P5 README

Pipeline 5 makes use of Independent Component Analysis (ICA) to separate the IRAC 1 images (3.6um) into two source maps, the first of which is identified with the stellar emission (s1), and the other one with dust (s2). The user is referred to Querejeta et al. (2014) for details on the motivation of the method and its operative details.

We impose two initial cuts on the sample to select only those galaxies that are suitable for the application of ICA. Excluded from our pipeline are all galaxies with initially low signal-to-noise (below S/N=10) and all galaxies with original colors [3.6]-[4.5]<0, indicative of limited dust contamination.

In the companion table (P5_table.txt), the second column indicates whether or not ICA was applied, according to the criteria above. '0' indicates that the pipeline was run; '1' means that no solution was calculated because the galaxy had [3.6]-[4.5]<0; '2' implies that it was not run due to low S/N; '3' means that there was some data problem (e.g. missing P3 information). Therefore, a P5/ subdirectory only exists for galaxies which fulfilled those conditions. For ease of use, all our maps are cutouts centered on the photometric center identified in P3.

For the galaxies with an ICA solution, the following files are provided:

  1. XXX.stellar.fits --- map of the 'cleaned' flux from old stars (s1), which can be effectively converted into a stellar mass map following the details in Querejeta et al. (2014).
  2. XXX.nonstellar.fits --- dust emission identified in the IRAC 1 image, what is identifies as s2 in the paper, which has been removed from the original image to produce file (1)
  3. XXX.ICAmask.fits --- regions in the mask locate pixels excluded from the analysis. The areas with positive integers are inherited from the P4 masks, although note that some P4 regions are removed in our pipeline. Regions labeled with negative integers are associated to our recursive ICA strategy (second ICA iteration). This mask applies to all our pipeline products.
  4. XXX.colormap.fits --- the map of [3.6]-[4.5] color for the cleaned stellar light map in file (1). This color map differs from the original color map (obtainable directly from IRAC 1 and IRAC 2) specifically in regions corrected with ICA (i.e. where nonstellar map (2) is non-zero). The color in these regions reflects the (constant) stellar color found by ICA. This can be useful, e.g., for assigning an optimal color-dependent M/L to convert the flux from (1) into an actual mass map.