This repository includes primarily work performed or initialized during Erik's August 2013 Edmonton workshop.

The goals:

1. Increase the number of reliable distance probability distribution functions (DPDFs) from ~600 via the EMAF method to, ideally, ~thousands.
2. Determine what a next possible catalog might include.

Not all sources have HCO+ detections. For those with HCO+ non-detections, or v2 sources unobserved in HCO+, some other method for acquiring velocities is useful.

Using a method described in the Distance Omnibus conference proceedings letter, we have extracted velocities for a significant number of sources. There is python code for the extraction component in python/spectrum_extraction/, but Tim did most of the work using IDL.

TODO: Describe method in more detail. A figure showing the rind & core regions would be very helpful here.

These plots are the most important. First, the HCO+ and ¹³ CO peak velocities match 85% of the time when using this method. This is about 5% better than Yancy did by just picking the velocity associated with the peak ¹³ CO directly, but the difference between ¹³ CO and HCO+ velocity has a dispersion of 0.5 km/s now, much smaller than the 2 km/s Yancy found. The agreement fraction is better for single-peaked ¹³ CO sources.

There may be hope of doing a step better, it is not yet clear. Yancy has suggested observing the non-matched sources deeply with the HHT to see if there are additional HCO+ lines on the multi-peaked ¹³ CO lines of sight.

This LV diagram shows the HCO+ peaks in black on a GRS latitude-integrated diagram, but with only the "rind-extracted" ¹³ CO spectra included.

Here are some additional plots comparing ¹³ CO and HCO+ spectra:

One intriguing possibility is resolving the distance, or at least the kinematic distance ambiguity (KDA), using the known size-linewidth relation (SLR) of clouds. We implemented two methods to measure the SLR.

Erik worked on implementing the PCA method introduced by Mark Heyer and Chris Brunt to measure the structure function of the molecular cloud.

He was able to successfully determine the distance to Perseus and Serpens to within ~10-20%, which is remarkable. Orion didn't match so well.

In order to apply this method to GRS clouds, we need to further develop a cloud masking procedure, which may just mean clumpfinding. Because the Galactic Plane is full of clouds, and the PCA approach functions on PPV data cubes, it is necessary to "mask out" the cloud of interest, otherwise the cloud signals will be confused.

Adam worked on using dendrograms to measure the SLR. Dendrograms effectively do clumpfind extractions with multiple different noise floors, preserving information about which clumps sit inside which clouds.

Dendrogramming worked, and produced sigma-R diagrams, but there is a great deal of uncertainty left in these, and they don't match predictions yet.

TODO: add figures