TMD:
Yes, latitude should definitely be considered. However, I think we can come up with a better criterion than |b|>1. It should really be a cut-off in z (linear distance above the plane), say |z| > 250 or 300 pc at the far distance implies the near distance. You might also consider computing near/far probabilities based on both size-linewidth and latitude and multiplying them together for a total probability near and far.
TSR:
Good point about the z cloud heights rather than latitude. I was skimming through Ellsworth-Bowers et al. (2013), which discusses KDA disambiguation at length, and the following section jumped out at me as relevant (especially regarding some of the Nessie figures Alyssa has shown)
I also like the idea of multiplying probabilities. I'll think carefully about how I want to evaluate the probability for the size-linewidth relation; I suspect I could use the 1-sigma scatter in the relation (measured from the non-KDA Outer Galaxy, probably in the Second Quadrant) to get a measure of that, and then assume Gaussian probabilities, but let me know if you have another recommendation.
identify scatter in R from the sigma-v - R fit in the second quadrant, and let that be the variance on a nominal gaussian used to calculate a "probability" of a certain size fitting the size-linewidth thing. (remember it's a power-law -- the uncertainty might better apply in log-space?)
Get the transformation to Galactic coordinates from the matrix thing above (Tim EB)
Figure out the variance/scale-height on the z-direction of molecular clouds, and use that to do a probability thing, too.
TMD: Yes, latitude should definitely be considered. However, I think we can come up with a better criterion than |b|>1. It should really be a cut-off in z (linear distance above the plane), say |z| > 250 or 300 pc at the far distance implies the near distance. You might also consider computing near/far probabilities based on both size-linewidth and latitude and multiplying them together for a total probability near and far.
TSR: Good point about the z cloud heights rather than latitude. I was skimming through Ellsworth-Bowers et al. (2013), which discusses KDA disambiguation at length, and the following section jumped out at me as relevant (especially regarding some of the Nessie figures Alyssa has shown)
from: http://adsabs.harvard.edu/abs/2013ApJ...770...39E so I would probably follow their prescription of calculating z in this case.
I also like the idea of multiplying probabilities. I'll think carefully about how I want to evaluate the probability for the size-linewidth relation; I suspect I could use the 1-sigma scatter in the relation (measured from the non-KDA Outer Galaxy, probably in the Second Quadrant) to get a measure of that, and then assume Gaussian probabilities, but let me know if you have another recommendation.