This site provides the basic EDGE data products for galaxies observed in both D and E arrays. The basic properties of the data cubes are summarized in this table.
13CO data products will be released at the end of 2020. Note that the 12CO and 13CO cubes have not been matched to the same resolution.
Please cite the CARMA EDGE data paper (Bolatto et al. 2017) if you use these data in presentations or publications.
Previous public data releases:
These are the "default" cubes generated in MIRIAD, where the cleaning is done over an unbiased region (the field of view which extends to where the RMS noise reaches 2.5 times its minimum value). The cubes were made with 20 km/s channels with a typical velocity range of 860 km/s. The signal cubes have been primary gain corrected, and masked beyond the 50% sensitivity region.
The 2D sensitivity pattern (normalized to 1 at the field center) is useful for plotting; for example, the value=2 contour shows the field of view over which the sensitivity falls to half of its best value. For the peak brightness temperature map, the sensitivity pattern has been divided out, so units are only correct at FOV center. This is proportional to a S/N map.
| CO | 13CO | |
| signal cube (Jy/beam) | data link | data link |
| RMS noise cube (Jy/beam) | data link | data link |
| 2D sensitivity pattern | data link | data link |
| 2D peak temperature image (K) | data link | data link |
These are the signal cubes but covering a larger field of view, and with the sensitivity pattern divided out so that the noise is constant across the field. These may be appropriate for some applications where varying noise across the field would pose a problem. Before making flux measurements, each plane of the noise-flattened signal cube should be multiplied by the corresponding 2D sensitivity pattern above.
| CO | 13CO | |
| noise-flattened signal cube (Jy/beam) | data link | data link |
To generate the moment maps, the cmmsk (signal) and cmnse (noise) cubes are uncompressed and unpacked into a single directory.
They are then processed through the maskmoment Python package (https://github.com/tonywong94/maskmoment) using the following script:
The "dilated" set of moment maps, with filenames of the form NGC5732.co.de20_dil.mom0.fits, uses a masking approach to exclude noise before taking moments. The mask is created by starting at high-significance peaks (3.5-sigma or greater in two consecutive channels) in the cube and expanding down to the surrounding 2-sigma contour. An additional "guard" band of 1 pixel in all directions around the mask is added to try to capture additional low-level emission.
The "smoothed" set of moment maps, with filenames of the form NGC5732.co.de20_smo.mom0.fits, uses pre-smoothing to enhance the signal to noise before creating a mask. The cube is spatially smoothed to a final resolution of 14", then the mask is created by starting at high-significance peaks (3.5-sigma or greater in two consecutive channels) in the
Finally a "straight" set of moment maps, with filenames of the form NGC5732.co.de20_str.mom0.fits, are calculated without masking and using the full velocity extent of the cube.
We also provide a set of CSV files containing the integrated spectrum for each galaxy and masking method. Each CSV file that is derived from a "smoothed" mask also provides a spectrum integrated over the projected 2D mask.
For moment 0, the smoothed mask is recommended because it will capture low-level flux better. For the more noise-sensitive higher moments, the dilated mask is recommended.
| Straight | Dilated | Smoothed | |
| mom0+error | CO data; 13CO data | CO data; 13CO data | CO data; 13CO data |
| mom1+error | CO data; 13CO data | CO data; 13CO data | CO data; 13CO data |
| mom2+error | CO data; 13CO data | CO data; 13CO data | CO data; 13CO data |
| peak SNR (no masking) | CO data; 13CO data | ||
| binary mask cube | CO data; 13CO data | CO data; 13CO data | CO data; 13CO data |
| data for flux spectra | CO data; 13CO data | CO data; 13CO data | CO data; 13CO data |
| plots of peak SNR | CO data; 13CO data |
An integrated flux can be obtained by summing the flux spectrum obtained using any of the three masking methods described above. A CSV file summarizing the measurements and 1-sigma errors (in Jy km/s) is provided here.