% as previously reported in the BGS₁ + BGS₂ at 100 µm. This is due to a more careful definition of resolved or marginally extended objects as those having significantly more flux between the baseline zero-crossings, f(z), than within the nominal detector size, f(t), in combination with a comparison of W25 and W50 measurements to point-source values. We should emphasize that the BGS₁ + BGS₂ used only W25 and W50 to determine whether sources are resolved, and always chose f(t) to estimate the flux for the R and M (U+) objects. The revised processing has resulted in significantly fewer objects with underestimated total fluxes in the RBGS compared to BGS₁ + BGS₂.