8.2 Properties of the input data 8.2.1 Source selection 8.3 Processing steps: the Apsis modules

8.2.2 Processing flags

Author(s): René Andrae, Orlagh Creevey

As mentioned in Section 8.2.1, we use flags in our two algorithms, Priam and FLAME, to indicate bad input photometry, missing astrometry, etc. After filtering data for the final catalogue release, the FLAME flags are now obsolete. The Priam flags are listed in Table 8.1. The flags are of the generic form XYABCDE, where each letter represents an integer digit. The first two, XY, identify the Apsis module (01 for a Priam flag). We strongly recommend the catalogue user to make use of these flags in order to filter out bad data.

Table 8.1: Definition of Priam processing flags, which have the format 01ABCDE. (Andrae et al. 2018).

position	value	meaning
A	0	parallax value is strictly positive ( $\varpi>0$ )
A	1	parallax value is non-positive ( $\varpi\leq 0$ ) such that extinction estimate does not work
	2	while $\varpi>0$ the parallax error is $\sigma_{\varpi}>1$ mas
B	0	both colours are close to the standard locus
	1	below standard locus, i.e., $(G_{\rm BP}-G)>0.1$ and $(G-G_{\rm RP})<((G_{\rm BP}-G)-0.1)^{0.4}-0.3$
	2	above standard locus, i.e., $(G-G_{\rm RP})>(2.5\cdot((G_{\rm BP}-G)+0.02))$ or
		$(G-G_{\rm RP})>(0.5\cdot((G_{\rm BP}-G)-1.0)+1.1)$
C	0	$G-G_{\rm RP}$ colour is inside union of $T_{\rm eff}$ and extinction training sets
	1	$G-G_{\rm RP}$ colour is smaller than union of $T_{\rm eff}$ and extinction training sets
	2	$G-G_{\rm RP}$ colour is larger than union of $T_{\rm eff}$ and extinction training sets
D	0	$G_{\rm BP}-G$ colour is inside union of $T_{\rm eff}$ and extinction training sets
	1	$G_{\rm BP}-G$ colour is smaller than union of $T_{\rm eff}$ and extinction training sets
	2	$G_{\rm BP}-G$ colour is larger than union of $T_{\rm eff}$ and extinction training sets
E	1	input data was gold photometry
E	2	input data was silver photometry

Concerning the Priam flags in Table 8.1, the digits C and D refer to colour ranges in the union of the $T_{\rm eff}$ and extinction training sets. These can be used to remove sources which are unlikely to get good results from Priam. The colour ranges for the separate models are as follows (so a colour selection could be used for the parameter of interest, instead of using these digits in the flags): For $T_{\rm eff}$ : $G_{\rm BP}-G$ : $-0.06$ to $4.38$ mag, $G-G_{\rm RP}$ : $-0.15$ to $2.08$ mag. For extinction: $G_{\rm BP}-G$ : $-0.12$ to $4.66$ mag, $G-G_{\rm RP}$ : $-0.20$ to $1.69$ mag.

As a first example, consider use of the temperature estimates (see Section 8.3.1). The best temperature estimates require good photometry whereas astrometry is irrelevant (since temperatures are estimated from colours). Therefore, from Table 8.1 only the flags 01A0001 and 01A0002 should be used, but all values of A are acceptable. As ${\cal R}$ has been derived from $T_{\rm eff}$ , it is strongly recommended that the user also use these flags to filter on ${\cal R}$ .

Concerning the use of extinction or reddening estimates (see Section 8.3.2), the filtering on best results has already been done during the Gaia DR2 catalogue production. Consequently, for $A_{\rm G}$ and $E(G_{\rm BP}-G_{\rm RP})$ only estimates with “clean” flags are provided and no further filtering is necessary.

Finally, note that these flags have been defined before the final filtering of data products for Gaia DR2. Consequently, some of these flags may never occur.