2 Data

As part of the incorporation of the OGS into the Canadian Galactic Plane Survey (CGPS, Taylor et al. 2003) the original OGS was reprocessed (Brunt & Ontkean 2003, in preparation). This reprocessing removed correlated noise signals (see Heyer et al. 1998), suppressed the effects of unclean reference positions and corrected incorrectly placed spectra from the first OGS release. The OGS data were convolved to 100  $.\!\!^{\prime\prime}$44 resolution and converted onto a standard CGPS coordinate grid prior to release as a CGPS data product. The typical sensitivity of the reprocessed OGS data is 0.17 K at 100  $.\!\!^{\prime\prime}$44 resolution. The reprocessed OGS data cubes, along with all of the other CGPS data sets, are publicly available through the Canadian Astronomy Data Centre (CADC).

As a cursory inspection of any of the OGS data cubes will reveal, the CO emission in the outer Galaxy is highly structured and quite extensive. To facilitate the comparison of the OGS data with other large databases a cloud catalogue was constructed (BKP) that describes the OGS data in terms of discrete lbv structures (i.e. CO "clouds''). BKP used the 100  $.\!\!^{\prime\prime}$44 resolution data on the original OGS 50  $.\!\!^{\prime\prime}$22 grid but after conversion onto the 0.824 km s^-1 CGPS spectroscopic grid. The BKP catalogue was generated using a two-phase object identification algorithm. In the first phase, all contiguous lbvstructures consisting of at least 4 voxels over which the observed radiation temperature exceeded 0.8 K were identified. The second phase of the algorithm further decomposed these (sometimes very large) structures into smaller regions of localized CO emission enhancements, using an enhanced version of the CLUMPFIND algorithm (Williams et al. 1994). This procedure leads to a high resolution discretization of the data which is critical for accurately associating CO emission with specific sources seen at other wavelengths. There are 14 592 objects contained in the BKP catalogue.

The IRAS point source catalogue (PSC) was used to obtain positional information on all of the IRAS PSC sources ("IRAS sources'' hereafter) in the OGS survey region. In total there are 6698 IRAS sources in the OGS region of which 4315 have detectable CO along the line of sight as accounted for in the BKP catalogue. Even in the outer Galaxy, the widespread distribution of CO emission means that for any given IRAS source there is a chance that non-associated CO emission could be detected along the line of sight and that multiple emission components along the line of sight will occur. The latter problem was recognized in WB89 and qualitative criteria were developed to determine which of the multiple CO components along the line of sight were more likely associated with the IRAS source. However, since full maps of the CO distribution around the IRAS source position were not available there was no way to quantitatively rank the quality of a given IRAS-CO association.

BKP developed a statistical source association method that exploits the spatial information in the OGS in order to discriminate between multiple CO detections. They examined the frequency with which 10⁶ randomly chosen lines of sight within the OGS region coincided with CO emission incorporated into the BKP catalogue, and occurred within an angular offset of $\delta r$ arcminutes from an object with peak temperature exceeding $T_{\rm p}$ kelvins. From the observed number of associations, $N_{\rm obs}({\leq}\delta r,~{\geq}T_{\rm p})$, they defined the expected number of associations, $N_{\rm E}({\leq}\delta r,~{\geq}T_{\rm p})$, that would be made towards a random position within the OGS boundaries:

$$N_{\rm E}({\leq}\delta r,~{\geq}T_{\rm p}) \;\;\; = \;\;\; N_{\rm obs}({\leq}\delta r,~{\geq}T_{\rm p}) / N_{\rm test}$$

where $N_{\rm test}$ = 10⁶.  BKP showed that 52% of lines of sight through the OGS region will encounter CO emission in excess of 0.8 K, so it is useful to evaluate the quality of an IRAS-CO association using BKP's derived $N_{\rm E}$ values. As one would expect, large $\delta r$ - low $T_{\rm p}$ "associations'' are more likely (i.e. have higher $N_{\rm E}$) in the random sample than low $\delta r$ - high $T_{\rm p}$ associations. For IRAS sources, we can examine the spatial offset and peak temperature of all possible CO associations and assign an $N_{\rm E}$ value accordingly. Chance associations will generally have high $N_{\rm E}$, while the close spatial coincidence (low $\delta r$) of a cloud with high $T_{\rm p}$ would have low $N_{\rm E}$, and indicate a physically interesting association.

The OGS data at each IRAS source position was examined for CO emission that is accounted for by the BKP catalogue. For each IRAS-CO coincidence, the $\delta r$-$T_{\rm p}$ values were determined and an $N_{\rm E}$ value was assigned. For IRAS sources with only one possible CO association, the $N_{\rm E}$ value provides a quantitative measure of the likelihood of a true association. Multiple associations along a given line of sight were ranked by their $N_{\rm E}$ value and the cloud with the lowest $N_{\rm E}$ was taken to be the most likely true IRAS-CO association (see e.g., Fig. 1). We did not discard the other possibilities; these were retained in an $N_{\rm E}$-ordered list for each IRAS source.

Figure 1: Multiple CO emission components. The upper left panel shows the CO spectrum towards IRAS 01522+6611 along with the $N_{\rm E}$ ranking of each of the three components along this line of sight. The other three panels show the corresponding CO maps at the velocity indicated above each panel. The white cross indicates the position of IRAS 01522+6611. $N_{\rm E}$ values are 0.0065, 0.3454, and 0.4877 for rank 1, 2, and 3 respectively.

Determining the absolute level of the $N_{\rm E}$ values where one would consider the IRAS-CO association to be true is more ambiguous and will often vary depending upon the sort of object one is looking at (e.g. Kerton 2002). Often additional information can be utilized to help determine the level at which the majority of the associations are non-random. Section 4.1 in this paper provides a concrete example of this for the case of an investigation of candidate zone-of-avoidance galaxies (ZOAGs).

In addition to the difficulty in establishing a useful absolute $N_{\rm E}$ level, there are three sources of confusion in this scheme. First a "type 0'' error arises when no CO detection is made towards an IRAS source due to the CO emission being either weak or very small in extent (e.g., WB89 298, see Sect. 3). A "type 1'' error arises when a spuriously low $N_{\rm E}$ value is assigned to a cloud that is not physically associated with the IRAS source; this will be of particular concern if a true physical association with a different cloud is masked. As an example of a type 1 error, consider a true physical association of an IRAS source with a far outer Galaxy molecular cloud. Since emission from the distant cloud is often weak, a chance association with a nearby cloud at comparably low $\delta r$ but higher $T_{\rm p}$ will result in the nearby cloud achieving a higher rank in the $N_{\rm E}$-ordered list. A "type 2'' error occurs when a true physical association of an IRAS source and a molecular cloud is assigned a spuriously high $N_{\rm E}$ value. The thresholding scheme used by BKP identified emission enhancements in the OGS data only if they were distinguishable from their surroundings by at least 0.8 K. This threshold is relatively high ($\sim$4.7$\sigma$) compared with the noise level in the data to avoid inclusion of spurious clouds in the catalog. However, such a high threshold can result in lack of precision on small scales, arising from less pronounced emission enhancements being incorporated into larger, composite structures; a particular example of this is given in Sect. 3. Since we retain all possible IRAS-CO associations in our list, no true associations are discarded but they may be hidden by the influence of type 1 and 2 errors on the assigned $N_{\rm E}$ and relative rank.

Table 1 contains the main result of this paper - an $N_{\rm E}$ ordered list of all the IRAS-CO associations within the OGS. The first column of the table contains a running number of IRAS-CO associations followed by the IRAS source name (Cols. 2 and 3), position in Galactic coordinates (Cols. 4 and 5) and flux density information (flux densities in Cols. 6-9 for 12, 25, 60 and 100 $\mu$m respectively; percentage errors for the respective flux densities in Cols. 10-13). Column 14 indicates if the following offsets refer to the CO cloud peak or centroid position. Columns 15 to 21 contain information about the CO cloud. The CO peak/centroid position in Galactic coordinates and the $V_{\rm lsr}$ is given in Cols. 15-17, followed by the peak temperature ($T_{\rm p}$; Col. 18), the CO temperature at the IRAS source position (Col. 19), the angular offset between the IRAS position and the CO peak (Col. 20) and the BKP number of the cloud (Col. 21). The BKP number can be used to access the BKP catalogue where more detailed information about the CO cloud can be found. Finally Cols. 22-24 provide the $N_{\rm E}$ value (Col. 22) the relative $N_{\rm E}$ ranking (Col. 23) and the WB89 catalogue number if applicable (Col. 24). Note that the WB89 number is indicated only if CO was detected towards the IRAS source by WB89. A more detailed description of the format and contents of the machine-readable table is given in the header of the electronic version.

 

 

Table 1: IRAS-CO associations.

This table is available only in electronic form at the CDS

http://cdsweb.u-strasbg.fr/cgi-bin/qcat?J/A+A/399/1083

Also provided in this paper is a listing of the IRAS sources within the OGS that do not have associated CO as accounted for by the BKP catalogue (see Table 2). The columns in Table 2 are the same as the first 13 columns of Table 1.

 

 

Table 2: IRAS sources with no CO associations.

This table is available only in electronic form at the CDS

http://cdsweb.u-strasbg.fr/cgi-bin/qcat?J/A+A/399/1083

The IRAS-CO association table contains a comprehensive account of the star-forming molecular ISM in the OGS region of the Galaxy that can be examined by itself or as a starting point for other investigations. In the next section we analyze the contents of the table and compare it to the results of the extensive WB89 targeted survey.